High Resolution Ct Of The Lung

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT OF THE LUNG II

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS Georgeann McGuinness, MD, and David P. Naidich, MD

With the emergence of effective antibiotic therapy in the middle of the twentieth century, postinfectious bronchiectasis in North America became a relatively uncommon diagnosis. A resurgence of interest in the diagnosis of bronchiectasis likely reflects the influence of enhanced diagnostic capabilities provided by advances in CT technique, in particular the advent of high-resolution CT (HRCT), and its ability to document mild or unsuspected disease. Since the introduction of CT into clinical practice in the 1970s, evaluation of the airways has shifted to use of CT as a first-line examination, with bronchoscopy relegated to a secondary role. Conventional CT allows a direct, noninvasive visualization of structural changes involving both large- and mediumsized airways. HRCT techniques now allow identification of findings associated with small airway disease, including both structural and physiologic abnormalities. Further advances in technology, specifically the advent of multislice detector CT, promises continued improvement in CT evaluation of the airways. Increasing awareness of a number of clinical settings associated with bronchiectasis also is partially responsible for increased interest in bronchiectasis. The association between bronchiectasis and HIV infection, and the association of bronchiectasis with the nonclassic form of Mycobacterium avium complex

(MAC) infection are newer clinical settings in which establishing the diagnosis of bronchiectasis is becoming increasingly important. The associations between bronchiectasis and allergic bronchopulmonary aspergillosis (ABPA) and cystic fibrosis (CF) are well-established settings where HRCT is recognized as a valuable clinical tool. This article discusses HRCT findings associated with medium-sized and small airways diseases, focusing on bronchiectasis of both infectious and noninfectious causes. Imaging techniques and characteristic HRCT findings in patients with bronchiectasis are reviewed. Bronchiolar diseases and asthma are covered in a separate article elsewhere in this issue. PATHOGENESIS OF BRONCHIECTASIS The most common established cause of bronchiectasis, defined as irreversible dilatation of the peripheral airways, is infection, often of a chronic or recurrent nature.2 Necrotizing bacterial infections, mycobacterial infection, and less commonly fungal infections, such as histoplasmosis, may result in bronchiectasis. Background conditions may predispose to infection. These include airway obstruction resulting from tumor, impacted mucus, foreign body, or stricture; impaired mucociliary clearance, such as cilia dysmotil-

From the Department of Radiology, New York University Medical Center, New York, New York

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

1

2

MCGUINNESS & NAIDICH

ity syndrome, bronchial cartilage deficiency, or CF; and generalized congenital and acquired immunodeficiency states. Depending on the underlying cause bronchiectasis may be either focal or diffuse (Table 1). A specific cause for bronchiectasis can be established, however, in less than 40% of cases. During acute inflammation the bronchial walls become thickened with inflammatory cells and edema. Ongoing cell-mediated immune response results in an infiltration of lymphocytes, antigen processing cells, and macrophages.9 Focal mucosal erosions may develop, resulting in peribronchial abscess formation. Chronic inflammation may result in destruction of the elastin layer of the bronchial wall, with thinning of the bronchial cartilage. Bronchomalacia results, with consequent hypertrophy of the muscular component of the bronchial wall. In severe bronchiectasis the ciliated epithelium is replaced by squamous or columnar epithelium. Additionally, with chronic inflammation bronchial artery neovascularization may occur, accounting for the high incidence of hemoptysis associated with bronchiectasis. Hemoptysis occurs in up to 50% of cases of bronchiectasis, and may be

the only clinical finding,2, 34 although recurrent infection and persistent cough, usually productive of sputum, are more usual symptoms. ACCURACY OF CT VERSUS BRONCHOGRAPHY Bronchography has been considered the gold standard for documenting the presence, severity, and extent of bronchiectasis. There are multiple risks and limitations to this procedure, however, beyond the obvious one of patient discomfort. Bronchography requires the use of local anesthesia and bronchographic medium, with the potential for allergic reactions and temporary impairment of ventilation. Evaluation is limited to the airways, with no visualization of the lung parenchyma or mediastinum, and only one lung may be evaluated at a time. Most importantly, despite meticulous technique, underfilling of bronchi at bronchography is common and diminishes the sensitivity of this examination.46 Consequently, it is now generally accepted that CT is the optimal noninvasive means of diagnosing bronchiectasis, obviating the need for bronchography. Bronchography may still

Table 1. ETIOLOGIES OF BRONCHIECTASIS Postinfectious

Bacterial Mycobacterial Viral (Swyer-James syndrome) Fungal (including ABPA)

Impaired mucociliary clearance

Dyskinetic cilia syndromes (Kartagener’s syndrome) Cystic fibrosis

Inherited cellular or molecular defects

Cystic fibrosis ␣1-antitrypsin deficiency

Congenital bronchial abnormalities

Mounier-Kuhn’s syndrome Williams-Campbell syndrome Bronchopulmonary sequestion

Acquired or inherited immune deficiency

AIDS Acquired hypogammaglobulinemia Selective or pangammaglobulinemia

Postobstructive

Intrinsic (neoplasm, foreign body, or stricture)

Post-transplantation

Chronic rejection (obliterative bronchiolitis) Graft-versus-host disease

Postinflammatory pneumonitis

Aspiration Inhalation of toxic gases

Collagen vascular diseases

Rheumatoid arthritis Sjo¨gren’s syndrome Ankylosing spondylitis Marfan syndrome Relapsing polychondritis Inflammatory bowel disease (UC; CD)

Miscellaneous

Asthma Yellow-nail syndrome Sarcoidosis

ABPA  Allergic bronchopulmonary aspergillosis; UC  ulcerative colitis; CD  Crohn’s disease.

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

rarely be performed in limited specific clinical settings, usually to evaluate postsurgical complications. High-resolution CT compares favorably with bronchography in the diagnosis of bronchiectasis.8, 37 Grenier et al8 found a sensitivity of 96% and a specificity of 93% when comparing HRCT using 1.5-mm-thick sections with bronchography. In this series CT underestimated the extent of disease in two lungs, although in only one of these was the CT interpreted as normal. In one of these patients bronchiectasis was missed bronchographically because of failure to opacify a dilated bronchus impacted with mucus, a diagnosis prospectively suggested at CT. More recently Young et al59 performed a retrospective comparison of the accuracy of CT and bronchography through a segment-by-segment analysis of 259 segmental bronchi in 19 patients. Despite the fact that high-resolution technique was used in only 11 of these cases the sensitivity of CT was 98% and the specificity was 99%, compared with bronchography. Bronchography, although imperfect, has been the benchmark in establishing the diagnosis because few comparisons between CT and pathologic features of bronchiectasis exist. One such study, by Kang et al,18 identified bronchiectasis in 87% of 47 surgically resected lobes in 22 patients with pathologic confirmation of the diagnosis. Because these patients underwent surgical resection for presumably severe disease the authors acknowledge that the detection rate of HRCT may be less with mild bronchiectasis. CT DIAGNOSIS OF BRONCHIECTASIS Even before the routine use of HRCT it was apparent that conventional CT could provide detailed characterization of airway pathology and consequently could play a role in diagnosing bronchiectasis. As predicted, HRCT has provided further accuracy in this regard. The key to the HRCT diagnosis of bronchiectasis is identification of an enlarged internal bronchial diameter. Failure of an airway to taper while progressing toward the lung periphery,28 and identification of airways in the extreme lung periphery, are other valuable CT signs. Indirect signs of bronchiectasis at HRCT include bronchial wall thickening, impaction, and focal air trapping.26 Normally, bronchi are not visible in the peripheral 2 cm of lung at CT, although with

3

current improved spatial resolution at HRCT normal bronchi within 1 cm of the pleural surfaces are now occasionally seen. This lack of visibility is consequent to an inability to identify the walls of small bronchi when of normal thickness. Maximal spatial resolution at HRCT is approximately 300 ␮m, corresponding to bronchi of approximately 1.5 to 2 mm in diameter (i.e., seventh-order through ninth-order bronchi).38, 57 The range of normal bronchial diameters at CT has not been determined to date, and the actual measurement of small bronchial is impractical and time consuming. Visual criteria for establishing the presence of bronchial dilatation are most often used.6, 12, 19, 26, 31, 33 Foremost among these is a comparison of the internal diameter of a bronchus to the diameter of the adjacent pulmonary artery branch. Normally the diameter of a bronchus is approximately equal to its companion pulmonary artery. Identifying a bronchus at any given level larger than its accompanying pulmonary artery remains the single most useful clue to the diagnosis. The accuracy of this technique has been validated in a number of studies comparing CT at bronchography in patients with bronchiectasis.5, 60 Limitations in the accuracy of this sign include oblique orientation of the bronchus and vessel to the plane of the section, and physiologic abnormalities resulting in variations in both the artery and bronchus size. For example, with regional hypoxia secondary vasoconstriction may result in an artery smaller than even a normal-sized accompanying bronchus (Fig. 1A), and in the setting of true bronchiectasis the disparity between the bronchial diameter and the pulmonary artery may be accentuated secondary to this decreased lung perfusion in regions of bronchiectasis.19 Paradoxically, bronchiectasis could be missed if there is concomitant arterial dilation, for instance if dilated arteries occurring in pulmonary hypertension obscure recognition of a dilated bronchus (Fig. 1B). Patients with asthma may also demonstrate bronchial dilatation without wall destruction, and transient bronchial dilatation may be identified in normal airways consequent to changes in positioning (Fig. 1C). Lynch et al28 questioned the reliability of this sign after finding that 59% of normal volunteers had at least one bronchus with an internal diameter exceeding that of its accompanying artery. This study was performed in a city located at an altitude 1 mile above sea

4

MCGUINNESS & NAIDICH

Figure 1. Pitfalls in the diagnosis of bronchiectasis: limitations of visual criteria. A, High-resolution CT (HRCT) image with 1.0 mm collimation through the midlungs of a 63-year-old woman with obliterative bronchiolitis and consequent mosaic lung perfusion. Note areas of relatively darker lung (open arrows) representing regions of hypoxic vasoconstriction consequent to local hypoxia. Within these regions, the constricted vessels appear smaller than their accompanying normal caliber bronchus (solid arrows). B, One-mm-thick section through the upper lungs of a 68-year-old man with bronchiectasis and pulmonary hypertension. Grossly dilated pulmonary artery branches are markedly larger than adjacent normal airways (curved arrows). Nearby bronchiectatic airways are larger than their companion pulmonary artery branches (open arrow). In the posterior right upper lobe, clearly dilated bronchi are approximately the same size as the adjacent dilated artery (solid straight arrow). Bronchiectasis could be missed in regions such as these if underlying arterial dilatation was not recognized. Supine (C) and prone (D) 1-mm-thick images through the lung bases in a 57-year-old man with asbestos exposure. In the supine position, the bronchi are the same size as adjacent arterial branches (arrow), whereas in the prone position, these same normal airways demonstrate transient, position-dependent dilation (arrows).

level, and may reflect the decreased ambient oxygen tension. The unreliability of this sign for patients living at high altitudes was corroborated by Kim et al,19 who found at least one bronchus larger than the accompanying artery in 53% of normal volunteers living at a high altitude, as compared with 12.5% of those living at sea level. When a dilated bronchus runs parallel to the plane of the CT section the walls of the bronchus can be recognized as parallel lines or ‘‘tram tracks.’’ Cylindrical bronchiectasis typically results in uniform airway dilation (Fig. 2A). In distinction, varicose bronchiecta-

sis is characterized by a beaded appearance when bronchi lie within the plane of the section, the result of focal segments with a variable degree of dilatation. Cystic bronchiectasis results in rounded cysts, sometimes referred to as a string of cysts or a cluster of cysts, depending on airway orientation (Fig. 2B). Retained fluid can create a pathognomonic radiographic appearance of multiple air–fluid levels. Lack of bronchial tapering, increasingly recognized as a critical means of diagnosing subtle bronchiectasis in particular,19, 28 is most easily appreciated if the bronchus is within the plane of the section (Fig.

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

5

Figure 2. CT diagnosis of bronchiectasis: visual criteria. A, 1.0-mm high-resolution image depicts cylindrical bronchiectasis in this 77-year-old woman. In cross section, a dilated bronchi forms a ‘‘signet ring’’ appearance with its companion arterial branch (solid arrows). Bronchiectasis in the plane of section appears as ‘‘tram tracks’’ (open arrow). Note relatively thin and thick walls of these bronchi; bronchi wall thickening is not a constant finding in bronchiectasis. B, Cystic bronchiectasis is evident in the posterior right upper lobe of this 51-year-old man at HRCT. These rounded structures might be mistaken for cysts or cavities if their tubular nature is not appreciated on consecutive images. Recognizing bronchiectasis nearby aids in the diagnosis. C, Failure of a bronchus to taper as it approaches the lung periphery is a useful sign to diagnose bronchiectasis, particularly when the bronchus runs within the plane of the section, and therefore, the ‘‘signet ring’’ sign will not be apparent. Nontapering bronchi are evident in both lower lobes of this 19-year-old male patient with cystic fibrosis (arrows), while more advanced bronchiectasis is present in the lingula.

2C). Accurate assessment of this finding for vertically or obliquely oriented airways is difficult in the absence of contiguous sections. Commonly the walls of dilated bronchi are thickened; it cannot, however, be overemphasized that this finding is neither requisite nor constant. When a dilated bronchus is oriented perpendicularly to the plane of the section it is recognized in cross-section as a ring structure, with an internal diameter larger than that of its accompanying pulmonary artery branch. Termed the signet ring sign, the wall of the dilated bronchus represents the band of the ring, whereas the signet is the soft-tissue density produced by the artery (see Fig. 2A).

Recognition of airways, both normal and abnormal, is achieved by identifying bronchial walls, facilitated by the presence of air within the bronchial lumen. If the lumen is opacified, rather than air filled, the appearance at CT is modified significantly. Mucoid impaction refers to the accumulation of inspissated secretions within an airway. Usually the affected airway is ectatic. Mucoid impaction is nonspecific, and quite commonly involves the peripheral airways in patients with typical postinflammatory bronchiectasis. It also may be seen in a wide variety of diseases including CF and ABPA, and congenital abnormalities, such as bronchial atresia and intrapulmonary sequestration. It also may be

6

MCGUINNESS & NAIDICH

associated with a focal proximal obstruction, as may occur secondary to acute inflammation and edema, endobronchial tumor, or bronchostenosis. An impacted bronchus is recognized as a nodular density in cross-section, or a tubular or branching structure if in the plane of the section (Fig. 3A). These usually are of low density, although calcification may occur if chronic. Larger fluid-filled airways may occasionally be confused with dilated vessels on a noncontrast examination, but in most cases the diagnosis is simplified by the presence of other foci of bronchiectasis. To avoid misinterpreting these linear and branching soft tissue structures as abnormal vessels the administration of intravenous contrast is useful to demonstrate a lack of enhancement, which would be expected in a vessel (see Fig. 3). Air trapping, a manifestation of small airways disease producing a mosaic attenuation pattern on HRCT, is increasingly recognized as a possible early marker of bronchiectasis. In a study of 70 patients with bronchiectasis identified in 52% of lobes evaluated, areas of decreased attenuation were identified in 20% of lobes at inspiration and in 34% of lobes at expiration. Although this finding was more prevalent in lobes with bronchiectasis, in 17% of lobes air trapping was identified in the absence of associated bronchiectasis (Fig. 4).11 These findings support speculation that bron-

chiolar disease causing air trapping may precede, and even lead to, the development of bronchiectasis. This relationship also has been noted in the setting of bronchiolitis obliterans (Fig. 5A and B) and bronchiectasis occurring in post-transplantation lungs or graft-versushost disease (Fig. 5C).22, 24 CT IMAGING TECHNIQUES The single most important technical factor influencing optimal evaluation of small airways is use of high-resolution technique, with slice collimation on the order of 1 to 2 mm, coupled with an edge-enhancing reconstruction algorithm. Optimal evaluation of small airways necessitates selecting a slice thickness commensurate with the size of peripheral airways. Significant bronchiectasis may be missed on conventional resolution 7- or 10mm-thick images. HRCT should be viewed as a sampling technique; usually the thin sections are obtained with a 10-mm gap between each section.8 In cases in which clinical or radiographic localizing signs are present, additional slices can be obtained, as needed, through specific regions of interest. Administration of intravenous contrast is rarely necessary. In select cases bolus administration of contrast can be useful. In patients with bronchiectasis secondary to a central ob-

Figure 3. Bronchial impaction: use of intravenous contrast. A, A dilated, branching bronchus evident on the 1-mm-thick high-resolution reconstruction in this 46-year-old man might be mistaken for a vessel given its solid nature (arrow). B, Images obtained 2 minutes after administration of intravenous contrast clearly demonstrate a lack of enhancement in this structure (straight arrow), while vessels increase in density (curved arrow). Density measurement before and after administration of intravenous contrast is a useful technique to confirm the nonvascular nature of a suspected impacted bronchus.

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

Figure 4. Bronchiectasis and air trapping. A, High-resolution image obtained with 1-mm collimation through the lower lungs at inspiration in this 72-year-old woman demonstrates mild bronchiectasis in the middle lobe, lingula, and left lower lobe (arrows). B, At expiration at the same level, air trapping is demonstrated in areas both with (solid arrows) and without (open arrows) bronchiectasis. Air trapping, a manifestation of small airway disease, may precede the development of bronchiectasis or occur in areas without gross airway dilatation.

Figure 5. Bronchiectasis and bronchiolitis obliterans. A, 1-mm collimated image through the lower lungs in a 26-year-old woman with progressive dyspnea in whom idiopathic bronchiolitis obliterans was diagnosed at biopsy 2 years earlier. Subtle areas of decreased lung attenuation (arrows) are evident even at inspiration. Central bronchial wall thickening is present in all lobes, as well as mild bronchiectasis in the middle lobe. B, High-resolution image obtained 2 years later at the same level depicts more pronounced mosaic attenuation, worsening middle lobe bronchiectasis, and new right lower lobe and lingula bronchiectasis. It has been suggested that obstructive disease in the small airways may lead to bronchiectasis, although a causal relationship remains speculative. C, One-mm collimated image through the lung bases in a 29-year-old woman after bone marrow transplantation during treatment for breast cancer demonstrates findings consistent with graft-versus-host disease. Areas of mosaic lung attenuation suggest bronchiolitis obliterans, whereas early bronchiectasis is evident in a characteristically central distribution (curved arrows).

7

8

MCGUINNESS & NAIDICH

structing tumor, for example, differentiation of tumor from impacted mucus (Fig. 6) and information regarding the vascularity of the tumor may be possible. In patients with bronchiectasis primarily involving the posterobasilar segments of the lower lobes recognition of a systemic vascular supply may aid in diagnosing pulmonary sequestration. Finally, in equivocal cases a lack of enhancement after administration of a bolus of intravenous contrast media can be useful for distinguishing

occluded, dilated bronchi from abnormal vessels, especially arteriovenous malformations (Fig. 3D). Indications for volumetric assessment of the small airways have been limited. Engeler et al7 found volumetric acquisition of sets of contiguous HRCT images at selected levels improved diagnostic accuracy in bronchiectasis, compared with analysis of the single initial HRCT image in each set. This was particularly true for bronchiectasis localized to the

Figure 6. Bronchiectasis and impaction consequent to central tumor. A, A 3-mm-thick image at the level of the truncus basalis to the left lower lobe in a 57-year-old man with a left hilar squamous cell tumor demonstrates on wide (lung) windows a tubular structure that could be mistaken for a vessel or tumor extension into a bronchus (arrow). B, After administration of intravenous contrast, the central mass demonstrates enhancement (open arrow), whereas the low-density tubular structure (curved arrow) does not. C, The same image with densitometry markers confirms enhancement of the central tumor to over 70 Hounsfield units (HU), an increase of over 40 HU from precontrast measurements (not shown), whereas the low-density tubular component continues to measure under 20 HU, unchanged from precontrast measurements (not shown) and consistent with postobstructive mucus impaction, rather than endobronchial tumor extension.

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

lung bases, where respiratory motion artifact can degrade images. Lucidarme et al25 studied 50 consecutive patient with suspected bronchiectasis, and compared traditional axial 1.5mm HRCT sections obtained at 10-mm intervals with contiguous 3-mm sections obtained with a single breath-hold volumetric acquisition. The volumetric studies identified bronchiectasis in four cases missed by HRCT, whereas although HRCT identified bronchiectasis in a few specific segments missed by volumetric CT, there were no cases in which the diagnosis of bronchiectasis was established by HRCT alone. More recently, the capabilities provided by multidetector volumetric CT scanning promise further potential advances in small airway imaging. Enhanced speed of data acquisition allows coverage of the entire thorax during a single suspended breath-hold, even with thinslice collimation, on the order of 1 mm. This allows retrospective reconstruction of thin sections at any level, including contiguous HRCT images as needed. Spatial resolution may be improved further with scanners capable of 0.5-mm sections, which generate near isotropic voxels, allowing true cross-sectional areas regardless of airway orientation. Postprocessing techniques include CT bronchographic images, which actually replicate the information provided by a conventional bronchogram (Fig. 7A and C), and multiplanar (Fig 7B) and three-dimensional reformatted images. Clinical applications for these emerging techniques are evolving as experience with, and investigation of, the capabilities of multidetector scanners continues. CLINICAL CORRELATIONS High-resolution CT contributes significantly to the diagnosis and management of the patient with bronchiectasis. HRCT is the imaging method of choice after chest radiography for examining the patient with suspected bronchiectasis. Once the diagnosis is established at HRCT an attempt is made to establish the cause of bronchiectasis when possible. In some cases the pattern and distribution of bronchiectasis is sufficiently characteristic to suggest a specific cause. It should be recognized, however, that in many cases the cause of the bronchiectasis remains unknown, and there is significant overlap in the appearance of bronchiectasis from both known and unknown causes. Reiff et al 43

9

studied the HRCT features of 168 patients with bronchiectasis of known cause compared with a larger group of patients with bronchiectasis of idiopathic cause. CT did identify features that occurred more frequently in certain groups of patients with an identifiable underlying cause of bronchiectasis, but these were not regarded as sufficiently diagnostic. In a similar study by Lee et al,21 experienced observers of CT scans in 108 patient with bronchiectasis of various causes could establish a correct first-choice diagnosis in only 45% of cases; more problematic still, a high confidence level was reached in only 9%, and of these a correct diagnosis was reached in only 35%. These investigators concluded that CT was of little value in diagnosing specific causes of bronchiectasis. In contrast, more recently Cartier et al4 assessed the accuracy of HRCT in determining the cause of bronchiectasis in a group of 82 patients with bronchiectasis of known causes. A correct diagnosis was reached by two independent observers in 61% of cases, including 68% of cases of CF, 67% of cases with tuberculosis, and 56% of cases of ABPA. The conclusion of these authors is that the pattern and distribution of abnormalities at HRCT are influenced by the underlying cause, but the studies still need to be interpreted within clinical context for best accuracy. High-resolution CT provides an assessment of the severity and distribution of the bronchiectasis. Bronchiectasis has traditionally been classified as cylindrical, varicose, or cystic, descriptive terms denoting progressive degrees of severity. Recently, Lynch et al28 correlated HRCT findings with clinical and spirometric findings and found that the type and degree of abnormality on CT correlates with the extent of physiologic impairment. In particular, patients with cystic bronchiectasis had a higher rate of Pseudomonas infection, and consequent purulent sputum. These patients also had more pronounced abnormalities at pulmonary function tests (PFT). Equally important is determination of disease extent. A number of quantitative scoring systems have been proposed, documenting good correlation between HRCT assessment of extent and severity of disease compared with radiographic, clinical, and functional evaluations.3, 14, 15, 47 This is important for accurate assessment of response to therapy and surgical planning, because surgery is only rarely performed in patients with involvement of multiple lung segments.2, 48

10

MCGUINNESS & NAIDICH

Figure 7. Multidetector-row CT and the airways. A, Volumetrically rendered CT bronchographic images of normal airways reconstructed from scan data obtained with contiguous 1.25-mm sections reconstructed every 0.5 mm during a single breath hold. This technique allows visualization of airways to approximately seventh to eighth order bronchi. B, Parasagittal multiplanar and volume rendered reconstructions, respectively, through the right lower lobe in a different patient than shown in A using identical scan technique (1.25-mm sections reconstructed every 0.5 mm). These reconstructions clearly demonstrate focal strictures (arrows), seen to better advantage using CT bronchography.

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

There are several clinical settings, either recently identified or well recognized, that are associated with bronchiectasis. These include a diffuse and accelerated form of bronchiectasis causing significant morbidity in HIV-positive and AIDS patients; atypical mycobacterial infections, especially in immunocompetent older women; CF; and ABPA. AIDS-Associated Airway Disease Lower respiratory tract infections, including bacterial pneumonia and bronchitis, have superseded Pneumocystis carinii pneumonia as the most common infection in the lungs of AIDS patients.29, 54 In particular, an increase in the incidence of infectious bronchitis, bronchiolitis, and bronchiectasis in AIDS has recently been recognized. As documented by the Pulmonary Complications of HIV Infection Study Group the predominant lower respiratory infection in over 1000 cohort members entered in the study with CD4 levels above 200 cells/mm3 was acute bacterial bronchitis.53 An accelerated, aggressive form of bronchiectasis in AIDS has recently been recognized.16, 30, 32, 52 Many times this is preceded by pneumonia or bronchitis, often caused by Streptococcus, Haemophilus, and Pseudomonas. In fact, a wide range of infectious agents has been described affecting the airways including mycobacterial and fungal infections. 32 Bronchiectasis may also be identified in the absence of documented prior infection, suggesting that the cause of bronchiectasis in this setting is multifactorial, and may partially reflect direct deleterious effects of HIV infection on the pulmonary immune system. Recently, a correlation has been shown between airway dilatation identified by CT and the presence of elevated levels of neutrophils on bronchoalveolar lavage. In a study comparing bronchoalveolar lavage findings in 50 HIVpositive subjects with 11 HIV-negative individuals, King et al20 showed that patients with bronchial dilatation on CT had significantly higher bronchoalveolar lavage neutrophil counts and significantly lower diffusing capacity. As noted by these authors, neutrophils are an important mediator of pulmonary damage, possibly because of the action of human neutrophil elastase. Bronchiectasis in AIDS has recently been associated with obstructive disease characterized by rapidly worsening airway obstruction

11

affecting both large- and small-caliber bronchi.1 With longer life expectancies consequent to protease inhibitor therapy, and effective prophylactic regimes for opportunistic infection, these airway changes may be increasingly significant causes of pulmonary morbidity in HIV and AIDS. High-resolution CT demonstrates multilobar bronchiectasis, which most often symmetrically involves the lower lobes. Often it is appreciated concurrently with typical HRCT signs of bronchitis-bronchiolitis (bronchial wall thickening, with and without mucus impaction, or a typical tree-in-bud pattern) (Fig. 8). Rapid development of bronchiectasis after a single episode of pneumonia or bronchitis may represent a distinctive feature of bronchiectasis in the HIV and AIDS patient (Fig. 9). Atypical Mycobacterial Infection Recent attention has focused on CT findings in patients with nontuberculous mycobacterial infections. Classic MAC infection occurs in older men with underlying lung disease, such as chronic obstructive pulmonary disease or interstitial disease. The radiographic pattern may be indistinguishable from postprimary tuberculosis, characterized by nodules, consolidation, and cavities, most often affecting the upper lobes and superior segments of the lower lobes. A new clinicopathologic subtype, termed nonclassic MAC, is seen in women 80% of the time, usually in their seventh to eighth decades. 13, 17, 50 Most of these patients are white (86%) and have no other predisposing risk factors.42 This pattern, seen in approximately 20% to 30% of MAC infections in the nonimmunocompromised population,35 is characterized by bronchiectasis and nodular densities, particularly affecting the middle lobe and lingua (Fig. 10).13, 36 These nodules have a distinct appearance, identifiable as clustered around peripheral vessels and airways, and frequently having a tree-in-bud configuration. Histologically this pattern is caused by either impaction of peripheral bronchioles or peribronchiolar granulomata.50 The finding of granulomas is especially significant in this population, because this constitutes evidence that bronchiectasis, instead of being a precursor, is likely the result of chronic infection.36, 51 Like the classic form of the disease, symptoms are insidious, most frequently a chronic cough.

12

MCGUINNESS & NAIDICH

Figure 8. Infectious small airway disease and AIDS. Diffuse bacterial bronchitis, bronchiolitis, and bronchiectasis in a 26-year-old male patient with AIDS. He had no history of prior opportunistic lung infection and presented with fever, dyspnea, and a productive cough. Cystic bronchiectasis is evident on this 1-mm-thick section in the superior segment of the left lower lobe (black arrow) with less severe bronchiectasis present in several other lobes (white arrows). Multilobar involvement is typical of infectious small airway disease in AIDS. Bronchiolar impaction peripherally is appreciated as clustered centrilobular small densities in a characteristic ‘‘treein-bud’’ configuration (arrowheads). Sputum culture was positive for mixed pyogenic organisms; symptoms resolved after antibiotic therapy.

The association of atypical, nontuberculous mycobacterial infections with bronchiectasis is well known. Hartman et al13 examined 62 patients with MAC infection by CT. Bronchiectasis was identified in 65% overall. Of particular interest was the strong association between bronchiectasis and MAC infections in elderly women in the absence of underlying malignancy or immune compromise. Most striking, compared with classic MAC infection in men with underlying chronic obstructive pulmonary disease or immune compromise in which bronchiectasis typically can be identified in approximately 20% of cases, Hartman et al13 found that the prevalence of bronchiectasis in their population of elderly women neared 100%. The high prevalence of bronchiectasis in elderly women with atypical mycobacterial infection is substantiated by Moore,36 who in addition showed in sequential CT studies that airway disease in these patients is frequently progressive. She suggests that bronchiectasis may not be a predisposing condition but rather a consequence of atypical mycobacterial infection, a theory supported by more recent work.39 Recently, it has been suggested that CT may be of further value in differentiating between

typical (e.g., Mycobacterium tuberculosis) and atypical (e.g., MAC) mycobacterial infection. Primack et al41 compared HRCT findings in tuberculosis and MAC infection in 77 immunocompetent patients and found that bronchiectasis was significantly more common and more extensive in patients with MAC, identifiable in 94% of the 32 patients compared with 27% of the 45 patients with typical reactivation tuberculosis. Furthermore, the distribution of bronchiectasis varied between the two groups, primarily involving the upper lobes in patients with M. tuberculosis, compared with patients with MAC in whom lobar predominance could not be identified (predictably, because this series included both classic and nonclassic patient subgroups). Based on these data, these investigators suggested that identification of characteristic patterns of bronchiectasis in select cases might allow presumptive diagnosis of atypical mycobacterial infection in smear-positive patients before definitive culture. The clinical use of these observations remains to be established. In the authors’ experience, it is not unusual for the initial diagnosis of MAC to be suggested first by cross-sectional imaging. In

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

Figure 9. Postinfectious bronchiectasis in AIDS. A, High-resolution 1-mm-thick section through the lower lungs in a 28-year-old woman with AIDS and bacterial airway infection. Bronchial wall thickening is apparent (open arrow), whereas peripheral centrilobular densities represent impacted bronchi with peribronchial inflammation (black arrows). Branching V- and Y-shaped peripheral densities are characteristic of impacted bronchioles (arrowheads). B, After antibiotic therapy, a high-resolution scan obtained at the same level 1 month later reveals the rapid development of bronchiectasis (arrows) concomitant with regression in inflammatory changes. This patient had had Pneumocystis carinii pneumonia (PCP) in the past, but this was the first documented bacterial infection.

13

14

MCGUINNESS & NAIDICH

Figure 10. Bronchiectasis and Mycobacterium avium complex (MAC) infection. A, HRCT at the level of the middle lobe bronchus in a 77-year-old woman with a chronic cough and multiple sputum cultures positive for MAC demonstrates air-filled (open arrows) and impacted (black arrows) ectatic bronchi in the right middle and lower lobe and lingula. Note the absence of consolidation or cavities. Predilection for middle lobe and lingula disease is characteristic of the ‘‘nonclassic’’ pattern of this infection. B, HRCT at the same level in this patient 16 months later, after multidrug antibiotic therapy, demonstrates regression in bronchial wall thickening, peribronchiolar inflammation, and impaction. Impacted bronchi remain in several areas. Note markedly worsened bronchiectasis in the lingula despite regression in inflammatory changes (white arrows).

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

part, this reflects a change in the epidemiologic patterns of disease, including recognition of atypical MAC infection in immunocompetent older women. Swensen et al, 50 using a combination of HRCT findings of bronchiectasis and small nodules, was able to predict cultures positive for MAC with a sensitivity of 80%, a specificity of 87%, and an accuracy of 86% in 63 patients with HRCTdiagnosed bronchiectasis in whom mycobacterial cultures were sent. Tanaka et al51 evaluated the usefulness of CT and bronchoscopy in establishing the diagnosis of MAC. Twenty-six patients with a CT pattern of clustered small nodules and bronchiectasis had MAC culture results from expectorated sputum compared with culture results from bronchoalveolar lavage fluid. Fifty percent of these patients suspected of having MAC pulmonary disease based on CT findings were culture-positive from the lavage fluid, versus a 23% positive yield from sputum. A characteristic CT pattern seems to be clinically useful in suggesting the diagnosis of MAC, which may prompt sputum evaluation for this infection. Additionally, the results of Tanaka et al51 and others17 suggest that recognition of this disease pattern may be useful in selecting patients for bronchoscopy, if clinical suspicion is high. Cystic Fibrosis Cystic fibrosis results from an autosomal recessive genetic defect, which causes abnormal chloride transport across epithelial membranes. In the lungs this leads to abnormally low water content of airway mucus, resulting in decreased mucus clearance, mucus plugging of airways, and an increased incidence of bacterial airway infection. Bronchial wall inflammation and consequent bronchiectasis are inevitable in patients with long-standing disease. CF is the most common cause of pulmonary insufficiency in the first three decades of life.58 The predominant HRCT finding in CF is bronchiectasis, which is typically panlobar (Fig. 11).14, 15 Proximal and parahilar bronchiectasis are identified in all of these cases; in about a third of cases it is limited to these areas, a pattern termed central bronchiectasis (see Fig. 11A).3, 10 Bronchial wall thickening and peribronchial interstitial thickening are also common, and may be seen in the absence of bronchiectasis in early disease.

15

Mucus plugging is reported in between a quarter and one half of cases.14 Branching or nodular centrilobular opacities, which reflect the presence of bronchiolar dilatation with associated mucus impaction, infection, or peribronchiolar inflammation, can be an early sign of disease.27 Mosaic perfusion may then also be identified, either surrounding mucusplugged airways or as a separate finding in as many as two thirds of patients, best demonstrated on expiratory scans.10, 49 The increased sensitivity of HRCT compared with radiography in detecting lung disease in patients with CF has been well described.3, 10, 27 Recent reports propose CT as a reliable alternative to routine radiographic and clinical methods for monitoring disease status and progression, and assessing response to treatment.3, 14, 45 Shah et al,45 using a modification of the scoring system proposed by Bhalla et al,3 reported findings in 19 symptomatic patients with adult CF at initial evaluation and following 2 weeks of therapy, compared with a control group of eight asymptomatic CF patients. Reversible findings included air–fluid levels in bronchiectatic cavities, centrilobular nodules, mucus plugging, and peribronchial thickening. Although the severity of bronchiectasis in this study correlated with forced vital capacity and forced expiratory volume in 1 second, there was no correlation between PFT parameters and either mucus plugging or centrilobular nodules. This suggests that PFTs were less sensitive than CT in identifying potentially reversible and treatable disease.45 These findings and related studies 15 lend support to the notion that HRCT should be incorporated into follow-up regimens of patients with CF. Allergic Bronchopulmonary Aspergillosis Occurring exclusively in asthmatics, ABPA represents a type I and III (IgE and IgG mediated) hypersensitivity reaction to the endobronchial growth of Aspergillus species. Characteristically seen in asthmatics with eosinophilia, and often associated with steroid therapy for asthma, the inflammation that results from the immune reaction to the presence of the fungus causes the consequent development of central bronchiectasis. The range of CT findings in patients with ABPA are well described (Fig. 12).23, 40, 44 The

16

MCGUINNESS & NAIDICH

Figure 11. Cystic fibrosis. A, Pan lobar bronchiectasis is identified on this 1-mm-thick section through the midlungs of a 19year-old male patient with cystic fibrosis. Note the predominately central distribution of the bronchiectasis (white arrows); in approximately one third of cases, bronchiectasis will be limited to these areas. Mosaic regions of hyperlucent lung likely reflect decreased perfusion locally (black arrows), a manifestation of bronchiolar impaction. Bronchiolar impaction is recognized in the left lower lobe (curved arrows). B, More advanced disease is evident in this 25-year-old man with cystic fibrosis. HRCT through the lower lungs depicts pan lobar bronchiectasis involving both central and peripheral lungs. Bronchial wall thickening is evident (white arrows), as is bronchial (open arrow) and bronchiolar (black arrows) impaction.

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS

17

Figure 12. Allergic bronchopulmonary aspergillosis (ABPA). A, High-resolution 1-mm-thick section through the lower lungs in this 58-year-old man with ABPA demonstrates the characteristic distribution of central bronchiectasis seen in this disease. Varicose bronchiectasis is well demonstrated in the left lower lobe (arrow). Three-mm-thick sections through the upper lungs in this same patient at a later date with wide (B) and narrow (C) windows. Mucus impaction is evident in dilated right upper lobe bronchi (arrows and asterisk in B). On the corresponding narrow windows at the same level, a large impacted bronchus is of increased density on this noncontrast examination (compare arrow in C with asterisk in B). This is a characteristic finding in ABPA and presumably reflects precipitation of calcium or metallic ions within viscous mucus in chronically impacted bronchi.

characteristic finding of central or proximal bronchiectasis can be identified in nearly 100% of cases. In their study of 23 patients with ABPA Panchal et al40 identified central bronchiectasis in 85% of lobes and 52% of lung segments using 4- and 8-mm-thick sections. Bronchiectasis is usually varicose or cystic in appearance (Fig. 12), and the frequent formation of mucus plugs containing fungus and inflammatory cells results in a characteristic pattern of mucoid impaction, atelectasis, or consolidation. Bronchial wall thickening is frequent, and air–fluid levels may be detected in dilated, cystic airways. Ancillary findings include evidence of peripheral airways disease, with mucus impaction in bronchioles resulting in a tree-inbud pattern, or mosaic attenuation because of bronchiolar obstruction with resulting air

trapping. 57 High CT attenuation numbers have been measured in the central impacted mucus (Fig. 12B and C). Presumably representing the presence of calcium or metallic ions within viscous mucus, the prevalence of this finding has been noted to be as high as 28% in one series,23 and when present should be considered characteristic. Although classically associated with central bronchiectasis this finding is neither sensitive nor specific for ABPA. Reiff et al,43 in a study of 168 patients with chronic sputum production, showed that patients with ABPA were significantly more likely to have central bronchiectasis (P ⬍ 0.005); however, the sensitivity of finding central bronchiectasis in this same study proved to be only 37%. Similarly, in a retrospective study of 82 consecutive patients with bronchiectasis of known etiologies, Car-

18

MCGUINNESS & NAIDICH

tier et al4 found that in only five (56%) of nine cases with the disease was the diagnosis of ABPA specifically suggested. Although the clinical features of ABPA significantly overlap those of asthmatics without complicating ABPA, including eosinophilia, bronchiectasis, serum precipitins, and positive skin tests for Aspergillus fumigatus in most cases, HRCT is accurate in establishing this diagnosis. Ward et al 55 retrospectively assessed the accuracy of CT in the diagnosis of ABPA in asthmatic patients. Comparing the CT findings in 44 patients with ABPA with 36 asthmatic controls, these authors noted a clear distinction in the frequency of a number of CT findings in patients with ABPA, including bronchiectasis in 95% of cases versus 29% of asthmatics, centrilobular nodules in 93% of cases versus 28% of asthmatics, and mucoid impaction in 67% of cases versus 4% of asthmatics. Additionally, patients with ABPA consistently had more severe and extensive disease compared with asthmatics. SUMMARY High-resolution CT is accepted as an accurate noninvasive means of diagnosing bronchiectasis. A wide spectrum of abnormalities may be identified at HRCT in patients with airway disease, including various distinctive patterns of bronchiectasis in specific clinical settings, such as ABPA, MAC infection, AIDS, and CF. Characteristic CT findings occasionally suggest a specific diagnosis that may not have been under clinical consideration. HRCT also provides significant clinical use in assessing the degree and extent of airway disease, and allows noninvasive monitoring of disease progression, regression, or response to therapy. References 1. Bard M, Couderc LJ, Saimot AG, et al: Accelerated obstructive pulmonary disease in HIV infected patients with bronchiectasis. Eur Respir J 11:771-775, 1998 2. Barker AF, Bardana EJ: Bronchiectasis: Update on an orphan disease. Am Rev Respir Dis 137:969-978, 1988 3. Bhalla M, Turcios N, Apointe V, et al: Cystic fibrosis: Scoring system with thin-section CT. Radiology 179:783-788, 1991 4. Cartier Y, Kavanagh PV, Johkoh T, et al: Bronchiectasis: Accuracy of high-resolution CT in the differentiation of specific disease. AJR Am J Roentgenol 173:47– 52, 1999

5. Cooke JC, Currie DC, Morgan MP: Role of computed tomography in the diagnosis of bronchiectasis. Thorax 42:272-277, 1987 6. Diederich S, Jurriaans E, Flower CDR: Interobserver variation in the diagnosis of bronchiectasis on highresolution computed tomography. Eur Radiol 6:801806, 1996 7. Engeler CE, Tashjian JH, Engeler CM, et al: Volumetric high-resolution CT in the diagnosis of interstitial lung disease and bronchiectasis: Diagnostic accuracy and radiation dose. AJR Am J Roentgenol 163:3135, 1994 8. Grenier P, Maurice R, Musset D, et al: Bronchiectasis: Assessment by thin-section CT. Radiology 161:95-99, 1986 9. Hansell DM: Bronchiectasis. Radiol Clin North Am 36:107-128, 1998 10. Hansell DM, Strickland B: High-resolution computed tomography in pulmonary cystic fibrosis. Br J Radiol 62:1-5, 1989 11. Hansell DM, Wells AU, Rubens MB, et al: Bronchiectasis: Functional significance of area of decreased attenuation at expiratory CT. Radiology 193:369-374, 1994 12. Hartman TE, Primack SL, Lee KS, et al: CT of bronchial and bronchiolar diseases. Radiographics 14:991– 1003, 1994 13. Hartman TE, Swensen SJ, Williams DE: Mycobacterium avium-intracellulare complex: Evaluation with CT. Radiology 187:23-26, 1993 14. Helbich TH, Heinz-Peer G, Eichler I, et al: Cystic fibrosis: CT assessment of lung involvement in children and adults. Radiology 213:537-544, 1999 15. Helbich TH, Heinz-Peer G, Fleischmann D, et al: Evolution of CT findings in patients with cystic fibrosis. AJR Am J Roentgenol 173:81-88, 1999 16. Holmes AH, Trotman-Dickenson B, Edwards AEA: Bronchiectasis in HIV disease. QJM 85:875-882, 1992 17. Huang JH, Kao PN, Adi V, et al: Mycobacterium avium-intracellulare pulmonary infection in HIV-negative patients without preexisting lung disease: Diagnostic and management limitations. Chest 115:10331040, 1999 18. Kang EY, Miller RR, Muller NL: Bronchiectasis: Comparison of preoperative thin-section CT and pathologic findings in resected specimens. Radiology 195:649-654, 1995 19. Kim JS, Muller NL, Park CS, et al: Bronchoarterial ratio on thin section CT: Comparison between high altitude and sea level. J Comput Assist Tomogr 21:306-311, 1997 20. King MA, Neal DE, Stjohn R, et al: Bronchial dilatation in patients with HIV infection: CT assessment and correlation with pulmonary function tests and findings at bronchoalveolar lavage. AJR Am J Roentgenol 168:1535-1540, 1997 21. Lee PH, Carr DH, Rubens MB, et al: Accuracy of CT in predicting the cause of bronchiectasis. Clin Radiol 50:839-841, 1995 22. Lentz D, Bergin CJ, Berry GJ, et al: Diagnosis of bronchiolitis obliterans in heart-lung transplantation patients: Importance of bronchial dilatation on CT. AJR Am J Roentgenol 159:463–467, 1992 23. Logan PM, Muller NL: High-attenuation mucous plugging in allergic bronchopulmonary aspergillosis. Can Assoc Radiol J 47:374-377, 1996 24. Loubeyre P, Reval D, Delignette A, et al: Bronchiectasis detected with thin-section CT as a predictor of chronic lung allograft rejection. Radiology 194:213– 216, 1995

CT OF AIRWAYS DISEASE AND BRONCHIECTASIS 25. Lucidarme O, Grenier P, Coche E: Bronchiectasis: Comparative assessment with thin-section CT and helical CT. Radiology 200:673-679, 1996 26. Lynch DA: Imaging of small airways disease. Clin Chest Med 14:623-634, 1993 27. Lynch DA, Brasch RC, Hardy KA, et al: Pediatric pulmonary disease: Assessment with high-resolution ultrafast CT. Radiology 76:243-248, 1990 28. Lynch DA, Newell J, Hale V, et al: Correlation of CT findings with clinical evaluations in 261 patients with symptomatic bronchiectasis. AJR Am J Roentgenol 173:53-58, 1999 29. Markowitz GS, Concepcion L, Factor SM, et al: Autopsy patterns of disease among subgroups of an inner-city Bronx AIDS population. J AIDS 13:48-54, 1996 30. McGuinness G, Gruden JF, Bhalla M, et al: AIDSrelated airway disease. AJR Am J Roentgenol 168:6777, 1997 31. McGuinness G, Naidich DP: Bronchiectasis: CT/clinical correlations. Semin Ultrasound CT MR 16:395419, 1995 32. McGuinness G, Naidich DP, Garay SM, et al: AIDS associated bronchiectasis: CT features. J Comput Assist Tomogr 17:260-266, 1993 33. McGuinness G, Naidich DP, Leitman BS, et al: Bronchiectasis: CT evaluation. AJR Am J Roentgenol 160:253-259, 1993 34. Millar A, Boothroyd A, Edwards D, et al: The role of computed tomography (CT) in the investigation of unexplained hemoptysis. Respir Med 86:39-44, 1992 35. Miller WT: Spectrum of pulmonary nontuberculous mycobacterial infection. Radiology 191:343-350, 1994 36. Moore EH: Atypical mycobacterial infection in the lung: CT appearance. Radiology 187:777-782, 1993 37. Munro NC, Cooke JC, Currie DC, et al: Comparison of thin section computed tomography with bronchography for identifying bronchiectatic segments in patients with chronic sputum production. Thorax 45:135-139, 1990 38. Murata K, Khan A, Rojas KA, et al: Optimization of computed tomography technique to demonstrate the fine structure of the lung. Invest Radiol 23:170-175, 1988 39. Obayashi Y, Fujita J, Suemitsu I, et al: Successive follow-up of chest computed tomography in patients with mycobacterium avium-intracellulare complex. Respir Med 93:11-15, 1999 40. Panchal N, Bhagat R, Pant C, et al: Allergic bronchopulmonary aspergillosis: The spectrum of computed tomography appearances. Respir Med 91:213-219, 1997 41. Primack SL, Logan PM, Hartman TE, et al: Pulmonary tuberculosis and mycobacterium avium-intracellulare: A comparison of CT findings. Radiology 194:413-417, 1995 42. Prince DS, Peterson DD, Steiner RM, et al: Infection with Mycobacterium avium complex in patients without predisposing conditions. N Engl J Med 321:863868, 1989

19

43. Reiff DB, Wells AU, Carr DH, et al: CT findings in bronchiectasis: Limited value in distinguishing between idiopathic and specific types. AJR Am J Roentgenol 165:261–267, 1995. 44. Sandhu M, Mukhopadhyay S, Sharma SK: Allergic bronchopulmonary aspergillosis: A comparative evaluation of computed tomography with plain chest radiography. Aust Radiol 38:288-293, 1994 45. Shah RM, Sexauer W, Ostrum BJ, et al: High-resolution CT in the acute exacerbation of cystic fibrosis: Evaluation of acute findings, reversibility of those findings, and clinical correlation. AJR Am J Roentgenol 169:375-380, 1997 46. Silverman PM, Godwin JD: CT/bronchographic correlations in bronchiectasis. J Comput Assist Tomogr 11:52-56, 1987 47. Smith IE, Jurriaans E, Diederich S, et al: Chronic sputum production: Correlation between clinical features and findings on high-resolution computed tomographic scanning of the chest. Thorax 51:914-918, 1996 48. Stanford W, Galvin JR: The diagnosis of bronchiectasis. Clin Chest Med 9:691-699, 1988 49. Stern RC: Current concepts: The diagnosis of cystic fibrosis. N Engl J Med 36:487-491, 1997 50. Swensen SJ, Hartman TE, Williams DE: Computed tomographic diagnosis of Mycobacterium avium-intracellulare complex in patients with bronchiectasis. Chest 105:49-52, 1994 51. Tanaka H, Amitani R, Niimi A, et al: Yield of computed tomography and bronchoscopy for the diagnosis of Mycobacterium avium complex pulmonary disease. Am J Respir Crit Care Med 155:2041-2046, 1997 52. Verghese A, Al-Samman M, Nabhan D, et al: Bacterial bronchitis and bronchiectasis in human immunodeficiency virus infection. Arch Intern Med 154:20862091, 1994 53. Wallace JM, Hansen NI, Lavange L, et al: Respiratory disease trends in the pulmonary complications of HIV infection study cohort. Am J Respir Crit Care Med 155:72-80, 1997 54. Wallace JM, Rao AV, Glassroth J, et al: Respiratory illness in persons with human immunodeficiency virus infection. Am Rev Respir Dis 148:1523-1529, 1993 55. Ward S, Heyneman L, Lee MJ, et al: Accuracy of CT in the diagnosis of allergic bronchopulmonary aspergillosis in asthmatic patients. AJR Am J Roentgenol 173:937-942, 1999 56. Webb WR: High-resolution computed tomography of obstructive lung disease. Radiol Clin North Am 32:745-757, 1994 57. Webb WR, Stein MG, Finkbeiner WE, et al: Normal and diseased isolated lungs: High-resolution CT. Radiology 166:81-87, 1988 58. Wood BP: Cystic fibrosis: 1997. Radiology 204:1-10, 1997 59. Young K, Aspestrand F, Kolbenstvedt: High resolution CT and bronchography in the assessment of bronchiectasis. Acta Radiol 32:439-441, 1991 Address reprint requests to Georgeann McGuinness, MD Department of Radiology New York University Medical Center 560 First Avenue New York, NY 10016

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT OF THE LUNG II

HIGH-RESOLUTION CT OF PERIPHERAL AIRWAYS DISEASES Gayle M. Waitches, DO, and Eric J. Stern, MD

Disorders of the peripheral small airways are ubiquitous and represent a large subgrouping of airway diseases. The term bronchiolitis applies to disorders affecting the small airways and may be caused by a wide spectrum of inflammatory and infectious processes including pulmonary infections, cigarette smoking, connective tissue and autoimmune disorders, toxic fume inhalation, drug toxicity, graft-versus-host disease, and lung transplantation. High-resolution CT (HRCT) plays a major role in the evaluation of diseased bronchioles, often showing characteristic imaging features. This article reviews anatomy, pathology, and the HRCT imaging of diseases of the peripheral airways. ANATOMY AND PHYSIOLOGY The complex function of the airways can be conceptually oversimplified by dividing the bronchial tree into three gross anatomically and functionally based zones: (1) conducting, (2) transitory, and (3) respiratory.10, 28 Originating at the trachea, the conducting zone continues for at least 16 generations, terminating at the terminal bronchioles. The function of the conducting zone as a transporter of oxygen is reflected in its structure; there are no alveoli to allow for gas exchange. Termed the anatomic dead space, airflow velocity is high and airflow volume is low. This

zone has the longest length but the smallest overall volume of the respiratory tract, with an estimated volume of approximately 150 mL. By contrast, the respiratory zone is comprised of the blind-ending alveolar sacs and is solely responsible for gas exchange. Airflow velocity is low and airflow volume is high in this zone to enhance efficient exchange of gases between air and blood.28 The overall volume of the respiratory zone approximates 3 L. The transitory zone reflects an anatomic bridge between the conducting and respiratory zones, providing a combined function of gas transport and exchange as the respiratory bronchioles progressively branch into the alveolar ducts toward the alveolar sacs. Reflecting differences in airflow volume and velocity, the airways also may be divided into central and peripheral airways. The diameter of a central airway generally is defined as larger than 2 to 3 mm, whereas that of a peripheral airway is less than 2 to 3 mm.10, 15 Hogg et al10 showed that despite the decreased radius of a smaller airway, airflow resistance is actually less because of the greater overall number of these peripheral bronchioles. Individual airflow resistance is minimized by the summation effect of some 200,000 respiratory bronchioles, with only approximately 25% of overall airway resistance caused by small airways. Because most airflow resistance occurs at the level of the larger airways, significant disease must be present

From the Department of Radiology, Harborview Medical Center, University of Washington School of Medicine, Seattle, Washington (GMW, EJS); and Adventist Medical Center, Portland, Oregon (GMW)

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

21

22

WAITCHES & STERN

in the peripheral airways to detect pathology on standard pulmonary function tests. This is where imaging can play a role in the diagnosis of small airways diseases. Understanding small airway diseases requires a review of pulmonary lobular anatomy and the HR CT appearance of the normal secondary pulmonary lobule. The secondary pulmonary lobule is a structural lung unit composed of a small number of acini (usually between 3 and 12). Measuring approximately 1 to 2.5 cm in diameter, the secondary pulmonary lobule is polyhedral in shape and is bordered by interlobular septa.15, 17 The central component of the lobule contains the pulmonary arteriole and terminal bronchiole. The pulmonary veins and lymphatic tissue are located along the periphery of the lobule within the interlobular septa (Fig. 1). The bronchioles are centrilobular structures located in the center of the secondary pulmonary lobule. Visualization of pulmonary lobular anatomy is determined by the size of the structures being imaged and the resolving capacity of the imaging modality. Because the fourthgeneration segmental bronchi are the smallest units that can be seen with chest radiography, this modality plays a limited role in the evaluation of small airway diseases.6, 27 HRCT of-

fers greatly improved resolution over chest radiography and conventional CT, detecting pulmonary lobular structures as small as 0.2 mm.15, 17, 25, 26 Bronchiolar visibility is determined primarily by wall thickness. The normal lobular terminal bronchiole wall measuring approximately 0.1 mm is beyond the resolving capabilities of HRCT.15, 17 Because the vasculature is not air filled, intralobular arteries measuring 0.2 mm in diameter are routinely seen with HRCT.17 Peripheral veins measuring 0.5 mm also are usually visible, although the peripheral interlobular septa measuring 0.1 mm in thickness are usually too small to be resolved with HRCT. Although an occasional peripheral septal line is seen on HRCT, the visualization of numerous clearly defined interlobular septa is abnormal.25, 26 The only structures of the secondary pulmonary lobule that can be seen routinely with HRCT are the central intralobular arteries and peripheral veins (Fig. 1). HIGH-RESOLUTION CT SCAN MANIFESTATIONS OF SMALL AIRWAYS DISEASE There are direct and indirect radiographic manifestations of peripheral airway diseases.

Figure 1. Normal high-resolution CT (HRCT) anatomy. Inspiratory HRCT scan through the upper lungs from a healthy patient shows the normal appearance of the lung architecture. Note that the secondary pulmonary lobules are not well defined except for the normal centrilobular arteries (arrows). The centrilobular airways and interlobular septa normally are not visualized.

HIGH-RESOLUTION CT OF PERIPHERAL AIRWAYS DISEASES

Direct manifestations are uncommon because most airways of the tracheobronchial tree are radiologically invisible. Those diseased bronchioles that are visible with HRCT appear as dilated air-filled, branching, ring-like, or tubular structures in the lung periphery.23 Enhanced visualization of the bronchioles in this location is caused by wall thickening and dilatation. When the airway is obliterated by intraluminal granulation tissue or submucosal or peribronchiolar fibrosis, nodular, linear, or branching peripheral opacities are also seen.23 Additional CT scan signs of small airways disease are indirect, reflecting a response of the subtended lung parenchyma to the airway insult. Subsegmental atelectasis occurring distal to an obstructed airway can produce a focal, regional, or wedge-shaped area of ground-glass opacity.23 Air-trapping or hyperinflation is another common manifestation of occlusion or stenosis of small airways.21, 23 Persistent aeration caused by collateral pathways, or hyperaeration from inability to expire air, produces the mosaic pattern of lung attenuation representative of air trapping. This air trapping is often only seen, or is more pronounced, on scans obtained at end-exhalation instead of the more conventional end-inspiration technique. Expiratory CT scanning provides both anatomic and physiologic information that is complementary to conventional suspended full-inspiration CT and pulmonary function testing. The extent and distribution of air trapping can be useful in indicating or directing further diagnostic work-up, such as transbronchial, thoracoscopic, or open lung biopsy. The presence of small peripheral centrilobular nodular parenchymal densities reflects another nonspecific indirect sign of small airways disease. These hazy nodular opacities are the peribronchiolar inflammatory sequela of diseased airways. These nodules appear as focal rounded areas of increased ground-glass attenuation measuring less than 1 cm in size.23 One or more of these direct and indirect signs of airways insult can be variably found in the CT scan evaluation of peripheral airway diseases.

23

chest radiograph can be normal, but nonspecific findings include hyperinflation, hypoinflation with multiple focal areas of atelectasis distal to obstructed airways, pneumomediastinum, interstitial pulmonary emphysema, and complicating pneumonia. Obstruction of small bronchi and bronchioles by inflammatory exudate and bronchiolar wall thickening from edema and smooth muscle hyperplasia produce the HRCT scan features of atelectasis, air trapping, centrilobular nodules, and bronchial wall thickening.18, 29 If air trapping is diffuse, it may be difficult to recognize and escape detection on CT scans obtained at full inspiration.18 Expiratory HRCT scanning offers greater sensitivity in the detection of asthmatic air trapping, the severity of which has been shown to correlate with the severity of asthma (Fig. 2).18, 29 The remainder of this article discusses the specific causes of bronchiolitis, using a classifying scheme based on radiologic features reported by Worthy and Mu¨ ller. 29 In this scheme, inflammation of the bronchioles is divided into five main groups: (1) bronchiolitis obliterans with intraluminal proliferation of granulation tissue polyps, (2), constrictive bronchiolitis, (3) panbronchiolitis, (4) cellular bronchiolitis, and (5) respiratory bronchiolitis (RB). BRONCHIOLITIS OBLITERANS WITH INTRALUMINAL PROLIFERATION OF GRANULATION TISSUE POLYPS Also termed proliferative bronchiolitis,5 this disorder is characterized by the presence of inflammatory or infectious exudative material within the lumen of the bronchioles and alveolar ducts. In most cases there is accompanying organizing pneumonia within the subtended distal components of the lobule (bronchiolitis obliterans with organizing pneumonia [BOOP]). Because the clinical, 4 physiologic, and imaging13, 15, 16, 29 features of this conditions are inconsistent with other forms of small airways disease, the CT features of this entity are discussed elsewhere in this issue.

ASTHMA CONSTRICTIVE BRONCHIOLITIS A reversible form of small airways obstruction, asthma affects both small and large airways. Asthma is very common, occurring in up to 5% of adults and 10% of children. The

Also known as obliterative bronchiolitis, constrictive bronchiolitis occurs when the bronchioles become concentrically obstructed be-

24

WAITCHES & STERN

Figure 2. Asthma. Paired inspiratory and expiratory HRCT scans through the upper lungs from a patient with asthma show a normal inspiratory scan and the typical heterogeneous pattern of lung attenuation of a small airways disease with patchy, multifocal air-trapping on the expiratory scan obtained at the same level.

cause of peribronchiolar or submucosal fibrous proliferation, with little accompanying active inflammatory granulation tissue.5, 7, 15, 29 Unlike the restrictive features of BOOP, pulmonary function tests usually show airflow obstruction, or mixed restriction and obstruction. Usually irreversible, the disease severity depends on the extent of bronchiolar involvement and airflow obstruction. 5, 7, 15, 29 Causes of constrictive bronchiolitis include previous childhood infections (particularly measles, adenovirus, and Mycoplasma); toxic fume inhalation; graftversus-host disease following bone marrow

transplantation; chronic rejection following lung or heart-lung transplantation; and autoimmune connective tissue disorders, such as rheumatoid arthritis treated with penicillamine therapy, and polymyositis.5, 7, 15, 20, 29 Constrictive bronchiolitis also occurs as a complication of the chronic pulmonary infections of cystic fibrosis, and has been reported rarely in association with inflammatory bowel disease.19, 29 Although unusual, this disease also can be idiopathic, occurring mainly in middle-aged women.5 Clinically, dyspnea is the most common complaint. The clinical criteria used for diagnosing constrictive bronchiolitis

HIGH-RESOLUTION CT OF PERIPHERAL AIRWAYS DISEASES

Figure 3. Common variable immunodeficiency syndrome. Coned-down inspiration CT scan through the right midlung from a patient with common variable immunodeficiency syndrome shows a combination of differing small and large airway findings including multifocal air-trapping most evident in combination with mild bronchiectasis and impacted small airways (tree-in-bud appearance). (Courtesy of Bruce Davidson, MD, Seattle, WA)

include evidence of severe airflow obstruction measured by pulmonary spirometry, with a forced expiratory volume in 1 second of less than 60% of predicted value, in the absence of

25

other pathologic causes of airway obstruction, such as emphysema or chronic bronchitis.24, 29 The HRCT findings of constrictive bronchiolitis are quite distinct from those of BOOP, and reflect the underlying airway obstruction. Hartman9 and Mu¨ller and Miller15 divided the CT features of this disease into direct and indirect signs. Broncholar wall thickening, the only direct evidence of disease, produces centrilobular branching opacities. Central bronchiectasis, mosaic pattern of lung attenuation, and expiratory air trapping are more common indirect HRCT findings of constrictive bronchiolitis (Figs. 3 and 4).5, 9, 15, 29 The mosaic pattern of lung attenuation produces heterogeneous, asymmetric lobular- or multilobularsized regions of hyperattenuated and hypoattenuated lung parenchyma, with associated increased or decreased vessel caliber size, respectively. Regions of decreased attenuation likely result from a combination of air trapping from bronchiolar luminal obliteration, and diminished perfusion caused by reactive hypoxic vasoconstriction within areas of poorly ventilated lung.5, 15, 29 Unlike patients with BOOP, patients with constructive bronchiolitis usually respond poorly to corticosteroid therapy. Swyer-James syndrome results from bronchiolitis obliterans secondary to an infectious insult to the small airways, typically occurring before 8 years of age, before alveolar development is complete. It should be em-

Figure 4. Obliterative bronchiolitis mosaic pattern. Expiratory HRCT scan through the lower lung from a patient with fume-related bronchiolitis obliterans shows the typical heterogeneous pattern of lung attenuation of a small airways disease with patchy, multifocal air-trapping.

26

WAITCHES & STERN

phasized that postinfectious bronchiolitis obliterans can occur at any age; this particular syndrome refers to that which occurs in childhood. It was first described as a radiographic entity by Swyer and James in 1953.22 The classic constellation of radiographic findings include hyperlucent lung, decreased lung volume, small ipsilateral hilum with decreased peripheral pulmonary vascularity, and air trapping. Adenovirus has been implicated as the primary agent. A wide variety of other organisms have been described, however, including measles, Mycobacterium tuberculosis, pertussis, and Mycoplasma pneumonia. Although the diagnosis can often be made by classic findings on inspiratory-expiratory chest radiographs, patients presenting with atypical radiographic or clinical findings can create a more difficult diagnostic dilemma. In the pediatric age group, CT scanning can help differentiate between several diagnostic possibilities including endobronchial foreign bodies, extrinsic compression of a bronchus by a congenital cyst or neoplasm, pulmonary agenesis or hypoplasia, compensatory emphysema, or congenital lobar emphysema. Expiratory HRCT scanning in patients with Swyer-James syndrome can show multifocal bilateral air trapping that is not seen or is underappreciated on conventional radiographs or CT scans. Patients may be entirely asymptomatic with this syndrome, but may also suffer from recurrent respiratory infections and hemoptysis caused by bronchiectasis.15, 29

CELLULAR BRONCHIOLITIS Cellular bronchiolitis describes the process of active inflammation involving the bronchiolar wall or lumen. Acute infectious bronchiolitis is an example of cellular bronchiolitis, including viral, mycoplasma, tuberculosis, or Aspergillus infection. 2, 15, 23 In infancy and childhood, respiratory syncytial virus is a common cause of acute infectious bronchiolitis. Although most episodes resolve without long-term sequelae, more severe cases of pulmonary parenchymal injury can lead to constrictive bronchiolitis (Swyer-James syndrome). Other cellular inflammatory processes that involve the bronchioles include asthma, aspiration pneumonia, chronic bronchitis, and hypersensitivity pneumonitis. High-resolution CT findings of cellular bronchiolitis include linear or nodular peripheral opacities, the so-called tree-in bud (TIB) appearance.3, 15, 23, 29 The linear and nodular opacities reflect the thickened, impacted, and distended bronchioles. Bronchiolitis and bronchiolectasis are nonspecific inflammatory processes of the small airways caused by many different insults; they are not specific diseases. The TIB pattern is a direct CT finding of this nonspecific bronchiolar disease, first used as a descriptor by Im et al12 for the CT findings of endobronchial spread of M. tuberculosis (Fig. 5). This pattern is analogous to the larger airway finger-in-glove appearance of bronchial impaction, but on a much smaller scale. The TIB pattern has become a

Figure 5. The ‘‘tree-in-bud’’ pattern. Expiratory CT scan through the midlung from this patient with active endobronchial tuberculosis shows multiple small, ill-defined bilateral nodules typical of the nonspecific tree-in-bud pattern. In this case, note the large right upper lobe granuloma as evidence of tuberculosis.

HIGH-RESOLUTION CT OF PERIPHERAL AIRWAYS DISEASES

popular descriptive term for many bronchiolar disease processes, all with similar appearances,31 although it is still often used inappropriately to imply a pathognomonic finding for tuberculosis. The list of diseases associated with the bronchioles potentially producing a TIB pattern at CT scanning is extensive. The more common disease processes can be grouped as follows: (1) infection, (2) immunologic disorders, (3) congenital disorders, (4) aspiration, and (5) idiopathic. 3 Indirect CT signs of bronchiolar disease include air trapping, especially with expiratory CT scanning, and subsegmental atelectasis. Nonspecific patchy regions of ground-glass attenuation or consolidation can be seen in cellular bronchiolitis caused by infections, and likely reflect accompanying bronchopneumonia.15

PANBRONCHIOLITIS Rare in North America and Europe, panbronchiolitis, also called diffuse panbronchiolitis, is an idiopathic inflammatory lung disease that is prevalent in the Asian population and most common in Japanese men. The disease is characterized by mononuclear cellular inflammation of the respiratory bronchioles. Patients present with nonspecific progressive dyspnea and productive cough. HRCT find-

27

ings include branching linear and nodular opacities, bronchiolectasis, bronchiectasis, mosaic perfusion, and air trapping (Fig. 6).1, 2, 9, 15, 29 The HRCT appearance can be very similar to that seen with dysmotile cilia syndromes or other causes of extensive bronchiolectasis. Although erythromycin therapy can be effective in some patients, overall longterm prognosis is poor.15

RESPIRATORY BRONCHIOLITIS AND RESPIRATORY BRONCHIOLITIS–INTERSTITIAL LUNG DISEASE Respiratory bronchiolitis, also known as smoker’s bronchiolitis, is present histologically in most smokers. 8 Characterized by mild chronic inflammation of the respiratory bronchioles and accumulation of pigmented macrophages within the respiratory bronchioles and adjacent alveoli, this disorder rarely is symptomatic and usually produces no chest radiographic abnormalities.8, 15, 29 Remy-Jardin et al 19 correlated the HRCT findings in asymptomatic heavy smokers with pathologic specimens from resected lung nodules. Although only 11 of 41 patients had abnormal HRCT findings (the most frequent finding being scattered areas of ground-glass attenua-

Figure 6. The ‘‘tree-in-bud’’ pattern. Inspiratory HRCT scan through the midlung from this patient with diffuse panbronchiolitis shows innumerable, small, illdefined nodules and branching structures throughout the lung, typical of the nonspecific tree-in-bud pattern of peripheral airways disease.

28

WAITCHES & STERN

tion and micronodular densities [Fig. 7]), 39 of these patients had histologic evidence of peribronchiolar and alveolar inflammation and pigmentation. They reported a low sensitivity and high specificity of HRCT scans in the detection of RB. Respiratory bronchiolitis–interstitial lung disease (RB-ILD) is a term usually used to describe patients who are symptomatic from respiratory bronchiolitis.8, 15, 29 These patients usually are heavy smokers with a long history of tobacco exposure. Patients with RB-ILD usually present with chronic cough, dyspnea, and restrictive lung function. HRCT scan findings in five patients with biopsy-proved RB-ILD described by Holt et al 11 include ground-glass or linear areas of increased attenuation, broad areas of atelectasis, emphysema, and linear and reticular septal thickening. One patient had normal findings. Another study by Yousem et al30 comparing the HRCT findings of RB-ILD with those of desquamative interstitial pneumonia showed that there are marked similarities between these two disorders. The distinguishing features of RB-ILD included less frequent areas of ground-glass attenuation, less severe symptomatology, and resolution of symptoms on cessation of tobacco exposure. The similarities of the two diseases suggest that they

may be related, with desquamative interstitial pneumonia representing a late or progressive form of RB-ILD.11, 30 Because of the nonspecific appearance of RB-ILD, diagnosis usually requires open lung biopsy. This disorder, however, should be suspected in symptomatic cigarette smokers with increased interstitial markings and restrictive lung function.

PITFALLS There are pitfalls in the HRCT diagnosis of bronchiolar disease. The findings of thickened lobular septa and adjacent peribronchial nodules present in various granulomatous processes can mimic the TIB pattern of bronchiolar disease on HRCT.3 These granulomatous diseases include sarcoidosis, silicosis, and Langerhans’ cell granulomatosis.3 Features that distinguish sarcoidosis from bronchiolitis include lymphadenopathy and perivenule nodularity.3 The lymphatic and hematogenous spread of pulmonary metastases produces a reticulonodular pattern that can also mimic bronchiolar disease. The lack of confluence of the nodular and linear opacities in metastatic disease allows for distinction.3

Figure 7. Respiratory bronchiolitis. Inspiratory HRCT scan through the upper lung from this heavy cigarette smoker shows innumerable, ill-defined, fuzzy nodules around all of the normal pulmonary vascular structures, typical for respiratory bronchiolitis. This abnormality usually is seen predominantly in the upper lungs. These nodules are more subtle than the tree-in-bud nodules.

HIGH-RESOLUTION CT OF PERIPHERAL AIRWAYS DISEASES

SUMMARY Bronchiolitis and bronchiolectasis are nonspecific inflammatory processes of the small airways that have a variable, but often characteristic, appearance on HRCT. Familiarity with the imaging features of these disorders is crucial in rendering an accurate radiographic diagnosis.

13. 14.

15. 16.

References 1. Akira M, Kitatani F, Yong-Sik I: Diffuse panbronchiolitis: Evaluation with HRCT. Radiology 168:433–438, 1988 2. Aquino SI, Gamsu G, Webb WR, et al: Tree-in-bud pattern: Frequency and significance on thin section CT. J Comput Assist Tomogr 20:594–599, 1996 3. Collins J, Blankenbaker, D, Stern EK: CT patterns of bronchiolar disease: What is ‘‘tree-in-bud’’? AJR Am J Roentgenol 171:365–370, 1998 4. Epler GR, Colby TV, McLoud TC, et al: Bronchiolitis obliterans organizing pneumonia. N Engl J Med 312:152–158, 1985 5. Garg K, Newell, JD, King Jr TE: Proliferative and constrictive bronchiolitis: Classification and radiologic features. AJR Am J Roentgenol 162:803–808, 1994 6. Gurney JW: Airways disease. In Greene R (ed): Syllabus: A Categorical Course in Diagnostic RadiologyChest Radiology. Oak Brook, IL, RSNA Publications, 1992, pp 99–112 7. Hansell DM, Rubens MB, Padley SP, et al: Obliterative bronchiolitis: Individual CT signs of small airways disease and functional correlation. Radiology 203:721–726, 1997 8. Hartman TE, Tazelaar HD, Swenson SJ, et al: Cigarette smoking: CT and pathologic findings of associated pulmonary diseases. Radiographics 17:377–390, 1997 9. Hartman TE, Primack SL, Lee KS, et al: CT of bronchial and bronchiolar diseases. Radiographics 14:991– 1003, 1994 10. Hogg JC, Macklem PT, Thurlbeck WM: Site and nature of airway obstruction in chronic obstructive lung disease. N Engl J Med 268:1355–1360, 1968 11. Holt RM, Schmidt RA, Godwin JD, et al: High resolution CT in respiratory bronchiolitis-associated interstitial lung disease. J Comput Assist Tomogr 17:46– 50, 1993 12. Im JG, Itoh H, Shim YS: Pulmonary tuberculosis: CT findings, early active disease and sequential change

17. 18.

19.

20.

21. 22. 23. 24. 25.

26. 27. 28. 29. 30.

29

with antituberculous therapy. Radiology 186:653– 660, 1993 Lee KS, Kullnig P, Hartman TE: COP: CT findings in 43 patients. AJR Am J Roentgenol 162:543–546, 1994 Morrish WF, Herman SJ, Weisbrod GL, et al: Bronchiolitis obliterans after lung transplantation: Findings at chest radiography and high-resolution CT. The Toronto Lung Transplant Group. Radiology 179:487– 490, 1991 Mu¨ller NL, Miller RR: Diseases of the bronchioles: CT and histopathologic findings. Radiology 196:3– 12, 1995 Mu¨ller NL, Staples CA, Miller RR: Bronchiolitis obliterans organizing pneumonia: CT features in 14 patients. AJR Am J Roentgenol 154:983–987, 1990 Murata K, et al: Centrilobular lesions of the lung: Demonstration by high-resolution CT and pathologic correlation. Radiology 161:641–645, 1986 Park CS, Mu¨ller NL, Worthy SA: Airway obstruction in asthmatic and healthy individuals: Inspiratory and expiratory thin-section CT findings. Radiology 203:361–367, 1995 Remy-Jardin M, Rem J, Boulenguez C, et al: Morphologic effects of cigarette smoking on airways and pulmonary parenchyma in healthy volunteers: CT evaluation and correlation with pulmonary function tests. Radiology 186:107–115, 1993 Skeens JI, Fuhrman CR, Yousem SA: Bronchiolitis obliterans in heart-lung transplantation patients: Radiologic findings in 11 patients. AJR Am J Roentgenol 153:253–256, 1989 Stern EJ, Frank MS: Small-airway diseases of the lungs: Findings at expiratory CT. AJR Am J Roentgenol 163:37–41, 1994 Swyer P, James GCW: A case of unilateral pulmonary emphysema. Thorax 8:133–136, 1953 Teel GS, Engeler CE, Tashijian JH, et al: Imaging of small airways disease. Radiographics 16:27–41, 1996 Turton CW, Williams G, Green M: COP in adults. Thorax 36:805–810, 1982 Webb WR: Chronic diffuse infiltrative lung disease: High resolution CT techniques and patterns. In Greene R (ed): Syllabus: A Categorical Course in Diagnostic Radiology-Chest Radiology. Oak Brook, IL, RSNA Publications, 1992, pp 67–76 Webb WR, Stein MG, Finkbeiner WE, et al: Normal and diseased isolated lungs: High resolution CT. Radiology 166:81–87, 1988 Webb WR: Radiology of obstructive pulmonary disease. AJR Am J Roentgenol 169:637–647, 1997 West JB: Respiratory Physiology: The Essentials. Baltimore, Williams and Wilkins, 1974 Worthy SA, Mu¨ller NL: Small airway diseases. Radiol Clin North Am 36:163–173, 1998 Yousem SA, Colby TV, Gaensler EA: Respiratory bronchiolotis-associated interstitial lung disease and its relationship to desquamative interstitial pneumonia. Mayo Clin Proc 64:1373–1380, 1989 Address reprint requests to Eric J. Stern, MD Department of Radiology University of Washington School of Medicine Box 359728 Seattle, WA 98104–9728

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT OF THE LUNG II

CT OF EMPHYSEMA John D. Newell Jr, MD

This article reviews the use of CT in the diagnosis and assessment of emphysema. CT has been used for the detection of emphysema since the first report of the CT findings in emphysema.10 CT provides direct visualization of emphysema and can clearly identify the three major types of emphysema. The advent of lung volume reduction surgery (LVRS) and lung transplant surgery has greatly increased the interest in using CT to detect and assess pulmonary emphysema. New ways to deliver ␣1-anti trypsin (A1A) to patients with ␣ 1 -antitrypsin deficiency (A1AD) and the possibility of retinoic acid therapy as a treatment for emphysema have further increased interest in CT of emphysema. Most cases of emphysema are induced by cigarette smoking. For this reason, patients referred for CT to assess the presence of emphysema may also have other cigarette smoke–induced lung disease including respiratory bronchiolitis; bronchogenic carcinoma; Langerhans, histiocytosis (eosinophilic granuloma); and desquamative interstitial pneumonitis (DIP). The reader is directed to other articles in this issue for the CT manifestations of other smoking-induced lung diseases. The radiologist also should seek evidence of atherosclerotic vascular disease and tumors in the lungs, esophagus, stomach, and cervical area in patients who have smoking histories. This article is divided into three major sections. The first section is a review of important medical background information on

emphysema. The second section covers the qualitative CT imaging features of the three major types of emphysema. The final section looks at the quantitative CT assessment of emphysema for the evaluation of patients for potential LVRS. This final section looks at the emerging role of quantitative CT in the evaluation of new drug therapies for the treatment of emphysema, including A1A replacement therapy in A1AD patients and retinoic acid therapy for the treatment of emphysema. EMPHYSEMA Definition Emphysema is defined by the American Thoracic Society as ‘‘permanent enlargement of the air spaces distal to the terminal bronchiole, accompanied by the destruction of their walls, and without obvious fibrosis.’’23 Emphysema, chronic bronchitis, and bronchiolitis comprise the clinical syndrome of chronic obstructive pulmonary disease. CT is helpful in trying to determine how much of the chronic obstructive pulmonary disease is secondary to emphysema. It can be quite difficult to exclude emphysema in a patient with chronic obstructive pulmonary disease by clinical and physiologic measures. Epidemiology Pulmonary emphysema is more common in polluted and industrialized countries and

From the Department of Radiology, University of Colorado Health Sciences Center; and Lung Imaging Center, National Jewish Medical and Research Center, Denver, Colorado

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

31

32

NEWELL

is quite common in autopsies.13 Some degree of emphysema, often mild, has been recorded in autopsies in up to 50% to 70% of patients. The peak prevalence of emphysema is about 70 years of age and emphysema is two to three times more frequent in men than in women.13 The National Health Interview Survey done in the United States indicated that 2 million Americans have emphysema and that 7.5 million Americans have chronic bronchitis.18 The 10-year mortality rate for chronic obstructive pulmonary disease is 50% and the mortality rate has risen 33% between the years 1979 and 1991. 18 Emphysema in the United States is estimated to kill 20,000 Americans per year and is the fourth leading cause of death in the United States.18 Etiology and Pathogenesis The underlying cause of emphysema is believed to be an inbalance in the activities of proteolytic and antiproteolytic enzymes in lung tissue.18, 22 Emphysema is produced as a result of the enzymatic destruction of lung elastin and collagen by neutrophil and macrophage elastases. This model is referred to as the elastase-antielastase hypothesis.22 The primary cause of this imbalance of proteolytic and antiproteolytic enzymes is cigarette smoking.18 It should be noted that only 10% to 15% of smokers develop emphysema.18 Cigarette smoke has been shown to produce an increase in the number of neutrophils in lung tissue and to stimulate the release of neutrophil elastase.18 Cigarette smoke also inhibits A1A and interferes with the repair of damaged elastin and collagen.18 Exposure to coal dust is also a recognized cause of emphysema, even in nonsmokers.18 It is believed that cigarette smoke has a longer clearance time from the upper lobes of the lungs, and this accounts for the increased incidence of emphysema in the upper lobes in smokers.11 Alpha-1 antitrypsin deficiency was first discovered as another cause of emphysema in 1963.18 A1AD is actually a primary liver disease in which the A1A enzyme cannot get out of the hepatocyte because of a point mutation that induces polymerization of the enzyme within the hepatocyte. This leads to a deficiency of the enzyme in plasma and it is not available in sufficient quantities to inactivate the neutrophil elastase in the lungs. The

abundance of activated neutrophil elastase leads to destruction of alveolar walls and a typical lower lobe panacinar emphyema in patients with A1AD. The PiZZ and Pi-phenotypes are usually the A1AD patients who have the lowest levels of enzyme and usually develop the severest disease. Normal individuals have the PiMM phenotype. Cigarette smoking increases the risk of developing emphysema in patients with reduced serum levels of A1A enzyme. The lower lobe distribution of emphysema is thought to be caused by the increased amount of blood flow and resultant increase in neutrophil activity in the lower lobes. The macrophages are concentrated in the centriacinar portion of the lung, in the area of the respiratory bronchiole, which is the area where cigarette smoking–induced emphysema begins. The macrophage and the neutrophil are likely both important in the development of emphysema. The reader is referred to the article by Snider22 for more details on the exact pathogenesis of emphysema. Intravenous drug users who inject oral drug preparations with a significant amount of talc as a binding agent develop upper lobe emphysema.18 Intravenous abuse of methylphenidate (Ritalin) has been shown to produce lower lobe panlobular emphysema.18 Severe centrilobular emphysema has been described in patients with HIV treated for Pneumocystis carinii pneumonia. Pathology Emphysema is defined as abnormal permanent enlargement of the airspaces distal to the terminal bronchial caused by alveolar wall destruction with minimal fibrosis.23 The pulmonary acinus is the respiratory airspaces arising from a single terminal bronchiole.23 The three anatomic types of emphysema are defined based on the portion of the pulmonary acinus that is involved.23 The three types of emphysema are centriacinar, panacinar, and distal acinar. Centriacinar Emphysema Centriacinar emphysema ([CAE] Fig. 1) involves the proximal portion of the acinus in the area of the respiratory bronchiole.23 There are two forms of this subtype of emphysema: (1) centrilobular emphysema and (2) focal emphysema.23 Centrilobular emphysema is

CT OF EMPHYSEMA

33

Figure 1. A, The normal pulmonary acinus. The terminal bronchiole leads into this structure with subsequent respiratory bronchioles, alveolar ducts, and alveolar sacs. B, The pulmonary acinus in a patient with central acinar emphysema. Note the marked dilatation of the respiratory bronchioles. (Courtesy of Salvador Beltran, artist; with permission.)

associated with cigarette smoking and airflow limitation. This lesion typically involves the upper lobes first.23 Focal emphysema is produced by exposure to coal dust and other mineral dusts that result in dilatation of the respiratory bronchiole. 23 There is a large amount of dust-laden macrophages in the respiratory bronchiole of affected patients.23

acinus, which is in close proximity to the interlobular septae and pleura, and also is referred to as paraseptal emphysema or subpleural emphysema. This form of emphysema often is associated with extensive bulla formation with only mild decreases in airflow and also is associated with spontaneous pneumothorax in young adults.23

Panacinar Emphysema

Irregular Emphysema

Panacinar emphysema ([PAE] Fig. 2) involves the pulmonary acinus in a diffuse and uniform way.23 This subtype of emphysema tends to be worse in the lower lobes and may accompany the CAE in cigarette smokers. It should be noted that as the emphysema becomes severe in a given patient, it becomes difficult to distinguish PAE from CAE. There is a common association of A1AD and lower lobe PAE.23

Airspace enlargement (irregular emphysema) may occur in association with tuberculosis, sarcoidosis, and Langerhans’ histiocytosis, diseases that produce a fibrotic response in the lung.23 This form of emphysema is often mixed in nature and may have some features of two or three of the recognized subtypes of emphysema previously noted.

Distal Acinar Emphysema

Clinical Manifestations

Distal acinar emphysema ([DAE] Fig. 3) involves primarily the alveolar ducts and sacs and does not affect the area of the respiratory bronchiole as does CAE.23 This form of emphysema involves the peripheral aspect of the

Patients with emphysema typically have nonproductive cough and progressive exertional dyspnea and relatively normal blood gases.13 Emphysema produces hyperinflation, overactive respiratory muscles, muscle wast-

34

NEWELL

section of the anatomy and there are no overlapping anatomic structures to confuse the observer. The second advantage is that CT scanning has a contrast to noise ratio that is 20 times greater than conventional radiography. The second advantage makes it much easier for CT to detect the subtle decreases in average lung density that are the hallmarks of emphysema on CT. The main observations on CT scans in emphysema are decrease in lung attenuation and decrease in the diameter and number of pulmonary vesels in the area of the emphysema. The technique recommended for the detection of emphysema is to use high-resolution CT (HRCT) with a collimation of 1-mm, high spatial frequency reconstruction algorithm, and a scan interval of 20 mm. The images are viewed with a window level of 600 to 700 H and a window width of 1500 to 1700 H. Figure 2. The pulmonary acinus in a patient with panacinar emphysema. Note the marked dilatation of the proximal respiratory bronchioles and the more distal alveolar ducts and sacs. (Courtesy of Salvador Beltran, artist; with permission.)

ing, high metabolic rate, and decreased PO2, but low to normal PCO2 until late in the disease.18 Pulmonary Function Tests Emphysema typically produces increases in lung volumes, decreases in airflow, and increases in diffusion measurements of carbon monoxide in the lung. There is typically an increase in total lung capacity (TLC), functional residual capacity, and residual volume. The forced expiratory volume in 1 second (FEV1) is decreased and the ratio of FEV1 to the forced vital capacity is reduced. There is an increase in the diffusion measurements of carbon monoxide (DLCO). There is also an increase in lung compliance and a decrease in exercise function.18 CT CT scanning has two distinct advantages over conventional radiography in assessing the presence and extent of emphysema. The first advantage is that the tomographic images provide a true two-dimensional cross

Centriacinar Emphysema Centriacinar emphysema involves the central portion of the pulmonary acinus, and because of this CAE presents on CT as focal areas of decreased attenuation resembling simple cysts in the lung with two noticable differences. The focal centriacinar lesion has no discernable wall and the centriacinar lesions usually have a focal arteriole at or near the center of the CAE lesion (Figs. 4 and 5). These two important features aid in distinguishing the CAE lesion from the focal lesions seen in eosinophilic granuloma and lymphangioleiomyoma and the congential cysts of the lungs. Centriacinar emphysema is usually caused by cigarette smoke and occurs first in the upper lungs. As the CAE process increases in severity the focal lesions that are initially separated by regions of normal lung tissue become confluent, usually first in the upper lobes and progressing inferiorly, so that it may become difficult to distinguish CAE from PAE. If there is evidence of coal workers’ pneumoconiosis or mixed dust exposure from both silica and coal inhalation the presence of CAE may be caused by the coal dust exposure. This has been termed focal emphysema. Exposure to coal dust and the absence of a smoking history should greatly increase the radiologist’s suspicion that the CAE is caused by pneumoconiosis.

CT OF EMPHYSEMA

35

Figure 3. The pulmonary acinus in a patient with distal acinar emphysema. Note the marked dilatation of the alveolar ducts and alveolar sacs. The graphic helps explain the subpleural distribution of distal acinar emphysema. (Courtesy of Salvador Beltran, artist; with permission.)

Panacinar emphysema looks distinctly different from CAE on CT scans of the thorax. PAE involves the entire acinus and presents

as a large area of decreased lung density or decreased attenuation on CT scans of the thorax (Figs. 5 and 6). There is no focality of the PAE lesion, but rather it is quite diffuse with poorly defined lateral margins. The pulmonary arteries and veins are decreased in diameter and number within the PAE lesion, but

Figure 4. High-resolution CT (HRCT) scan using 1 mm collimation through the right upper lobe in a patient with early central acinar emphysema. Note the many small discrete areas of decreased density without any discernible wall. A small central arteriole can be seen in many of the lesions.

Figure 5. HRCT scan using 1 mm collimation through the left upper lobe in a patient with advanced smoking-induced emphysema. Almost all of the lung has been replaced with emphysema and it is difficult to distinguish central acinar emphysema from panacinar emphysema at this point.

Panacinar Emphysema

36

NEWELL

Figure 6. HRCT scan using 1 mm collimation through the right lower lobe in a patient with ␣-1 antitrypsin deficiency (A1AD) and early panacinar emphysema. The panacinar emphysema lesion is seen in the anterior and medial segments of the right lower lobe, posterior to the right major fissure.

some evidence of pulmonary vasculature persists in the lesion. PAE usually develops in the lower lobes. The patient may have a history of A1AD or cigarette smoking. Patients are often relatively young, less than 50 years old. There may be associated CAE lesions in the upper lobes in PAE patients. This reverse distribution of the severity of emphysema in PAE as opposed to CAE (i.e., in PAE the lower lobes are more severely affected than the upper lobes) should help make the diagnosis of PAE rather than CAE. In very advanced cases of CAE and PAE it may be impossible to tell the difference in these two anatomic subtypes of emphysema. The main reason for trying to make a distinction between PAE and CAE is that PAE is often associated with A1AD and therapeutic intervention with A1A enzyme replacement therapy, smoking cessation programs, and patient education are very important. It also may be difficult to distinguish PAE, especially early PAE, from diffuse air trapping caused by constrictive bronchiolitis. A1A enzyme levels and A1AD phenotype determination are helpful in this setting to try to make the distinction between PAE and constrictive bronchiolitis. Expiratory HRCT images to complement the normal inspiratory HRCT may be helpful in distinguishing early PAE from constrictive bronchiolitis or severe

asthma. Constrictive bronchiolitis may show a mosaic pattern of attenuation on the expiratory HRCT images that cannot be appreciated on the inspiratory HRCT images. Mosaic pattern of attenuation is almost always associated with constrictive bronchiolitis rather than severe asthma (unpublished data) and mosaic air trapping is not seen in PAE. The presence of bronchiectasis has been described in patients with A1AD (Figs. 7 and 8). In one series of A1AD patients, bronchiectasis was identified in 43% of the patients studied.14 The presence of bronchiectasis in a patient with diffuse decrease in lower lobe lung density does not by itself make the distinction between A1AD and air trapping secondary to bronchiolitis, because cylindrical bronchiectasis is also a finding in patients with bronchiolitis. Distal Acinar Emphysema Distal acinar emphysema involves the peripheral, subpleural, aspect of the pulmonary acinus and presents as focal areas of subpleural emphysema. These lesions have an upper lobe predominance and a propensity to produce a spontaneous pneumothorax in an otherwise asymptomatic patient. CT scans of DAE characteristically show focal areas of decreased density or lung atten-

Figure 7. Spiral CT image using 7 mm collimation through the right and left lower lobes in an A1AD patient. There is advanced panacinar emphysema throughout both lower lobes. Cylindrical bronchiectasis is present in the left lower lobe (arrow).

CT OF EMPHYSEMA

37

Figure 8. A, HRCT image using 1 mm collimation through the upper lobes in an A1AD patient. This patient had extensive cylindrical and varicoid bronchiectasis in all five lobes of the lungs. There is extensive bronchial wall thickening present and a focal area of consolidative pneumonia in the left upper lobe. There is relatively mild emphysema present. The emphysema has features of central acinar emphysema, panacinar emphysema, and distal acinar emphysema. B, HRCT scan using 1 mm collimation through the lower lobes in the same patient as A. Extensive bronchiectasis again is noted with mild emphysema. Focal pneumonia also can be seen in the left lower lobe.

uation in the subpleural areas of the lungs with an upper lobe predominance (Fig. 9). The focal DAE lesions have a thin wall and no interior vessels. These two features help distinguish DAE from CAE. DAE can be diagnosed on CT by noting the subpleural location, thin wall, and absence of any interior vessels in the DAE lesion. The DAE lesions are subpleural in location and are seen adjacent to the lateral chest wall, mediastinum, and minor and major fissures. The DAE lesions are often asymptomatic but can grow to very large sizes, greater than 10 cm in diameter. When a single DAE lesion grows to occupy more than one third of the volume of the affected lung it usually compresses the adjacent lung enough to compromise ventilation of the normal adjacent lung tissue and resection of this bulla may be quite successful in treating the patient.

Irregular Emphysema CT scans that are obtained in patients with tuberculosis, sarcoidosis, Langerhans’ histiocytosis, or other fibrotic conditions may have areas of emphysema associated with the fibrotic lesions in the lung. These areas of em-

physema may arise in lung that was unaffected by emphysema before the onset of fibrosis (Fig. 10). The irregular emphysema may have features of CAE, PAE, or DAE in a given patient. The association with fibrotic lung injury and its greater severity in the area of fibrotic lung injury on CT scanning should enable the radiologist to make the diagnosis of irregular emphysema in most cases. Correlation with Pulmonary Function Testing There have been several studies that looked at the correlation of semiquantitative scoring of emphysema on CT, pathology, and pulmonary function tests.1, 5, 20 CT correlated better with pathology specimens than did pulmonary function testing but there was good correlation between CT and pulmonary function testing in patients with emphysema.1 CT correlated better with pulmonary function tests than did plain radiography.20 Qualitative scoring of inspiratory spiral CT scans correlated best with pulmonary function tests for emphysema, and quantitative measures of the ratio of inspiratory to expiratory CT density correlated best with measures of air trapping on pulmonary function tests.5

38

NEWELL

Figure 9. A, HRCT image using 1 mm collimation through the upper lobes in a patient with distal acinar and central acinar emphysema. The distal acinar emphysema has replaced all of the normal parenchyma in the right lung at this level. Less severe distal acinar emphysema and some mild central acinar emphysema can been seen in the left lung. B, HRCT image using 1 mm collimation through the upper lobes in the same patient as A but at a more caudal position. This image shows a large lesion of distal acinar emphysema (bulla) in the medial posterior right upper lobe. Extensive central acinar emphysema is present in both lungs along with some milder distal acinar emphysema.

Figure 10. A, This patient had underlying infection with tuberculosis and has developed irregular emphysema associated with the fibrotic response to the lung infection. HRCT image through the left upper lobe using 1 mm collimation shows mainly distal acinar emphysema. Several nodules can be seen in the left lung. B, HRCT image at a more caudal level through the upper lobes shows more nodules and less emphysema. The emphysema at this level has features of both distal acinar emphysema and panacinar emphysema.

CT OF EMPHYSEMA

Klein et al15 found that HRCT examination of the lungs can detect clinically significant CAE in patients with abnormal DLCO but no airflow changes. This study highlighted the sensitivity of HRCT to detect early, clinically significant disease. Gurney et al12 conducted a CT study on 59 smokers. Functional emphysema was defined to be present if the pulmonary function tests showed a DLCO of 75% of predicted value or less and an FEV1 of 80% of predicted or less. In this study both subjective and objective measures of emphysema were obtained. Twenty-five (40%) of the patients had HRCT evidence of emphysema but no functional evidence of emphysema. This showed the sensitivity of CT for detecting emphysema. Emphysema in the lower lobes had better correlation with abnormal pulmonary function tests than did emphysema in the upper lobes. The authors pointed out that upper lobe emphysema is relatively silent. Fifteen of the 59 patients had functional emphysema but no evidence of emphysema on HRCT scans. It is not clear if these patients really had emphysema, because no biopsy was done, but more recent work has shown that pulmonary function testing may suggest the presence of emphysema when only small airway disease is present.8 Quantitative CT Analysis of Emphysema A major advantage that CT scanning of the lungs provides is access to lung attenuation values, measured in Hounsfield units, that compose the digital image data. The digital data from each image can be analyzed to try and detect the presence and severity of emphysema by several means. The mean lung density can be computed on each image and a comparison made with the average normal mean lung density. The image histogram curve can be obtained and measures of skewness and kurtosis can be looked at as another means of detecting and assessing the presence of emphysema. One of the more widely accepted methods of using histogram curve analysis is looking at the lung attenuation value where 15% of the lung is less dense than this particular value. This value is expressed in terms of grams per liter, which refers to grams of lung per liter of lung volume.4 This value shifts toward a lower value as the emphysema progresses. This approach avoids some of the problems associ-

39

ated with a fixed threshold value in which higher values of lung attenuation may be shifting downward but are not counted as emphysema until they fall below a certain threshold (i.e., 910 H). The amount of abnormal lung tissue can be determined by using a threshold technique for image segmentation. This is accomplished by setting a lower threshold for normal lung attenuation, typically 910 H. The total area of volume of lung is calculated first using a lower threshold value of 1000 H and an upper threshold value typically between 600 and 200 H. The amount of emphysematous lung is determined by calculating the percentage of lung that is less than 910 H. This technique was described initially as the ‘‘density mask’’ technique by Mu¨ller et al in 1988.17 The technique has been extended in a couple of ways. The threshold may be decreased to 960 H to detect more severe emphysema only. For milder emphysema, the threshold may be as high as 856 H.2 In addition to looking at the amount of abnormal lung as percent of total lung, one can apply a technique now known as CT morphometry.2 In this technique prediction equations were developed by comparing histologic tissue samples with CT scans from parts of lung that had varying amounts of CAE. A number of indices were developed from this approach, including the volume of severe emphysema present (⬎ 10.2 mL gas per gram of tissue); the amount of mild-moderate emphysema present (6 to 10.2 mL gas per gram of tissue); and also measures of surface area to volume ratio and surface area. These new quantitative CT techniques have enabled investigators to perform point in time quantitative CT studies on patients with emphysema and longitudinal quantitative CT studies. Surgical Treatment of Emphysema The surgical treatment of emphysema involves two basic strategies. Bilateral and single lung transplantation can be done for the treatment of severe emphysema in patients less than 65 years of age. This often involves patients with A1AD-induced emphysema. A1AD patients with severe emphysema may undergo bilateral lung transplantation up to age 55 and single lung transplantation up to age 65.25 In older patients, typically greater than 65 years of age, LVRS can be performed. There is still controversy about the effective-

40

NEWELL

ness of LVRS and there is an ongoing randomized multicenter study in the United States that is trying to address the issues of patient selection and benefit from LVRS. This study is called the National Emphysema Treatment Trial (NETT) (Fig. 11). CT has been used to measure the TLC of the transplanted lung and the native lung in nine emphysema patients who underwent single lung transplantation.6 The study obtained CT examinations at a mean of 326 days before transplantation and at mean intervals of 239, 588, and 932 days after transplantation. Using quantitative CT the authors were able to show that the TLC in the native and transplanted lungs did not change significantly after transplant surgery. There has been concern that in single lung transplantations in patients with emphysema the emphysematous native lung continues to hyperinflate and ultimately compress and compromise the transplanted lung. This CT study suggests this is not the case. CT has been used in selecting patients for LVRS. The ideal patient for LVRS seems to be one with a CT study of the lungs that suggests predominant upper lobe emphysema (see Fig. 10) with sparing of the lung bases and with evidence of hyperinflation (flattened diaphragm) on chest radiography. CT is used in the selection of patients for surgery in the NETT study and was used in a recently pub-

lished randomized study looking at the surgical and medical benefits of treating people with LVRS or medical therapy.7 In this study, 174 subjects were assessed for entrance into the study and 24 were randomly assigned to surgical therapy and 24 to medical therapy. After 6 months in this study the median FEV1 had increased 70 mL in the surgical group and decreased 80 mL in the medical group (P0.02). Similar improvements in the surgical patients and deterioration in the medical patients were noted in the median shuttle walking distance and in the quality of life scale. In the 24 patients who went to surgery, the CT scans showed generalized emphysema in 14 patients, upper lobe predominant emphysema in 8 patients, and lower lobe predominant emphysema in 2 patients. Five of the 19 surgical patients did not improve following LVRS. None of these five patients resumed smoking and each had similar baseline characteristics to the other patients who did show improvement with the exception of their CT scans. The five patients who did not improve after surgery had more diffusely distributed emphysema as assessed by CT scanning than the other 19 surgical patients. It should be noted that even though there was a significant improvement in the surgically treated patients compared with the nonsurgically treated patients, the annual decline in FEV1 was similar in both groups after ran-

Figure 11. A, This HRCT scan using 1 mm collimation was obtained on a patient referred for lung volume reduction surgery evaluation as part of the National Emphysema Treatment Trial (NETT). The image demonstrates advanced upper lobe emphysema, which at this point appears to be panacinar. It was estimated that more than 75% of the upper third of each lung was affected by emphysema. B, This is a 6-month follow-up of the patient in A after the patient underwent bilateral lung volume reduction surgery through a midline thoracotomy. The image demonstrates a typical surgical staple line in the right upper lobe, with substantial residual emphysema.

CT OF EMPHYSEMA

domization. The FEV1 continued to decline by 100 mL per year in both groups. This suggests that surgery provides a one-time benefit to carefully selected patients but does not subsequently modify the natural history of the disease. Another important role of CT is its ability to identify patients with one or more large bullae. It is well known that marked improvement in the FEV1 may result from the removal of large bullae and it is important to exclude these patients in any credible study that looks at the value of LVRS in the treatment of diffuse emphysema.7 Geddes et al7 did take care to exclude patients with large bullae who were treated surgically outside the study, and so their results are probably more reliable in terms of FEV1 improvements than other studies that did not exclude patients with large bullae. CT morphometry, one of the quantitative techniques described previously, has been reported to provide a close relationship between the baseline severity of emphysema and subsequent improvements in emphysema as measured by CT morphometry and maximal cardiopulmonary exercise following LVRS.19 The authors report in this study that the decrease in TLC following LVRS was entirely accounted for in the decrease in the amount of severe emphysema as determined by CT morphometry. The surface area to volume ratio and surface area determined from CT morphometry also increased significantly following LVRS in this study.19 Both qualitative and quantitative CT measures of emphysema are being done in the NETT trial, which runs another 2 years. The results of this study should be quite helpful in trying to evaluate the efficacy of LVRS. Enzyme Replacement Therapy for ␣1-Antitrypsin Deficiency– Induced Emphysema There is great interest in using CT scanning to study the therapeutic effects of enzyme replacement therapy in patients with A1AD. Intravenous enzyme therapy has been around a long time but there are emerging inhaled and oral preparations of prolastin and newer gene therapy techniques for A1AD. CT scanning is considered the most promising noninvasive technique for detecting and following emphysema therapeutic trials of these new therapies. A promising study that looked at using CT scanning in the assessment of patients with

41

A1AD was recently published by Dirksen et al.3 The authors investigated whether the restoration of balance between neutrophil elastase and A1A could prevent the progression of emphysema in 26 Danish and 30 Dutch exsmokers with A1AD. The A1AD patients all had PI*ZZ phenotype for A1AD. The patients had a moderate decrease in FEV1 ranging between 30% and 80% of predicted. The patients participated in a double-blind randomized trial of intravenous infusion either of A1AD replacement therapy or of albumin over at least 3 years. In this study there was no significant difference in the decline of FEV1 between the placebo group and those receiving A1A. Using the 15th percentile point of the lung density histogram, however, the loss of lung tissue measured by CT was 2.6 g/ L/y for placebo, as compared with 1.5 g/ L/y for A1A infusion (P0.07). A statistical power analysis showed that this reaches statistical significance with a trial of 130 patients. This contrasted to the power analysis for the FEV1 measurements that showed one needs 550 patients to show a 50% reduction in the annual decline of the FEV1. The major significance of the study was showing how quantitative CT scanning could significantly decrease the study duration or the number of patients required to show a positive respone to A1A treatment compared with a study using FEV1 measurements. This is especially important in rare diseases like A1AD, where it is difficult to recruit enough patients into a study and the life expectancy of the patients is often very short compared with the normal population. CT scanning can help researchers design and implement studies to evaluate the potential of newer methods of treatment for A1AD including oral and inhaled A1A replacement therapy and emerging gene therapies for the treatment of A1AD. Retinoic Acid Treatment of Emphysema There has been excitement around the potential to use retinoic acid in the treatment of emphysema patients.16, 21, 24 Sklan et al21 reported in 1990 the inhibition of the activity of human leukocyte elastase by lipids. They reported that two lipids in particular, oleic acid and retinoic acid, were especially effective in inhibiting the activity of leukocyte elastase. The investigators reported that the intrapulmonary instillation of oleic or retinoic acid reduced lung injury caused by human

42

NEWELL

leukocyte elastase in an emphysema mouse model. Massaro and Massaro16 reported in 1997 that elastase-induced emphysema in the rat could be reversed by the treatment of the rat with retinoic acid. The retinoic acid works by increasing the number of alveoli in the rat lung. These exciting reports of a new treatment for emphysema have led to a multicenter study to examine the effects of retinoic acid in the treatment of patients with emphysema. CT scanning will be used as one of the critical ways of detecting emphysema and monitoring any improvement or progression of the emphysema in patients treated with retinoic acid and in the control group of patients who do not receive retinoic acid. References 1. Bergin C, Muller N, Nichols DM, et al: The diagnosis of emphysema: A computed tomographic-pathologic correlation. Am Rev Respir Dis 133:541–546, 1986 2. Coxson HO, Rogers RM, Whittall KP, et al: A quantification of the lung surface area in emphysema using computed tomography. Am J Respir Crit Care Med 159:851–856, 1999 3. Dirksen A, Dijkman JH, Madsen F, et al: A randomized clinical trial of alpha(1)-antitrypsin augmentation therapy. Am J Respir Crit Care Med 160:1468– 1472, 1999 4. Dirksen A, Friis M, Olesen KP, et al: Progress of emphysema in severe alpha 1-antitrypsin deficiency as assessed by annual CT. Acta Radiol 38:826–832, 1997 5. Eda S, Kubo K, Fujimoto K, et al: The relations between expiratory chest CT using helical CT and pulmonary function tests in emphysema. Am J Respir Crit Care Med 155:1290–1294, 1997 6. Estenne M, Cassart M, Poncelet P, et al: Volume of graft and native lung after single-lung transplantation for emphysema. Am J Respir Crit Care Med 159:641–645, 1999 7. Geddes D, Davies M, Koyama H, et al: Effect of lungvolume-reduction surgery in patients with severe emphysema. N Engl J Med 343:239–245, 2000 8. Gelb AF, Zamel N, Hogg JC, et al: Pseudophysiologic emphysema resulting from severe small-airways disease. Am J Respir Crit Care Med 158:815–819, 1998 9. Gierada DS, Yusen RD, Villanueva IA, et al: Patient selection for lung volume reduction surgery: An objective model based on prior clinical decisions and quantitative CT analysis. Chest 117:991–998, 2000

10. Goddard PR, Nicholson EM, Laszlo G, et al: Computed tomography in pulmonary emphysema. Clin Radiol 33:379–387, 1982 11. Gurney JW: Pathophysiology of obstructive airways disease. Radiol Clin North Am 36:15–27, 1998 12. Gurney JW, Jones KK, Robbins RA, et al: Regional distribution of emphysema: Correlation of high-resolution CT with pulmonary function tests in unselected smokers. Radiology 183:457–463, 1992 13. Hansell DM: Disease of the airway. In Armstrong P, Wilson AG, Dee P, et al (eds). Imaging of Diseases of the Chest, ed 3. London, Mosby, 2000, pp 893–947 14. King MA, Stone JA, Diaz PT, et al: Alpha 1-antitrypsin deficiency: Evaluation of bronchiectasis with CT. Radiology 199:137–141, 1996 15. Klein JS, Gamsu G, Webb WR, et al: High-resolution CT diagnosis of emphysema in symptomatic patients with normal chest radiographs and isolated low diffusing capacity. Radiology 182:817–821, 1992 16. Massaro GD, Massaro D: Retinoic acid treatment abrogates elastase-induced pulmonary emphysema in rats [published erratum appears in Nat Med 3:805, 1997]. Nat Med 3:675–677, 1997 17. Mu¨ller NL, Staples CA, Miller RR, et al: ‘‘Density mask.’’ An objective method to quantitate emphysema using computed tomography. Chest 94:782– 787, 1988 18. O’Driscoll MC, Chan ED, Fernandez E: Imaging of emphysema. In Lynch DA, Newell JD Jr, Lee JS (eds): Imaging of Diffuse Lung Disease. Hamilton, B.C. Decker, 2000, pp 99–233 19. Rogers RM, Coxson HO, Sciurba FC, et al: Preoperative severity of emphysema predictive of improvement after lung volume reduction surgery: Use of CT morphometry. Chest 118:1240–1247, 2000 20. Sanders C, Nath PH, Bailey WC: Detection of emphysema with computed tomography: Correlation with pulmonary function tests and chest radiography. Invest Radiol 23:262–266, 1988 21. Sklan D, Rappaport R, Vered M: Inhibition of the activity of human leukocyte elastase by lipids particularly oleic acid and retinoic acid. Lung 168:323– 332, 1990 22. Snider GL: Emphysema: The first two centuries— and beyond. A historical overview, with suggestions for future research: Part 2. Am Rev Respir Dis 146:1615–1622, 1992 23. Snider GL, Kleinerman J, Thurlbeck WM, et al: The definition of emphysema: Report of a National Heart, Lung, and Blood Institute, Division of Lung Diseases workshop. Am Rev Respir Dis 132:182–185, 1985 24. Tepper J, Pfeiffer J, Aldrich M, et al: Can retinoic acid ameliorate the physiologic and morphologic effects of elastase instillation in the rat? Chest 117:242S– 244S, 2000 25. Trulock EP, Cooper JD: Lung transplantation for alpha1AT-deficiency emphysema. In Crystal RG (ed): Alpha 1-Antitrypsin Deficiency: Biology, Pathogenesis, Clinical Manifestations, Therapy, vol 88. New York, Marcel Dekker, 1996, pp 387–404 Address reprint requests to John D. Newell Jr, MD National Jewish Medical and Research Center 1400 Jackson Street Denver, CO 80206 e-mail: [email protected]

HIGH-RESOLUTION CT OF THE LUNG II

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE Masanori Akira, MD

CT has an increasing role in the radiologic evaluation of occupational lung disease. High-resolution CT (HRCT) is more sensitive than chest radiography in the depiction of parenchymal abnormalities in asbestosis, silicosis, and other pneumoconioses. The most common HRCT finding of pneumoconiosis is centrilobular nodular or branching areas of high attenuation, which are bronchiolar lesions. Interstitial fibrosis may be manifested as traction bronchiectasis, honeycombing, or more confluent areas of hyperattenuation. HRCT-pathologic correlations have been demonstrated in asbestosis and coal workers’ pneumoconiosis (CWP), but larger prospective studies with HRCT–pathologic correlation in each type of pneumoconiosis are required. SILICOSIS Silicosis is a fibrotic disease of the lungs caused by inhalation of dust containing free crystalline silica. Three different types of tissue reaction have been distinguished: (1) chronic, (2) accelerated, and (3) acute. The most usual form of silicosis occurs after many years of exposure to relatively low levels of dust. It is rare for the chest radiograph to become positive before 20 years of exposure. Silicosis is characterized by the presence of small discrete nodules exclusively distributed

in the upper zones of the lung, with a posterior predominance (Fig. 1).13, 49 These nodules tend to coalesce and form massive fibrotic lesions. When nodules are 10 mm or less in size, the disease is termed simple silicosis or simple pneumoconiosis. Complicated pneumoconiosis is defined by the presence of nodules measuring 1 cm or more. Progressive massive fibrosis (PMF) is seen as irregularly shaped masses, most frequently in apical and posterior segments of upper and lower lobes, with peripheral parenchymal distortion. Cavitation can occur from ischemic necrosis or accompanying tuberculosis (Fig. 2). Various types of calcifications in PMF can be observed; punctate, linear, or massive. Hilar and mediastinal lymph nodes tend to enlarge and calcify in an eggshell, punctate, or massive form. Exposure to high concentrations of silica over a period of as little as 5 years results in a more rapidly progressive form of the disease (accelerated silicosis). In accelerated silicosis, the major features of the disease are identical to those of the chronic disease. HRCT scans at the early stage demonstrate irregular nonseptal small linear opacities in addition to nodular opacities diffusely distributed throughout both lungs (Fig. 3). Serial CT scans demonstrate the progression of the disease from a diffuse linear pattern to a confluence of nodules with or without PMF.49 Acute silicosis, also called silicoproteinosis,

From the Department of Radiology, National Kinki Chuo Hospital for Chest Disease, Sakai, Osaka, Japan

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

43

44

AKIRA

Figure 1. High-resolution CT (HRCT) scan of silicosis. Nodules are predominantly posterior in distribution. Confluence of nodules is seen.

Figure 2. Complicated silicosis. HRCT scan shows bilateral masses of progressive massive fibrosis (PMF) with cavitation. Gas–fluid level is identified in the area of PMF on the left.

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

45

Figure 3. HRCT scan obtained in a 55-year-old man with accelerated silicosis who had worked as a stoneworker for 5 years. His radiographic signs were observed within 3.6 years of first exposure. Scan demonstrates irregular nonseptal small linear opacities in addition to confluent nodular opacities diffusely distributed throughout both lungs in an intermediate reticular pattern mixed with areas of low attenuation. Small areas of low attenuation with a central dot also are found, suggesting focal emphysema.

occurs in subjects exposed to very high concentrations of silica over periods of as little as a few weeks. Sandblasting is the most common cause of acute silicosis. Pathologically, there is a proteinaceous and cellular alveolar exudate similar in character to that of alveolar proteinosis. The presence of the interstitial pneumonitis differentiates this disease from pulmonary alveolar proteinosis. Radiographic findings are similar to those in alveolar proteinosis. HRCT scans show a diffuse groundglass or alveolar pattern. The nodular pattern is absent.49 An HRCT scan is more sensitive than chest radiography in detecting lung parenchymal changes suggestive of silicosis and early confluence of small opacities in the lung.12 It has been shown that the reduced levels of lung function in patients with silicosis correlated with superimposed emphysema rather than the nodular profusion, and that emphysema associated with silicosis is easily detected on HRCT but not on the radiograph.13, 30 It is suggested that silicosis, in the absence of PMF, does not cause significant emphysema.30 MIXED DUST PNEUMOCONIOSIS Mixed dust pneumoconiosis is defined as a pneumoconiosis caused by concomitant exposure to silica and less fibrogenic dusts, such as

iron, silicates, and carbon. The silica usually is at a lower percentage of the total lung dust than when silicotic nodules occur. Microscopically the mixed dust fibrotic nodule is characterized by a stellate shape and has a central hyalinized collagenous zone surrounded by linearly and radially arranged collagen fibers admixed with dust-containing macrophages. Generally, as the proportion of silica increases, the number of silicotic nodules increases in proportion to the mixed dust nodules.24 The radiologic findings of mixed dust pneumoconiosis include a mixture of small rounded and irregular opacities (Fig. 4). Honeycombing also is seen.51 COAL WORKERS’ PNEUMOCONIOSIS In CWP the miners are exposed to a dust that is a mixture of coal, kaolin, mica, and silica. The inhalation of coal mine dust may lead to the development of CWP and silicosis.54 The characteristic lesion of CWP is the collection of closely packed dust-laden macrophages around the respiratory bronchiole, which contains little collagen (coal dust macule). They usually are associated with dilatation of respiratory bronchioles, so-called focal emphysema. Another type of lesion is the fibrotic nodule. Based on the classification es-

46

AKIRA

Figure 4. HRCT scan in a 60-year-old man with pathologically proven mixed dust pneumoconiosis. Scan shows multiple bullae, traction bronchiectasis, and areas of ground-glass attenuation with intralobular interstitial thickening. Pathologic examination obtained by autopsy showed mixed dust nodules and no silicotic nodules.

tablished by the College of American Pathologists,31 micronodules are lesions up to 7 mm in diameter, and lesions from 7 to 20 mm are defined as macronodules. PMF is defined as nodules at least 2 cm in diameter. Using the International Labour Office 1980 international classification of radiographs of pneumoconiosis,27 small rounded opacities smaller than 1.5 mm in diameter are classified as p, those between 1.5 and 3 mm are q, and those between 3 and 10 mm are r. PMF or large opacities are defined radiologically as a lesion of 1 cm or greater in longest diameter. The HRCT appearance in patients with radiographic type-p pneumoconiosis is characterized by tiny branching lines or ill-defined punctate opacities, usually in a centrilobular location (Fig. 5). In some patients, small nonperipheral areas of low attenuation with a central dot are found.5 Previous studies have shown that these tiny opacities correspond to irregular fibrosis around and along the respiratory bronchioles or dust macule with dilatation of respiratory bronchioles.5, 49 Small areas of low attenuation with a central dot have been shown to correspond to focal dust emphysema. Focal emphysema has been found most commonly in pneumoconiosis with ptype changes. Opacities of the q and r types are characterized by sharply demarcated, rounded nodules or contracted nodules.5 Micronodules can be detected in the subpleural areas, and confluence of subpleural

micronodules may simulate pleural plaques and is referred to as pseudoplaques (Fig. 6). Pseudoplaques are observed with high frequency in CWP, graphite pneumoconiosis, sarcoidosis, and pulmonary lymphangitic carcinomatosis. 48 Remy-Jardin et al 49 demon-

Figure 5. HRCT scan in a coal miner with radiographic type p pneumoconiosis. Tiny branching lines in a centrilobular location (arrows) and a few well-defined nodules are seen. Subpleural nodules also are seen.

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

47

GRAPHITE PNEUMOCONIOSIS

Figure 6. HRCT scan at the level of upper lobes shows subpleural micronodules along the concavity of the chest wall and a pseudoplaque (arrows). Parenchymal micronodules with confluence are seen.

Graphite pneumoconiosis is defined as pneumoconiosis in workers exposed to graphite dust at the workplace. Graphite exposure produces roentgenographic and pathologic changes similar to those of coal miners. The common HRCT findings of graphite pneumoconiosis are small nodular hyperattenuating areas, interlobular septal thickening, and the prevalence of large opacities (PMF). Bullae are frequently found. Other less common CT manifestations include emphysema, areas of ground-glass attenuation, focal decreased attenuation, and bronchiectasis. Small nodular hyperattenuating areas are classified into two patterns: (1) ill-defined tiny hyperattenuating areas that appear either as fine branching structures or as a few dots clustered together (p-type pneumoconiosis); or (2) well-defined, discrete nodules (Fig. 7). Both types of nodules usually are in a centrilobular location. Some nodules are present along interlobular septa or pleural surface. Ill-

strated that subpleural micronodules in CWP corresponded to several microscopic features: macules, nodules, localized visceral pleural thickening, or subpleural lymph nodes. The appearance of PMF in CWP is similar to that in silicosis; however, histopathologically PMF of coal or other carbonaceous pneumoconiosis is distinguished from a silicotic conglomeration by the excess of black dust, and by the absence of individually identifiable silicotic nodules with their distinctive whorled pattern in the aggregate.45 Cavitation, usually caused by ischemic necrosis, may be present. Diffuse interstitial pulmonary fibrosis (DIPF) is sometimes found in the lungs of coal miners. Its appearances and distribution are often similar to those of idiopathic DIPF. The overall incidence of DIPF is approximately 18%.38 Coal workers with rheumatoid disease may develop nodules rapidly even after relatively low exposures to dust (Caplan’s syndrome). Caplan’s syndrome or rheumatoid pneumoconiosis consists of rounded discrete rheumatoid nodules ranging in size from 0.5 to 5 cm in diameter, superimposed on pneumoconiotic opacities. They may cavitate and calcify. Systemic sclerosis (scleroderma) is reported to occur with unusual frequency in coal miners and men with silicosis (Erasmus’ syndrome).49

Figure 7. HRCT scan obtained in a 53-year-old man employed in the graphite industry for 12 years. Two types of small nodules are seen: ill-defined tiny branching areas of hyperattenuation (arrows) and well-defined nodules (arrowhead). (From Akira M: Uncommon pneumoconioses: CT and pathologic findings. Radiology 197:403–409, 1995; with permission.)

48

AKIRA

defined hyperattenuating areas are abundant, and there are comparably fewer discrete nodules. HRCT–pathologic correlation has revealed that ill-defined hyperattenuating areas corresponded to macular lesions along the walls of bronchioles, which are often dilated, and that the discrete nodules correspond to larger macular or nodular lesions.4 Reticular hyperattenuating areas similar to findings in DIPF are a predominant finding in some cases with graphite pneumoconiosis (Fig. 8). Microscopically, the fibrosis is pigmented interstitial fibrosis or unpigmented or mildly pigmented interstitial pneumonia suggestive of usual interstitial pneumonia.38 ASBESTOSIS Asbestosis is defined as diffuse interstitial fibrosis of the lung as a consequence of exposure to asbestos dust. The fibrotic process of asbestosis begins in the peribronchiolar areas of the lung, then progresses to involve adjacent alveoli, showing the classic features of asbestosis. Concerning alveolar wall thickening by fibrosis, recent studies indicate that alveolar deposition of asbestos fibers is also of pathogenic importance. There is evidence that asbestos fibers can penetrate directly through airway walls to reach the parenchyma.17, 21

The common HRCT manifestations of asbestosis are interstitial lines (thickened intralobular core lines and irregular thickening of interlobular septa); subpleural curvilinear lines; parenchymal bands; and honeycombing.1, 2, 7, 8 Thickened intralobular core lines arise a few millimeters from the pleura but seldom attack it. Some of them appear as fine branching structures. Interlobular lines are 1 to 2 cm in length and are seen in the subpleural parenchyma, extending peripherally toward the pleural surface and contacting it. The thickened intralobular lines have been shown pathologically to be caused by peribronchiolar fibrosis with subsequent involvement of the alveolar ducts. Thickened interlobular lines have been shown to represent either fibrotic or edematous thickening of the interlobular septa.7 Subpleural curvilinear lines are defined as linear areas of increased attenuation within 1 cm of the pleura and parallel to the inner chest wall (Fig. 9). The pathologic correlate of subpleural lines represents peribronchiolar fibrotic thickening combined with flattering and collapse of the alveoli caused by fibrosis.7, 57 The distance of the subpleural curvilinear lines from the inner chest wall is never more than 1 cm but is mostly less than 0.5 cm in patients with asbestosis. Subpleural curvilinear lines are considered to be caused

Figure 8. HRCT scans obtained in a 60-year-old man with a history of exposure to graphite dust for 23 years. Scans show irregular fibrotic areas of high attenuation with traction bronchiectasis and bronchiolectasis. Areas of ground-glass attenuation, interlobular septal thickening, and centrilobular nodular and branching opacities also are seen.

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

Figure 9. HRCT scan shows subpleural curvilinear line (arrows) parallel to the inner chest wall at a distance of approximately 1.5 mm from the pleura.

by some different pathologic processes. Kubota et al 33 postulated that the curvilinear lines represented plate-like atelectasis in the corticomedullary junction of the lung. The curvilinear lines, in their cases with idiopathic pulmonary fibrosis, were observed at distances of about 1.5 to 2 cm from the inner chest wall. Naidich et al42 described a similar curvilinear line in CT findings of interstitial diseases and interpreted the line to be caused by the thickened secondary lobular septa. Pilate et al46 reported that lipoidal filling subpleural curvilinear line was seen after lymphography. They suggested that a subpleural curvilinear line might correspond to the normal subpleural lymphatic network. Also, Arai et al10 reported two cases of transient subpleural curvilinear line caused by pulmonary congestion. Parenchymal bands are linear densities from 2 to 5 cm in length coursing through the lung, usually to contact the pleura (Fig. 10). Parenchymal bands represent fibrosis along the bronchovascular sheath or interlobular septa with distortion of the parenchyma, often caused by traction of the thickened pleura.7 Parenchymal bands caused by trac-

49

tion of the thickened pleura may appear with or without asbestosis. Ground-glass opacities are common in idiopathic interstitial pneumonias and uncommon with asbestosis. In asbestosis, ground-glass opacity may be seen mainly in the subpleural zones and generally in association with other architectural abnormalities. In some cases of asbestosis honeycombing is absent or slight even when severe fibrosis is present. Aberle et al2 described that honeycombing was found in 5 of 29 cases with asbestosis. Al-Jarad et al9 also reported that crescentic fine reticular pattern with small cysts was found in 4 of 24 patients with asbestosis, whereas it was found in 15 of 18 patients with cryptogenic fibrosing alveolitis. The earliest findings in asbestosis on HRCT have been shown to be subpleural isolated dots connected with the most peripheral branch of the pulmonary artery and located a few millimeters from the pleura (Fig. 11). Paired serial CT images revealed that these subpleural dots or branching structures increased in number, and the confluence of dots created pleural-based nodular irregularities

Figure 10. HRCT scan of a patient with asbestos exposure shows parenchymal bands extending from the pleura to contact the pleural surface (arrows). Scattered areas of thickened septal and core lines are present.

50

AKIRA

Figure 11. Early findings of asbestosis. A, Prone HRCT scan in a subject with asbestos exposure shows subpleural dots or branching structures (arrows). Some dots are connected with the most peripheral branches of the pulmonary artery and are located a few millimeters from the pleura (From Akira M, Yokoyama K, Yamamoto S, et al: Early asbestosis: Evaluation with thin-section CT. Radiology 178:409–416, 1990; with permission.). B, Postmortem radiograph of an inflation-fixed lung obtained from a patient with asbestosis. Nodular opacities can be seen a few millimeters from the pleural surface (arrows). The increased opacity attached to the pleura is seen in the upper right corner. C, Histologic section reveals that subpleural dot-like lesions correspond histologically to peribronchiolar fibrosis with sequential involvement of the alveolar ducts (From Akira M, Yamamoto S, Yokoyama K, et al: Asbestosis: Thin-section CT-pathologic correlation. Radiology 176:389–394, 1990; with permission.)

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

and subpleural curvilinear lines. Later, hazy patches of ground-glass opacity appear around the subpleural parenchymal abnormalities or in the central parenchyma.8 Pleural plaques represent circumscribed areas of fibrous thickening, typically of the parietal pleura, caused by the deposition of paucicellular collagenous tissue with a laminar or basket-weave pattern. Pleural plaques are the radiologic hallmark of asbestos exposure. Diffuse pleural thickening is seen less frequently than pleural plaques following exposure to asbestos. Diffuse pleural thickening involves the visceral pleura and is frequently the result of prior asbestos exudative effusions. McLoud et al39 defined diffuse pleural thickening as a smooth uninterrupted pleural density extending over at least one fourth of the chest wall, with or without obliteration of the costophrenic angle. On CT scans, Lynch et al34 defined diffuse pleural thickening as a continuous sheet of pleural thickening more than 5 cm wide, more than 8 cm in craniocaudal extent, and more than 3 mm thick (Fig. 12). Thickening of the interlobar fissures is another manifestation of diffuse pleural thickening. Diffuse pleural thickening has been shown to cause restrictive pulmonary function and reduced diffusing capacity.29, 50

Figure 12. Diffuse pleural thickening related to asbestos exposure. HRCT scan at a wide window (level, –400 Hounsfield units [H]; width, 2000 H) shows diffuse pleural thickening (arrowheads).

51

The term rounded atelectasis refers to the deformity of peripheral lung and the associated bending of small bronchovascular structures secondary to overlying pleural adhesions. The most helpful features in the diagnosis of rounded atelectasis with CT are (1) contiguity to areas of diffuse pleural thickening, (2) a lentiform or wedge-shaped outline, (3) evidence of volume loss in the adjacent lung, and (4) a characteristic ‘‘comet tail’’ of vessels and bronchi sweeping into the margins of the mass.35 CT is more sensitive than the chest radiograph for detection of pleural plaques. HRCT is more sensitive than the chest radiograph and conventional CT for detection of the early changes of asbestosis. It is also shown that asbestosis can be present histopathologically with a normal or near normal HRCT scan.23 HRCT findings in patients with asbestosis are nonspecific. The similar findings may occur in patients with a variety of underlying diseases or conditions unrelated to asbestosis.11 Gamsu et al23 studied HRCT scans of 24 patients and six lungs obtained at autopsy with asbestos exposure in shipyards or construction. Histopathologic asbestosis was present in 28 of 30 patients or lungs. The HRCT scans were assessed in two ways: (1) a subjective semiquantitative scoring method and (2) a method using a cumulation of the different HRCT features of asbestosis. On the cumulative method, interstitial lines (thickening of the interlobular septa and centrilobular core structures), parenchymal bands, subpleural curvilinear lines, honeycombing, subpleural nodules, and architectural distortion were used. These had to be bilateral or on several scans in one hemithorax to be considered present. With the subjective semiquantitative HRCT severity score, asbestosis was suggested in 64% of cases with asbestosis. The semiquantitative grade of asbestosis on HRCT was significantly associated with the histopathologic fibrosis severity score. With the cumulative method, any one type of abnormality was present in 88% of cases with asbestosis, two types in 78%, and three in 56%. To include only cases with asbestosis, three different abnormalities had to be present. In patients with suspected asbestosis, prone HRCT scans should be performed. Atelectasis is commonly seen in the dependent lung in both healthy and diseased subjects, resulting in dependent density or subpleural line. Prone HRCT scans allow dependent density

52

AKIRA

to be differentiated from true disease. A true abnormality remains unchanged regardless of whether it is dependent or nondependent, whereas normal dependent density disappears when the patient is prone.1 A wide window suitable for evaluating the pleura and parenchyma simultaneously is recommended by some investigators.2, 22 TALCOSIS Talc is a hydrated magnesium silicate with the formula Mg 3 Si 4 O 10 (OH) 2 , although calcium, aluminum, and iron are always present in variable amounts. Talcosis can produce several radiographic appearances: a nodular pattern resembling silicosis or mixed dust fibrosis, including, in some cases, large opacities; a diffuse interstitial pattern, simulating asbestosis; or a combination of nodular and linear patterns (Fig. 13). Histologically, three distinct types of lesions may occur: (1) illdefined nodular lesions, (2) DIPF, (3), and foreign body granulomas.28 The development of a foreign body granulomatous reaction fol-

Figure 13. HRCT scan obtained in a 68-year-old man employed in the talc industry for 20 years. Ill-defined small nodules diffusely distributed throughout the lung are seen. Many tiny branching structures are seen (arrows). A few nodules attach to the pleura. Most nodules seldom attach to the pleura or to the outer margin of a large bronchus or vessels, although they occur near the pleura and vessels.

lowing injection of talc-containing drugs is well recognized. Padley et al44 described the CT appearances in three patients with pulmonary talcosis resulting from chronic intravenous drug abuse. They included widespread ground-glass attenuation in one case, and confluent perihilar masses with areas of high attenuation similar to that of progressive massive fibrosis in two cases. Ward et al56 described the CT appearances of 12 patients with talcosis associated with intravenous abuse of oral medications. The predominant abnormalities consisted of a diffuse fine nodular pattern (N  2); a combination of nodules and lower lobe panacinar emphysema (N  3); and ground-glass attenuation (N  2). Emphysema was the only abnormality seen in the remaining five patients. Lower lobe panacinar emphysema was more common in methylphenidate abusers than in nonmethylphenidate drug abusers. WELDER’S LUNG Electric-arc and oxyacetylene welders may inhale finely divided particles of iron oxide during the course of their work. Iron oxide dust is not fibrogenic in human and animal lungs. Some workers exposed to metallic iron or iron oxide fume may also have had significant exposure to other dusts, however, such as quartz, cristobalite, or asbestos, so that siderosis may be complicated by the presence of mixed dust fibrosis or asbestosis. The typical chest radiographic finding in arc welders’ siderosis consists of micronodules that are most prominent in the middle third of the lungs in the perihilar regions. The micronodules do not reflect reactive fibrosis but, rather, radiopaque accumulations of iron particles that lie within macrophages, aggregated along the perivascular and peribronchial lymphatic vessels.16 The most common CT findings in the arc welders are ill-defined micronodules diffusely distributed in the lung (Fig. 14). Some of the micronodules appear as fine branching lines. In less affected lung, micronodules show centrilobular distribution. In more affected lung, gathering of micronodules and fine branching lines form a fine network pattern or area of groundglasss attenuation. Other CT findings in welder’s lung include emphysema, focal decreased attenuation area distributed multilobularly,

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

53

upper lobe fibrosis with some associated pulmonary nodularity. In the remaining five cases CT showed varying degrees of reticulonodular interstitial change. In one of these five CT also showed subpleural cysts of bleb formation. In another patient CT demonstrated widespread honeycomb cysts. Newman et al43 studied 28 patients with biopsyproved beryllium disease and concluded that HRCT was more sensitive than chest radiography in detection of beryllium disease, but the diagnosis was missed in up to 25% of cases with histologic proof. The most common CT abnormalities were parenchymal nodules and septal lines. The nodules were usually well defined and were often distributed along the bronchovascular bundles or interlobular septa, similar to the distribution seen in sarcoidosis. Figure 14. HRCT scan obtained in a 60-year-old man employed as an arc welder for 30 years. Numerous illdefined micronodules and fine branching opacities diffusely distribute through the lung. In a more affected lung, gathering of micronodules and fine branching lines form a fine network pattern.

ALUMINUM LUNG Inhalation of dusts containing metallic and oxidized aluminum has been reported to be

honeycombing, bronchiectasis, and conglomerate masses with areas of high attenuation (Fig. 15). The mass with areas of high attenuation corresponded histologically to organizing pneumonia with siderosis.4 BERYLLIOSIS Beryllium disease is a multisystem disorder caused by exposure to dust, fumes, or aerosols of beryllium metal or its salts. Beryllium disease occurs in two forms: (1) an acute toxic chemical pneumonitis, occurring after brief exposure to extremely high levels of airborne beryllium; and (2) a chronic granulomatous pulmonary disease as a result of a berylliumspecific, cell-mediated immune response. The former is extremely rare and follows large accidental exposures. Beryllium disease is pathologically indistinguishable from sarcoidosis. With the use of a blood test of the Tlymphocyte response to beryllium, called the beryllium lymphocyte transformation test, beryllium disease can be identified at early stages. Harris et al25 described CT appearances in eight patients with chronic berylliosis. In two patients the CT scan showed widespread nodularity. In one patient CT showed linear,

Figure 15. Organizing pneumonia with siderosis. HRCT scan at mediastinal windows (level, 30 H; width, 300 H) shows pulmonary mass with high-attenuation material. The transbronchial biopsy specimen from the mass showed organization with siderosis.

54

AKIRA

associated with development of pulmonary fibrosis. At present the cause of aluminuminduced fibrosis remains unresolved. It may be caused by concomitant exposure to free silica and not to the aluminum oxide in some cases. A significant retention of aluminum fibers in biologic samples from subjects working in the aluminum industry is identified.55 It is thought that metallic aluminum powder and aluminum oxide can cause lung disease after high-level occupational exposure. Desquamative interstitial pneumonia,26 a granulomatous lung reaction,19 and pulmonary alveolar proteinosis, 40 which develop after exposure to fumes from aluminum welding, also have been described. Radiographically and pathologically, the fibrosis usually is most severe in the upper lung zones. The author previously described HRCT findings in six cases of aluminum pneumoconiosis. They could be categorized into three forms: (1) predominantly reticular fibrosis (N  2); (2) predominantly nodular fibrosis (N  2); and (3) upper lung fibrosis (N  2). Fibrosis in five of the six patients occurred predominantly in the upper zone. In two patients with predominantly reticular fibrosis, HRCT findings were similar to those of usual interstitial pneumonia, and honeycomb formation was found (Fig. 16). In one of two patients with predominantly nodular fibrosis, HRCT depicted ill-defined centrilobular nodules diffusely throughout both lungs. In the other patient with predominantly nodular fi-

brosis, HRCT findings were similar to those of simple silicosis with type q or r.4 Recently, Kraus et al32 described the early stage of aluminum dust–induced changes that could be diagnosed for the first time using HRCT in a worker exposed long-term to high levels of aluminum dust. The HRCT findings were characterized by small, centrilobular, nodular opacities and slightly thickened interlobular septae. HARD METAL LUNG DISEASE Hard metal is produced by metallurgical blending of tungsten and carbon, with cobalt used as a binder. It is suggested that cobalt rather than tungsten is responsible for hard metal disease. Three main respiratory effects are produced by exposure to hard metal: (1) reversible airways obstruction, (2) hypersensitivity pneumonitis or alveolitis, and (3) pulmonary fibrosis.41 The typical histologic patterns of pulmonary fibrosis caused by hard metal consist of interstitial fibrosis and interstitial inflammation that may resemble findings in desquamative interstitial pneumonia or giant cell interstitial pneumonia. Hard metal pulmonary fibrosis is rare and the report of the CT appearances is limited. The HRCT findings in two hard metal workers were bilateral airspace consolidation or ground-glass attenuation, which were mainly panlobular and multilobular in character,

Figure 16. HRCT scans obtained in a 52-year-old man with a history of exposure to aluminum for 7 years. Reticular hyperattenuation diffusely distributes throughout the lungs. Traction bronchiectasis, dilated air bronchiolograms, and honeycomb cysts are identified in the areas of increased attenuation.

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

with parenchymal distortion (Fig. 17). Traction bronchiectasis and dilated air bronchiolograms were evident in the consolidation. The lobular areas of attenuation were characterized by the distinct margins produced by interlobular septae. Small branching areas of increased attenuation were present to some degree, corresponding to peribronchiolar fibrosis histologically.4

HYPERSENSITIVITY PNEUMONITIS Hypersensitivity pneumonitis is an immunologic disorder characterized by granulomatous interstitial pneumonia associated with exposure to occupational or environmental antigens. Disease entities include farmer ’s lung, bird-fancier’s lung, mushroom worker’s lung, bagassosis, air conditioner lung, cheese washer ’s lung, malt worker ’s lung, maple bark disease, pituitary snuff taker ’s lung, woodworker’s lung, and so forth. A unique form of hypersensitivity pneumonitis, summer-type hypersensitivity pneumonitis, in which signs and symptoms appear in the summer and subside spontaneously in midautumn, is seen in Japan.6 The clinical manifestations of hypersensitivity pneumonitis are classically divided into three syndromes: (1) acute, (2) subacute, and (3) chronic. Significant overlap, however, of-

55

ten exists between syndromes. Acute hypersensitivity pneumonitis typically occurs 4 to 6 hours after exposure and includes fever, chill, cough, and dyspnea. A subacute form of the disease occurs weeks to months after continued exposure. The symptoms in the subacute disease are similar to, but are less severe than, those in the acute disease. Longterm, low-level exposure may lead to insidious chronic disease.37, 52 Rarely, after repeated acute episodes, chronic pulmonary fibrosis may occur (Fig. 18). The HRCT characteristics of hypersensitivity pneumonitis vary with the stage of the disease. In the acute stage, the CT findings are diffuse heterogeneous or homogeneous opacities. In the subacute stage, they include small centrilobular ground-glass or nodular opacities and patchy air-space opacification. In transition from acute to subacute disease, poorly defined air-space opacities may be replaced by well-defined reticular or nodular opacities.6, 37 Lobular areas of low attenuation may be seen in the affected lung. These areas of decreased attenuation are thought to be areas of air trapping. Expiratory CT scans are useful for detecting air trapping.53 In the chronic stage, HRCT findings include irregular fibrotic opacities, traction bronchiectasis, and honeycombing mimicking idiopathic pulmonary fibrosis. The presence of centrilobular nodules on HRCT helps distinguish chronic

Figure 17. HRCT scans obtained in a 56-year-old man with a history of exposure to hard metal for 14 years. Multilobular ground-glass attenuation and consolidation with shrinkage are seen. The lobules adjacent to the lobular areas of hyperattenuation are hyperaerated. Irregular thickening of interlobular septae, nodular irregularity of pleural surface, and a few centrilobular branching lines also are seen.

56

AKIRA

Figure 18. Initial (A) and follow-up (B) HRCT scans at the level of right bronchus intermedius in a 39-year-old woman with summer-type hypersensitivity pneumonitis. She repeated acute episodes in the summer for several years. A, Diffusely distributed centrilobular nodules and patchy ground-glass opacity are seen. B, Follow-up scan obtained 6 years later demonstrates progression of parenchymal changes, forming honeycomb cysts, traction bronchiectasis, and bullae. Note multilobular areas of low attenuation.

hypersensitivity pneumonitis from idiopathic pulmonary fibrosis.14 Lobular areas of low attenuation may be common in chronic hypersensitivity pneumonitis and uncommon with idiopathic pulmonary fibrosis. The fibrosis of chronic hypersensitivity pneumonitis is often situated predominantly in the upper or middle lung zones or shows no zonal predominance.3, 36 The HRCT is more sensitive and more accurate than plain radiography and conventional CT in the evaluation of parenchymal abnormalities in hypersensitivity pneumonitis. HRCT can show perenchymal abnormalities in patients with hypersensitivity pneumonitis who have a normal chest radiograph.6, 37 Follow-up CT scans may show micronodular abnormalities even after the chest radiograph returns to normal.6 Underlying airway pathology in patients with diseases, such as emphysema or SwyerJames syndrome, may influence the distribution of hypersensitivity pneumonitis.47 In patients with centrilobular emphysema, areas of ground-glass attenuation that surround emphysema may mimic fibrosis with honeycombing and centrilobular nodules may be ambiguous. Hypersensitivity pneumonitis is uncommon in smokers.

CHEMICAL PNEUMONITIS Chemical pneumonitis is caused by exposure to toxic fumes of such gases as sulfur dioxide, ammonia, chlorine, phosgene, oxides of nitrogen, and ozone. Workers in a variety of occupations are exposed to toxic agents in the form of fumes and vapors. A highly soluble gas, such as ammonia or sulfur dioxide, is absorbed in the upper respiratory tract. A less soluble gas, such as nitrogen dioxide, is not removed in the upper passages and reaches the more peripheral areas of the respiratory tree. Inhalation of the toxic agent causes direct irritation and inflammation of the tracheobronchial tree. A large exposure may result in pulmonary edema. Ammonia is a colorless, highly soluble, extremely irritant alkaline gas with a characteristic pungent odor. Because of its high solubility ammonia causes chemical burns of eyes, skin, oropharynx, and upper respiratory tract. The severity of the injury is directly related to the concentration of the ammonia and the duration of exposure. Those injured can be divided into three groups depending on the severity of respiratory injury. The mild group may present with inflammation of the conjunctiva and upper respiratory passages. They may also complain of pain and

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE

hoarseness but are not in respiratory distress. In the moderate injury group these signs are exaggerated. Dysphagia, a productive cough, and some degree of respiratory distress usually are present. Those in the severe group frequently have been overcome by the ammonia and are unable to remove themselves from the area of contamination. They are in severe respiratory distress and have evidence of pulmonary edema.15 Close et al18 separated 12 patients exposed to anhydrous ammonia as a result of the same accident into two groups according to history and clinical course. One group of patients who sustained exposure to high concentrations of ammonia over a short period of time manifested upper airway obstruction and required early intubation or tracheostomy. These patients recovered with few pulmonary sequelae. The second group of patients who were exposed to lower concentrations of gas over a prolonged period of time did not manifest upper airway obstruction; however, significant long-term pulmonary sequelae were manifested. Bronchiectasis or bronchiolitis obliterans have developed in several survivors. Chlorine is a heavy irritating gas with a characteristic odor. Chlorine is intermediate in solubility and affects the lower respiratory tract more often than does ammonia. Acute exposure of humans and animals to high concentrations of chlorine gas is known to produce bronchiolar and alveolar-capillary

57

damage that is associated clinically with necrotizing bronchiolitis, bronchitis, and pulmonary edema. HRCT scans in patients who are exposed to chlorine gas demonstrate diffuse centrilobular nodular areas of ground-glass attenuation or patchy areas of ground-glass attenuation (Fig. 19). Nitrogen dioxide may be encountered in a wide variety of industrial situations. Silo filler’s disease is an acute lung injury caused by inhalation of nitrogen dioxide in or near an agricultural silo. The sudden death from exposure to silo gas is caused by nitrogen oxide or by asphyxiation. Delayed symptoms begin several hours to days after the episode. The injury is diffuse alveolar damage and pulmonary edema. Sometimes, bronchiolitis obliterans occurs 2 to 6 weeks after exposure.20 Bronchiolitis obliterans may appear as hyperinflation, mosaic patterns, centrilobular nodules, or bronchiectasis on HRCT. Expiratory HRCT may be useful in the detection of bronchiolitis obliterans. SUMMARY The most common of the pneumoconioses are silicosis, CWP, and asbestosis. The former two are characterized by the presence of small nodular opacities predominantly distributed in the upper zones of the lung. The small nodular opacities are classified into two pat-

Figure 19. HRCT scan in a patient who was exposed to chlorine gas. Centrilobular nodular areas of ground-glass attenuation are distributed diffusely throughout the lungs. The transbronchial lung biopsy showed edematous thickening of alveolar walls with cellular infiltration and intra-alveolar fibrin exudation.

58

AKIRA

terns on HRCT: (1) ill-defined fine branching lines and (2) well-defined discrete nodules. Asbestosis demonstrates thickened interlobular and intralobular lines, subpleural dot-like or curvilinear opacities, and honeycombing on HRCT, predominantly distributed in the bases of the lungs. Although HRCT findings of other pneumoconioses are variable and nonspecific, there are predominant and characteristic findings for each type of pneumoconiosis. HRCT is useful in achieving more accurate categorization of the parenchymal changes in each type of pneumoconiosis. References 1. Aberle DR, Gamsu G, Ray CS, et al: Asbestos-related pleural and parenchymal fibrosis: Detection with thin-section CT. Radiology 166:729–734, 1988 2. Aberle DR, Gamsu G, Ray CS, et al: High-resolution CT of benign asbestos-related disease: Clinical and radiographic correlation. AJR Am J Roentgenol 151:883–891, 1998 3. Adler BD, Padley SPG, Muller NL, et al: Chronic hypersensitivity pneumonitis: High-resolution CT and radiographic features in 16 patients. Radiology 185:91–95, 1992 4. Akira M: Uncommon pneumoconioses: CT and pathologic findings. Radiology 197:403–409, 1995 5. Akira M, Higashihara T, Yokoyama K, et al: Radiographic type p pneumocontosis: Thin-section CT. Radiology 171:117–123, 1989 6. Akira M, Kita N, Higashihara T, et al: Summer-type hypersensitivity pneumonitis: Comparison of highresolution CT and plain radiographic findings. AJR Am J Roentgenol 158:1223–1228, 1992 7. Akira M, Yamamoto S, Yokoyama K, et al: Asbestosis: Thin-section CT-pathologic correlation. Radiology 176:389–394, 1990 8. Akira M, Yokoyama K, Yamamoto S, et al: Early asbestosis: Evaluation with thin-section CT. Radiology 178:409–416, 1990 9. Al-Jarad N, Strickland B, Pearson HC, et al: High resolution computed tomographic assessment of asbestosis and cryptogenic fibrosing alveolitis: A comparative study. Thorax 47:645–650, 1992 10. Arai K, Takashima T, Matsui O, et al: Transient subpleural curvilinear shadow caused by pulmonary congestion. J Comput Assist Tomogr 14:87–88, 1990 11. Bergin CJ, Castellino RA, Blank N, et al: Specificity of high-resolution CT findings in pulmonary asbestosis: Do patients scanned for other indications have similar findings. AJR Am J Roentgenol 163:551–555, 1994 12. Bergin R, Ostiguy G, Fillion R, et al: Computed tomography scan in the early detection of silicosis. Am Rev Respir Dis 144:697–705, 1991 13. Bergin R, Ostiguy G, Fillion R, et al: CT in silicosis: Correlation with plain films and pulmonary function tests. AJR Am J Roentgenol 146:477–483, 1986 14. Buschman DL, Gamsu G, Waldron JA Jr, et al: Chronic hypersensitivity pneumonitis: Use of CT in diagnosis. AJR Am J Roentgenol 159:957–960, 1992 15. Caplin M: A Ammonia-gas poisoning: Forty-seven cases in a London shelter. Lancet 2:95–96, 1941

16. Charr R: Respiratory disorders among welders. Am Rev Tuberc Pulm Dis 71:877–884, 1955 17. Churg A: Occupational lung disease. In Thurlbeck WM, Churg A (eds): Pathology of the Lung, ed 2. New York, Thieme, 1995, pp 851–929 18. Close LG, Catlin FI, Cohn AM: Acute and chronic effects of ammonia burns of the respiratory tract. Arch Otolaryngol 106:151–158, 1980 19. De Vuyst P, Dumortier P, Schandene L, et al: Sarcoidlike lung granulomatosis induced by aluminum dusts. Am Rev Respir Dis 135:493–497, 1987 20. Douglas WW, Hepper NG, Colby TV: Silo-filler’s disease. Mayo Clin Proc 64:291–304, 1989 21. Filipenko D, Wright JL, Churg A: Pathologic changes in the small airways of the guinea pig after amosite asbestos exposure. Am J Pathol 119:273–279, 1985 22. Friedman AC, Fiel SB, Radecki PD, et al: Computed tomography of benign pleural and pulmonary parenchymal abnormalities related to asbestos exposure. Semin Ultrasound CT MR 11:393–408, 1990 23. Gamsu G, Salmon CJ, Warnock ML, et al: CT quantification of interstitial fibrosis in patients with asbestosis: A comparison of two methods. AJR Am J Roentgenol 164:63–68, 1995 24. Gibbs AR: Pathological reactions of the lung to dust. In Morgan WKC, Seaton A (eds): Occupational Lung Diseases, ed 3. Philadelphia, WB Saunders, 1995, pp 127–157 25. Harris KM, McConnochie K, Adams H: The computed tomographic appearances in chronic berylliosis. Clin Radiol 47:26–31, 1993 26. Herbert A, Sterling G, Abraham J, et al: Desquamative interstitial pneumonia in an aluminum welder. Hum Pathol 13:694–699, 1982 27. International Labour Office: Guidelines for the use of ILO international classification of radiographs of pneumoconiosis, revised ed. International Labour Office Occupational Safety and Health Series No. 22 (rev 80). Geneva, International Labour Office, 1980 28. Jones RN, Weill H, Parkes WR: Disease related to non-asbestos silicates. In Parkes (ed): Occupational Lung Disorders, ed 3. London, Butterworths, 1994, pp 536–570 29. Kee ST, Gamsu G, Blanc P: Causes of pulmonary impairment in asbestos-exposed individuals with diffuse pleural thickening. Am J Respir Crit Care Med 154:789–793, 1996 30. Kinsella M, Muller NL, Vedal S, et al: Emphysema in silicosis. Am Rev Respir Dis 141:1497–1500, 1990 31. Kleinerman J, Green F, Harley RA, et al: Pathology standards for coal workers’ pneumoconiosis. Arch Pathol Lab Med 103:375–432, 1979 32. Kraus T, Schaller KH, Angerer J, et al: Aluminum dust-induced lung disease in the pyro-powder-producing industry: Detection by high-resolution computed tomography. Int Arch Occup Environ Health 73:61–64, 2000 33. Kubota H, Hosoya T, Kato M, et al: Plate-like atelectasis at the corticomedullary junction of the lung: CT observation and hypothesis. Radiat Med 1:305–310, 1983 34. Lynch DA, Gamus G, Aberle DR: Conventional and high resolution computed tomography in the diagnosis of asbestos-related diseases. Radiographics 9:523– 551, 1989 35. Lynch DA, Gamus G, Ray CS, et al: Asbestos-related focal lung masses: Manifestations on conventional and high-resolution CT scans. Radiology 169:603– 607, 1988

HRCT IN THE EVALUATION OF OCCUPATIONAL AND ENVIRONMENTAL DISEASE 36. Lynch DA, Newell JD, Logan PM, et al: Can CT distinguish hypersensitivity pneumonitis from idiopathic pulmonary fibrosis? AJR Am J Roentgenol 165:807–811, 1995 37. Matar LD, McAdams P, Sporn TA: Hypersensitivity pneumonitis. AJR Am J Roentgenol 174:1061–1066, 2000 38. McConnochie K, Green PHY, Vallyathan V, et al: Interstitial fibrosis in coal workers—experience in Wales and West Virginia. Ann Occup Hyg 32(suppl 1):553–560, 1988 39. McLoud TC, Woods BO, Carrington CB, et al: Diffuse pleural thickening in an asbestos-exposed population: Prevalence and causes. AJR Am J Roentgenol 144:9–18, 1985 40. Miller RR, Churg AM, Hutcheon M, et al: Pulmonary alveolar proteinosis and aluminum dust exposure. Am Rev Respir Dis 130:312–315, 1984 41. Morgan WKC: Other pneumoconioses. In Morgan WKC, Seaton A (eds): Occupational Lung Diseases, ed 3. Philadelphia, WB Saunders, 1995, pp 407–456 42. Naidich DP, Zerhouni EA, Siegelman SS: Computed Tomography of the Thorax. New York, Raven, 1984 43. Newman LS, Buschman DL, Newell JD, et al: Beryllium disease: Assessment with CT. Radiology 190:835–840, 1994 44. Padley SG, Adler BD, Staples CA, et al: Pulmonary talcosis: CT findings in three cases. Radiology 186:125–127, 1993 45. Parkes WR: Pneumoconiosis associated with coal and other carbonaceous materials. In Parkes WR (ed): Occupational Lung Disorders, ed 3. London, Butterworths, 1994, pp 340–410 46. Pilate I, Marcelis S, Timmerman H, et al: Pulmonary asbestosis; CT study of subpleural curvilinear shadow [letter]. Radiology 164:584, 1987

59

47. Primack SL, Remy-Jardin M, Remy J, et al: Highresolution CT of the lung: Pitfalls in the diagnosis of infiltrative lung disease. AJR Am J Roentgenol 167:413–418, 1996 48. Remy-Jardin M, Beuscart R, Sault MC, et al: Subpleural micronodules in diffuse infiltrative lung diseases: Evaluation with thin-section CT scans. Radiology 177:133–139, 1990 49. Remy-Jardin M, Remy J, Farre I, et al: Computed tomographic evaluation of silicosis and coal workers’ pneumoconiosis. Radiol Clin North Am 30:1155– 1175, 1992 50. Schwartz DA, Galvin JR, Dayton CS, et al: Determinants of restrictive lung function in asbestos-induced pleural fibrosis. J Appl Physiol 68:1932–1937, 1990 51. Shida H, Chiyotani K, Honma K, et al: Radiologic and pathologic characteristics of mixed dust pneumoconiosis. Radiographics 16:483–498, 1996 52. Silver SF, Muller NL, Miller RR, et al: Hypersensitivity pneumonitis: Evaluation with CT. Radiology 173:441–445, 1989 53. Small JH, Flower CDR, Traill ZC, et al: Air-trapping in extrinsic allergic alveolitis on computed tomography. Clin Radiol 51:684–688, 1996 54. Stark P, Jacobson F, Shaffer K: Standard imaging in silicosis and coal workers, pneumoconiosis. Radiol Clin North Am 30:1147–1154, 1992 55. Voisin C, Fisekci F, Buclez B, et al: Mineralogical analysis of the respiratory tract in aluminium oxideexposed workers. Eur Respir J 9:1874–1879, 1996 56. Ward S, Heyneman LE, Reittner P, et al: Talcosis associated with IV abuse of oral medications: CT findings. AJR Am J Roentgenol 174:789–793, 2000 57. Yoshimura H, Hatakeyama M, Otsuji H, et al: Pulmonary asbestosis: CT study of curvilinear shadow. Radiology 158:653–658, 1986 Address reprint requests to Masanori Akira, MD Department of Radiology National Kinki Chuo Hospital for Chest Disease 1180 Nagasone-cho Sakai City, Osaka 591–8555 Japan e-mail: [email protected]

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT OF THE LUNG II

HIGH-RESOLUTION CT OF DRUG-INDUCED LUNG DISEASE Jeremy J. Erasmus, MD, H. Page McAdams, MD, and Santiago E. Rossi, MD

Numerous agents, including cytotoxic and noncytotoxic drugs, have the potential to affect the lungs adversely. Since the initial reports in the early 1970s, drug-induced lung injury has become a common cause of both acute and chronic lung disease.6, 34, 35 Recognition of drug-induced lung disease can be difficult because the clinical manifestations are often nonspecific and may be attributed to infection, radiation pneumonitis, or recurrence of the underlying disease. Early recognition is important because undiagnosed injury can be progressive and fatal, whereas cessation of therapy can result in reversal of drug-induced lung injury. Because the radiologic manifestations of drug-induced lung injury frequently reflect the underlying histopathologic process, knowledge of these findings and of the drugs most frequently involved can facilitate diagnosis and institution of appropriate treatment. This article reviews the common histopathologic manifestations of drug-induced lung injury including diffuse alveolar damage (DAD), nonspecific interstitial pneumonia (NSIP), bronchiolitis obliterans organizing pneumonia (BOOP), eosinophilic pneumonia (EP), and pulmonary hemorrhage; the corresponding high-resolution CT (HRCT) manifestations; and the drugs most frequently involved.

HISTOPATHOLOGY Drug-induced lung injury can occur as a result of direct cytotoxic injury, oxidative injury, injury caused by intracellular deposition of phospholipid, or immune-mediated injury.7, 8, 30, 33 The lung’s response to injury is limited and the most common histopathologic manifestations of pulmonary toxicity are as follows. Diffuse Alveolar Damage Diffuse alveolar damage is a descriptive term for the sequence of events that occurs after acute, severe lung injury and that results in necrosis of type I pneumocytes and alveolar endothelial cells.27 It is a common manifestation of drug-induced lung injury and the drugs that most frequently cause DAD are bleomycin, busulfan, carmustine, cyclophosphamide, melphalan, mitomycin, and gold salts.19, 28 Histopathologically, DAD is divided into two phases: (1) an acute exudative phase and (2) an organizing reparative phase.19, 30 The exudative phase is most prominent in the first week after lung injury and is characterized by edema and intra-alveolar membranes composed of necrotic epithelial cells and

From the Department of Radiology, MD Anderson Cancer Center, Houston, Texas (JJE); Department of Radiology, Duke University Medical Center, Durham, North Carolina (HPM); and Department of Radiology, Fundacion Dr ‘‘Enrique Rossi,’’ Buenos Aires, Argentina (SER)

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

61

62

ERASMUS et al

Figure 1. Carmustine (BCNU)-induced diffuse alveolar damage (DAD) in a 54year-old man with lymphoma, progressive dyspnea, and decreased diffusing capacity for carbon monoxide (DLCO). High-resolution CT (HRCT) shows bilateral scattered ground-glass opacities and diffuse consolidation in the lower lobes, findings that are typical of early DAD. Transbronchial biopsy of left lower lobe revealed prominent hyaline membranes, alveolar septal thickening, and hyperplastic change of type II pneumocytes, findings that are characteristic of the early exudative phase DAD.

protein-rich fluid.19, 27, 30 The reparative phase typically occurs after 1 or 2 weeks and is characterized by proliferation of type II pneumocytes and fibrosis.19, 27, 30, 37 Depending on the severity of the lung injury, fibrosis can progress to honeycomb lung. Fibrosis is, however, often reversible and most patients who survive have minimal residual functional abnormalities.27 High-resolution CT in early DAD usually shows scattered or diffuse ground-glass opacities. Severe injury typically manifests as bilateral, symmetric consolidation (Fig. 1). Busulfan-related lung injury in particular has a tendency to manifest as diffuse, confluent consolidation because of extensive desquamation of cells into the alveolar spaces.35 Although fibrosis usually is not radiologically visible in the early phase of DAD, HRCT performed during the late, reparative phase can show fibrosis (Fig. 2).

Nonspecific Interstitial Pneumonia Nonspecific interstitial pneumonia is a term used to describe interstitial inflammation and fibrosis that does not fulfill the diagnostic criteria for usual interstitial pneumonia, desquamative interstitial pneumonia, or acute

interstitial pneumonia. It is the interstitial pneumonia most commonly associated with drug-induced lung disease and the drugs most frequently implicated are amiodarone, methotrexate, and carmustine.7, 17, 19, 24, 27, 30, 37 Gold salts and chlorambucil are less common causes of drug-induced NSIP.27 Histopathologically NSIP is characterized by scattered infiltrates of mononuclear inflammatory cells, mild interstitial fibrosis, and reactive hyperplastic type II pneumocytes.19 Nonspecific interstitial pneumonia usually manifests radiographically as diffuse heterogeneous opacities. 7, 20, 27 HRCT shows scattered areas of ground-glass opacity, focal areas of consolidation, and irregular linear opacities, usually in a basal distribution (Fig. 3).21 Bronchiolitis Obliterans Organizing Pneumonia Bronchiolitis obliterans organizing pneumonia is a term used to describe inflammatory infiltration and fibrosis that predominantly involves the distal airways. The drugs that are most commonly associated with this type of lung injury are bleomycin, methotrexate, cyclophosphamide, and gold salts.19, 27, 35 Amiodarone, nitrofurantoin, penicillamine,

HRCT OF DRUG-INDUCED LUNG DISEASE

63

Figure 2. Bleomycin-induced DAD in a 39-year-old man with testicular malignancy, nonproductive cough, and dyspnea. HRCT shows scattered ground-glass opacities and thickening of interlobular septa. Minimal architectural distortion and bronchiectasis suggest fibrosis caused by late-stage DAD. Note right pleural effusion. Transbronchial biopsy revealed interstitial and alveolar duct fibrosis and prominent reactive change in hyperplastic type II pneumocytes, findings that are characteristic of the late proliferative phase of DAD.

and sulfasalazine are less common causes of drug-induced BOOP. 35 Histopathologically BOOP is characterized by fibroblastic plugs (Masson bodies) within the respiratory bronchioles, alveolar ducts and adjacent alveolar spaces, and accumulation of lipid-laden macrophages in the distal air spaces.10

High-resolution CT usually shows scattered areas of ground-glass opacities or consolidation bilaterally that is often peripheral in distribution (Figs. 4 and 5). Poorly defined nodular opacities, centrilobular nodules, branching opacities, and bronchial dilation are common associated findings.10

Figure 3. Topotecan-induced nonspecific interstitial pneumonia (NSIP) in a 68year-old man with acute myeloid leukemia, dyspnea, and fever. HRCT reveals scattered ground-glass opacities and thickening of interlobular septa. Transbronchial biopsy of the right upper lobe revealed mild, immature fibrosis and mononuclear interstitial infiltrate, consistent with NSIP.

64

ERASMUS et al

Figure 4. Topotecan-induced bronchiolitis obliterans organizing pneumonia (BOOP) in a 45-year-old woman with small cell lung cancer and increasing dyspnea. HRCT shows peripheral ground-glass opacities and consolidation in the right lung. Wedge resection biopsy of the right upper lobe revealed scattered interstitial inflammation and occlusion of terminal bronchioles and alveolar ducts with plugs of loose connective tissue, findings that are consistent with BOOP.

Eosinophilic Pneumonia Eosinophilic pneumonia is a well-recognized manifestation of drug-induced lung injury. The drugs that are most commonly associated with this type of lung injury are peni-

cillamine, sulfasalazine, nitrofurantoin, paraaminosalicylic acid, and nonsteroidal anti-inflammatory drugs.19, 27, 30, 35 Histopathologically, EP is characterized by the accumulation of eosinophils and macrophages in the alveolar spaces.30 There is also usually an accompa-

Figure 5. Bleomycin-induced BOOP in a 44-year-old woman with Hodgkin’s lymphoma, nonproductive cough, dyspnea, and decreased DLCO. HRCT shows groundglass opacities and scattered areas of focal consolidation peripherally. Transthoracic biopsy confirmed BOOP and patient improved symptomatically and radiologically after discontinuation of bleomycin and corticosteroid therapy.

HRCT OF DRUG-INDUCED LUNG DISEASE

65

Figure 6. Nonsteroidal anti-inflammatory drug-induced eosinophilic pneumonia in a 76-year-old woman with cough and fever. A, Posteroanterior (PA) chest radiograph shows bilateral, peripheral heterogeneous opacities. B, HRCT shows scattered areas of consolidation with a peripheral distribution. Transbronchial biopsy of the left lower lobe showed filling of the alveoli with eosinophils and macrophages consistent with eosinophilic pneumonia.

nying infiltrate composed of eosinophils, lymphocytes, and plasma cells within the alveolar septa and adjacent interstitium. HRCT typically shows ground-glass opacities and consolidation that is typically peripheral and upper lobe in distribution (Fig. 6).22

with this type of lung injury are anticoagulants, amphotericin B, high-dose cyclophosphamide, mitomycin, cytarabine, and penicillamine.19, 27, 35 HRCT usually shows bilateral, scattered, or diffuse ground-glass opacities (Fig. 7).31

Pulmonary Hemorrhage

SPECIFIC DRUGS

Diffuse pulmonary hemorrhage is an uncommon complication of drug therapy. The drugs that are most commonly associated

Chemotherapeutic Drugs Chemotherapeutic drugs are an important cause of drug-induced lung injury, affecting

66

ERASMUS et al

Figure 7. Cytarabine-induced pulmonary hemorrhage in a 30-year-old man with acute leukemia, severe dyspnea, and decreased DLCO. HRCT shows scattered ground-glass opacities and consolidation. Transbronchial biopsy of the right upper lobe showed organizing hemorrhage and mild interstitial fibrosis.

up to 10% of patients receiving chemotherapy.7, 26, 27, 33, 38 Cyclophosphamide, busulfan, carmustine, bleomycin, and methotrexate are the most common chemotherapeutic drugs that cause lung injury. Recently, there has been increasing use of the taxoid derivatives

paclitaxel and docetaxel and other new chemotherapeutic agents, such as gemcitabine, topotecan, and vinorelbine, to treat malignancies of the breast, lung, and ovary. Preliminary experience indicates that many of these agents also can cause lung injury (Fig. 8).9, 32

Figure 8. Carbotaxol-induced lung disease in a 62-year-old man with small cell lung cancer, progressive dyspnea, and fever. HRCT shows areas of consolidation, thickening of interlobular septa, architectural distortion, and bronchiectasis. Diagnosis of drug-induced lung injury was based on clinical history, presentation, and exclusion of infection. Patient improved symptomatically following institution of corticosteroid therapy.

HRCT OF DRUG-INDUCED LUNG DISEASE

Cyclophosphamide, an immunosuppressive alkylating agent, is used to treat a wide range of malignancies and benign conditions, such as glomerulonephritis and Wegener ’s granulomatosis. Despite its widespread use, cyclophosphamide-induced lung injury is relatively rare.1 There are two distinct forms of cyclophosphamide-induced lung injury: (1) acute onset and (2) chronic onset. Acute-onset toxicity usually occurs during the first 6 months of therapy. Affected patients present with dyspnea, cough, and fever.26 Discontinuation of therapy is typically associated with a good prognosis.27, 35 Chronic-onset toxicity occurs several months or years after prolonged treatment with cyclophosphamide. Clinical onset is insidious with progressive dyspnea and cough.26 DAD is the most histopathologic common manifestation of cyclophosphamide-induced lung disease. 4, 26, 27, 39 NSIP and BOOP are uncommon manifestations of injury.27 Busulfan, an immunosuppressive alkylating agent, is used in the treatment of myeloproliferative disorders, primarily chronic myelogenous leukemia. The incidence of busulfan-induced lung disease is approximately 4% and toxicity typically occurs in patients who have received a total dose of more than 500 mg.1, 7, 14 Lung injury can, however, occur at low doses if the patient has had prior cytotoxic therapy.7, 36 DAD is the most common

67

histopathologic manifestation of busulfan-induced lung disease.16, 23, 27 NSIP is less common and typically occurs only after prolonged administration of high doses of busulfan.12, 27 Carmustine, a nitrosourea, is frequently used to treat central nervous system malignancies, lymphoma, melanoma, and gastrointestinal malignancies.1 There is a direct relationship between cumulative dose and lung injury with the incidence increasing to 50% if the cumulative dose is more than 1.5 g/m2.7, 25 Lung injury can, however, occur at low doses if the patient has had thoracic radiation. DAD is the most common histopathologic manifestation of carmustine-induced lung disease; NSIP is a less common manifestation (Fig. 9).17, 18 Bleomycin, an antineoplastic antibiotic, is used primarily in the treatment of lymphomas and squamous and testicular carcinomas. Bleomycin-induced lung injury occurs in up to 5% of treated patients, although there is a markedly increased risk if the total cumulative dose is more than 450 to 500 units. 7 The risk of developing lung injury is also increased in the elderly; in patients receiving oxygen, concomitant chemotherapeutic drugs, or thoracic irradiation; and in patients in whom bleomycin is readministered within 6 months of discontinuation. 7 Bleomycininduced lung disease usually manifests clini-

Figure 9. Acute bischloroethyl-nitrosourea (BCNU)-induced lung disease in a 23year-old woman with an astrocytoma, dyspnea, and decreased DLCO. HRCT filmed with narrow window settings (level  675, window  650) shows ground-glass opacities bilaterally. Diagnosis of drug-induced lung injury was presumptive as sputum cultures were negative for infection and patient’s symptoms resolved with cessation of BCNU therapy and administration of corticosteroids.

68

ERASMUS et al

cally as slowly progressive dyspnea and cough developing over several weeks to months.6 Although mortality rates up to 25% have been reported, early detection may result in an improvement in prognosis because mild injury can respond to discontinuation of bleomycin. 6 It has consequently been suggested that HRCT may be useful in evaluating patients who have normal radiographs but have clinical findings suggestive of lung injury.1, 7, 27 DAD is the most common histopathologic manifestation of bleomycin-induced lung disease, with NSIP and BOOP being less common (Fig. 10).33 Bleomycin has been reported to manifest occasionally as nodules that mimic metastases. These nodules are usually peripheral in location, 5 mm to 3 cm in diameter, and either sharply or poorly marginated (Fig. 11).1, 7, 14, 15 Methotrexate, a folic acid analogue, is used to treat a wide range of malignant and benign conditions including lung and breast carcinoma, osteosarcoma, advanced stage nonHodgkin’s lymphoma, psoriasis, and severe rheumatoid arthritis. Methotrexate-induced lung injury occurs in up to 10% of treated patients and toxicity is not related to the du-

ration of therapy or total cumulative dose.7 Although both acute and chronic presentations have been described, the typical manifestation of methotrexate-induced lung injury is subacute with symptoms usually occurring within months of starting therapy.7 The prognosis is good with most patients improving despite continuation of therapy.7, 38 NSIP is the most common histopathologic manifestation of methotrexate-induced lung disease, with BOOP and DAD being less common (Fig. 12).5, 11 Pleural effusions can occur in association with manifestations of lung injury because of development of acute pleuritis.40, 41 Cardiovascular Drugs Amiodarone, an iodinated benzofuran derivative, is used to treat ventricular tachyarrhythmias and up to 10% of treated patients develop lung injury.8, 20, 28 Most cases of amiodarone-induced lung injury occur when the daily dose is greater than 400 mg and injury is in fact rare in patients receiving less than 400 mg daily.8, 28 The risk of lung injury is also increased in elderly patients.28 Amiodarone-

Figure 10. Bleomycin-induced lung disease in a 68-year-old man with nonseminomatous germ cell malignancy, dyspnea, and decreased DLCO. CT shows subtle linear and ground-glass opacities peripherally (arrows). Diagnosis of bleomycin-induced lung injury was based on clinical findings, exclusion of infection, and the temporal relationship to bleomycin therapy. Patient improved symptomatically after discontinuation of bleomycin and treatment with corticosteroids.

HRCT OF DRUG-INDUCED LUNG DISEASE

Figure 11. Bleomycin-induced lung disease in a 35-year-old man with nonseminomatous germ cell malignancy, cough, dyspnea, and decreased DLCO. Chest CT shows small, poorly defined pulmonary nodules (arrows) new from prior CT. Transthoracic needle aspiration biopsy was negative for malignancy, and nodules resolved after cessation of bleomycin therapy.

Figure 12. Methotrexate-induced lung disease in a 41-year-old woman with rheumatoid arthritis, dyspnea, and decreased DLCO. HRCT shows scattered ground-glass opacities with associated thickening of interiobular septa. Lung biopsy showed NSIP consistent with methotrexate-induced lung disease.

69

70

ERASMUS et al

Figure 13 Amiodarone-induced lung disease in a 77-year-old man on a 400-mg daily dose of amiodarone presenting with progressive dyspnea. HRCT shows consolidation, ground-glass opacities, and thickening of interiobular septa. Mild architectural distortion and bronchiectasis suggest fibrosis. Bronchoalveolar lavage excluded infection. Transbronchial biopsy was nondiagnostic. Autopsy specimens revealed NSIP consistent with amiodaroneinduced lung disease.

induced lung disease usually occurs within 1 to 10 months of initiation of therapy, although acute onset of injury occurs in up to one third of patients.1 Although death has been

reported in some patients with acute, severe lung injury, the prognosis is usually good with most patients improving after discontinuation of therapy.13 NSIP is the most common

Figure 14. Amiodarone-induced lung disease in a 71-year-old man with a history of ventricular arrhythmia and cough. Non–contrast enhanced CT shows focal masslike opacities in both lungs. Note high attenuation in right upper lobe opacity consistent with amiodarone-induced lung injury.

HRCT OF DRUG-INDUCED LUNG DISEASE

71

Figure 15. Nitrofurantoin-induced lung disease in a 77-year-old woman with chronic urinary tract infection, progressive dyspnea, and cough. HRCT shows focal, poorly marginated opacities and mild bronchiectasis (arrows). Transbronchial biopsy showed findings of NSIP and radiologic findings improved after discontinuation of nitrofurantoin.

histopathologic manifestation of amiodaroneinduced lung disease with BOOP being less common (Fig. 13).20 Pleural effusions can occur in association with manifestations of lung injury because of development of pleural inflammation.13 A unique feature of amiodarone-induced lung disease is the occurrence of focal, homogeneous opacities that are typically peripheral in location and of high-attenuation on CT (82 to 174 H) (Fig. 14).1, 8, 24 Although this finding of high-attenuation lesions is helpful if present, it is relatively uncommon. Antimicrobials Nitrofurantoin is an antiseptic used to treat urinary tract infections. Nitrofurantoin-induced lung injury is uncommon, although many cases have been reported because the drug is frequently prescribed. 8 Acute and chronic drug-induced injury has been described. Acute nitrofurantoin-induced lung injury usually occurs within 2 weeks of initiation of therapy and is most consistent with a hypersensitivity reaction, although pulmonary hemorrhage has been reported as a histopathologic finding.1–3 Patients present with dyspnea, cough, fever, and eosinophilia. 27 Prognosis is good with most patients recovering after discontinuation of therapy. 29

Chronic nitrofurantoin-induced lung injury is less common, usually occurs after months or years of nitrofurantoin use, and typically manifests clinically with insidious onset of dyspnea and cough. NSIP is the most common histopathologic manifestation of chronic toxicity (Fig. 15).5 SUMMARY Drug-induced pulmonary toxicity is increasing and early diagnosis is important because of the associated morbidity and mortality. Diagnosis is often difficult and is usually based on a history of drug therapy and exclusion of infection, radiation pneumonitis, and recurrence of the underlying disease. Although HRCT findings are frequently nonspecific, diagnosis can be facilitated by an understanding of the most common histopathologic and radiologic manifestations of drug-induced lung injury and knowledge of the drugs usually involved.

References 1. Aronchick JM, Gefter WB: Drug-induced pulmonary disorders. Semin Roentgenol 30:18-34, 1995 2. Boggess KA, Benedetti TJ, Raghu G: Nitrofurantoininduced pulmonary toxicity during pregnancy: A re-

72

3. 4. 5. 6. 7. 8. 9. 10. 11. 12.

13. 14. 15. 16. 17. 18. 19. 20. 21.

22.

ERASMUS et al port of a case and review of the literature. Obstet Gynecol Surv 51:367-370, 1996 Bucknall CE, Adamson MR, Banham SW: Non fatal pulmonary haemorrhage associated with nitrofurantoin. Thorax 42:475-476, 1987 Burke DA, Stoddart JC, Ward MK, et al: Fatal pulmonary fibrosis occurring during treatment with cyclophosphamide. BMJ 285:696, 1982 Cannon GW: Methotrexate pulmonary toxicity. Rheum Dis Clin North Am 23:917-937, 1997 Cooper JA Jr, Matthay RA: Drug-induced pulmonary disease. Dis Mon 33:61-120, 1987 Cooper JA Jr, White DA, Matthay RA: Drug-induced pulmonary disease. Part 1: Cytotoxic drugs. Am Rev Respir Dis 133:321-340, 1986 Cooper JA Jr, White DA, Matthay RA: Drug-induced pulmonary disease. Part 2: Noncytotoxic drugs. Am Rev Respir Dis 133:488-505, 1986 Eisenhauer EA, Vermorken JB: The taxoids: Comparative clinical pharmacology and therapeutic potential. Drugs 55:5-30, 1998 Epler GR, Colby TV, McLoud TC, et al: Bronchiolitis obliterans organizing pneumonia. N Engl J Med 312:152-158, 1985 Everts CS, Westcott JL, Bragg DG: Methotrexate therapy and pulmonary disease. Radiology 107:539-543, 1973 Feingold ML, Koss LG: Effects of long-term administration of busulfan: Report of a patient with generalized nuclear abnormalities, carcinoma of vulva, and pulmonary fibrosis. Arch Intern Med 124:66-71, 1969 Fraire AE, Guntupalli KK, Greenberg SD, et al: Amiodarone pulmonary toxicity: A multidisciplinary review of current status. South Med J 86:67-77, 1993 Ginsberg SJ, Comis RL: The pulmonary toxicity of antineoplastic agents. Semin Oncol 9:34-51, 1982 Glasier CM, Siegel MJ: Multiple pulmonary nodules: Unusual manifestation of bleomycin toxicity. AJR Am J Roentgenol 137:155-156, 1981 Heard BE, Cooke RA: Busulphan lung. Thorax 23:187-193, 1968 Holoye PY, Jenkins DE, Greenberg SD: Pulmonary toxicity in long-term administration of BCNU. Cancer Treat Rep 60:1691-1694, 1976 Iacovino JR, Leitner J, Abbas AK, et al: Fatal pulmonary reaction from low doses of bleomycin: An idiosyncratic tissue response. JAMA 235:1253-1255, 1976 Kay JM: Drug-induced lung disease. In Hasleton PS (ed): Spencer’s Pathology of the Lung. New York, McGraw-Hill, 1996, pp 551–595 Kennedy JI, Myers JL, Plumb VJ, et al: Amiodarone pulmonary toxicity: Clinical, radiologic, and pathologic correlations. Arch Intern Med 147:50-55, 1987 Kim TS, Lee KS, Chung MP, et al: Nonspecific interstitial pneumonia with fibrosis: High-resolution CT and pathologic findings. AJR Am J Roentgenol 171:1645-1650, 1998 Kim Y, Lee KS, Choi DC, et al: The spectrum of

23. 24.

25. 26.

27.

28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40.

41.

eosinophilic lung disease: Radiologic findings. J Comput Assist Tomogr 21:920-930, 1997 Kirschner RH, Esterly JR: Pulmonary lesions associated with busulfan therapy of chronic myelogenous leukemia. Cancer 27:1074-1080, 1971 Kuhlman JE, Teigen C, Ren H, et al: Amiodarone pulmonary toxicity: CT findings in symptomatic patients [published erratum appears in Radiology 178:287, 1991]. Radiology 177:121-125, 1990 Litam JP, Dail DH, Spitzer G, et al: Early pulmonary toxicity after administration of high-dose BCNU. Cancer Treat Rep 65:39-44, 1981 Malik SW, Myers JL, DeRemee RA, et al: Lung toxicity associated with cyclophosphamide use: Two distinct patterns. Am J Respir Crit Care Med 154:18511856, 1996 Meyers JL: Pathology of drug-induced lung disease. In Katzenstein AA, Askin FB (eds): Surgical Pathology of Non-neoplastic Lung Disease. Philadelphia, WB Saunders, 1997, pp 81–111. Myers JL, Kennedy JI, Plumb VJ: Amiodarone lung: Pathologic findings in clinically toxic patients. Hum Pathol 18:349-354, 1987 Padley SP, Adler B, Hansell DM, et al: High-resolution computed tomography of drug-induced lung disease. Clin Radiol 46:232-236, 1992 Pietra GG: Pathologic mechanisms of drug-induced lung disorders. J Thorac Imaging 6:1-7, 1991 Primack SL, Miller RR, Muller NL: Diffuse pulmonary hemorrhage: Clinical, pathologic, and imaging features. AJR Am J Roentgenol 164:295-300, 1995 Ramanathan RK, Reddy VV, Holbert JM, et al: Pulmonary infiltrates following administration of paclitaxel. Chest 110:289-292, 1996 Rosenow EC III, Limper AH: Drug-induced pulmonary disease. Semin Respir Infect 10:86-95, 1995 Rosenow EC III. The spectrum of drug-induced pulmonary disease. Ann Intern Med 77:977-991, 1972 Rosenow EC III, Myers JL, Swensen SJ, et al: Druginduced pulmonary disease: An update. Chest 102:239-250, 1992 Schallier D, Impens N, Warson F, et al: Additive pulmonary toxicity with melphalan and busulfan therapy. Chest 84:492-493, 1983 Smith GJ: The histopathology of pulmonary reactions to drugs. Clin Chest Med 11:95-117, 1990 Sostman HD, Matthay RA, Putman CE, et al: Methotrexate-induced pneumonitis. Medicine (Baltimore) 55:371-388, 1976 Spector JI, Zimbler H, Ross JS: Early-onset cyclophosphamide-induced interstitial pneumonitis. JAMA 242:2852-2854, 1979 Urban C, Nirenberg A, Caparros B, et al: Chemical pleuritis as the cause of acute chest pain following high-dose methotrexate treatment. Cancer 51:34-37, 1983 Walden PA, Mitchell-Weggs PF, Coppin C, et al: Pleurisy and methotrexate treatment. BMJ 2:867, 1977 Address reprint requests to Jeremy J. Erasmus, MD Department of Radiology MD Anderson Cancer Center 1515 Holcombe Boulevard Houston, TX 77098 e-mail: [email protected]

HIGH-RESOLUTION CT OF THE LUNG II

0033–8389/02 $15.00
.00

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS Jin Mo Goo, MD, and Jung-Gi Im, MD

Pulmonary tuberculosis (TB) is a major cause of morbidity and mortality worldwide, resulting in the greatest number of deaths caused by any one single infectious agent. The incidence of TB in the United States decreased continuously until 1985, when the trend reversed, mainly because of the increasing number of patients with AIDS.27 Radiologic manifestations of pulmonary TB can vary according to several host factors, including prior exposure to TB, age, and underlying immune status. In patients with normal immune function, radiologic manifestations can be categorized logically into two distinct forms: (1) primary TB, infection in individuals without prior exposure, and (2) postprimary TB, infection in individuals with prior exposure and acquired specific immunity. Chest radiography is the mainstay in the evaluation of pulmonary TB. CT is generally required to detect fine lesions overlooked on chest radiographs,12 to define equivocal lesions, or to evaluate complications. Moreover, high-resolution CT (HRCT) is useful in understanding the pathologic process of the disease and in determining disease activity in selected cases.12, 17, 24, 25 This article describes the characteristic CT findings of various forms of pulmonary TB and nontuberculous mycobacterial (NTMB)

infection according to immune status of the patients, and assesses the role of CT in the diagnosis and management of pulmonary TB. TUBERCULOSIS IN IMMUNOCOMPETENT HOSTS Primary Tuberculosis Primary TB is acquired by the inhalation of airborne organisms. In primary TB, because patients are not sensitized to tubercle bacilli, immune reaction for localization of the lesion and killing of the bacilli is lacking.15 Inability to localize the lesion results in a wide area of airspace consolidation accompanied by infiltration of neutrophils and occurrence of caseation necrosis. Until the cellular immune response develops, infection can progress locally and spread beyond the area of primary focus. Lymphatic spread of organisms to hilar and mediastinal lymph nodes occurs, along with a frequent, usually subclinical, hematogenous spread. Healing of larger parenchymal lesions may leave fibrous scars or persistent nodules known as tuberculomas, both of which may calcify. Typical CT findings of primary TB are dense, homogeneous air-space consolidations.15, 22 The middle lobe, the lower lobes,

From the Department of Radiology, Seoul National University College of Medicine, Seoul, South Korea

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

73

74

GOO & IM

or less frequently the anterior segment of an upper lobe are most commonly involved,5, 25 because of their greatest ventilation. Primary TB can, however, involve any lobe of the lung because the infection is established by inhalation of the droplets.25 Resolution of the parenchymal focus is typically slow, however, requiring 6 months to 2 years for complete clearing.28 Antituberculous chemotherapy speeds the resolution of radiographic findings, although paradoxical worsening in the first 3 months of therapy is not uncommon.28 HRCT may be useful to differentiate transient radiographic progression from true progression of pulmonary TB.1 According to Akira et al,1 dominant CT findings of transient radiographic progression ipsilateral or contralateral to the lesion were ground-glass opacity or consolidation, whereas the dominant CT findings of true worsening of TB were macronodules and centrilobular nodules, often with cavitation. Lymphadenopathy is the radiologic hallmark of primary TB. Although enlarged nodes occur in up to 92% of pediatric cases,22, 28 the prevalence of lymphadenopathy decreases with increasing age.28 The right paratracheal and hilar stations are the most common sites of nodal involvement in primary TB, although any combination, including bilateral hilar or isolated mediastinal lymphadenopathy, also may occur.19, 22 Lymphadenopathy usually is seen in association with parenchymal consolidation or atelectasis.30 On CT scans after intravenous injection of contrast media, enlarged lymph nodes typically show central low attenuation, which represents caseation necrosis, and peripheral rim enhancement, which represents inflammatory hypervascularity in granulomatous tissue (Fig. 1).19, 22, 26 TB lymphadenopathy typically resolves at a slower rate than the associated parenchymal disease and it usually resolves without significant radiographic sequelae, although nodal calcification may result.28 A cavitary lesion may be encountered in 8% to 29% of cases of primary TB.5, 41 Cavitation, however, is more likely a manifestation of progressive primary TB or of postprimary TB. Acute bronchogenic spread of TB occurs from breakdown of a lobar infection or from rupture of an infected lymph node into the bronchus. Typical CT findings in bronchogenic spread of pulmonary TB are centrilobular branching linear structure (tree-in-bud lesions); relatively poorly defined centrilobular nodules 2 to 3 mm in size; acinar shadows 4

Figure 1. Primary tuberculosis (TB) in a 12-year-old girl. Contrast-enhanced CT scan shows conglomerate, enlarged mediastinal lymph nodes containing central necrotic low-attenuation and peripheral rim enhancement.

to 10 mm in size; and large lobular consolidations.17, 25, 26

Postprimary Tuberculosis Postprimary TB results from reactivation of a previously dormant primary infection in 90% of cases35; a minority of cases represent a continuation of the primary disease.35 Exogenous reinfection rarely occurs. Despite the development of specific immunity resulting in healing with fibrosis of the granulomas, viable organisms often survive. Reactivation of dormant bacilli occurs during periods of immunodepression, malnutrition, and debilitation, or as a result of aging. Pulmonary lesions tend to be localized forming predominantly nodular lesions rather than wide areas of consolidation. The capability to localize the tuberculous lesion is provided by sensitization of the host by previous infection. Unlike primary TB, which is often an acute and selflimited disease, postprimary TB is typically a chronic, slowly progressive disease with high morbidity and mortality if not adequately treated. Although the radiographic findings of postprimary TB may overlap those of primary TB, distinguishing features include a predilection for the upper lobes, absence of

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS

lymphadenopathy, and a propensity for cavitation. Parenchymal involvement in postprimary TB most commonly manifests radiographically as heterogeneous opacities. Parenchymal opacities of postprimary TB are located in the apical and posterior segments of the upper lobes and less frequently in the superior segment of the lower lobes, often associated with cavitation. This upper lung predilection of postprimary disease is likely caused by either the high oxygen tension or the decreased lymphatic clearance of these regions.30 Cavitation is the hallmark of postprimary TB and begins from the oldest portion of the lesions, which are centrilobular in location (Fig. 2).17 As the disease progresses, several centrilobular cavities of 2 to 4 mm in size coalesce to form a larger cavity.17, 16 Bronchogenic spread of disease occurs when an area of caseous necrosis liquefies and communicates with the bronchial tree. Cavitation leads to expectoration of large numbers of bacilli with endobronchial spread to previously unaffected areas of the lung. The presence of a cavity is an important sign that indicates active disease. Cavitation in single or multiple sites is evident radiographically in 22% to 45% of cases of postprimary TB.17, 41 The thickness and sharpness of the walls of the cavities depends on the stage of disease; the outer margin of the cavity is usually indistinct and

75

the walls are generally thick in the stage of active caseous air-space consolidation. With antituberculous chemotherapy, the thickness of the wall decreases and the outer margin of the cavity becomes more distinct. Air–fluid levels have been reported to occur in 9% to 21% of tuberculous cavities (Fig. 3A).41 When communication of a cavity with a bronchus is closed, usually by fibrosis, the cavity disappears leaving a nodule or fibrotic bands (Fig. 3B). Cavitary lesions tend to leave more fibrotic change than those of noncavitary lesion.17 When tuberculous involvement of the bronchial wall results in sloughing off of the necrotic wall, so-called bronchiectatic cavities are formed. Persistent sterile cavitation is rare. Bronchogenic dissemination is the most common means of spread in postprimary TB.15, 17 Such a bronchogenic spread may occur in the absence of radiographically demonstrable cavitation. Radiographically, bronchogenic spread manifests as multiple, fluffy, 5- to 10-mm nodules distributed in a segmental or lobar distribution.15 HRCT is extremely helpful in understanding the route of spread and in understanding the pathologic nature of the tuberculous lesions. In the series by Im et al,17 centrilobular nodules or tree-in-bud lesions (branching linear structure 2 to 4 mm in diameter) on CT scans were the most common (95%) and earliest CT

Figure 2. Postprimary TB with a cavity in a 67-year-old woman. Contrast radiography of the lung specimen shows consolidation of a secondary lobule (arrows) containing a centrilobular cavity. (From Im JG, Itoh H, Shim YS, et al: Pulmonary tuberculosis: CT findings—early active disease and sequential change with antituberculous therapy. Radiology 186:653–660, 1993; with permission.)

76

GOO & IM

Figure 3. Postprimary TB in a 23-year-old woman. A, Initial high-resolution CT (HRCT) shows a thick-walled cavity with an air–fluid level in the right upper lobe (arrow). There are nodules and lobular consolidation in the right upper lobe. B, Follow-up (HRCT) after 6 months of antituberculous therapy reveals obliteration of the cavity with residual fibrotic bands (arrow). Also note partial resolution of the nodules and area of consolidation.

finding of bronchogenic dissemination (Fig. 4). These lesions had sharp margins and relatively high attenuation. With HRCT– pathologic correlation, centrilobular lesions have been shown to represent caseous necrosis within and around terminal and respiratory bronchioles (Fig. 5).17 Other HRCT find-

ings in decreasing order of frequency include bronchial wall thickening (73%); 5- to 8-mm poorly defined nodules (61%); cavity (51%); and lobular consolidation (41%). 17 Lobular and lobar consolidation on CT scans corresponded with large areas of air-space consolidation on pathologic examination (Fig. 6). Lobular consolidation consists of centrally located granulomas that contain caseation necrosis and marginal nonspecific inflammation. The areas of nonspecific inflammation appear relatively loose as compared with areas of caseation necrosis. In another study, ground-glass pattern, poorly marginated nodules, and septal thickening were present only before treatment.37 Hilar or mediastinal lymphadenitis is unusual for postprimary TB.

Tuberculoma

Figure 4. Bronchogenic spread of pulmonary TB in a 35-year-old woman. HRCT shows the thickening of the posterior segment of the right upper lobar bronchial wall (arrowheads) and multiple centrilobular nodular and branching linear structures at the bronchial territory (arrows). (From Im JG, Itoh H, Han MC: CT of pulmonary tuberculosis. Semin Ultrasound CT MR 16:420–434, 1995; with permission.)

Tuberculoma is defined as a round or oval granuloma usually less than 3 cm in diameter.30, 38 Most of these lesions remain stable for a long time, and many calcify.30 Enlargement of a nodule, or the development of a new one, suggests either reactivation of TB or a new process, such as carcinoma. Satellite lesions, which are small, discrete opacities in the immediate vicinity of the main lesion, may be identified in as many as 80% of cases (Fig. 7). 38 Calcified hilar lymphadenopathy supports the diagnosis.

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS

Figure 5. Bronchogenic spread of pulmonary TB in a 50-year-old woman. Radiograph of a lung section shows centrilobular branching linear lesions (arrowheads) that fill the bronchiolar lumen with terminal clubbing caused by peribronchiolar extension. Note also the contiguous branching linear lesions with a tree-in-bud appearance produced by caseous material within the respiratory bronchioles and alveolar ducts (arrows). (From Im JG, Itoh H, Shim YS, et al: Pulmonary tuberculosis: CT findings—early active disease and sequential change with antituberculous therapy. Radiology 186:653–660, 1993; with permission.)

Figure 6. Postprimary TB in a 34-year-old man. HRCT shows the approximately 2-mm-thick centrilobular branching linear structure (white arrow), which corresponds to caseous necrosis within the bronchiole shown in Figure 5. Also note that multiple discrete centrilobular nodules 2- to 3-mm in diameter (white arrowheads) are seen. Some secondary lobules are consolidated entirely, with (black arrowhead) and without (black arrows) a patent bronchiole. The bar at the top indicates 10 mm. (From Im JG, Itoh H, Shim YS, et al: Pulmonary tuberculosis: CT findings—early active disease and sequential change with antituberculous therapy. Radiology 186:653–660, 1993; with permission.)

77

78

GOO & IM

their distribution. Although the nodules are relatively uniform in size, the nodules in the upper lung zone tend to be larger than those in the lower lung zone.34 As the infection progresses, the miliary nodules can coalesce. Cavitation may occasionally occur in the nodules.35 Thickening of the intralobular and interlobular septa occurs in 40% to 45%.13, 34 Diffuse or localized ground-glass opacity is seen in 9% to 92% of the patients (Fig. 9).13, 34 Recognition of diffuse ground-glass opacity is important because sometimes it heralds acute respiratory distress syndrome.15 Healing of Tuberculous Lesions

Figure 7. Tuberculoma in a 43-year-old man. CT scan shows a mass with multiple satellite nodules (arrows) in the right upper lobe. Note the lobular hyperlucency distal to the mass.

The tuberculomas on CT scans usually are regular and smooth in outline but may have a rough edge (Fig. 7).25 Some irregularity or focal loss of definition may occur because of adjacent fibrotic changes. Cavitation in the lesion can be seen on CT scans. The lesions usually are low in attenuation and show no or minimal enhancement with administration of contrast medium.26 Calcification is found in 20% to 30% of the lesions and is usually nodular and diffuse. 25 Pulmonary tuberculoma can cause an increase in 2-[fluorine 18]fluoro-2-deoxy-D-glucose uptake at positron emission tomography.6

Healing of tuberculous lesions may occur in three different ways: (1) resolution, (2) fibrosis, and (3) calcification.15, 16 If lesions heal before necrosis has developed, complete resolution may occur; however, it is unusual. In almost all patients, caseation necrosis develops, which is surrounded by a varying amount of collagen tissue during the healing process, leaving residual fibrosis.15, 16 Dystrophic calcification is common in necrotic and fibrotic lesions. Because architectural distortion (fibrosis) and calcification are features found in both healed and active disease, radiographic determination of disease activity based on their presence is unreliable. On follow-up CT examinations, gradual

Miliary Tuberculosis In 2% to 6% of primary TB, massive, lymphohematogenous dissemination of tubercle bacilli results in clinical and radiographic evidence of miliary TB.25, 35, 41 Miliary TB can occur in postprimary TB when the immune mechanism of the host is overwhelmed. The characteristic CT findings of miliary TB consist of innumerable, 1- to 3-mm nodules scattered throughout both lungs (Fig. 8).13, 34 Typical miliary lesions may not be visible on chest radiograph until 3 to 6 weeks after hematogenous dissemination.3 The small nodules bear no relationship to the airways in

Figure 8. Miliary TB in a 53-year-old man. HRCT shows innumerable, 1- to 3-mm nodules in an even distribution throughout the lung.

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS

79

12 months of treatment. Those lesions result in various degrees of fibrosis.17 The findings of inactive or stable pulmonary TB on HRCT scans include areas of irregular bands, linear or stellate lesions, and calcified nodules along with distortion of bronchovascular bundles, bronchiectasis, and paracicatricial emphysema (Fig. 10).12, 17, 25 Emphysema results mainly from traction or from obstruction of air passages caused by fibrotic stenosis of the bronchioles.17 Areas of low attenuation can also occur as a result of air trapping caused by bronchiolar obstruction (Fig. 11). Chronic Infection and Destroyed Lung

Figure 9. Acute respiratory distress syndrome in a 36year-old woman with miliary TB. HRCT shows extensive ground-glass opacities, miliary nodules, and areas of airspace consolidation.

disappearance of lobular consolidation, poorly defined nodules, and centrilobular nodules or tree-in-bud lesions may be seen.17, 25 Resolution of lobular consolidation begins usually at the periphery, with eventual transformation into a poorly defined nodule, followed by the appearance of a centrilobular nodule or tree-in-bud lesions. Centrilobular nodules or tree-in-bud lesions visible on initial CT scans are no longer present after 5 to

Complete destruction of the whole or a major part of a lung can occur. It may result from a progressive primary infection or from inadequately treated prolonged postprimary TB. Unilateral involvement of an upper lobe is most frequently observed. Severe fibrosis with upper lobe volume loss, hilar retraction, and secondary tracheomegaly is seen.25 Bronchiectasis is almost always associated (see Fig. 10). Fungus ball (aspergilloma) is not an uncommon complication in chronic destructive pulmonary TB. It appears as intracystic or intracavitary soft tissue lesion associated with an internal sponge-like or marginal curvilinear air shadow (air meniscus) on CT scan (Fig. 12).

Figure 10. Chronic infection with a destroyed left lung in a 55-yearold woman. HRCT shows destruction of the left lung with multiple bronchiectatic cavities. Note also stellate lesion with parenchymal distortion in the right lower lobe.

80

GOO & IM

Figure 11. Postprimary TB in a 23-year old woman. A, Initial HRCT shows multiple compact small nodular and branching linear lesions in the left lower lobe with localized sparing of the normal lung (arrowheads). Note also the larger, poorly defined nodules (arrow). B, Follow-up CT scan obtained 9 months after A was obtained, and after antituberculous chemotherapy, shows air trapping with distortion of the bronchovascular arrangement at the previous bronchiolar lesions in A, leaving a localized area of normal lung (arrowheads) that corresponds with the normal lung shown in A. Note the absence of vascular distortion in the normal lung (arrow) in contrast to the surrounding hypoattenuated area. (From Im JG, Itoh H, Lee KS, et al: CT–pathology correlation of pulmonary tuberculosis. Crit Rev Diagn Imaging 36:227–285, 1995; with permission.)

Figure 12. Fungus balls in a 54-year-old woman. HRCT scan shows multiple low-attenuated masses with internal sponge-like shadows and air meniscus in the upper lobes (arrows). There is a calcified nodule in the right upper lobe.

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS

An apical cap is commonly observed on the chest radiograph in patients with chronic destructive lesions of the upper lobe. Hypertrophied extrapleural fat is mainly responsible for the apical cap, and to a lesser degree, pleural thickening with subpleural atelectasis and fibrotic lung.20 Thick apical cap is indicative of the chronicity of the inflammatory process.20 Tuberculous Airway Diseases The most common cause of inflammatory stricture of the bronchus is TB. Tracheobronchial TB has been reported in 10% to 20% of all patients with pulmonary TB.26, 32 CT findings of TB involving the trachea and proximal bronchi depend on the stage of the disease. In cases of active tuberculous tracheobronchitis, a circumferential narrowing of the lumen associated with thickening of the wall is seen (Fig. 13). The thickened airway wall usually is reversible with antituberculous therapy.4, 23, 32 Other CT findings include bronchial obstruction by a peribronchial cuff of soft tissue, by tuberculous lymphadenopathy, or by an intraluminal polypoid mass caused by granuloma formation.

81

In contrast to the active stage, CT findings of treated or healed tracheobronchial TB are smooth narrowing of the lumen and minimal thickening of the wall.4, 23, 32 The left main bronchus is involved more frequently in the fibrotic stage of the disease, whereas both main bronchi are equally involved in the active stage of the disease.23, 32 This fact may reflect an anatomic difference between the right and left main bronchus and impaired lymphatic drainage in the left main bronchial wall.32 PULMONARY TUBERCULOSIS IN IMMUNOCOMPROMISED PATIENTS Impaired host immunity has been regarded as a predisposing factor in TB. Known risk factors for development of active TB include conditions that are associated with defects in cell-mediated immunity, such as malnutrition; drug and alcohol abuse; malignancy; coexistent medical conditions (e.g., chronic renal failure, diabetes mellitus, and silicosis); and corticosteroid or other immunosuppressive therapy.14 When the cellular immune response of patients is lowered, lobar or segmental pneumonia may break down into

Figure 13. TB airway disease in a 29-year-old woman. A, Contrast-enhanced CT scan shows irregular luminal narrowing with wall thickening of the right upper lobar bronchus (arrows). Bronchoscopy revealed occlusion of the right upper lobar bronchus with caseation necrosis. B, Right anterior oblique view of a shaded surface display of the tracheobronchial tree demonstrates irregular narrowing of the right main bronchus (arrowhead) and occlusion of the right upper lobar bronchus (arrow).

82

GOO & IM

multiple cavities. 35 Nonsegmental distribution of lesions also occurs.14 Worldwide, TB is the major opportunistic infection of HIV-infected individuals. In the study by Goodman,7 nearly 25% of patients with AIDS had TB. HIV infection is the most potent known risk factor for reactivation of latent Mycobacterium tuberculosis infection. Because HIV depletes T lymphocytes in patients with AIDS, local delayed-type hypersensitivity reaction is deficient, lacking in the ability to form granuloma, kill the bacilli, and localize the disease.2 Patients with AIDS are extremely vulnerable to tuberculous infection and the radiologic manifestations can be different from those of immunocompetent host because of altered immunity. Disseminated disease is more common in AIDS patients than in the immunologically competent host. In comparison with HIVseronegative patients, HIV-seropositive patients have a lower prevalence of parenchymal consolidation, cavitation, and postprimary pattern, and a higher prevalence of lymphadenopathy (50% to 74%); pleural effusions (21% to 38%); and miliary (17%) disease.10, 27 Radiologic manifestations of HIV-associated TB are dependent on the level of CD4 Tlymphocyte depletion.2, 7, 21 Patients with relatively intact cellular immune function present with symptoms similar to non–HIV-infected individuals, and TB generally remains localized to the lung. At severe levels of immunosuppression, CT findings are those of primary TB, regardless of prior exposure to TB (Fig. 14).2, 21 In patients with advanced HIV disease (CD4 T-lymphocyte count ⬍ 200/mm3), pulmonary TB is often accompanied by extrapul-

monary involvement, which most commonly takes the form of lymphadenitis, pleural effusion, or miliary disease.2, 21 Tuberculous lymphadenitis in the HIVseropositive population also may be associated with central low attenuation and peripheral enhancement on contrast-enhanced CT scans.27, 36 These findings are basically similar to those in patients without AIDS,19, 27 except that patients with severe immunosuppression tend to reveal more extensive necrosis, extranodal extension, and esophagonodal fistula.18 The esophagonodal fistula most commonly occurs at the subcarinal region and it can be recognized by the presence of abnormal gas within the mediastinum (Fig. 15).18 The presence of esophagonodal fistula does not alter treatment; antituberculous chemotherapy is the only effective treatment. NONTUBERCULOUS MYCOBACTERIAL INFECTIONS Depending on the specific organism and host, NTMB can cause a variety of human infections. Pulmonary manifestations are most common and usually are caused by Mycobacterium avium-intracellulare and M. kansasii.31, 40 M. xenopi, M. fortuitum, and M. chelonae are unusual pathogens that cause a spectrum of pulmonary, cervical lymph node, cutaneous, and soft tissue infections, with the potential for disseminated disease. Disseminated mycobacterial infection after bacille Calmette-Gue´rin vaccination has been reported.9 The NTMB are ubiquitous organisms that constitute part of the normal environmental

Figure 14. Pulmonary TB in a patient with AIDS. Contrastenhanced CT scan shows extensive air-space consolidation in the left upper lobe and an enlarged lymph node in paraaortic area (arrow).

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS

83

Figure 15. Esophagonodal fistula in a 28-year-old man. Contrast-enhanced CT scan shows amorphous gas collection (arrows) in the right paratracheal and retrotracheal area. Note multiple enlarged mediastinal lymph nodes and a left pleural effusion.

flora. Despite the high rate of exposure to these organisms there is a low rate of clinical infection.31 Delays in diagnosis are frequent because the findings may be subtle. Diagnosis of pulmonary NTMB infection is often difficult because isolation of the organism from sputum or bronchoalveolar lavage fluid may represent airway colonization, not infection. The best criteria for NTMB pulmonary disease are repeated positive sputum or bronchoalveolar lavage fluid cultures combined with clinical and radiographic signs of infection.31, 40 The NTMB pneumonia has a spectrum of clinical and radiographic manifestations that are sometimes indistinguishable from M. tuberculosis. In general, the radiographic appearance is not influenced by the specific mycobacterial species.31 The natural history is one of gradual progression, although a stable clinical course is frequently observed despite prolonged, persistent positive sputum cultures. Two main groups of patients with distinct demographic and clinical presentations have been identified. The first group are typically white men in their 60s and 70s who have underlying chronic lung disease, such as chronic obstructive pulmonary disease or pulmonary fibrosis.31 The characteristic radiographic findings are linear nodular opacities that are indistinguishable from postprimary TB.31, 40 These linear nodular lesions can remain radiographically stable for many years, although slow progression is possible. A marked fibrotic response associated with volume loss of the upper lobe with tracheal deviation and hilar

elevation occurs in approximately one third of patients.31, 40 Bronchiectasis is often present in areas of maximal disease. Cavitation occurs commonly (80% to 95%) and is frequently associated with pleural thickening (37% to 56%).31 Endobronchial spread occurs in 40% to 70% of cases31 and results in unilateral or bilateral diffuse scattered heterogeneous opacities. These nodular areas of increased opacity range from 5 to 15 mm in diameter and have a centrilobular distribution at CT (Fig. 16).33 Miliary disease is rare in immunocompetent patients. Lymphadenopathy and pleural effusion are uncommon.31, 40 The second group includes middle aged and elderly white women without underlying lung disease. Characteristic radiographic findings are scattered bronchiectasis and multiple centrilobular nodules of varying size, usually less than 1 cm (Fig. 17).11, 31, 33 These nodules are thought to represent NTMB granulomas and are indicative of infection, not colonization.31 The nodules tend to involve airways and are much better seen on HRCT scans than on conventional radiographs; often, the nodules have a ‘‘tree-in-bud’’ appearance.11, 31 The most common location for bronchiectasis is the middle lobe and lingula (Fig. 17).31, 33 There is no dominant apical distribution in contrast to postprimary TB. Serial CT has demonstrated worsening of bronchiectasis as the disease progresses,31, 33 suggesting that bronchiectasis may not only predispose to but also be caused by NTMB infection. Occasionally, NTMB infection results in solitary or multiple nodules, which are usually incidentally detected in asymptomatic pa-

84

GOO & IM

Patients with AIDS demonstrate features of NTMB infection that differ from those of other immunocompromised patients. Unlike normal hosts, the portal of entry in patients with AIDS is thought to be the gastrointestinal tract. The chest radiograph is often normal in AIDS patients with NTMB infection. Mediastinal and hilar lymphadenopathy are the most common radiographic findings. Cavitary lesions, atelectasis, and pleural effusions are distinctly uncommon. ROLE OF CT IN DIAGNOSIS AND MANAGEMENT OF TUBERCULOSIS Detection and Characterization of Parenchymal Tuberculosis

Figure 16. Nontuberculosis mycobacterial (NTMB) infection in a 65-year-old woman. HRCT shows multiple centrilobular nodules and branching linear opacities in the right upper lobe. Sputum cultures were repeatedly positive for Mycobacterium avium-intracellulare (MAI).

tients.31 Patients with achalasia have a radiographic appearance that most often resembles aspiration pneumonia. 31 CT demonstrates nonspecific patchy air-space disease. NTMB infection in achalasia is almost always associated with M. fortuitum–chelonei complex rather than with other atypical mycobacteria.31

Many patients with pulmonary TB do not need CT in the initial diagnosis of pulmonary TB. CT, which has a better accuracy than chest radiography in the diagnosis of primary TB, can allow prompt diagnosis25, 36 and help to start the treatment. CT can help identify or confirm the presence of findings that may be used to suggest the diagnosis of TB when the radiographic findings are inconspicuous but when TB is suspected clinically. Notably, chest radiographs remain normal in up to 15% of patients who have proved primary TB.41 Chest radiography also is an insensitive screening method for the detection of early stage of miliary TB and HIV-associ-

Figure 17. NTMB in a 58-year-old woman. HRCT shows bronchiectasis in the right middle lobe and lingular division of the left upper lobe. Note centrilobular nodules and segmental hypoattenuation in the right lower lobe (arrows). Cultures grew MAI.

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS

ated TB.3, 25, 41 Normal chest radiographs in AIDS patients with culture-positive sputum or bronchoscopy specimens have been reported in 14% to 40% of cases.8, 27 In primary TB, CT is used to help identify or confirm the presence of lymphadenopathy and to detect subtle parenchymal sites of primary infection that may be inconspicuous on chest radiographs and may suggest the diagnosis of TB (see Fig. 1). The extent of tuberculous involvement is clearly documented, occasionally showing involvement of a lobe not previously suspected on chest radiograph. CT is a useful adjunct to direct bronchoscopic visualization because it can accurately depict the bronchial abnormality.4, 26 On CT scans, involved bronchi are stenosed or obstructed, with a peribronchial cuff of soft tissue, bronchial wall thickening, or peribronchial lymphadenopathy (see Fig. 13).4, 26 CT is helpful in the diagnosis of miliary TB in patients who have fever of unknown origin and a chest radiograph that is normal. In the series by Woodring et al,41 the radiographic diagnosis of TB was initially correct in only 49% of cases, 34% for the diagnosis of primary TB and 59% for the diagnosis of postprimary TB. With CT, the diagnosis of pulmonary TB was correct in 91% of patients and TB was correctly excluded in 76% of patients.24 Among CT abnormalities suggestive of bronchogenic spread, only acinar nodules can be visualized on chest radiograph.29 Acinar nodules may appear as satellite nodules nearby a consolidation or excavated area or be located in lobar regions not involved by main foci of TB.37 In the series by Im et al,17 tree-in-bud lesions were the most common characteristic finding on CT obtained in patients with newly disseminated pulmonary TB (see Fig. 4). Because these lesions are not identified on chest radiographs, CT with HRCT technique can be recommended in case of suspicion of reactivation TB. Chest CT is more sensitive than chest radiography in the detection of small cavities, particularly ones in the apices, lung bases, and paramediastinal and retrocardiac locations. The cavitation typically occurs within areas of consolidation, although isolated cavities are sometimes seen. In the series by Im et al17 of 41 patients with active TB, the prevalence of cavities demonstrated on initial CT scan was 58%, whereas the prevalence of cavities demonstrated on radiographs was only 22%. Moreover, in cases of extensive fibrosis

85

with architectural distortion, HRCT can clearly distinguish cavities from paracicatricial emphysema or cystic bronchiectasis (see Fig. 10).17 This most often concerns the apical or posterior segment of the upper lobes. In case of suspected aspergilloma, CT is more sensitive and allows detection with exact location of occult or small aspergillomas (see Fig. 12). Rasmussen’s aneurysms and hypertrophied bronchial arteries can be diagnosed with a spiral volumetric acquisition with fine collimation and intravenous administration of contrast media.39 Determination of Disease Activity In addition to the diagnosis of tuberculous infection, HRCT is useful in determining disease activity. Determination of disease activity in patients with pulmonary TB usually depends on the detection of acid-fast bacilli on culture of sputum or bronchoalveolar lavage fluid. Acid-fast bacilli, however, are found in a limited number of patients (20% to 55%) with active pulmonary TB, in particular in patients with AIDS.25 Disease activity cannot be assessed accurately by chest radiography. Although active infection is most often associated with exudative lesions or cavitation on chest radiographs,30 fibroproductive lesions also indicate active disease.41 CT scans, especially HRCT scans, are superior to chest radiographs for assessing the disease activity of postprimary TB.12, 17, 25, 28 Parenchymal lesions having tree-in-bud lesions, centrilobular nodules, acinar shadows, and large lobular consolidations are considered active (see Figs. 3 to 6). Cavities, the most important radiologic evidence of activity, can be well depicted on CT scans (see Fig. 3A). In the series by Lee et al,24 80% of patients with active disease and 89% of those with inactive disease were correctly differentiated by this technique. The most frequent CT patterns of active TB were centrilobular branching linear opacities (92%), lobular consolidation (62%), acinar nodule (61%), cavity (36%), and ground-glass attenuation (35%). Areas of well-defined centrilobular nodule (95%), bronchiectasis (86%), and irregular lines (77%) were the most frequent CT findings observed in inactive pulmonary TB (see Fig. 10).24 In difficult cases, serial CT examinations may be necessary to visualize increase, stability, or decrease of the lesions. Total disappear-

86

GOO & IM

also helpful in the diagnosis and evaluation of the extent of cold abscess in the chest wall, which is a complication of tuberculous pleurisy.25 In cases of empyema with an air–fluid level, a bronchopleural fistula can be diagnosed with a volumetric acquisition and fine collimation (Fig. 18). Any fibrothorax is also correctly evaluated. In tuberculous pericarditis, CT may enable diagnosis at an early stage by showing the nature and amount of effusion, the thickened pericardium, and associated pulmonary parenchymal tuberculous lesions or mediastinal lymphadenitis abutting the pericardium. References

Figure 18. Bronchopleural fistula in a 76-year-old woman. CT scan demonstrates dilated airway (arrow) that communicates directly with an air collection in the right pleural space. Note irregular thickening of the visceral and parietal pleura.

ance of some lesions, such as centrilobular nodules, can be assessed with CT after treatment. Extrapulmonary Disease CT can help identify or confirm the presence of lymphadenopathy and lead to the diagnosis of tuberculous mediastinitis by showing the lymph nodes of central low attenuation with peripheral rim enhancement on contrast-enhanced scans (see Fig. 1). In patients with AIDS and lymphadenopathy containing central necrotic low attenuation, starting of antituberculous chemotherapy can be justified (see Fig. 14).36 Other mediastinal abnormalities, such as tuberculous fibrosing mediastinitis, a rare entity, also are clearly recognized with CT. CT can detect a small quantity of free pleural fluid not visible on chest radiographs. A true pleural thickening suspected on chest radiograph is clearly differentiated from a chronic loculated effusion and from adjacent parenchymal lesions. In all cases of tuberculous pleurisy, CT may help to detect focal areas of subpleural nodule or cavitation. It is

1. Akira M, Sakatani, Ishikawa H: Transient radiographic progression during initial treatment of pulmonary tuberculosis: CT findings. J Comput Assist Tomogr 24:426–431, 2000 2. Barnes PF, Bloch AB, Davidson PT, et al: Tuberculosis in patients with human immunodeficiency virus infection. N Engl J Med 324:1644–1650, 1991 3. Berger HW, Samortin TG: Miliary tuberculosis: Diagnostic methods with emphasis on the chest roentgenogram. Chest 58:586–589, 1970 4. Choe KO, Jeong HJ, Sohn HY: Tuberculous bronchial stenosis: CT findings in 28 cases. AJR Am J Roentgenol 155:971–976, 1990 5. Choyke PL, Sostman HD, Curtis AM, et al: Adult onset of pulmonary tuberculosis. Radiology 148:357– 362, 1983 6. Goo JM, Im JG, Do KH, et al: Pulmonary tuberculoma evaluated by means of FDG PET: Findings in 10 cases. Radiology 216:117–121, 2000 7. Goodman PC: Pulmonary tuberculosis in patients with acquired immunodeficiency syndrome. J Thorac Imaging 5:38–45, 1990 8. Greenberg SD, Frager D, Suster B, et al: Active pulmonary tuberculosis in patients with AIDS: Spectrum of radiographic findings including a normal appearance. Radiology 193:115–119, 1994 9. Han TI, Kim IO, Kim WS, et al: Disseminated BCG infection in a patient with severe combined immunodeficiency. Korean J Radiol 1:114–117, 2000 10. Harries AD: Tuberculosis in human immunodeficiency virus infection in developing countries. Lancet 335:387–390, 1990 11. Hartman TE, Swensen SJ, Williams DE: Mycobacterium avium-intracellulare complex: Evaluation with CT. Radiology 187:23–26, 1993 12. Hatipoglu ON, Osma E, Manisali M, et al: Highresolution computed tomographic findings in pulmonary tuberculosis. Thorax 51:397–402, 1996 13. Hong SH, Im JG, Lee JS, et al: High-resolution CT findings of miliary tuberculosis. J Comput Assist Tomogr 22:220–224, 1998 14. Ikezoe J, Kakeuchi N, Johkoh T, et al: CT appearance of pulmonary tuberculosis in diabetic and immunocompromised patients: Comparison with patients who had no underlying disease. AJR Am J Roentgenol 159:1175–1179, 1992 15. Im JG, Itoh H, Han MC: CT of pulmonary tuberculosis. Semin Ultrasound CT MR 16:420–434, 1995

CT OF TUBERCULOSIS AND NONTUBERCULOUS MYCOBACTERIAL INFECTIONS 16. Im JG, Itoh H, Lee KS, et al: CT-pathology correlation of pulmonary tuberculosis. Crit Rev Diagn Imaging 36:227–285, 1995 17. Im JG, Itoh H, Shim YS, et al: Pulmonary tuberculosis: CT findings-early active disease and sequential change with antituberculous therapy. Radiology 186:653–660, 1993 18. Im JG, Kim JH, Han MC, et al: Computed tomography of esophagomediastinal fistula in tuberculous mediastinal lymphadenitis. J Comput Assist Tomogr 14:89–92, 1990 19. Im JG, Song KS, Kang HS, et al: Mediastinal tuberculous lymphadenitis: CT manifestations. Radiology 164:115–119, 1987 20. Im JG, Webb WR, Han MC, et al: Apical opacity associated with pulmonary tuberculosis: High-resolution CT findings. Radiology 178:727–731, 1991 21. Jones BE, Young SM, Antoniskis D, et al: Relationship of manifestations of tuberculosis to CD4 cell counts in patients with human immunodeficiency virus infection. Am Rev Respir Dis 148:1292–1297, 1993 22. Kim WS, Moon WK, Kim I, et al: Pulmonary tuberculosis in children: Evaluation with CT. AJR Am J Roentgenol 168:1005–1009, 1997 23. Kim Y, Lee KS, Yoon JH, et al: Tuberculosis of the trachea and main bronchi: CT findings in 17 patients. AJR Am J Roentgenol 168:1051–1056, 1997 24. Lee KS, Hwang JW, Chung MP, et al: Utility of CT in the evaluation of pulmonary tuberculosis in patients without AIDS. Chest 110:977–984, 1996 25. Lee KS, Im JG: CT in adults with tuberculosis of the chest: Characteristic findings and role in management. AJR Am J Roentgenol 164:1361–1367, 1995 26. Lee KS, Song KS, Lim TH, et al: Adult-onset pulmonary tuberculosis: Findings on chest radiographs and CT scans. AJR Am J Roentgenol 160:753–758, 1993 27. Leung AN, Brauner MW, Gamsu G, et al: Pulmonary tuberculosis: Comparison of CT findings in HIV-seropositive and HIV seronegative patients. Radiology 198:687–691, 1996 28. Leung AN, Muller NL, Pineda PR, et al: Primary

29.

30. 31. 32. 33. 34. 35. 36.

37. 38. 39.

40. 41.

87

tuberculosis in childhood: Radiographic manifestations. Radiology 182:87–91, 1992 Long R, Maycher B, Dhar A, et al: Pulmonary tuberculosis treated with directly observed therapy: Serial changes in lung structure and function. Chest 113:933–643, 1998 McAdams HP, Erasmus J, Winter JA: Radiologic manifestations of pulmonary tuberculosis. Radiol Clin North Am 33:655–678, 1995 Miller WT Jr: Spectrum of pulmonary nontuberculous mycobacterial infection. Radiology 191:343–350, 1994 Moon WK, Im JG, Yeon KM, et al: Tuberculosis of the central airways: CT findings of active and fibrotic disease. AJR Am J Roentgenol 169:649–653, 1997 Moore EH: Atypical mycobacterial infection in the lung: CT appearance. Radiology 187:777–782, 1993 Oh YW, Kim YH, Lee NJ, et al: High-resolution CT appearance of miliary tuberculosis. J Comput Assist Tomogr 18:862–866, 1994 Palmer PES: Pulmonary tuberculosis: Usual and unusual radiographic presentations. Semin Roentgenol 14:204–243, 1979 Pastores SM, Naidich DP, Aranda CP, et al: Intrathoracic adenopathy associated with pulmonary tuberculosis associated with human immunodeficiency virus infection. Chest 103:1433–1437, 1993 Poey C, Verhaegen F, Giron J, et al: High-resolution chest CT in tuberculosis: Evolutive patterns and signs of activity. J Comput Assist Tomogr 21:601–607, 1997 Sochochky S: Tuberculoma of the lung. Am Rev Tubercle Pulmonary Dis 78:4030-410, 1958 Song JW, Im JG, Shim YS, et al: Hypertrophied bronchial artery at thin-section CT in patients with bronchiectasis: Correlation with CT angiographic findings. Radiology 208:187–191, 1998 Woodring JH, Vandiviere HM: Pulmonary disease caused by nontuberculous mycobacteria. J Thorac Imaging 5:64–76, 1990 Woodring JH, Vandiviere HM, Fried AM, et al: Update: The radiographic features of pulmonary tuberculosis. AJR Am J Roentgenol 146:497–506, 1986 Address reprint requests to Jung-Gi Im, MD Department of Radiology Seoul National University Hospital 28, Yongon-dong, Chongro-gu Seoul 110-744 South Korea e-mail: [email protected]

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT OF THE LUNG II

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE Jerald P. Kuhn, MD, and Alan S. Brody, MD

Since a review in 1993 of high-resolution CT (HRCT) in children,43 the technique has become more widely used, but there is still limited experience with many of the uncommon and rare childhood pulmonary diseases. Even with the common diseases, such as asthma, cystic fibrosis, bronchopulmonary dysplasia (BPD), and infection, there is no consensus regarding when and how to use HRCT. This article provides a foundation for differential diagnosis based on an approach using anatomic categories of disease defined by CT findings. CT findings have been organized into those suggesting disease of the large and small airways; alveolar disorders, both interstitial and air-space; and diseases involving the peripheral (septal) interstitial tissues. Also discussed are diseases associated with pulmonary nodules, vascular disorders, pulmonary fibrosis, and fatal neonatal lung diseases. INDICATIONS The most common indications for pediatric HRCT are listed in Table 1. Virtually any time a child has a pulmonary parenchymal abnormality requiring CT, a relatively thin-slice technique should be used in combination with the edge-enhancing algorithm. For some conditions, such as metastatic disease, all of both lungs need to be imaged. For other con-

ditions, such as an anomaly or localized mass, it may be possible to restrict the examination to the region of interest to limit radiation to the patient. Although these are not strictly HRCT studies, and are not discussed, the principles of interpretation are the same. TECHNIQUE Image Quality Four components impact on the quality of HRCT images in children: (1) motion, (2) lung Table 1. INDICATIONS FOR PEDIATRIC HRCT 1. Normal chest radiograph with serious or unexplained symptom Fever in an immune compromised patient Unexplained dyspnea, wheezing, or severe or atypical asthma 2. Abnormal but nonspecific chest radiograph Ill-defined nodules, opacities or suspicion of interstitial disease Guide for biopsy 3. Detection of bronchiectasis 4. Detection of sequelae of infection Bronchiolitis obliterans or bronchiectasis 5. Cystic fibrosis Staging Evaluation of therapy Research 6. Bronchopulmonary dysplasia Evaluation of severity of disease

From the Department of Radiology, State University of New York at Buffalo School of Medicine, Buffalo, New York (JPK); and Departments of Radiology and Pediatrics, Children’s Hospital Medical Center, Cincinnati, Ohio (ASB)

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

89

90

KUHN & BRODY

volume, (3) patient size, and (4) CT technical factors. Technique should be chosen after consideration of the first three factors. Although technique is the primary determinant of dose, it has the least impact on image quality. Motion. The greatest impediment to obtaining high-quality CT studies in children is respiratory and gross body motion. In children who cannot follow instructions, typically those under 3 years old, motion must be controlled by using restraints, sedation, and the shortest possible scan time available. Sedation is used in varying degrees in different departments and many protocols are in use.17 With a 0.1-second scan time of electron beam CT (EBCT)43 and the speed of the new multisection helical CT scanners, sedation is not usually necessary.68 Sedation is still required in many uncooperative children between the ages of 6 months and 6 years. Children less than 6 months old usually lie still, frequently falling asleep after a feeding. Sedation is rarely required in children over 6 years old. Chloral hydrate is frequently used in children less than 2 years old. Pentobarbital is a commonly used drug in older children. In most cases young children are imaged during quiet respiration. Posterior atelectasis is usually less striking than in adults,74 but is present in nearly all patients undergoing general anesthesia. 74, 77 Prone imaging is rarely necessary. Lung Volume Lung volume during quiet respiration is lower than during maximum inflation, but the appearance during quiet breathing is closer to a full inspiratory image than an expiratory image. No comparison between breathing and breath-hold images has been performed with HRCT, but for helical CT an experimental study with pathologic correlation found comparable image quality.12 Inspiratory and expiratory images are very helpful in evaluating abnormal lung attenuation in children and adults. With EBCT a dynamic scan can be obtained during the course of a normal respiratory cycle (Fig. 1).43 If a child is too young to cooperate, a decubitus scan can serve to evaluate possible air trapping on the dependent side.48 By about 6 years of age, most children can cooperate to produce a good quality inspiratory and expiratory HRCT. This degree of cooperation usually requires preparation and a member of the CT team in the scan room. Practicing inspiratory

and expiratory maneuvers before entering the scan room is helpful. Parents can sometimes be enlisted to help their child cooperate. It is unusual for a child to follow verbal instructions from the control room for HRCT with inspiratory and expiratory images until about 12 years old. CT Technique The factors varied in pediatric HRCT are slice thickness, scan time, tube kilovoltage, and tube current. Slice thickness is usually the thinnest available. Slice spacing should be chosen based on the suspected pathology. If chest radiographs are normal, and a diffuse interstitial process is suspected, four or five slices may be sufficient. If the detection of a few small cysts is of diagnostic importance, more slices should be obtained. In general the authors use 10-mm slice spacing in children over 10 years old, 7-mm spacing in children from 2 to 10 years old, and 5-mm spacing for children less than 2 years of age. Early reports of HRCT emphasized the potentially high radiation dose from this technique.60 Since that time it has been recognized that HRCT can be done using a lower patient dose than conventional CT.57 Low-dose techniques have further reduced HRCT dose.2 The authors have calculated that in infants, four images at 40 mA can be obtained using the same dose as a two-view chest radiograph. Dose is directly proportional to the product of scan time and tube current. The dose increase caused by increasing kilovoltage is not linear, and is greater than often appreciated. An increase from 120 to 140 kilovoltage (peak) increases dose by approximately 40%. Pediatric patients are rarely large enough to warrant the use of increased kilovoltage (peak). Scan time can be shortened by more rapid gantry rotation or by decreasing beam rotation to less than a full 360. In general, the fastest scan time that uses full rotation should be used. The use of partial scans should be evaluated at each site. As long as full scanner rotation is used, the effect of decreasing scan time is the same as an equal decrease in tube current. Tube current should be adjusted to provide the lowest dose consistent with adequate diagnostic quality. Several authors have advocated doses as low as 20 to 40 mA for HRCT in adults.57, 90 Because of their smaller size, the potential for decreasing dose is particularly

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

91

Figure 1. Dynamic electron beam CT (EBCT) study in a 5-year-old boy with bronchiolitis obliterans. The image shows marked air-trapping and mosaic perfusion pattern. The graph shows virtually no air exchange in the hyperlucent lung regions (A). The ground-glass regions (B and C) show a normal variation of approximately 150 Hounsfield units (H) between inspiration and expiration.

great in children. A technique chart that relates tube current to weight is appropriate. Forty mA provides good quality images in infants. Children rarely require more than 100 mA.2 Other Considerations A high-frequency reconstruction algorithm (bone or lung) should be used to increase edge enhancement and improve visualization of parenchymal detail. The authors routinely also reconstruct images using a standard (soft tissue) reconstruction algorithm to review mediastinal structures. The use of the smallest

field of view possible optimizes spatial resolution. SPECIAL HIGH-RESOLUTION CT TECHNIQUES Electron Beam CT Electron beam CT allows routine use of a 0.1-second scan time that is short enough to stop respiratory motion artifact. Drawbacks include spatial resolution inferior to that of helical scanners and the lack of widespread availability of EBCT.

92

KUHN & BRODY

Stop Ventilation Technique HRCT in young children is limited both by patient motion and by an inability to obtain inspiratory and expiratory images. The stop ventilation technique uses conscious sedation and mask ventilation to provide motion-free images at inspiration and expiration.47 If a sedated child is given several deep breaths, the child’s respiration pauses for 8 to 15 seconds. During this pause inspiratory images are obtained by administering positive pressure. Expiratory images are obtained by allowing the elastic recoil of the lung to decrease lung volume when the airway is open to atmospheric pressure. Using this technique high-quality HRCT images can be obtained in young children (Fig. 2). Expertise in respiratory management and additional experience are required to make this technique reliable and effective. HIGH-RESOLUTION CT SIGNS OF DISEASE Airways Investigation of diseases of the pediatric airways is perhaps the most important use of HRCT because this group of diseases is so common in children. HRCT signs of diseases involving the airways are listed in Table 2. Diseases primarily affecting the bronchi are diagnosed by finding bronchial wall thickening or bronchiectasis. The abnormalities can range from subtle and equivocal to grossly obvious. Minimal bronchial wall thickening, whether caused by infection, infiltration, or edema, can be difficult to diagnose with certainty on CT. Similarly, distinguishing between mild, reversible bronchial dilatation and minimal cylindric bronchiectasis is difficult. Many cases can be recognized using the criteria for bronchiectasis used in adults but equivocal cases do exist. Recognition of more marked bronchial dilatation is simple but there can be inconsistency in trying to decide if the bronchiectasis is cylindric or varicose or when varicose, if it is severe enough to be considered cystic. Despite these limitations, it is generally agreed that CT is the most accurate imaging procedure for the study of bronchiectasis. Diseases associated with pediatric bronchiectasis13 are listed in Table 3. Bronchial dilatation and bronchiectasis are uncommon in

childhood asthma, although their true incidence is not known. In cystic fibrosis, disease begins in the small airways, but bronchiectasis is almost invariably present as the child ages.26 Bronchiectasis is typically worse in the upper lobes than in the lower lungs and often more severe on the right than on the left side. The HRCT findings in cystic fibrosis are well documented and include not only progressively more severe bronchiectasis, but evidence of small airway disease including focal and generalized air trapping and centrilobular opacities (Figs. 3 and 4).25, 26, 32 Mucoid impaction of bronchi and bronchioles can be widespread but is often reversible. The role of HRCT in cystic fibrosis is still undergoing study,72 but CT can document the presence and extent of disease earlier and more accurately than chest radiography. Long,46 using the stop ventilation technique and careful measurements, showed early bronchial wall thickening and airway dilatation in a group of young (mean age 2.5) children with cystic fibrosis. CT scoring systems have been developed to quantitate the severity of disease52 and have been shown to correlate well with pulmonary function tests.61 It has been shown that CT can evaluate response to therapy and should be considered as an accurate way to evaluate efficacy of experimental therapies.10 Signs of bronchiolar (small airway) disease are in Table 2. The presence of air trapping suggests disease in the smaller airways. In infancy, these tiny airways easily become obstructed by inflammatory edema or exudate and contribute disproportionately to airway resistance resulting in marked air trapping that can be either focal or diffuse.20 Diffuse air trapping can be easily overlooked because the lungs are uniformly involved. In a child breathing quietly or breathholding at less than total vital capacity, visible anterior or posterior junction lines indicate that significant air trapping is present.43 Attenuation of the lung is lower than normal, often with values in the 900 H range, rather than 600 to 750 H,84 but unless actually measured can escape visual detection. Focal air trapping is more easily recognized, but requires a dynamic or expiratory scan for optimal diagnosis.48 Focal air trapping is often accompanied by a mosaic perfusion pattern that results when blood is shunted from a hyperinflated, hypoxic region to adjacent normal lung.86 Air trapping occurs in diseases that affect the smaller airways: bronchiolitis, asthma,

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

Figure 2. A, A 4-month-old patient who had postgastric pull-up surgery and chronic aspiration pneumonia. A 10-mm helical CT slice at 240 milliampere-second (mAs) was made at an outside hospital. A dilated, fluidfilled esophagus is present, but no pulmonary abnormalities are evident. Compare with B. B, Image using a ‘‘stop ventilation’’ technique with a 1mm slice at 80 mAs. This image shows numerous, small, centrilobular opacities caused by chronic aspiration. Also a number of normal bronchi are visible that are not seen on the other study. This image was obtained with about one-tenth the dose of the outside study.

93

94

KUHN & BRODY

Table 2. HRCT SIGNS OF PEDIATRIC AIRWAY DISORDERS Bronchial disease Bronchial wall thickening Compare with other regions of lung Visibility of bronchus further peripheral than normal Bronchiectasis Bronchus larger than accompanying vessel Lack of tapering Bronchus visible within 1 cm of pleural surface Bronchiolar disease Diffuse air trapping Decreased lung attenuation Prominent anterior and posterior junction lines Focal air trapping Low attenuation accentuated on expiration Decrease vessel size Mosaic perfusion pattern Bronchial wall thickening or bronchiectasis Centrilobular opacities Strongly suggest infection

cystic fibrosis, and viral lower respiratory infection. Other causes of generalized hyperinflation include large left-to-right shunts, diabetic acidosis, aspirin intoxication, and other causes of metabolic acidosis. There is no indication to image these conditions with CT unless a complication or an unusual associated finding is present. Bronchiolitis is a common viral lower respiratory tract infection that is most frequent in infancy. In some cases wheezing is the only symptom, but respiratory distress can be severe and, in rare instances, death can result. The only indication for CT is to exclude some other cause for the infant’s respiratory distress. CT shows findings of hyperaeration and focal disturbances in aeration and a mosaic perfusion pattern.43 Bronchial asthma is a common, recurrent pulmonary disease characterized by wheez-

Table 3. CAUSES OF PEDIATRIC BRONCHIECTASIS Cystic fibrosis Ciliary dyskinesia Allergic bronchopulmonary aspergillosis Infection Chronic, recurrent infection Aspiration syndromes AIDS Other immune deficiency syndromes Sequelae of infection Bronchiolitis obliterans Obstruction Foreign body Neoplasm (rare) Congenital anomaly Williams-Campbell syndrome

ing, paroxysmal cough, and dyspnea caused by hyperreactivity of the tracheobronchial airways resulting in increased resistance to airflow. CT scans are rarely obtained in children with uncomplicated asthma. The authors have examined a few such children who were suspected of having other conditions and found peribronchial thickening and focal, peripheral regions of air trapping, a pattern similar to that described in bronchiolitis obliterans. Asthmatic children with superimposed infection often have segmental or subsegmental regions of centrilobular opacities not apparent on chest radiography.43 CT may be indicated in some cases of atypical or refractory asthma. Occasionally, an unsuspected cause for wheezing is diagnosed. The authors have encountered BPD, bronchiectasis, bronchiolitis obliterans, and congenital pulmonary anomalies in children thought clinically to be asthmatic.43 In a series of 16 children with atypical asthma examined by HRCT, Nuhoglu et al64 found nine with minor abnormalities (mostly fibrotic scars and linear atelectasis) and three with major pathology including bronchiectasis and right middle lobe atelectasis. CT is also useful to diagnose allergic bronchopulmonary aspergillosis that is an infrequent complication of long-standing asthma.78 It has been shown that HRCT can assess airway reactivity29 in addition to diagnosing small airways disease, so it is likely that CT will have use as a research tool and for more accurate clinical diagnosis in atypical asthma. In addition to having generalized air trapping, patients with asthma, bronchiolitis, and viral lower respiratory infection have focal air trapping that resolves between bouts of illness. Constrictive or obliterative bronchiolitis (OB), however, is characterized by irreversible focal air trapping. OB is diagnosed pathologically by the presence of concentric fibrosis of the submucosal and peribronchial tissues of the terminal and respiratory bronchioles resulting in bronchial obliteration or severe narrowing.19 OB is commonly present distal to areas of bronchiectasis but it is not clear whether it is a cause or a result of bronchiectasis, or merely a related condition. 24 During childhood, there are two important causes of OB. The first is a complication of organ transplantation particularly lung and heart-lung transplants, although it has also been associated with bone marrow transplantation.73 Bronchiolitis obliterans is a late complication thought to represent a form of

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

Figure 3. A 9-year old patient with minimal changes of cystic fibrosis. The lungs are slightly hyperexpanded and of decreased attenuation. Slight bronchial wall thickening is present.

Figure 4. An 18-year-old patient with advanced cystic fibrosis. Changes of cystic bronchiectasis with air–fluid levels are obvious in the right lower lobe as well as diffuse bilateral bronchiectasis of all degrees, mucoid impaction, centrilobular opacities, air-trapping, and mosaic perfusion in the left lower lobe. On the right, there is a chest tube and subcutaneous emphysema from an earlier pneumothorax.

95

96

KUHN & BRODY

chronic rejection. It has a prevalence of about 50% and is a leading post-transplant cause of morbidity and mortality. It rarely develops sooner than 3 months after surgery and in most cases does not appear until near the end of the first postoperative year. Plain radiographic findings are not helpful in establishing the diagnosis. Transbronchial biopsy is diagnostic but often is falsly negative because of the patchy nature of the abnormality. HRCT is an important diagnostic aid with a reported sensitivity of 74% to 91% and a specificity of 67% to 94%. The most useful findings are dilatation of lower lobe bronchi, focal air trapping on expiratory CT, and mosaic perfusion pattern.44, 45, 87 The second important cause of OB in childhood is infection. Prior pneumonia, early in life with adenovirus, especially types 21 and 7, has been shown to be an important cause of this condition.11 Kim et al36 reported that 38% of children hospitalized with mycoplasma pneumonia were shown to have sequelae on CT that were consistent with OB. Other predisposing causes of OB include measles, influenzal viral infections, rheumatoid arthritis, systemic lupus erythematosus, and inhalation of toxic gases. Interestingly, OB has not been noted as a complication of respiratory syncytial viral infection.55 The true prevalence of OB following pulmonary infection in infancy or early childhood is not known, but the authors see several cases a year with typical CT findings, although lung biopsy is rarely performed. The article by Kim et al36 suggests that OB may be a fairly common complication of mycoplasma infection. It is apparent that infection in early life can scar the lungs much as it does the kidney; what is still unknown is how often this sequence of events occurs. The clinical findings of OB are nonspecific but include prolonged recovery from an initial infection with recurrent bouts of wheezing, pneumonia, and atelectasis. Fixed airway obstruction is often detected if pulmonary function tests can be performed. Chest radiograph is normal in less severe cases but shows intermittent pneumonia, hyperinflation, and peribronchial thickening in the sicker infants. Unilateral hyperlucent lung (Swyer-James syndrome) is a unique manifestation of OB with involvement asymmetric enough to allow initial recognition by radiography. When CT is performed on these children, bilateral changes of OB are found in about 50% of cases.40, 49, 54

CT findings of OB are those of small airways disease including focal air trapping, especially on expiration; mosaic perfusion pattern; bronchial wall thickening; and bronchiectasis (Figs. 1 and 5). 50, 81, 89 These CT findings suggest a diagnosis of OB, but unless air trapping and bronchiectasis are both present, the findings have to be interpreted with caution because reactive airways disease, bronchiolitis, and acute viral pneumonia also produce focal areas of air trapping and a mosaic perfusion pattern. If bronchiectasis is not present, a follow-up CT study showing no change in the areas of air trapping after an interval of 6 months to a year strongly suggests the diagnosis. The last CT finding to be discussed as an indication of small airway disease is the presence of centrilobular opacities (CLO). These result when the terminal bronchiole becomes dilated and fluid-filled allowing it to become visible on CT. CLOs are described as having a ‘‘tree-in-bud appearance’’ or branching Ys or dots in the center of the secondary pulmonary lobule.4 Their presence strongly suggests bronchiolar infection.21, 50 The authors have noted this pattern frequently in children with mycoplasma pneumonia, aspiration pneumonia (see Fig. 2B), cystic fibrosis (see Fig. 4), and occasionally in asthmatics. Bronchial wall disease, focal air trapping, ground-glass opacities (GGO), and consolidation are frequent concomitant abnormalities (Fig. 6). CLOs are rarely found uniformly throughout both lungs and their presence should be suspected if one sees too many dots near the lung periphery. Usually, comparison with other lung regions confirms the suspected focal pathology. Alveolar Disorders Disease at the alveolar level is manifest by the presence of GGO. GGO is defined as an increase in normal lung attenuation that does not obscure the underlying parenchymal detail (Figs. 1 and 5B). As discussed previously, GGO may occur in normal lung from shunting of blood from adjacent hypoxic or oligemic segments. GGO is also produced by alveolar disease. This is either the result of thickening of the alveolar septal walls; a reflection of interstitial disease; or opacification of alveolar air-spaces with fluid, exudate, or hemorrhage conditions often associated with frank areas of consolidation. GGOs are non-

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

Figure 5. A 5-year-old boy with biopsy-proved bronchiolitis obliterans from pneumonia in infancy. A, Inspiration reveals loss of volume on the right with upper lobe atelectasis and markedly decreased attenuation and vascularity in the right lung. Bronchial wall thickening is noted on the left with a barely perceptible mosaic perfusion pattern. B, Expiration reveals left sided air-trapping, and mosaic perfusion pattern is much more apparent.

97

98

KUHN & BRODY

Figure 6. A 3-mm 0.1-second EBCT image reveals centrilobular opacities and marked bronchial wall thickening are seen throughout the right upper lobe in this 9-year-old girl with mycoplasma pneumonia. Incidental note is made of the azygos vein abutting the right paravertebral surface.

specific and their presence can implicate airway, vascular, air-space, or interstitial disease from a wide range of causes (Table 4). Presence of diffuse, relatively uniform GGO suggests diffuse interstitial disease. Focal GGOs are more often caused by air-space disease or vascular redistribution. Often the clinical picture or associated CT findings, such as the crazy paving pattern of pulmonary alveolar proteinosis (PAP) (Fig. 7) helps to narrow the differential diagnosis, but it is not unusual that biopsy or bronchoalveolar lavage (BAL) is required to determine the underlying cause. Pulmonary hemorrhage produces GGOs that tend to be poorly marginated, somewhat fluffy or micronodular in appearTable 4. CAUSES OF GROUND-GLASS OPACITY IN CHILDHOOD Interstitial diseases associated with GGO in children Chronic pneumonitis of infancy Desquamative interstitial pneumonitis Nonspecific interstitial pneumonitis Air-space disorders associated with GGO Diffuse alveolar damage (ARDS) Hemorrhage Vasculitis Exudate (pneumonias) Pneumocystis carinii pneumonia Pulmonary edema Pulmonary alveolar proteinosis (both types) Lymphocytic interstitial pneumonia Leukemic infiltrate Alveolar capillary dysplasia

ance, and usually multifocal. Patients present with anemia, hemoptysis, or sudden onset of cough or dyspnea. Hemorrhage and pulmonary edema are widespread in diffuse alveolar damage, typified by acute respiratory distress syndrome (ARDS). Less common diseases causing pulmonary hemorrhage include idiopathic pulmonary hemorrhage (hemosiderosis); Goodpasture’s syndrome; systemic lupus erythematosus; and, less commonly, the other collagen vascular diseases. Diseases associated with pulmonary vasculitis and hemorrhage include Wegener’s granulomatosis and Churg-Strauss syndrome. The most common pulmonary presentation of Wegener ’s granulomatosis in childhood seems to be ill-defined, bilateral nodules that may cavitate (Fig. 8). Imaging of these disorders is nonspecific and diagnosis requires BAL, immunologic tests, or biopsy. The idiopathic, diffuse interstitial lung diseases that are so often diagnostic dilemmas in adults are much less common in children. There is still controversy about the classification of these rare disorders in children and only a few reports of their CT appearances. 39, 51 Katzenstein et al 35 described an entity called chronic pneumonitis of infancy (CPI). This condition develops at an average age of about 4 months in infants who are normal at birth. Most cases have a progressive downhill course. They believe many childhood cases of what were previously labeled usual interstitial pneumonia or desquamative interstitial pneumonia actually repre-

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

Figure 7. Pulmonary alveolar proteinosis in a 12-year-old girl. A 6-mm 0.1-second EBCT image reveals bilateral ground-glass opacities and a marked, smooth interlobular septal thickening (crazy-paving) appearance.

Figure 8. Wegener’s granulomatosis in 15-year-old boy presenting with renal failure and pulmonary opacities. A 6-mm EBCT 0.1-second image reveals bilateral nodules, several of which have margins of ground-glass attenuation suggestive of vasculitis. Several cavitated lesions were noted on other images.

99

100

KUHN & BRODY

Figure 9. A 3-mm 0.1-second EBCT image of a 4-week-old patient with chronic pneumonitis of infancy. Diffuse, severe, ground-glass opacities are present without apparent septal thickening.

sented CPI. The authors have seen one proved case in which CT showed initially diffuse GGO (Fig. 9) with progressive changes of severe, chronic interstitial fibrosis, which resulted in the death of the infant following attempted lung transplantation. Katzenstein et al35 believe most of the remaining cases of idiopathic interstitial pneumonias of childhood fit in the category they call nonspecific interstitial pneumonitis, which is characterized pathologically by mild chronic inflammatory changes in the alveolar septum and a lack of an intralveolar exudate. The CT appearances of this condition, which may have multiple causes, have not been fully described. Desquamative interstitial pneumonia in childhood is still poorly understood. Some cases may represent chronic pneumonitis of infancy and there is some apparent overlap with surfactant B deficiency and congenital alveolar proteinosis. Reported CT findings of desquamative interstitial pneumonia 51 include diffuse GGO with progression to pulmonary fibrosis. Usual interstitial pneumonitis either does not exist in children or is extremely rare.62 Signs of diseases affecting the peripheral portion of the interstitium of the lung are listed in Table 5. The first is interlobular septal thickening. The interlobular septa are part of the peripheral interstitium of the lung and form the margins of the secondary pulmonary lobule.85 They contain pulmonary lymphatics and venous structures, are best developed in the periphery of the lung anteriorly, and are usually not visualized on a normal

pediatric chest radiograph. A few septa may normally be visible on chest CT. As septae become thickened, they are recognized as Kerley B lines on a radiograph and on CT as thin, 1- to 2-cm lines, generally perpendicular to the pleural surface. When numerous they form a fine mesh-like pattern of lines (Fig. 10). They are contiguous with the interlobar septa (the pleural fissures) and the bronchovascular bundles, and often these structures become thickened, making recognition of septal pathology more obvious. Thickening of the septa may be either smooth, nodular, or irregular.33 Smooth septal thickening is much more common in children. Conditions causing smooth septal thickening are listed in Table 6. Because many of these disorders are associated with either engorged veins, lymphatics or interstitial edema, areas of GGO are often also present. Cardiogenic pulmonary edema is not studied by CT for diagnostic purposes, but may be detected serendipitously as may edema related to renal disease or iatrogenic fluid overload. Early edema is Table 5. HRCT SIGNS OF PERIPHERAL INTERSTITIAL (SEPTAL) DISORDERS Smooth septal thickening Reticular pattern Regular bronchovascular bundle thickening Prominent interlobular septal (fissural thickening) Nodular septal thickening May be associated with pulmonary nodules Irregular or nodular bronchovascular bundle thickening

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

101

Figure 10. Unilateral, (right-sided) smooth, interlobular septal thickening in a 12-year-old girl who has right-sided partial anomalous pulmonary venous return.

manifest by slight increase in lung attenuation, GGO, and septal thickening.30 One of the most striking causes of septal thickening is pulmonary alveolar proteinosis. This condition, which may present in childhood, is of unknown cause.1 Pathologic examination reveals that alveoli are filled with a periodic acid–Schiff (PAS)-positive material, perhaps from type 2 pneumocytes. Interstitial fibrosis is absent in early cases, although septal edema is present and may be striking on CT (see Fig. 7). Pulmonary lymphangiomatosis (lymphangiectasia) can be limited to one lung and is associated with marked septal and bronchovascular bundle thickening. Patients often also have a pleural effusion and a mediastinal mass or infiltration.82, 88 Other rare con-

Table 6. CAUSES OF SMOOTH SEPTAL THICKENING Fluid overload Cardiac, renal, or iatrogenic Diabetic keto-acidosis Pulmonary vein atresia or obstruction Pulmonary lymphangiectasia (lymphangiomatosis) Associated with mediastinal mass or pleural effusion Rare diseases Pulmonary alveolar proteinosis Pulmonary capillary hemangiomatosis Pulmonary alveolar microlithiasis Pulmonary hemosiderosis Gaucher disease Niemann-Pick disease

ditions presenting with septal thickening are pulmonary alveolar microlithiasis, pulmonary capillary hemangiomatosis, and pulmonary hemosiderosis.16, 27, 39, 51, 75 These conditions usually also have a small, nodular component. Nodular septal thickening is usually associated with an infiltrative process, often malignant, that spreads along the pulmonary lymphatics resulting not only in nodular septal thickening but thickening of the bronchovascular bundle and often with pulmonary nodules, masses, or areas of consolidation. Diseases associated with this pattern include metastatic sarcomas; lymphoma, (especially large cell); and rarely neuroblastoma. In most cases the nature of the underlying disease is already well established clinically. Large cell lymphoma is one condition that can present relatively suddenly with an abnormal chest film before the diagnosis is apparent (Fig. 11). Irregular septal thickening is associated with pulmonary fibrosis and is discussed later.

DISORDERS ASSOCIATED WITH PULMONARY NODULES OR SMALL MASSES Pulmonary nodules can be classified in a number of ways: by attenuation that may be

102

KUHN & BRODY

Figure 11. Histiocytic (large cell) lymphoma in a 17-year-old boy. Axial EBCT scan with 3-mm collimation through the upper lung zones shows a marked interlobular septal pattern and nodular thickening of the bronchovascular bundles, present bilaterally, but much more marked in the left upper lobe.

ground-glass, soft tissue, calcific, or air-containing; by size; or by location (Table 7). It can be difficult to determine if a CT image showing too many dots is caused by multiple small nodules, vessels, or CLOs. Using a slightly thicker slice (3 to 5 mm) allows easier recognition of vessels. Nodules tend to be diffuse and CLOs tend to be localized, associated with bronchial wall thickening, and possess a branching or Y pattern not observed with nodules.

Table 7. PULMONARY NODULES IN CHILDHOOD Small, ill-defined centrilobular nodules Hypersensitivity pneumonia LIP Pulmonary capillary hemangiomatosis Pulmonary hemosiderosis Follicular bronchiolitis Small soft tissue nodules Metastatic disease Miliary tuberculosis Granulomatous and fungal infections Langerhans’ cell histiocytosis Pulmonary hemosiderosis Larger nodules or masses Metastases Recurrent respiratory papillomatosis Lymphoma Lymphomatoid granulomatosis Fungal infection Bronchiolitis obliterans with organizing pneumonia Vasculitis Septic emboli Arteriovenous malformations Pulmonary artery aneurysms

Ill-defined centrilobular nodules of groundglass attenuation are associated with hypersensitivity pneumonitis, which in children is usually acute and caused by exposure to one of many bird antigens. Vasculitis can produce similar appearances, although the nodules usually are larger. Similar nodules have also been described in pulmonary capillary hemangiomatosis.16 Small soft tissue nodules occur in pulmonary lymphoid hyperplasia, including AIDS-associated LIP and rarely follicular bronchiolitis. LIP is an AIDS-defining illness if it occurs in a child under 13 years. Up to 30% to 40% of children with AIDS develop LIP. Reticulonodular infiltrates and nodules from miliary to a few millimeters in size, GGO, and bronchovascular bundle thickening are characteristic. Hilar adenopathy and thymic cysts are other common findings. Small cysts and areas of bronchiectasis occur and may be related to partial bronchiole obstruction from lymphocytic infiltration.18, 53 Follicular bronchiolitis is characterized by the presence of enlarged lymphoid aggregates along distal bronchi and bronchioles. This condition is also thought to be part of the spectrum of pulmonary lymphoid proliferation. It usually has its onset before 6 months of age and is characterized on CT by small nodules, bronchial wall thickening, and sometimes by striking findings of focal air trapping suggestive of bronchiolitis obliterans (Fig. 12).38, 66, 71 Langerhans’ cell histiocytosis presents with small, soft tissue, centrilobular

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

103

Figure 12. A 2-year-old patient with biopsy-proved follicular bronchiolitis. Wide-spread air-trapping and mosaic perfusion pattern. Nodules have been reported in this condition but are not apparent in this patient.

nodules some of which coalesce and cavitate (Fig. 13). Thymic cysts and calcifications are associated findings sometimes present on CT.9, 39, 79, 80 Pulmonary involvement occurs in from 10% to 50% of cases, but lung disease has not been shown to have an adverse prognostic effect. 22 As this disease progresses,

however, large pulmonary cysts can form and produce pneumothoraces or progressive pulmonary failure. Miliary tuberculosis has innumerable 1- to 3-mm soft tissue nodules scattered diffusely throughout both lungs. 37 Pulmonary alveolar microlithiasis is a rare familial disease characterized by micronodu-

Figure 13. A 3-mm EBCT image at 0.1 second reveals a 16-month-old patient with Langerhans’ cell histiocytosis. Multiple, small, thin-walled cysts and scattered, small soft tissue nodules are noted throughout both lungs.

104

KUHN & BRODY

lar calcifications, septal thickening, and GGOs. Punctate and linear pleural calcifications are also present.27, 75 Surprisingly, these nodules do not always appear calcified on CT51 Larger, more sharply defined pulmonary nodules can result from metastatic disease. Wilms’ tumor and the sarcomas are the most frequent causes, but lymphoma and rarely neuroblastoma may also develop nodules. It is unusual to have pulmonary metastases as the initial manifestation of a pediatric malignancy. Lymphomatoid granulomatosis is a rare angiocentric, angiodestructive lymphoproliferative disorder that may be a precursor of T-cell lymphoma.34 Its pulmonary presentations range from a solitary nodule34 to large, bilateral conglomerate opacities69 Other soft tissue nodules that tend to be large and may cavitate include (1) septic emboli, a unique variant of which is Lemierre syndrome, which is pharyngitis, jugular venous thrombosis, and septic emboli15, 42; (2) fungal infection, especially with candida and aspergillus, which can produce multiple peripheral nodules; and (3) recurrent respiratory papillomatosis,6, 42 previously called juvenile laryngeal papillomatosis, which can seed down the airway from its origin in the larynx and present as pulmonary nodules, which often cavitate. Vasculitis, especially Wegener’s granulomatosis, can produce multiple pulmo-

nary nodules, often with ground-glass margins and occasional cavitation (see Fig. 8).14, 23, 59 Bronchiolitis obliterans with organizing pneumonia (BOOP), also called cryptogenic organizing pneumonia, is one of the lung’s most common, nonspecific, reparative reactions. Pathologic examination reveals a sharply demarcated cellular fibrotic reaction that plugs the air spaces and small airways. BOOP is uncommon in children but has been seen in patients recovering from malignancy and as a manifestation of drug reaction, especially bleomycin toxicity.28, 56 CT demonstrates a subpleural distribution of opacities that are often nodular in nature and may be accompanied by ground-glass attenuation (Fig. 14). BOOP is often reversible and may respond to steroid therapy.

Vascular Disorders Diseases affecting the pulmonary circulation are listed in Table 8. Secondary signs of vascular disease have been discussed previously and include septal thickening (pulmonary venous engorgement); GGO (capillaritis or edema), mosaic perfusion pattern (oligemia and vascular redistribution); and nodules (vasculitis). Primary signs of vascular diseases include visualization of vessels in an

Figure 14. Biopsy-proved bronchiolitis obliterans with organizing pneumonia (BOOP) in an 18-year-old patient with post–bone marrow transplant for leukemia. Large, ill-defined soft tissue masses are present primarily in the periphery of the right lower lobe.

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

Table 8. PULMONARY VASCULAR DISORDERS Large vessels Anomalies Anomalous venous return Aberrant arterial supply Left to right shunt Congestive failure Pulmonary hypertension Aneurysms Associated with tetrology of Fallot with absent valve Behc¸et’s syndrome Hepatopulmonary syndrome Decreased vessel size Right-to-left shunt Acute chest syndrome in sickle cell disease Bronchiolitis obliterans Pulmonary embolism Secondary signs Smooth interlobular septal thickening Mosaic perfusion pattern Nodules and small masses

anomalous position, such as total or partial anomalous pulmonary venous return. Abnormalities of size include a pulmonary artery that may be absent, hypoplastic, or rarely aneurysmally dilated. The pulmonary arteries are larger than their accompanying bronchus if there is a left to right shunt, vascular redistribution, or pulmonary hypertension. Pulmonary arteriovenous malformation occur mainly in patients with Osler-Weber-Rendu disease and may manifest as a single large mass or multiple small nodules.70 They are bilateral in about two thirds of cases. Hepatopulmonary syndrome is associated with right to left shunting but prominent pulmonary vascularity, especially in the lower lobes.58 Behc¸et’s syndrome is a rare systemic vasculitis that can manifest with large pulmonary artery aneurysms.65 Atresia or obstruction of pulmonary veins presents with smooth interlobular septal thickening, and in more severe cases with areas of GGO. Decreased vessel size can be seen in acute chest syndrome in patients with sickle cell disease.7 Pulmonary thromboembolism is uncommon in children but has the same CT appearances as described in adults. Pulmonary embolism in children usually is associated with an indwelling catheter,3 malignancy, sepsis, or other predisposing factor.

Pulmonary Fibrosis End stage pulmonary fibrosis is fortunately rare in children. Findings of fibrosis include

105

honeycombing, cysts, traction bronchiectasis, and parenchymal bands. Large cysts form in end-stage cystic fibrosis; most are caused by cystic bronchiectasis, but large subpleural bullae and pneumatoceles can also develop.83 Extensive cysts can occur in end-stage Langerhans’ cell histiocytosis particularly in the upper lung zones. BPD can range in appearance from a few subpleural areas of hyperlucency, possibly caused by small airways disease, to extensive lung destruction, generally more severe anteriorly and in the upper lobes.5, 31, 43, 67 Frequent findings include parenchymal bands, multifocal areas of decreased attenuation, and perfusion. Reduced bronchial caliber has been reported,31 but bronchial dilatation also occurs (Fig. 15). Similar changes develop in pediatric survivors of ARDS. Interstitial fibrosis can occur in collagen vascular disease, particularly systemic sclerosis.76 Interstitial fibrosis progressing to end-stage pulmonary fibrosis is unusual in children but can be seen in chronic pneumonitis of infancy (Fig. 16), and nonspecific interstitial pneumonitis. Focal areas of hyperlucency, usually associated with a mass or disturbance of normal pulmonary anatomy, occur with congenital malformations, such as bronchial atresia, sequestration, and congenital cystic adenomatoid malformation (CCAM). Postinfectious pneumatoceles and loculated interstitial pulmonary emphysema are transient causes of what initially might seem to be destructive lung disease.32, 41 FATAL NEONATAL LUNG DISORDERS Surfactant B deficiency has been identified as a cause of congenital pulmonary alveolar proteinosis.95 This condition presents at birth with radiographic findings similar to respiratory distress syndrome. The CT appearance in two cases was similar to that described in the adult form of alveolar proteinosis described previously with diffuse GGO and marked septal thickening (Fig. 17).61a This disease also is usually fatal unless lung transplantation is performed. A second rare neonatal condition that has a recognizable CT pattern is alveolar capillary dysplasia. In this disorder there is apparent malalignment of the pulmonary vessels with the capillaries. Clinical presentation is marked pulmonary hypertension. CT shows diffuse GGO with ‘‘dark bronchus’’ sign. Finally, there is a newly reported condition,

106

KUHN & BRODY

Figure 15. A 3-mm EBCT image shows bronchopulmonary dysplasia in a 6-year-old patient. A, Section through the lung bases reveals hyperaerated lower lobes especially on the left. Mild bronchial dilatation is noted. This condition had developed since an earlier study at age 4. B, On a higher level there are typical changes of bronchopulmonary dysplasia with linear, fibrotic changes medially, some loss of upper lobe lung volume, and primarily peripheral areas of air-trapping. Small, triangular, pleural-based opacities are present at both levels.

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

Figure 16. End stage pulmonary fibrosis in the same patient as Figure 9 now at age 4 months. The lungs are essentially destroyed with severe small honeycomb cysts and marked septal thickening throughout. The patient died following a lung transplant.

Figure 17. A 3-mm EBCT image showing congenital alveolar proteinosis caused by surfactant B deficiency in a 3-week-old patient. Note the similarity in appearance to Figure 7 with ground-glass opacities and smooth interlobular septal thickening.

107

108

KUHN & BRODY

Figure 18. Persistent tachypnea of infancy. A 3-month-old infant with tachypnea since birth. Chest radiography revealed nonspecific central haziness. A, HRCT shows findings of bilateral GGO in a somewhat dependent distribution. Biopsy revealed pulmonary neuroendocrine cell hyperplasia. B, Follow-up HRCT 15 months later using stop-ventilation technique shows improvement, although the basic pattern of GGO persists. (Courtesy of Dr. R. Cohen, Oakland, CA.)

persistent tachypnea of infancy, that still is poorly defined.15a, 71a These patients present in infancy with chronic tachypnea and hypoxemia. Radiographic findings are nonspecific and may be normal or show mild interstitial disease and peribronchial thickening. Some CT scans have been normal, others (Fig. 18) show GGO suggestive of edema. Pathologic findings characteristically include pulmonary neuroendocrine cell hyperplasia. Affected infants show gradual improvement over a course of months to years. SUMMARY High-resolution CT in children remains a technically challenging procedure, both to perform optimally and to interpret correctly. Although much remains to be learned about its optimal application, it is apparent that often confusing or nonspecific chest radiographs are clarified and a much clearer understanding is being gained about the diagnosis and evolution of both common and unusual pediatric lung diseases. As new therapies become available for these disorders, and CT becomes faster and easier to perform, it will become increasingly used not only for more accurate diagnosis but also for better evaluation of effects of therapy.

3.

4. 5.

6. 7. 8. 9. 10.

11.

12.

13.

References 1. Albafouille V, Sayegh N, De Coudenhove S: CT scan patterns of pulmonary alveolar proteinosis in children. Pediatr Radiol 29:147–152, 1999 2. Ambrosino MM, Roche KJ, Genieser NB, et al: Ap-

14. 15.

plication of thin-section low-dose chest CT (TSCT) in the management of pediatric AIDS. Pediatr Radiol 25:393–400, 1995 Andrew M, David M, Adams M, et al: Venous thromboembolic complications (VTE) in children: First analyses of the Canadian registry of VTE. Blood 83:1251–1257, 1994 Aquino SL, Gamsu G, Webb WR, et al: Tree-in-bud pattern: Frequency and significance on thin section CT. J Comput Assist Tomogr 20:594–599, 1996 Aquino SL, Schechter M, Chiles C, et al: High resolution inspiratory and expiratory CT in older children and adults with bronchopulmonary dysplasia. AJR Am J Roentgenol 173:963–967, 1999 Bauman N, Smith R: Recurrent respiratory papillomatosis. Pediatr Clin North Am 43:1385–1401, 1996 Bhalla M, Abboud MR, McLoud TC, et al: Acute chest syndrome in sickle cell disease: CT evidence of microvascular occlusion. Radiology 187:45–49, 1993 Bhalla M, Turcios N, Aponte V, et al: Cystic fibrosis scoring system with thin section CT. Radiology 179:783–788, 1991 Brauner MW, Grenier P, Tijani K: Pulmonary Langerhans’ cell histiocytosis: Evolution of lesions on CT scan. Radiology 204:497–502, 1997 Brody AS, Molina PL, Klein JS, et al: High resolution computed tomography of the chest in children with cystic fibrosis: Support for use as an outcome surrogate. Pediatr Radiol 29:731–735, 1999 Chang AB, Masel JP, Masters B: Post-infectious bronchiolitis obliterans: Clinical, radiological and pulmonary function sequelae. Pediatr Radiol 28:23– 29, 1998 Coakley F, Cohen MD, Johnson MS, et al: Detection of pulmonary metastases with pathological correlation: Effect of breathing on the accuracy of spiral CT. Pediatr Radiol 27:576–579, 1997 Coleman L, Kramer SS, Markowitz RI, et al: Bronchiectasis in children. J Thorac Imaging 10:268–279, 1995 Connolly B, Manson D, Ebeshard A, et al: CT appearance of pulmonary vasculitis in children. AJR Am J Roentgenol 167:901–904, 1996 DeSena S, Rosenfeld D, Santos S, et al: Jugular

HIGH-RESOLUTION CT OF PEDIATRIC LUNG DISEASE

15a. 16.

17. 18. 19.

20. 21.

22. 23.

24.

25. 26. 27.

28.

29. 30.

31.

32.

33.

thrombophlebitis complicating bacterial pharyngitis (Lemierre syndrome). Pediatr Radiol 26:141–144, 1996 Deterding R, Hay T, Langston C, et al: Persistent tachypnea of infancy. Am J Respir Crit Care Med 155:A715, 1997 Dufour B, Maitre S, Humbe M, et al: High-resolution CT of the chest in four patients with pulmonary capillary hemangiomatosis or pulmonary venoocclusive disease. AJR Am J Roentgenol 171:1321– 1324, 19?? Eglehoff J, Ball W Jr: Safety and efficacy of sedation in children using a structured sedation program. AJR Am J Roentgenol 168:1259–1262, 1997 Fishback N, Koss M: Update on lymphoid interstitial pneumonitis. Curr Opin Pulm Med 2:429–433, 1996 Garg K, Lynch DA, Newell JD, et al: Proliferative and constrictive bronchiolitis: Classification and radiologic features. AJR Am J Roentgenol 162:803– 808, 1994 Griscom NT: Diseases of the trachea, bronchi and smaller airways. Radiol Clin North Am 31:605–615, 1993 Gruden J, Webb WR, Warnock M: Centrilobular opacities in the lung on high resolution CT: Diagnostic consideration and pathologic correlation. AJR Am J Roentgenol 162:569–574, 1994 Ha S, Helms P, Fletcher M, et al: Lung involvement in Langerhans’ cell histiocytosis: Prevalence, clinical features and outcome. Pediatrics 89:466–469, 1992 Haliloglu M, Karcaaltincaba M, Toru M, et al: CT presentation of Wegener ’s granulomatosis in a child: Rapidly progressive changes of pulmonary nodules to cavities. Eur J Radiol 35:12–14, 2000 Hansell DM, Wells AU, Rubens MB, et al: Bronchiectasis: Functional significance of areas of decreased attenuation at expiratory CT. Radiology 193:369– 374, 1994 Helbich TH, Heinz-Peer G, Eichler I, et al: Cystic fibrosis: CT assessment of lung involvement in children and adults. Radiology 213:537–544, 1999 Helbich TH, Heinz-Peer G, Fleischmann D, et al: Evolution of CT findings in patients with cystic fibrosis. AJR Am J Roentgenol 173:81–88, 1999 Helbich TH, Wojnarovsky C, Wunderbaldinger P, et al: Pulmonary alveolar microlithiasis in children: Radiographic and high-resolution CT findings. AJR Am J Roentgenol 168:63–65, 1997 Helton K, Kuhn J, Fletcher BD, et al: Bronchiolitis obliterans-organizing pneumonia (BOOP) in children with malignant disease. Pediatr Radiol 22:270– 274, 1992 Herold CJ, Brown RH, Mitzner W, et al: Assessment of pulmonary airway reactivity with high resolution CT. Radiology 181:369–374, 1991 Hoffman WH, Locksmith JP, Burton EM, et al: Interstitial pulmonary edema in children and adolescents with diabetic ketoacidosis. J Diabetes Complications 12:314–320, 1998 Howling SJ, Northway WH Jr, Hansell DM, et al: Pulmonary sequelae of bronchopulmonary dysplasia survivors: High-resolution CT findings. AJR Am J Roentgenol 174:1323–1326, 2000 Jabra AA, Fishman EK, Shehata BM, et al: Localized persistent pulmonary interstitial emphysema: CT findings with radiographic-pathologic correlation. AJR Am J Roentgenol 169:1381–1384, 1997 Kang EY, Grenier P, Laurent F, et al: Interlobular septal thickening: Patterns at high-resolution computed tomography. J Thorac Imaging 11:260–264, 1996

109

34. Karnack I, Ciftci A, Talim B, et al: Pulmonary lymphomatoid granulomatosis in a 4 year old. J Pediatr Surg 34:1033–1035, 1999 35. Katzenstein AL, Gordon L, Oliphant M, et al: Chronic pneumonitis of infancy: A unique form of interstitial lung disease occurring in early childhood. Am J Surg Pathol 19:439–477, 1995 36. Kim CK, Chung CY, Kim JS, et al: Late abnormal findings on high-resolution computed tomography after mycoplasma pneumonia. Pediatrics 105:372– 378, 2000 37. Kim WS, Moon WK, Kim IO, et al: Pulmonary tuberculosis in children: Evaluation with CT. AJR Am J Roentgenol 168:1005–1009, 1997 38. Kinane B, Mansell A, Zwerdling R, et al: Follicular bronchitis in the pediatric population. Chest 104:1183–1186, 1993 39. Koh DM, Hansell DM: Computed tomography of diffuse interstitial lung disease in children. Clin Radiol 55:659–667, 2000 40. Kuhn J: Diseases of the airways. In Silverman F, Kuhn J (eds): Caffey’s Pediatric X Ray Diagnosis, vol 1. Philadelphia, Mosby Yearbook, 1993, pp 477–484 41. Kuhn J: Normal lung and anomalies. In Silverman F, Kuhn J (eds): Caffey’s Pediatric X-Ray Diagnosis, vol 1. Philadelphia, Mosby Yearbook 1993, pp 454–455 42. Kuhn JP: High resolution CT of diffuse parenchymal disorders. In RSNA Special Course in Pediatric Radiology. Chicago, 1999, pp 57–67 43. Kuhn JP: High-resolution computed tomography of pediatric pulmonary parenchymal disorders. Radiol Clin North Am 31:533–551, 1993 44. Lau D, Siegel MJ, Hildebolt C, et al: Bronchiolitis obliterans syndrome: Thin section CT diagnosis of obstructive changes in infants and young children after lung transplantation. Radiology 208:783–788, 1998 45. Lee EL, Gotway M, Reddy G, et al: Early bronchiolitis obliterans following lung transplantation: Accuracy of expiratory thin section CT for diagnosis. Radiology 216:472–477, 2000 46. Long FR: Structural changes in the airways of infants with cystic fibrosis. Pediatr Pulmonol 20:295, 2000 47. Long FR, Castile RG, Brody AS, et al: Lungs in infants and young children: Improved thin-section CT with a noninvasive controlled-ventilation technique–initial experience. Radiology 212:588–593, 1999 48. Lucaya J, Garcia-Pena P, Herrera L, et al: Expiratory chest CT in children. AJR Am J Roentgenol 174:235– 241, 2000 49. Lucaya J, Gartner S, Garcia-Pena P, et al: Spectrum of manifestations of Swyer-James MacLeod syndrome. J Comput Assist Tomogr 22:592–597, 1998 50. Lynch DA, Brasch RC, Hardy KA, et al: Pediatric pulmonary disease: Assessment with high-resolution ultrafast CT. Radiology 176:243–248, 1990 51. Lynch DA, Hay T, Newell JD, et al: Pediatric diffuse lung disease: Diagnosis and classification using high resolution CT. AJR Am J Roentgenol 173:713–718, 1999 52. Maffessanti M, Canduss M, Brizzi F, et al: Cystic fibrosis in children: HRCT findings and distribution of disease. J Thorac Imaging 11:27–38, 1996 53. Marks M, Haney P, McDermott M, et al: Thoracic disease in children with AIDS. Radiographics 16:1349–1362, 1996 54. Marti-Bonmati L, Ruiz Perales F, Catala F, et al: CT findings in Swyer-James syndrome. Radiology 172:477–480, 1989 55. Massie R, Armstrong D: Bronchiectasis and bronchio-

110

56.

57. 58. 59. 60. 61. 61a. 62.

63.

64.

65. 66.

67.

68. 69. 70. 70a. 71. 72.

KUHN & BRODY litis obliterans post respiratory syncytial virus infection: Think again. J Paediatr Child Health 35:497– 498, 1999 Mathew P, Bozeman P, Krance RA, et al: Bronchiolitis obliterans organizing pneumonia (BOOP) in children after allogeneic bone marrow transplantation. Bone Marrow Transplant 13:221–223, 1994 Mayo JR, Jackson SA, Muller M, et al: High resolution CT of the chest: Radiation dose. AJR Am J Roentgenol 160:479–481, 1993 McAdams HP, Erasmus JJ, Crockett R, et al: The hepatopulmonary syndrome: Radiologic findings in 10 Patients. AJR Am J Roentgenol 166:1379–1385, 1996 McHugh K, Manson D, Eberhard B, et al: Wegener’s granulomatosis in childhood. Pediatr Radiol 21:552– 555, 1991 Mu¨ller L: Clinical value of high-resolution CT in chronic diffuse lung disease. AJR Am J Roentgenol 191:1163–1170, 1991 Nathanson I, Conboy K, Murphy S, et al: Ultrafast computed tomography of the chest in cystic fibrosis: A new scoring system. Pediatr Pulmonol 11:81–86, 1991 Newman B, Kuhn JP, Kramer SS, et al: Congenital surfactant protein B deficiency–emphasis on imaging. Pediatr Radiol 31:327–331, 2001 Nicholson A, Kim HC, Corrin B, et al: The value of classifying interstitial pneumonitis in childhood accoring to defined histological patterns. Histopathology 33:203–211, 1998 Nogee L, deMello D, Dehner L, et al: Brief report: Deficiency of pulmonary surfactant protein B in congenital alveolar proteinosis. N Engl J Med 328:406–410, 1993 Nuhoglu Y, Bahceciler N, Yuksel M, et al: Thorax high resolution computerized tomography findings in asthmatic children with unusual clinical manifestations. Ann Allergy Asthma Immunol 82:311–314, 1999 Numan F, Islak C, Berkmen T, et al: Behcet disease: Pulmonary arterial involvement in 15 cases. Radiology 192:465–468, 1994 Oh YW, Effmann EL, Redding G, et al: Follicular hyperplasia of bronchus-associated lymphoid tissue causing severe air trapping. AJR Am J Roentgenol 172:745–747, 1999 Oppenheim C, Mamou-Mani T, Sayegh N, et al: Bronchopulmonary dysplasia: Value of CT in identifying pulmonary sequelae. AJR Am J Roentgenol 163:169–172, 1994 Pappas J, Donnelly LF, Frush DP: Reduced frequency of sedation in young children with multisection helical CT. Radiology 215:897–899, 2000 Pearson A, Kirpalani H, Ashcroft T, et al: Lymphomatoid granulomatosis in a 10 year old boy. BMJ 286:1313–1314, 1983 Pick A, Deschamps C, Stanson AW: Pulmonary arteriovenous fistula: Presentation, diagnosis, and treatment. World J Surg 23:1118–1122, 1999 Pye C, Fan L, Langston C: Pulmonary neuroendocrine cell hyperplasia in persistent tachypnea of infancy. Mod Pathol 11:4, 1998 Reittner P, Fotter R, Lindbichler F, et al: HRCT features in a 5-year-old child with follicular bronchiolitis. Pediatr Radiol 27:877–879, 1997 Santamaria F, Grillo G, Guidi G, et al: Cystic fibrosis:

73.

74. 75. 76. 77.

78. 79. 80. 81. 82. 83.

84. 85. 86.

87.

88.

89.

90.

When should high-resolution computed tomography of the chest be obtained? Pediatrics 101:908–913, 1998 Sargent MA, Cairns RA, Murdoch MJ, et al: Obstructive lung disease in children after allogeneic bone marrow transplantation: Evaluation with high-resolution CT. AJR Am J Roentgenol 164:693–696, 1995 Sargent MA, McEachern A, Jamieson DH, et al: Atelectasis on pediatric chest CT: Comparison of sedation techniques. Pediatr Radiol 29:509–513, 1999 Schmidt H, Lorcher U, Kitz R, et al: Pulmonary alveolar microlithiasis in children. Pediatr Radiol 26:33–36, 1996 Seely JM, Jones LT, Wallace C, et al: Systemic sclerosis: Using high-resolution CT to detect lung disease in children. AJR Am J Roentgenol 170:691–697, 1998 Serafini G, Cornara G, Cavalloro F, et al: Pulmonary atelectasis during paediatric anaesthesia: CT scan evaluation and effect of positive endexpiratory pressure (PEEP). Paediatr Anaesth 9:225–228, 1999 Shah A, Pant C, Bhagat R, et al: CT in childhood allergic bronchopulmonary aspergillosis. Pediatr Radiol 22:227–228, 1992 Shaker KG, Umali CB, Fraire AE: Langerhans’ cell histiocytosis of the lung in association with mediastinal lymphadenopathy. Pathol Int 45:762–766, 1995 Smets A, Mortele K, de Praeter G, et al: Pulmonary and mediastinal lesions in children with Langerhans’ cell histiocytosis. Pediatr Radiol 27:873–876, 1997 Stern EJ, Frank MS: Small-airway diseases of the lungs: Findings at expiratory CT. AJR Am J Roentgenol 163:37–41, 1994 Swensen SJ, Hartman TE, Mayo JR, et al: Diffuse pulmonary lymphangiomatosis: CT findings. J Comput Assist Tomogr 19:348–352, 1995 Tomashefski JF, Bruce M, Stern R: Pulmonary air cysts in cystic fibrosis: Relationship of pathologic features to radiologic findings and history of peneumothorax. Hum Pathol 16:253–261, 1985 Vock P, Malanowski D, Tschaeppeler H, et al: Computed tomographic lung density in children. Invest Radiol 22:627–631, 1987 Weibel E, Crystal R: Structural organization of the pulmonary interstitium. In Crystal R, West J (eds): The Lung, Vol 1. New York, Raven, 1991, pp 369–380 Worthy S, Muller M, Hartman TE, et al: Mosaic attenuation pattern on thin section CT scans of the lung: Differentiation among infiltrative lung, airway, and vascular diseases as a cause. Radiology 205:465– 470, 1997 Worthy S, Park CS, Kim J, et al: Bronchiolitis obliterans after lung transplantation: High resolution CT findings in 15 patients. AJR Am J Roentgenol 169:673–677, 1997 Wunderbaldinger P, Paya K, Partik B: CT and MR imaging of generalized cystic lymphangiomatosis in pediatric patients. AJR Am J Roentgenol 174:827– 832, 2000 Zhang L, Irion K, da Silva Porto N, et al: Highresolution computed tomography in pediatric patients with postinfectious bronchiolitis obliterans. J Thorac Imaing 14:85–89, 1999 Zwirewich C, Mayo JR, Muller M: Low dose, high resolution CT of lung parenchyma. Radiology 180:413–417, 1991 Address reprint requests to Jerald P. Kuhn, MD 160 Bryant Street Buffalo, NY 14222 e-mail: [email protected]

HIGH-RESOLUTION CT OF THE LUNG II

0033–8389/02 $15.00
.00

CT OF PULMONARY THROMBOEMBOLIC DISEASE Kavita Garg, MD

Although a clinical description of pulmonary embolism (PE) was reported in the early 1800s, it was von Virchow who first described the connection between venous thrombosis and PE. The first radiographic description of PE was reported in 1922 by Wharton and Pierson.61 Since that time, imaging has played an important role in the diagnosis of PE. For many years, ventilation-perfusion scintigraphy has been the main imaging modality used in the evaluation of patients with suspected PE. More recently, with the widespread availability of faster scanners, CT has emerged as another important diagnostic test.

PATHOPHYSIOLOGY Three primary influences predispose to thrombus formation, the so-called Virchow’s triad: (1) endothelial injury, (2) stasis or turbulence of blood flow, and (3) blood hypercoagulability. Conditions associated with an increased risk of thrombosis are listed in Table 1. An embolus is a detached intravascular solid, liquid, or osseous mass that is carried by the blood to a site distant from its point of origin. Virtually 99% of all emboli represent some part of a dislodged thrombus, hence the commonly used term thromboembolism. Rare forms of emboli include droplets of fat; bubbles of air or nitrogen; atherosclerotic debris (cholesterol emboli); tumor fragments;

bits of bone marrow; or foreign bodies, such as bullets, mercury, or radiation seeds. Unless otherwise specified, however, an embolus should be considered to be thrombotic in origin. Inevitably, emboli lodge in vessels too small to permit further passage, resulting in partial or complete vascular occlusion. The potential consequence of such thromboemTable 1. CONDITIONS ASSOCIATED WITH AN INCREASED RISK OF THROMBOSIS Primary (genetic) Mutations in factor V Antithrombin III deficiency Protein C or S deficiency Fibrinolysis defects Secondary (acquired) High risk for thrombosis Prolonged bed rest or immobilization Myocardial infarction Tissue damage (surgery, fracture, or burns) Cancer Prosthetic cardiac valves Disseminated intravascular coagulation Lupus anticoagulant Low risk for thrombosis Atrial fibrillation Cardiomyopathy Nephrotic syndrome Hyperestrogenic states Oral contraceptive use Sickle cell anemia Smoking Data from Mitchell RN, Kumar V: Hemodynamic disorders, thrombosis, and shock. In Kumar V, Cotran RS, Robbins SL (eds): Basic Pathology, ed 6. Philadelphia, WB Saunders, 1997, pp 60–80.

From the Department of Radiology, Veterans Affairs Medical Center, University of Colorado, Denver, Colorado

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

111

112

GARG

bolic events is the ischemic necrosis of distal tissue, known as infarction. More than 90% of all pulmonary emboli arise from thrombi within the large deep veins of the lower legs, typically originating in the popliteal vein and larger veins above it. The pathophysiologic consequences of thromboembolism in the lung depend largely on the size of the embolus, which in turn dictates the size of the occluded pulmonary artery, and on the cardiopulmonary status of the patient. There are two important consequences of PE: (1) an increase in pulmonary artery pressure caused by blockage of flow and possibly vasospasm caused by neurogenic mechanisms or release of mediators (e.g., thromboxane A2 and serotonin); and (2) ischemia of the downstream pulmonary parenchyma. Occlusion of major vessels or more than 60% of the arterial bed results in sudden increase in pulmonary artery pressures; diminished cardiac output; right-sided heart failure (acute cor pulmonale); or even death. Usually hypoxemia develops as a result of multiple mechanisms. If smaller vessels are occluded, the result is less catastrophic and the event may be clinically silent.37

pnea with or without associated anxiety, pleuritic chest pain, and hemoptysis are common but nonspecific symptoms of PE. Any of these symptoms may also develop with several other conditions, such as pneumonia, exacerbated chronic obstructive lung disease, congestive heart failure, or lung cancer. Lightheadedness and syncope may be caused by PE but may also result from several other entities that cause hypoxemia or hypotension. Clinical findings alone, however, are not a reliable guide to the diagnosis of PE, as is underscored by the high incidence of unsuspected PE in autopsy series.20 Physical examination of the patient with PE may reveal tachypnea, tachycardia, fever, and pleuritic rub, all of which are nonspecific symptoms. Hypoxemia is common in acute PE, but is not universally present. Even the alveolar–arterial difference may be normal in rare cases of PE, particularly in younger patients without concomitant lung disease.35, 54 Nonspecific electrocardiographic abnormalities may develop in acute PE, including Twave changes, ST segment abnormalities, and left or right axis deviation.

FREQUENCY It is estimated that deep venous thrombosis (DVT) and PE are associated with 300,000 to 600,000 hospitalizations a year and that as many as 50,000 individuals die each year as a result of PE in the United States. PE is predominantly a disease of older age. The increase in venous thromboembolic events with age is greater among men than women. In patients younger than 55 years, the incidence of PE is higher in women. The overall age- and sex-adjusted annual incidence of venous thromboembolism is reported to be 117 per 100,000 (DVT, 48 per 100,000; PE, 69 per 100,000).52 Acute PE is a common and potentially lifethreatening disorder. Treatment is highly effective but not without complications, and the diagnosis of PE requires a high degree of certainty. Conversely, if untreated, PE can be fatal. Treatment reduces mortality from 30% to less than 10%. CLINICAL DETAILS Pulmonary embolism should be considered whenever unexplained dyspnea occurs. Dys-

CHEST RADIOGRAPHY The chest radiograph is abnormal in most cases of PE but the findings are nonspecific. Common radiographic findings include atelectasis, pleural effusion, parenchymal opacification, and elevation of a hemidiaphragm. Classic radiographic findings of pulmonary infarction, such as a wedge-shaped pleuralbased triangular opacity with apex pointing toward the hilus (Hampton’s hump) or decreased vascularity (Westermark’s sign), are suggestive of PE but are infrequent. Prominent central pulmonary arteries, cardiomegaly (especially right heart), and pulmonary edema are other findings and in the appropriate clinical setting could be consistent with acute cor pulmonale. A normal-appearing chest radiograph in a patient with severe dyspnea and hypoxemia without evidence of bronchospasm or cardiac shunt strongly suggests PE. Generally, chest radiography cannot be used conclusively to prove or exclude PE; however, this modality and electrocardiography may be useful for determining alternative diagnoses.

CT OF PULMONARY THROMBOEMBOLIC DISEASE

CT SCAN The contrast-enhanced CT is increasingly being used as the initial radiologic study, especially in patients with abnormal chest radiographs in whom scintigraphy is more likely to be nondiagnostic, in the diagnosis of PE (Fig. 1).18, 22 Not only does CT show emboli directly like pulmonary angiograms, it is also noninvasive, cheaper, and widely available like ventilation-perfusion scintigraphy. In one study cost-effectiveness of CT was analyzed.58 The authors used a model with various diagnostic algorithms consisting of combinations of scintigraphy, ultrasound, D-dimer assay, conventional angiography, and spiral CT. The results of this study showed that all of the best strategies included spiral CT and its use improved cost-effectiveness in the diagnostic

113

work-up of suspected PE. Moreover, CT is the only test that can provide significant additional information or an alternate diagnosis, which is a clear advantage of CT when compared with either pulmonary angiography or scintigraphy.7, 18, 19 MR imaging is the only imaging study that has potential to challenge CT for the diagnosis of acute thromboembolic disease.53 At this time, however, MR imaging has not been widely applied to the diagnosis of PE, although early results show potential.13, 25, 26, 31, 36 Technique In the evaluation of patients with suspected PE, initial unenhanced axial CT of the chest is performed to analyze lung parenchyma,

Figure 1. Algorithm for evaluation of thromboembolic disease. PE  pulmonary embolism; DVT  deep venous thrombosis; CXR  chest X-ray; V/Q  ventilation-perfusion.

114

GARG

airways, and the chest wall followed by contrast-enhanced spiral CT (CT pulmonary angiogram). Lung images are reconstructed with a high-spatial-frequency algorithm and photographed at a window width of 1500 H and level of 550 H. Because calcifications are seen on lung window, it may not be necessary to photograph unenhanced images at mediastinal window (width of 400 H and level of 40 H). A low-dose technique may be adequate for unenhanced CT in most patients. CT pulmonary angiography technique for a single-detector helical CT scanner includes 3-mm collimation, 2-mm reconstruction interval, pitch of 2, and an average acquisition time of 24 seconds. These parameters allow thinnest collimation possible in a single breath-hold of the target volume. On a multidetector scanner, the appropriate protocol includes 1.25-mm collimation, with a pitch of 6 and 7.5 cm/s table feed. Images should be reconstructed at 1.25-mm thickness and at 0.6- to 1-mm intervals. Images are acquired from aortic arch to the diaphragm either in craniocaudal or caudocranial direction. Patients on a respirator are imaged either during forced apnea or while breathing at minimal tidal volume and respiratory rate. It is important that patients maintain strict apnea while they are scanned. Although interpretable images might still be obtained while the patient is gently breathing, a confident evaluation of pulmonary arteries is usually limited only to central vessels (Fig. 2).

Iodinated contrast media is administered as a bolus with an automated injector. The injection is monitored carefully by a registered nurse or a physician. Generally, a high volume of undiluted or diluted contrast (100 to 150 mL) at a high-flow rate (3 to 5 mL/s) is used. The scan delay time is related to the flow rate and intravenous access site. A 15to 17-second scan delay is generally adequate in most patients with administration of contrast at 4 mL/s rate by 18- or 20-gauge antecubital intravenous catheter. The delay time is decreased or increased when more central or peripheral intravenous access is used. In patients with suspected abnormal hemodynamics a timing study can be performed to determine the appropriate scan acquisition delay for contrast enhancement. Because DVT and PE are part of the same disease process, CT venography can be easily performed after CT pulmonary angiography without additional contrast. This study requires only a few extra minutes and allows ‘‘one-stop imaging’’ for both PE and DVT. Investigators have described varying techniques for CT venography. 6, 17, 32, 33, 65 A 3minute delay after the start of injection before acquiring venography images is adequate for most patients. A longer delay of up to 4 minutes is suggested in older patients with peripheral arterial disease and abnormal hemodynamics.17 Thinner slices at narrower intervals in a reasonable time are possible with multidetector CT. Currently, with single-detector CT, 5- or 10-mm-thick axial sections can be obtained at 20-mm intervals from the knees to the mid abdomen to evaluate femoropopliteal veins, pelvic veins, and the inferior vena cava. This allows for required anatomic coverage in a reasonable time without missing assessment of longer venous segments. Imaging of calf veins and the clinical significance of isolated small calf deep venous thrombi are controversial.24, 29 Image Interpretation

Figure 2. Acute pulmonary embolism (PE) in a 53-yearold man who was unable to hold his breath. CT angiogram shows intraluminal filling defects in left lower lobe artery and lingular artery that are distended (arrows). In spite of relatively poor opacification and breathing artifacts, central emboli were seen clearly on multiple contiguous images.

Technically good quality studies with optimum opacification of vessels and knowledge of bronchovascular anatomy are key to an accurate interpretation of CT scans for the evaluation of PE. A systematic approach to identify all vessels is important. The bronchovascular anatomy has been described based on segmental anatomy of lungs.3, 28 The segmental arteries are seen near the accompa-

CT OF PULMONARY THROMBOEMBOLIC DISEASE

Figure 3. Acute PE in a 56-year-old man. Axial CT angiogram shows intraluminal defect distending the right middle lobe artery (arrow).

nying branches of the bronchial tree and are situated either medially (in the upper lobes) or laterally (in the lower lobes, lingula, and the right middle lobe). The segmental arteries that have a vertical course (those to the upper and lower lobes) are better evaluated than arteries to the lingula, middle lobe, and to the superior segments of the lower lobes, which run obliquely with respect to axial images (Fig. 3). Reconstruction of images in sagittal plane or with the long axis of the artery can sometimes help to confirm a true intraluminal filling defect seen on axial images. Source images (directly acquired axial images) are technically the best in quality; however, diagnostic quality multiplanar reconstructed images can be obtained especially with the data set acquired with the thinnest collimation

115

possible and without motion artifacts. This should be possible more frequently if not routinely with images acquired with multidetector scanners. Using multiplanar reconstructions on a workstation allows faster, easier, and more confident evaluation of CT angiograms. The images can be zoomed and if needed the windowing can be changed to assess relationship of the artery and bronchus. Modified mediastinal window setting referenced to the right or left main pulmonary artery attenuation may improve depiction of small pulmonary emboli, which may otherwise be poorly seen or obscured with standard window settings.4 In the author’s practice although every alternate image of CT pulmonary angiogram is still photographed, the interpretation is primarily done at a workstation. When PE is identified, it is characterized as acute or chronic. An embolus is considered acute if it is situated centrally within the vascular lumen or if it occludes a vessel (vessel cut-off sign) (Figs. 2 to 5). An acute PE commonly distends the involved vessel (see Figs. 2 and 3). An embolus is considered chronic if it is eccentric and contiguous with vessel wall with or without calcification (Fig. 6). Resolving embolus may result in reduction in the arterial diameter by less than 50%. Recanalization within the thrombus and an arterial web (Fig. 7) are other uncommon findings of chronic PE. Pulmonary embolism is further characterized as central or peripheral depending on the location or the arterial branch involved. Central vascular zones include the main pul-

Figure 4. Acute PE in a 50-year-old woman. Axial CT angiogram shows small emboli in segmental and subsegmental arteries of lower lobes bilaterally (arrows). Bilateral pleural effusions and a triangular pleural based consolidation in left lower lobe also were noted.

116

GARG

Figure 5. Small nonoccluding pulmonary embolus in a 48-year-old man with cardiomyopathy. Axial CT angiogram shows a tiny intraluminal filling defect (arrow) in right posterior basal segmental artery.

monary artery, the left and right main pulmonary arteries, the anterior trunk, the right and left interlobar arteries, the left upper lobe trunk, the right middle lobe artery, and the right and left lower lobe arteries. Peripheral vascular zones include the segmental and subsegmental arteries of the right upper lobe, right middle lobe, right lower lobe, left upper lobe, lingula, and the left lower lobe. The rest of the chest is evaluated as routine. A recent study evaluated the value of lung parenchymal and pleural findings in 88 patients with clinical suspicion of acute PE who

Figure 6. Chronic PE in a 67-year-old man with known pulmonary arterial hypertension. Axial CT angiogram shows an eccentric mural thrombus (arrowhead) along the anterior wall of the right main pulmonary artery.

underwent CT pulmonary angiography. 7 Wedge-shaped pleural-based consolidation and linear bands were significantly more often seen in patients with PE than those without PE. The authors concluded that these parenchymal findings may suggest further investigations when results of spiral CT are inconclusive in diagnosis of PE. More importantly lung parenchymal findings may provide an alternate diagnosis or show additional findings in up to 50% of patients with suspected PE.18 Commonly made alternate diagnoses include emphysema (clinically exacerbation of chronic obstructive pulmonary disease); aspiration pneumonia; pulmonary edema (or acute respiratory distress syndrome); interstitial lung disease; lung cancer or other malignancy; and hypersensitivity pneumonitis. The CT criteria for diagnosis of DVT are the tomographic equivalent of the classic findings of DVT described on conventional venography or sonography.2, 8, 9, 10, 15, 30, 41, 60, 63 Pitfalls and Artifacts A number of interpretive pitfalls exist in assessing CT angiography, but their recognition is easier as the radiologist gains experience with interpretations. The pitfalls, especially those related to volume averaging of perivascular tissue, branching points, and nonvertical vessels, can be limited by using a trackball on a workstation and by having knowledge of vascular anatomy. The lymphatic and connective tissue seen more commonly adjacent to central vessels is located between the artery and the bronchus (Fig. 8). In severely tachypneic patients, breathing artifacts are commonly observed especially in obliquely oriented arteries because of the variable position of the vessel in the section width on two contiguous slices. Overall, 2% to 4% of CT pulmonary angiography examinations are reported to be nondiagnostic because of severe motion artifacts (severe dyspnea), which is similar to what has been reported with conventional angiography. Extensive parenchymal opacification, a large pleural effusion, and other congenital or acquired causes of asymmetric pulmonary vascular resistance and slow flow may lead to pseudofilling defects, similar to those previously described on conventional angiography and scintigraphy.5, 48 Whenever possible the artifact caused by slow flow can be con-

CT OF PULMONARY THROMBOEMBOLIC DISEASE

117

Figure 7. A and B, Acute and chronic PE in a 78-year-old man with a history of pulmonary hypertension and acute shortness of breath after total hip arthroplasty. Axial CT angiogram shows a spiral linear filling defect (arrow) in a left lower lobe artery indicative of a web. An intraluminal filling defect in the right lower lobe artery (arrowhead), which extended on multiple contiguous images inferiorly, is consistent with acute embolus.

firmed easily with a repeat study of the target vasculature with additional contrast and a longer delay. The scanning delay used may limit the quality of arterial opacification at the beginning or end of the scanned volume. An inadequate delay before initiation of imaging whether technical or caused by pathologic causes like upper extremity central DVT or impaired cardiac contractility may cause inadequate enhancement of the pulmonary arteries on the first few slices; on the other hand, if the scanning delay is too long, there is not enough contrast left for the later slices.43 CT venograms should be added to CT angiogram in intubated patients or in patients who cannot hold their breath and who are at high risk for thromboembolic disease, be-

Figure 8. Perivascular mildly enlarged lymph nodes. Coronal reconstructed image from spiral CT data shows prominent extraluminal soft tissue (curved arrow) interposed between artery and the bronchus (not well seen on this window).

cause if thrombus is seen in deep veins, then follow-up examinations (sonography or CT) or more invasive tests like catheter angiography can be obviated (Figs. 9 and 10).

Diagnostic Accuracy of CT Pulmonary Angiography In most cases when spiral CT is positive for PE, the emboli are multiple with intraluminal filling defects in the larger central arteries and in the segmental and subsegmental vessels. An apparent filling defect in a single segmental or especially subsegmental vessel can be challenging. One should keep in mind all the pitfalls discussed previously before diagnosing subsegmental embolus. The sensitivity of spiral CT for central PE is up to 100%; however, it is reported to be variable and generally much lower (Table 2), in the 53% to 91% range, when subsegmental vessels are analyzed. There is also a wide range of the reported incidence of isolated subsegmental PE from 5% in the PIOPED study57 to 36% in another study.39 Moreover, the true significance of small emboli has not been proved conclusively. It is generally believed that small thromboemboli may have clinical significance in patients with limited cardiopulmonary reserve. Pulmonary angiography demonstrates subsegmental vessels in much more detail than CT, although superimposition of the small vessels remains a limiting factor resulting in only 45% interobserver

118

GARG

Figure 9. Acute deep venous thrombosis in a 65-year-old man with suspected PE. CT venogram shows an intraluminal filling defect in the right greater saphenous vein at its confluence with common femoral vein (arrow).

Figure 10. Acute deep venous thrombosis in a 63-year-old man. Axial CT venogram shows a central intraluminal defect distending the right distal superficial vein.

Table 2. ACCURACY OF SPIRAL CT PULMONARY ANGIOGRAPHY

Authors

No. of Patients

Sensitivity (%)

Specificity (%)

CT Technique (Collimation)/ Anatomic Level

Remy-Jardin et al46 Goodman et al21 Remy-Jardin et al44 Mayo et al34 Garg et al19 Drucker et al12 Qanaldi et al40

42 20 75 142 26 47 157

100 86 91 87 67 53–60 90

96 92 78 95 100 81–97 94

5 mm/segmental 5 mm/segmental 5 and 3 mm/segmental 5 mm/segmental 3 mm/subsegmental 5 mm/segmental *2.7 mm/subsegmental

*Dual-section spiral CT.

CT OF PULMONARY THROMBOEMBOLIC DISEASE

agreement rate for isolated subsegmental PE.11 The agreement rate for CT pulmonary angiography is much higher, in the range from 75% to 96%.14, 16, 34 This is also significantly higher than the agreement rate (k value, 0.61 and 0.39) reported for scintigraphy. Clinicians in general have more confidence in CT when CT shows emboli than when a CT scan is interpreted as negative for PE. Two recent studies addressed clinical validity of CT being interpreted as negative for PE.18, 23 Both studies reported uneventful clinical outcome in patients who were not treated and in whom CT scans were interpreted as negative for CT. A negative predictive value of 99% was estimated in both studies. The outcome was similar to that reported for patients clinically suspected of having PE but without emboli on pulmonary arteriography.38 This indicates that although some small emboli may be missed at spiral CT, it does not seem that the subsequent morbidity from PE is high. DIAGNOSTIC ACCURACY OF CT VENOGRAPHY Few investigators have assessed the sensitivity and specificity of CT venography compared with bilateral leg sonography. CT and sonographic findings correlated exactly in the femoropopliteal deep venous system in one study,33 whereas two CT venography studies out of 68 had false-positive findings of DVT because of flow artifacts in another study. In the latter study the authors reported a sensitivity and specificity of 100% and 97%, respectively.17 Interobserver agreement variability in the interpretation of CT venograms results mainly because of the technical quality of the study similar to what has been observed with CT pulmonary angiography studies. The flow artifact and relatively poor opacification of the veins seen more commonly in patients with marked peripheral vascular disease is the main pitfall, which can lead to misinterpretation of CT venograms.16 FOLLOW-UP OF ACUTE THROMBOEMBOLIC DISEASE Several investigators have reported changes within arteries, bronchi, and lung parenchyma after acute PE.45, 47, 51, 55 In a study 62 patients with acute PE diagnosed by CT

119

or angiography underwent spiral CT after a mean of 11 months. At follow-up 48% had complete resolution and 52% showed residual arterial changes. Patients with more extensive acute PE more often showed incomplete resolution of acute PE and development of chronic PE occurred in 13%. 42 In a study group of 19 patients with acute PE identified with spiral CT, resolving clots were found in 68% (13 of 19 patients) at 6-week follow-up.59 The development of chronic changes subsequent to acute PE has been reported with a frequency varying between 2% and 18% of cases by scintigraphy. In a study of 55 patients with suspected chronic thromboembolic pulmonary hypertension, spiral CT was compared with MR imaging with conventional angiography and surgical correlation.1 Central vessel disease was assessed more accurately with CT than with MR imaging or angiography. Segmental vessels were also assessed more accurately with CT compared with MR imaging. The authors concluded that CT may be a useful alternative to angiography for diagnosis of chronic thromboembolism but it may not be sufficient for selecting candidates for surgery in all cases. Residual changes from acute DVT may be seen in up to 50% of sonograms at 6 months and may persist indefinitely (Fig. 11). The chronicity of DVT may not always be assessable even with serial examinations. The potential for overtreatment exists if old clot is misdiagnosed as acute DVT. Alternatively, acute clot superimposed on the background of chronic DVT must be recognized, or the patient may be undertreated. CT venography and sonography play complementary roles in the differentiation of acute from chronic DVT (Table 3).

Table 3. PARAMETERS ASSOCIATED WITH ACUTE VERSUS CHRONIC DVT Parameter Vein size Calcification Wall thickening or web Compressibility Valvular competence Collateral veins Flow

Acute DVT

Chronic DVT

Increased Absent Absent

Decreased Present Present

None Present Absent None

Partial Absent Present Some (recanalization)

DVT  Deep venous thrombosis.

120

GARG

Figure 11. Chronic deep venous thrombosis in a 57-year-old man with suspected PE. CT venogram shows calcified thrombi in iliac veins bilaterally (arrows).

FUTURE CONSIDERATIONS Many institutions now include CT as their initial diagnostic examination for PE. With the availability of faster multidetector CT scanners and the capability of acquiring CT venograms for combined evaluation of both PE and DVT, the role of CT will become even more significant. Advances in CT technology combined with faster user-friendly workstations may result in better display and visualization strategies of CT data for pulmonary angiography. The feasibility of MR imaging has been reported by few investigators; however, its role in most patients is limited to those who have impaired renal function or other contraindications for iodinated contrast. Newer blood pool contrast agents and respiratory navigators may enhance the role of MR imaging in the diagnosis of PE.

7. 8.

9. 10. 11.

12. 13.

14.

References 1. Bergin CJ, Sirlin CB, Hauschildt JP, et al: Chronic thromboembolism: Diagnosis with helical CT and MR imaging with angiographic and surgical correlation. Radiology 204:695–702, 1997 2. Bettmann MA, Robbins A, Braun SD, et al: Contrast venography of the leg: Diagnostic efficiency, tolerance, and complication rates with ionic and nonionic contrast media. Radiology 165:113–116, 1987 3. Boyden EA: Segmental Anatomy of the Lungs. New York, McGraw-Hill, 1955 4. Brink JA, Woodard PK, Horesh L, et al: Depiction of pulmonary emboli with spiral CT: Optimization of display window settings in a porcine model. Radiology 204:703–708, 1997 5. Brookstein JJ, Silver TM: The angiographic differential diagnosis of acute pulmonary embolism. Radiology 110:25–33, 1974 6. Cham MD, Yankelevitz DF, Shaham D, et al: Deep

15. 16.

17.

18.

19.

venous thrombosis: Detection by using indirect CT venography. Radiology 216:744–751, 2000 Coche EE, Muller NL, Kim K, et al: Acute pulmonary embolism: Ancillary findings at spiral CT. Radiology 207:753–758, 1998 Cogo A, Lensing AWA, Wells P, et al: Noninvasive objective tests for the diagnosis of clinically suspected deep-venous thrombosis. Haemostasis 25:27– 39, 1995 Cronan JJ: Venous thromboembolic disease: The role of US. Radiology 186:619–630, 1993 Cronan JJ, Dorfman GS, Scola FH, et al: Deep venous thrombosis: US assessment using vein compression. Radiology 162:191–194, 1987 Diffin DC, Leyendecker JR, Johnson SP, et al: Effect of anatomic distribution of pulmonary emboli on interobserver agreement in the interpretation of pulmonary angiography. AJR Am J Roentgenol 171:1085–1089, 1998 Drucker EA, Rivitz SM, Shepard JO, et al: Acute pulmonary embolism: Assessment of helical CT for diagnosis. Radiology 209:235–241, 1998 Erdmann WA, Peshock RM, Redman HC, et al: Pulmonary embolism: Comparison of MR images with radionuclide and angiographic studies. Radiology 190:499–508, 1994 Ferreti GR, Bosson JL, Buffaz PD, et al: CT in 164 patients with intermediate probability at ventilationperfusion scintigraphy and normal results at duplex US of legs. Radiology 205:453–458, 1997 Fraser JD, Anderson DR: Deep venous thrombosis: Recent advances and optimal evaluation with US. Radiology 211:9–24, 1999 Garg K, Kemp JL, Russ PD, et al: Thromboembolic disease: Variability of interobserver agreement in the interpretation of CT venography obtained in addition to CT pulmonary angiography. AJR Am J Roentgenol 176:1043–1047, 2001 Garg K, Kemp JL, Wojcik D, et al: Thromboembolic disease: Comparison of combined CT pulmonary angiography and venography with bilateral leg sonography in 70 patients. AJR Am J Roentgenol 175:997– 1001, 2000 Garg K, Sieler H, Welsh HC, et al: Clinical validity of helical CT being interpreted as negative for pulmonary embolism: Implications for patient management. AJR Am J Roentgenol 172:1627–1631, 1999 Garg K, Welsh CH, Feyerabend AJ, et al: Pulmonary

CT OF PULMONARY THROMBOEMBOLIC DISEASE

20.

21.

22. 23.

24. 25. 26. 27.

28. 29. 30. 31.

32.

33.

34.

35. 36. 37.

embolism: Diagnosis with spiral CT and ventilationperfusion scanning-correlation with pulmonary angiographic results or clinical outcome. Radiology 208:201–208, 1998 Goldhaber SZ, Hennekens CH, Evans DA, et al: Factors associated with correct antemortem diagnosis of major pulmonary embolism. Am J Med 73:822–826, 1982 Goodman LR, Curtin JJ, Mewissen MW, et al: Detection of pulmonary embolism in patients with unresolved clinical and scintigraphic diagnosis: Helical CT versus angiography. AJR Am J Roentgenol 164:1369–1374, 1995 Goodman LR, Lipchik RJ: Diagnosis of acute pulmonary embolism: Time for a new approach. Radiology 199:25–27, 1996 Goodman LR, Lipchik RJ, Kuzo RS, et al: Subsequent pulmonary embolism: Risk after a negative helical CT pulmonary angiogram-prospective comparison with scintigraphy. Radiology 215:535–542, 2000 Gottlieb RH, Widjaja J, Mehra S, et al: Clinically important pulmonary emboli: Does calf vein US alter outcomes? Radiology 211:25–29, 1999 Grist TM, Sostman HD, MacFall JR, et al: Pulmonary angiography with MR imaging: Preliminary clinical experience. Radiology 189:523–530, 1993 Gupta A, Frazer CK, Ferguson JM, et al: Acute pulmonary embolism: Diagnosis with MR angiography. Radiology 210:353–359, 1999 Hurst DR, Kazerooni EA, Stafford-Johnson D, et al: Diagnosis of pulmonary embolism: Comparison of CT angiography and MR angiography in canines. J Vasc Interv Radiol 10:309–318, 1999 Jackson CL, Huber JF: Correlated applied anatomy of bronchial tree and lungs with system of nomenclature. Dis Chest 9:319–326, 1943 Kakkar VV, Howe CT, Flanc C, et al: Natural history of postoperative deep venous thrombosis. Lancet 2:230–232, 1969 Lensing AWA, Prandoni P, Buller HR, et al: Lower extremity venography with iohexol: Results and complications. Radiology 177:503–505, 1990 Loubeyre P, Revel D, Douek P, et al: Dynamic contrast-enhanced MR angiography of pulmonary embolism: Comparison with pulmonary angiography. AJR Am J Roentgenol 162:1035–1039, 1994 Loud PA, Grossman ZD, Klippenstein DL, et al: Combined CT venography and pulmonary angiography: A new diagnostic technique for suspected thromboembolic disease. AJR Am J Roentgenol 170:951–954, 1998 Loud PA, Katz DS, Klippenstein DL, et al: Combined CT venography and pulmonary angiography in suspected thromboembolic disease: Diagnostic accuracy for deep venous evaluation. AJR Am J Roentgenol 174:61–65, 2000 Mayo JR, Remy-Jardin M, Muller NL, et al: Pulmonary embolism: Prospective comparison of spiral CT with ventilation-perfusion scintigraphy. Radiology 205:447–452, 1997 McFarlane MJ, Imperiale TF: Use of the alveolararterial oxygen gradient in he diagnosis of pulmonary embolism. Am J Med 96:57–62, 1994 Meany JFM, Weg JG, Chenevert TL, et al: Diagnosis of pulmonary embolism with magnetic resonance angiography. N Engl J Med 336:1422–1427, 1997 Mitchell RN, Kumar V: Hemodynamic disorders, thrombosis, and shock. In Kumar V, Cotran RS, Robbins SL (eds): Basic Pathology, ed 6. Philadelphia, WB Saunders, 1997, pp 60–80

121

38. Novelline RA, Baltarowich OH, Athanasoulis CA, et al: The clinical course of patients with suspected pulmonary embolism and a negative pulmonary arteriogram. Radiology 126:561–567, 1978 39. Oser RF, Zuckerman DA, Guttierez FR, et al: Anatomic distribution of pulmonary emboli at pulmonary arteriography: Implications for cross-sectional imaging. Radiology 199:31–35, 1996 40. Qanaldi SD, Hajjam ME, Mesurolle B, et al: Pulmonary embolism: Prospective evaluation of dual-section helical CT versus selective pulmonary arteriography in 157 patients. Radiology 217:447–455, 2000 41. Raghavendra BN, Rosen RJ, Lam S, et al: Deep venous thrombosis: Detection by high-resolution realtime ultrasonography. Radiology 152:789–793, 1984 42. Remy-Jardin M, Louvegny S, Remy J, et al: Acute central thromboembolic disease: Post-therapeutic follow-up with spiral CT angiography. Radiology 203:173–180, 1997 43. Remy-Jardin M, Remy J: Spiral CT angiography of the pulmonary vasculature. Radiology 212:615–636, 1999 44. Remy-Jardin M, Remy J, Deschildre F, et al: Diagnosis of pulmonary embolism with spiral CT: Comparison with pulmonary angiography and scintigraphy. Radiology 200:699–706, 1996 45. Remy-Jardin M, Remy J, Louvegny S, et al: Airway changes in chronic pulmonary embolism: CT findings in 33 patients. Radiology 203:355–360, 1997 46. Remy-Jardin M, Remy J, Wattinne L, et al: Central pulmonary thromboembolism: Diagnosis with volumetric CT with single-breath-hold technique-comparison with pulmonary angiography. Radiology 185: 381–387, 1992 47. Roberts HC, Kauczor HU, Schweden F, et al: Spiral CT of pulmonary hypertension and chronic thromboembolism. J Thorac Imaging 12:118–127, 1997 48. Sagel SS, Greenspan RH: Nonuniform pulmonary arterial perfusion: Pulmonary embolism. Radiology 99:541–548, 1970 49. Schwickert HC, Schweden FJ, Schild HH, et al: Pulmonary arteries and lung parenchyma in chronic pulmonary embolism: Preoperative and postoperative CT findings. Radiology 191:351–357, 1994 50. Sharma R, Mao J, Garg K: Can acute deep venous thrombosis be differentiated from chronic venous thrombosis on CT venography performed in addition to CT pulmonary angiography [abstract]. Radiology 217(P):646, 2000 51. Sheerick AD, Swensen SJ, Hartman TE: Mosaic pattern of lung attenuation on CT scans: Frequency among patients with pulmonary artery hypertension of different causes. AJR Am J Roentgenol 169:79–82, 1997 52. Silverstein MD, Heit JA, Mohr DN, et al: Trends in the incidence of deep venous thrombosis and pulmonary embolism: A 25-year population-based study. Arch Intern Med 158:585–593, 1998 53. Smith TP: Pulmonary embolism: What’s wrong with this diagnosis? AJR Am J Roentgenol 174:1489–1497, 2000 54. Stein PD, Goldhaber SZ, Henry JW: Alveolar-arterial oxygen gradient in the assessment of acute pulmonary embolism. Chest 107:139–143, 1995 55. Stern EJ, Muller NL, Swensen SJ, et al: CT mosaic pattern of lung attenuation: Etiologies and terminology. J Thorac Imaging 10:294–297, 1995 56. Teigen CL, Maus TP, Sheedy PF, et al: Pulmonary embolism: Diagnosis with contrast-enhanced elec-

122

57.

58. 59.

60.

GARG

tron-beam CT and comparison with pulmonary angiography. Radiology 194:313–319, 1995 The PIOPED investigators: Value of the ventilation/ perfusion scan in acute pulmonary embolism: Results of the Prospective Investigation of Pulmonary Embolism Diagnosis (PIOPED). JAMA 263:2753–2759, 1990 Van Erkel, van Rossum AB, Bloem J, et al: Spiral CT angiography for suspected pulmonary embolism: A cost-effectiveness analysis. Radiology 201:29–36, 1996 Van Rossum AB, Pattynama PM, Ton ET, et al: Spiral CT appearance of resolving clots at 6 week followup after acute pulmonary embolism. J Comput Assist Tomogr 22:413–417, 1998 Vogel P, Laing FC, Jeffrey RB, et al: Deep venous thrombosis of the lower extremity: US evaluation. Radiology 163:747–751, 1987

61. Wharton LR, Pierson JW: Minor forms of pulmonary embolism after abdominal operations. JAMA 79:1904–1910, 1922 62. Wheeler HB, Anderson FA Jr: Diagnostic methods for deep vein thrombosis. Haemostasis 25:6–26, 1995 63. White RH, McGahan JP, Daschbach MM, et al: Diagnosis of deep-vein thrombosis using duplex ultrasound. Ann Intern Med 111:297–304, 1989 64. Worthy SA, Muller NL, Hartman TE, et al: Mosaic attenuation pattern on thin-section CT scans of the lung: Differentiation among infiltrative lung, airway and vascular diseases as a cause. Radiology 205:465– 470, 1997 65. Yankelevitz DF, Gamsu G, Shah A, et al: Optimization of combined CT pulmonary angiography and lower extremity CT venography. AJR Am J Roentgenol 174:67–69, 2000 Address reprint requests to Kavita Garg, MD Veterans Affairs Medical Center 1055 Clermont Street University of Colorado Denver, CO 80220–3808 e-mail: [email protected]

0033–8389/02 $15.00
.00

HIGH-RESOLUTION CT OF THE LUNG II

THE SOLITARY PULMONARY NODULE Johnsey L. Leef III, MD, and Jeffrey S. Klein, MD

The evaluation of a solitary pulmonary nodule (SPN) is a common diagnostic dilemma that has become more prevalent with the increasing use of helical CT.21 The ultimate goal of imaging in the evaluation of SPNs is to accurately distinguish benign from potentially malignant lesions. This has practical importance because the ultimate goal of imaging is to avoid referring a patient with a benign SPN for unnecessary surgical resection, while being certain not to characterize a small malignant SPN incorrectly that may represent resectable (i.e., curable) early stage lung cancer as benign. Although most SPNs prove to be benign, the distinction of benign from malignant lesions can be difficult. One of the most important but understated aspects of SPN evaluation is to be certain that a focal opacity detected radiographically actually represents a solitary intrapulmonary lung nodule. Once this has been determined, demographic features including patient age, smoking history, history of prior malignancy, and environmental exposures are important in guiding evaluation because these factors influence the posttest probability of malignancy independent of imaging characteristics of the lesion in question. An important characteristic of the lesion that can often be discerned from a review of prior radiographs is an assessment of growth rate that helps determine the likelihood of

malignancy. Following this preliminary assessment, assuming the risk of malignancy remains significant, most patients undergo thin-section CT for detailed analysis of size, internal density, and morphologic features of the SPN.28 Adhering to strict criteria for defining benign SPNs leaves most lesions indeterminate. These patients require more detailed nodule evaluation with contrast-enhanced CT or positron emission tomography (PET); some ultimately require biopsy or resection for definitive diagnosis. DEFINITION An SPN is defined as a single round intraparenchymal opacity, at least moderately well-marginated and no greater than 3 cm in maximum diameter. This size limitation is based on the fact that most solitary lung lesions larger than 3 cm in diameter (termed masses) are malignant, whereas most lesions less than 3 cm are benign. In addition, the detection of benign patterns of calcification in a solitary lung lesion allows the confident diagnosis of a benign lesion only when it is less than 3 cm.12 It is important when considering a radiographically detected SPN to be certain that the density in question is truly solitary, lies within the lung, and represents a nodule. As

From the Department of Radiology, Fletcher Allen Health Care, and University of Vermont College of Medicine, Burlington, Vermont

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

123

124

LEEF & KLEIN

many as 50% of patients with suspected SPNs detected radiographically actually prove to have multiple nodules on CT evaluation. 6 This is particularly important because the presence of multiple lung nodules is suggestive of metastatic or granulomatous disease and requires a different approach than the SPN. Even when an SPN is detected as an incidental finding on helical CT or in the course of screening for lung cancer, a detailed analysis of both lungs, particularly with thinsection collimation (i.e., ⬍ 5 mm) and overlapping reconstructions viewed on a workstation, can allow detection of multiple lesions that may alter the diagnostic evaluation.25 The intrapulmonary localization of a nodular opacity seen on only a single radiographic projection can likewise be difficult. For a lesion to be considered intrapulmonary on conventional radiographs it must be a discrete opacity completely circumscribed by aerated lung on orthogonal projections. Pseudonodules caused by EKG pads or other devices on the patient’s skin surface can mimic intrapulmonary lesions on frontal radiographs. Cutaneous lesions including moles, nipple shadows, hemangiomas, neurofibromas, and lipomas that protrude from the patient’s skin and are surrounded by atmospheric air also may appear intrapulmonary on a single radiographic projection (Fig. 1). Careful examination of the skin surface usually readily identifies these lesions; if necessary these can

be confirmed easily by repeat frontal and oblique radiographs or limited chest fluoroscopy following placement of localizing metallic markers. Other lesions that may mimic a SPN include sclerotic bone lesions, such as bone islands; healing rib fractures; and spinal osteophytes. Although these usually are easily recognized after review of prior radiographs, detailed views of the area in question, chest fluoroscopy, or CT may be necessary in selected cases. Likewise, mediastinal and pleural lesions that are pedunculated and project into the lung (i.e., pleural plaques) can appear as SPNs when viewed en face. A tangential view of the region of contact between the lesion and the mediastinum or pleura, however, usually shows an indistinct border that forms obtuse angles. Many apparent focal nodular opacities seen radiographically actually represent vascular structures that are tortuous, seen en face, or are superimposed on other vascular structures that lie in the same sagittal plane. These usually can be identified by review of prior radiographs taken at slightly different obliquities or by performing chest fluoroscopy; CT should be reserved for equivocal cases. Occasionally, a mucocele within a dilated bronchus appears as a SPN. This diagnosis is usually readily apparent by detailed review of thin-section CT images that show a tubular or branching lesion of water attenuation; residual air between the mucocele and bron-

Figure 1. Skin lesion mimicking intrapulmonary nodule. A, Frontal chest radiograph demonstrates a vague nodular opacity overlying the right midlung. B, Lateral radiograph shows that the density represents a cutaneous lesion projecting from the posterior chest wall. Physical examination confirmed the presence of a cutaneous hemangioma.

THE SOLITARY PULMONARY NODULE

chial wall is seen sometimes. Other presumed SPNs are found to represent linear parenchymal scars seen en face. Unless this is easily recognized on current or prior radiographic studies, thin-section CT is often necessary to display the two-dimensional linear nature of the opacity and help distinguish this lesion from a spherical nodule. DIFFERENTIAL DIAGNOSIS The differential diagnosis of a SPN is extensive and includes neoplastic, infectious, inflammatory, vascular, traumatic, and congenital conditions (Table 1). Most benign SPNs are granulomas, hamartomas, or intrapulmonary lymph nodes, whereas bronchogenic carcinomas represent most malignant SPNs. CLINICAL DATA Once a SPN has been definitively identified, a detailed investigation almost invariably ensues. Specific clinical features affect the likelihood of benignancy or malignancy, however, and in conjunction with the imaging characteristics of the lesion can impact both the diagnostic approach and choice of therapeutic options. Bayesian analysis allows for a

more precise determination of the likelihood of malignancy by combining radiographic findings with clinical information (specifically age, smoking history, and symptoms) to calculate mathematically the probability of malignancy of a specific SPN. Bayes’ theorem uses an odds-ratio formula where the odds of malignancy are divided by the odds of malignancy plus one. The equation for determining the overall odds of malignancy is calculated by multiplying the patients’ prior odds of malignancy by the radiographic likelihood ratio of malignancy by the clinical likelihood ratio of malignancy. In this equation, the prior odds of malignancy are the prevalence of malignancy for a given population divided by the prevalence of benign disease of that population.14 The likelihood of malignancy ratios are intuitive measures of diagnostic information provided by radiographic test results or clinical findings.1, 14 Clinical information or radiographic test results strongly suggestive of malignancy have a likelihood ratio much greater than one, whereas those findings suggestive of benignity have a likelihood ratio close to zero, and information or test results that are considered to contain no diagnostic information have a likelihood ratio of one.14 The clinical factors to consider in evaluating the likelihood of malignancy include

Table 1. DIFFERENTIAL DIAGNOSIS OF A SOLITARY PULMONARY NODULE Neoplasm

Benign Malignant

Infection

Granuloma Septic embolus Abscess Round pneumonia Parasitic

Inflammatory Vascular

Airway

Connective tissue Sarcoidosis (rare) Arteriovenous malformation Hematoma Pulmonary infarct Pulmonary artery aneurysm Pulmonary venous varix Congenital lesion Mucocele Infected bulla

125

Hamartoma Inflammatory pseudotumor Bronchogenic carcinoma Carcinoid tumor Lymphoma (Non-Hodgkin’s) Metastasis Mycobacteria Fungi Bacteria (anaerobes, Staphylococcus, gram-negative) Nocardia Pneumococcus Echinococcus Dirofilaria (dog heartworm) Wegener’s granulomatosis Rheumatoid (necrobiotic) nodule

Bronchogenic cyst Bronchial atresia

126

LEEF & KLEIN

patient age; smoking history; symptoms; comorbid conditions (particularly severe emphysema); history and type of prior malignancy; and environmental exposures. For example, a 30-year-old nonsmoking patient without a history of malignancy found to have a smooth round SPN has a high likelihood of a benign lesion and radiologic followup to ensure stability is a reasonable and costeffective approach. Conversely, one could argue that a spiculated SPN in a 50-year-old cigarette smoker should proceed to immediate resection because the likelihood of lung cancer is so high that the results of diagnostic studies are unlikely to impact the patient’s management.33 As in many areas of clinical practice, semiquantitative techniques, such as bayesian analysis of SPNs, have not gained widespread use because of the large number of variables to be estimated and differences in local practice, particularly the availability and expertise with advanced imaging techniques, such as PET. CHARACTERISTICS OF SOLITARY PULMONARY NODULES Size There is no size criterion that reliably distinguishes benign from malignant SPNs. In

general, smaller nodules are more likely to be benign and larger lesions, particularly those exceeding 3 cm in diameter, are more likely to be malignant. Although 80% of benign SPNs are less than 2 cm in diameter, small size is not necessarily reliable evidence of benignity because 15% of malignant nodules are less than 1 cm in diameter (Fig. 2) and approximately 42% are less than 2 cm in diameter. 9 As shown in the prevalence data from the Early Lung Cancer Screening Project, a growing number of smaller malignant SPNs will likely be detected if lung cancer screening with low-dose helical CT gains widespread acceptance. 15 For these reasons, SPN size is helpful information but does not allow definitive distinction of benign from malignant SPNs. Growth The absence of growth over at least a 2-year period is a reliable indicator of benignity.23 To establish firmly the absence of growth, sequential films obtained with comparable radiographic technique are necessary. Limited thin-section CT should be performed for more accurate lesion measurement if there is any question whether a SPN has increased in size. The use of doubling time, which for spherical lesions is defined as a 25% increase in diameter, is based on the observation that benign

Figure 2. Small solitary pulmonary nodule (SPN) represents non–small cell lung carcinoma. A, Prone CT scan through lower lobes shows a 12-mm irregular nodule in the superior segment of the right lower lobe. B, Scan during needle placement for biopsy shows needle approaching edge of lesion with a small pneumothorax. Biopsy yielded adenocarcinoma.

THE SOLITARY PULMONARY NODULE

lesions have doubling times of less than 30 days or greater than 450 days. SPNs with doubling times between 30 and 450 days require further evaluation. The use of helical CT for evaluating SPN growth rates, particularly for small lesions, has recently received greater attention given advances in scanner technology, the ability to obtain nodule volumes using helical acquisition techniques and computer-aided evaluation of data sets, and the increasing number of small lesions detected on CT performed for other indications or to screen for lung cancer presenting as a SPN. A recent paper describes the use of a software program that segments SPNs and calculates nodule volume to within a 3% accuracy. In a small series of

127

patients with SPNs and a definitive diagnosis, the doubling time as determined by change in nodule volume on repeat CT scans was less than 177 days for all malignant nodules and greater than 396 days for all benign lesions (Fig. 3).34 It is likely that computer-aided analysis of nodule volume will become an important part of the evaluation of small SPNs by helping to determine growth rates that allow distinction of benign from malignant SPNs. The technique may prove useful for irregularly shaped small lesions and lesions that grow in a cephalocaudal fashion that is difficult to discern on review of axial images given differences in scan planes and patient positioning on follow-up CT examinations.

Figure 3. CT-derived volumetric analysis of SPN growth. A, Scan shows a small left upper lobe nodule. Boxes at right show detailed sequential axial reconstructions through nodule. B, Repeat thinsection CT scan with sequential reconstructions shows questionable interval change in nodule shape. Illustration continued on following page

128

LEEF & KLEIN

Figure 3 (Continued). Three-dimensional shaded surface displays of nodule obtained from volumetric acquisition in various projections from June (C) and October (D) with nodule volume calculated shows an interval increase in nodule volume. Biopsy revealed adenocarcinoma. (From Henschke CI, Yankelevitz DF: CT screening for lung cancer. Radiol Clin North Am 383:487–495, 2000; with permission.)

Density and Internal Characteristics Calcification The presence of a specific pattern of macroscopic calcification within a SPN as seen on conventional radiographs or CT is indicative of a benign lesion. These patterns include central, laminated, diffuse, or popcorn calcification (Fig. 4).30 Central, solid (Fig. 5), and laminated forms of calcification (Fig. 6) are found in association with prior granulomatous infection, most commonly histoplasmosis or tuberculosis. Popcorn calcification usually represents the chondroid calcification in a pulmonary hamartoma. Eccentric or amorphous calcification can represent a calcified granuloma engulfed by a malignancy or dys-

trophic malignant calcification, respectively, and should not be taken as evidence of benignancy (Fig. 7). 20 Although the presence and pattern of calcification can sometimes be determined on conventional radiographs, approximately one third of noncalcified SPNs have calcification on CT.35 For this reason, definitive identification of calcium usually requires thin-section CT using 1- to 3-mm collimated scans reconstructed with a high spatial frequency algorithm. In lesions where calcification is not visible on thin-section CT scans, quantitative CT densitometry can be performed to determine the attenuation value of the nodule. Quantitative CT densitometry can be performed by comparing the CT density of the SPN with that of a microscopically calcified prosthetic reference nodule placed

THE SOLITARY PULMONARY NODULE

129

Figure 4. Benign patterns of calcification within SPNs. (From Klein JS, Wand A: Pulmonary neoplasms. In Brant WE, Helms CA (eds): Fundamentals of Diagnostic Radiology, ed 2. Baltimore, Williams and Wilkins, 1999; with permission.)

within a chest phantom model.22 Most thoracic radiologists no longer use a reference CT phantom for this purpose and instead calculate the average attenuation value through the central portion of the nodule. A value of 200 H or greater from at least 10% of the central pixels or throughout the nodule is reliable evidence of microscopic calcification and indicates benignity.17 The absence of calcium is of little value in distinguishing benign from malignant SPNs because 38% to 63% of benign nodules, two thirds (67%) of carcinoid tumors, and as many as 94% of all lung cancers do not contain appreciable calcium.9

most common cause of a SPN, is a benign neoplasm composed of disorganized epithelial and mesenchymal elements normally found in the bronchus or lung. Up to 50% of hamartomas have fat that can be detected on thin-section CT (Fig. 8), with 30% showing calcification or ossification that is often popcorn in appearance.27 These lesions typically develop in middle-aged adults and demonstrate slow growth. In patients with characteristic findings on thin-section CT, pathologic confirmation is usually unnecessary and radiographic follow-up to confirm stability over a 2-year period is recommended, although transthoracic needle biopsy can be performed if a definitive diagnosis is deemed necessary.

Fat The identification of fat within a SPN with smooth or lobulated margins is indicative of benignity. A pulmonary hamartoma, the third

Cavitation Although cavitation can occur in necrotic malignant SPNs, particularly squamous cell

130

LEEF & KLEIN

Figure 5. Calcified SPN from tuberculosis. A, Frontal chest radiograph shows a peripheral right upper lobe nodule. B, Thin-section CT scan shows a completely calcified nodule representing prior tuberculosis.

THE SOLITARY PULMONARY NODULE

Figure 6. Laminated calcification in histoplasmoma. A, Scout view from CT scan shows a sharply marginated peripheral right upper lobe nodule. B, Thin-section CT scan shows laminated calcification. The patient had a history of prior histoplasma infection.

131

132

LEEF & KLEIN

Figure 7. Dystrophic calcification in lung cancer. A, Frontal chest radiograph shows a large left lower lobe mass. B, Unenhanced thin-section CT scan shows eccentric punctate calcifications. Biopsy revealed adenocarcinoma.

THE SOLITARY PULMONARY NODULE

133

Figure 8. Fat within pulmonary hamartoma. A, Detailed view of a thin-section CT scan through the right middle lobe shows a smooth, sharply defined 15-mm oval nodule. B, Scan at mediastinal windows shows scattered fat attenuation within the nodule characteristic of a hamartoma.

carcinoma, inflammatory lesions, such as abscesses, infectious granulomatous lesions, Wegener’s granulomatosis, and pulmonary infarcts, can also cavitate. The thickness of the cavity wall can be helpful in distinguishing benign from malignant lesions. Cavities with a greatest wall thickness less than 5 mm are almost always benign, whereas most of those with a maximal wall thickness greater than 15 mm are malignant (Fig. 9).31, 32

Air Bronchograms or Bubbly (Cystic) Lucencies The assessment of the internal density of a noncalcified SPN as evaluated with thin-section CT can provide useful information. Homogeneous nodule attenuation is observed more frequently in benign (55%) than malignant lesions (20%).36 Internal inhomogeneity, particularly the presence of air bronchograms

Figure 9. Cavitatary lung cancer. CT scan through the upper lobes shows a right upper lobe mass with central cavitation. The thickest portion of the cavity wall measures 22 mm. Transthoracic biopsy showed squamous cell carcinoma.

134

LEEF & KLEIN

or cystic or ‘‘bubbly’’ lucencies within a SPN, is highly suggestive of an adenocarcinoma, particularly the localized form of bronchioloalveolar cell carcinoma (Fig. 10). Other malignancies, however, such as lymphoma and benign lesions including organizing pneumonia, pulmonary infarcts, and mass-like sarcoidosis, can produce a similar appearance. Margins A detailed assessment of the margins of a SPN as depicted on thin-section CT can provide useful information. Although smooth, well-defined margins most often indicate a benign nodule (Figs. 8 and 11), 21% of malignant nodules have a smooth well-defined margin.26, 35 Alternatively, a lobulated margin may reflect uneven growth of a SPN and can indicate malignancy (Fig. 12), although 25% of benign nodules, particularly hamartomas, are lobulated.36 A nodule that has an ill-defined margin and demonstrates an irregular or spiculated contour most often represents a malignancy. Spiculated margins result from cicatrization (scarring) of the lung interstitium surrounding a lesion (Fig. 13). It is important to recognize, however, that whereas a spiculated margin is highly suspicious for lung cancer, cicatrization producing a spiculated nodule can also be seen in benign inflammatory processes, such as lipoid pneu-

monia, organizing pneumonia, tuberculoma, and the mass-like lesions of progressive massive fibrosis seen in complicated silicosis. There are several other features of the margins of SPNs as depicted on thin-section CT that are of diagnostic value. The presence of small satellite nodules surrounding the periphery of a smooth dominant nodule is strongly suggestive of a granulomatous infection (Fig. 14).4 The pleural tail is a linear opacity seen extending from the edge of a peripherally situated SPN to indent the pleura. Although this finding can be associated with lung cancer (see Fig. 13), in particular bronchioloalveolar cell carcinoma, it is also seen with peripheral granulomas and its presence is of limited diagnostic value.35 The halo sign refers to the presence of groundglass opacity surrounding a nodule or mass. When seen in a neutropenic patient, this finding is highly suggestive of an angioinvasive opportunistic infection, particularly aspergillosis. 3 The detection of feeding and draining vessels entering the hilar aspect of a round or lobulated SPN is diagnostic of a pulmonary arteriovenous malformation, which is seen more commonly in patients with hereditary hemorrhagic telangiectasia (Rendu-Osler-Weber syndrome). CT NODULE ENHANCEMENT The technique of contrast enhancement of SPNs on CT has developed as an extension

Figure 10. Cystic lucencies in bronchioloalveolar cell carcinoma. Thin-section CT scan through left upper lobe shows an irregularly marginated nodule containing small cystic lucencies. Biopsy revealed bronchioloalveolar cell carcinoma.

THE SOLITARY PULMONARY NODULE

Figure 11. Smoothly marginated benign SPN. A, Coned-down view of right lung shows a nodule at the right base. B, Thin-section CT scan shows a smoothly marginated middle lobe nodule. Biopsy revealed a granuloma caused by Cocciodioides immitis infection.

Figure 12. Lobulated margin in malignant SPN. Coned-down view of right upper lobe shows a 20-mm nodule with a lobulated margin. Biopsy revealed adenocarcinoma.

135

136

LEEF & KLEIN

Figure 13. Spiculated margin in malignant SPN. CT scan at the level of the aortic arch shows a spiculated right upper lobe nodule with a pleural tail extending posteriorly to distort the major fissure. Biopsy showed adenocarcinoma.

of the observation that malignant lung tumors are relatively hypervascular compared with benign lesions. CT nodule enhancement after intravenous contrast administration is now accomplished easily with the use of power injectors and rapid-scan acquisition with helical CT. The technique involves thincollimated (3 mm) spiral acquisitions through a SPN between 6 and 30 mm in diameter before and after intravenous contrast injection. Scans obtained each minute for 4 minutes after contrast injection are compared with baseline unenhanced scans. An enhance-

ment value is then determined by calculating the mean attenuation value through the center of the nodule at peak contrast enhancement and subtracting the baseline value. A recent prospective multicenter study has shown that an enhancement value of less than 15 H is virtually diagnostic of a benign lesion (i.e., the test has a high sensitivity for malignancy) (Fig. 15). 29 The rare false-negative study is usually caused by central necrosis or a mucin-producing malignant neoplasm, such as a bronchioloalveolar cell carcinoma (Fig. 16). An enhancement value exceeding 15 H is

Figure 14. Satellite nodules in granulomatous SPN. Prone CT scan before transthoracic biopsy of a radiographically detected SPN shows a cluster of small nodules in the posterior segment of the right upper lobe with a larger central nodule (arrow). Biopsy revealed granulomatous inflammation with a specific organism identified.

THE SOLITARY PULMONARY NODULE

Figure 15. Benign nodule by contrast-enhanced SPN on CT scan. Frontal (A) and lateral (B) chest radiographs in a 54-year-old man show a subtle right lower lobe nodule seen best on the lateral film. C, Sequential contrast-enhanced scans obtained at baseline (upper left) and every minute after contrast injection (upper left, lower left, lower right, respectively) show enhancement of 5 H from baseline. Biopsy revealed a granuloma.

137

138

LEEF & KLEIN

Figure 16. False-negative contrast-enhanced CT nodule study. Sequential contrast-enhanced scans obtained at baseline (upper left) and every minute after contrast injection (upper right, lower left, lower right, respectively) shows virtually no enhancement of right upper lobe nodule. Biopsy and surgical resection revealed mucinous subtype of bronchioloalveolar cell carcinoma.

nonspecific because both inflammatory lesions and malignancies can enhance (Fig. 17); this limits the positive predictive value of the test to 68%. Because of the rapid image acquisition with spiral CT, CT nodule enhancement can be performed following routine scanning of the chest without need for additional contrast administration and little additional scan time and radiation dose. The technique requires meticulous attention to detail, however, and may be less accurate in larger SPNs (i.e., those ⬎ 2 cm) because these lesions are more often necrotic and may produce false-negative examinations. The technique may prove to be most useful for the evaluation of probably benign SPNs when transthoracic needle biopsy is unavailable, cannot be performed, or is nondiagnostic and the patient is a poor surgical candidate. POSITRON EMISSION TOMOGRAPHY There is a growing experience in the use of PET using the radiopharmaceutical fluoro-2deoxy-D-glucose (FDG) in the evaluation of focal lung lesions including SPNs. FDG up-

take in focal lesions is measured semiquantitatively by calculating a standardized uptake ratio. When assessing lesions greater than 10 mm in diameter, FDG-PET has a sensitivity of 94% to 96% and a specificity of 87% to 88% for the evaluation of SPNs (Fig. 18).7, 9, 10 The low false-negative rate of PET makes this a useful adjunct to thin-section CT in excluding malignancy and allows clinical follow-up of probably benign lesions. As with CT nodule enhancement, false-positive cases are usually seen in granulomas with active inflammation. The limited availability and high expense of FDG-PET are the main obstacles to widespread use in the evaluation of SPNs. Given the increasing availability of mobile PET scanners and the reimbursement for solitary nodule evaluation and lung cancer staging currently supported by Medicare, it is likely that the use of thoracic and whole-body PET will expand well beyond academic research centers and become a mainstay of solitary nodule evaluation and lung cancer staging. TECHNETIUM 99m DEPREOTIDE SCINTIGRAPHY The affinity of malignant neoplasms including both small cell and non–small cell lung

THE SOLITARY PULMONARY NODULE

Figure 17. Positive contrast-enhanced CT nodule study. A, Coned-down view of CT scan through right upper lobe shows an SPN with a lobulated and spiculated margin. B, Sequential contrastenhanced scans obtained at baseline (upper left) and every minute after contrast injection (upper left, lower left, lower right, respectively) shows enhancement of 41 H from baseline. Biopsy revealed adenocarcinoma.

139

140

LEEF & KLEIN

Figure 18. Positron emission tomography (PET) of SPN. A, Frontal chest radiograph shows a round right lower lung mass. B, Unenhanced CT scan shows a middle lobe mass with an irregular margin peripherally. C, Axial PET image shows intense uptake. Biopsy proved bronchogenic carcinoma. (Courtesy of Edward F. Patz, MD, Department of Radiology, Duke University Medical Center, Durham, NC.)

cancer for peptide analogues of somatostatin has led to at least two studies investigating the use of this agent in the noninvasive evaluation of SPNs. In a recent multicenter study using this agent commercially available as Neotect (Diatide, Londonderry, NH), 96.6% of malignant SPNs and masses were correctly identified (Fig. 19), although the specificity of the agent for distinguishing benign from malignant lesions was only 73%.2 Because the availability of single photon emission CT is currently greater than that of FDG-PET, this agent may have some clinical use in selected patients, most typically nonsurgical candidates with irregularly shaped lesions that are not amenable to CT nodule enhancement or percutaneous biopsy.

PATHOLOGIC DIAGNOSIS OF SOLITARY PULMONARY NODULES Transthoracic Needle Biopsy Image-guided transthoracic needle biopsy has become the semi-invasive procedure of choice for definitive characterization of peripheral SPNs. The procedure is most often performed under CT guidance and has been shown to have a sensitivity of over 90% for malignancy in most series, particularly when expert cytopathology is used (see Fig. 2).16, 18 The ability of transthoracic needle biopsy to obtain a specific benign diagnosis for focal lung lesions is limited by difficulty in aspirating diagnostic material from sclerotic granu-

THE SOLITARY PULMONARY NODULE

141

Figure 19. Positive technetium medronate (Tc99m) depreotide scintigram of lung. A, CT scan through upper lobes shows irregular mass peripherally in the right upper lobe. B, Axial single photon emission computed tomography (SPECT) scan at the same level as A shows increased uptake in the lesion. Biopsy proved non–small cell carcinoma.

lomas. Recently, the use of small-gauge (⬍ 20gauge) cutting biopsy needles has provided histologic material from SPNs and can significantly improve the yield for benign lesions to 80% or greater.11 Bronchoscopy In a patient with a SPN and findings suggesting central airway involvement (i.e., hemoptysis or bronchus entering hilar aspect of nodule on thin-section CT), bronchoscopy with brushings, washings, and endobronchial or transbronchial biopsy or transbronchial needle aspiration is the initial diagnostic procedure of choice, and in such situations can obtain a diagnosis in up to 80% of lesions.5 The diagnostic yield from bronchoscopy dif-

fers significantly with the size and location of the lesion and the experience of the bronchoscopist, however, with yields as low as 28% in one series.24 A cooperative bronchoscopic and transthoracic approach has been shown to be effective for peripheral nodules,13 and conventional or CT fluoroscopy helps guide accurate placement of transbronchoscopic brushes and needles to improve diagnostic yield. Video-Assisted Thoracoscopic Surgery This technique, usually performed by surgeons in the operating room under general anesthesia with single lung ventilation, can be used as both a diagnostic and therapeutic

142

LEEF & KLEIN

procedure for SPNs. The indications for video-assisted thoracoscopic surgery resection of SPNs differ considerably between institutions, but it has been shown to have high diagnostic accuracy with less morbidity than thoracotomy.19 Thoracotomy Thoracotomy for resection of SPNs is usually limited to lesions with a high likelihood of malignancy when lobectomy and nodal resection are necessary for definitive lung cancer staging and treatment.8 Additional indications include a deeply situated nodule or other lesion not amenable to thoracoscopic localization and resection or when limited pulmonary reserve necessitates limited lung resection to preserve pulmonary function. References 1. Black WC, Armstrong P: Communicating the significance of radiologic test results: The likelihood ratio. AJR Am J Roentgenol 147:1313, 1986 2. Blum J, Handmaker H, Lister-James J, et al: A multicenter trial with a somatostatin analog 99mTc Depreotide in the evaluation of solitary pulmonary nodules. Chest 117:1232, 2000 3. Blum U, Windfuhr M, Buitrago-Tellez C, et al: Invasive pulmonary aspergillosis: MRI, CT, and plain radiographic findings and their contribution for early diagnosis. Chest 106:1156, 1994 4. Carucci LR, Maki DD, Miller WT: Clustered pulmonary nodules: Highly suggestive of benign disease. J Thorac Imaging 16:103–105, 2001 5. Chechani V: Bronchoscopic diagnosis of solitary pulmonary nodules and lung masses in the absence of endobronchial abnormality. Chest 109:620, 1996 6. Costello P, Anderson W, Blume D: Pulmonary nodule: Evaluation with spiral volumetric CT. Radiology 179:875, 1991 7. Dewan NA, Gupta NC, Redepenning LS, et al: Diagnostic efficacy of PET-FDG imaging in solitary pulmonary nodules. Chest 104:997, 1993 8. Edwards WM, Cox RS Jr, Garland LH: The solitary nodule (coin lesion) of the lung: An analysis of 52 consecutive cases treated by thoracotomy and a study of preoperative diagnostic accuracy. AJR Am J Roentgenol 88:1020, 1962 9. Erasmus JJ, Connolly JE, McAdams HP, et al: Solitary pulmonary nodule. Part I. Morphologic evaluation. Radiographics 20:43, 2000 10. Erasmus JJ, Patz EF Jr: Positron emission tomography imaging in the thorax. Radiol Clin North Am 20:715, 1999 11. Fraser RS: Transthoracic needle aspiration: The benign diagnosis. Arch Pathol Lab Med 115:751, 1991 12. Fraser RS, Mu¨ller NL, Coleman N, et al: Pulmonary carcinoma. In Fraser and Pare’s: Diagnosis of Diseases of the Chest, ed 4. Philadelphia, WB Saunders, 1999, p 1069

13. Gasparini S, Ferretti M, Bichi Secchhi E, et al: Integration of transbronchial and percutaneous approach in the diagnosis of peripheral pulmonary nodules or masses: Experience with 1027 consecutive cases. Chest 108:131, 1995 14. Gurney JW, Lyddon DM, McKay JA: Determining the likelihood of malignancy in solitary pulmonary nodules with Bayesian analysis. Part I and II. Appl Radiol 186:405, 1993 15. Henschke CI, McCauley DI, Yankelevitz DF, et al: Early lung cancer action project: Overall design and findings from baseline screening. Lancet 354:99, 1999 16. Khouri NF, Stitik FP, Erozan YS, et al: Transthoracic needle aspiration biopsy of benign and malignant lung lesions. AJR Am J Roentgenol 144:281, 1985 17. Klein JS, Wand A: Pulmonary neoplasms. In Brant WE, Helms CA (eds): Fundamentals of Diagnostic Radiology, ed 2. Baltimore, Williams & Wilkins, 1999, p 377 18. Lee SI, Shephard JL, Boiselle PM, et al: Role of transthoracic needle biopsy in patient treatment decisions. Radiology 201(P):269, 1996 19. Mack MJ, Hazebrigg SR, Landreneau RJ, et al: Thoracoscopy for the diagnosis of the indeterminate solitary pulmonary nodule. Ann Thorac Surg 56:825, 1993 20. Mahoney MC, Shipely RT, Corcoran HL, et al: CT demonstration of the calcification in carcinoma of the lung. AJR Am J Roentgenol 154:255, 1990 21. Midthun DE, Swenson SJ, Jett JR: Approach to the solitary pulmonary nodule. Mayo Clin Proc 68:378, 1993 22. Naidich DP, Zerhouni EA, Siegelman SS: Focal lung disease. In Computed Tomography and Magnetic Resonance of the Thorax, ed 2. New York, Raven Press, 1991, p 303 23. Nathan MH, Collins VP, Adams RA: Differentiation of benign and malignant pulmonary nodules by growth rate. Radiology 79:221, 1962 24. Reichenberger F, Weber J, Tamm M, et al: The value of transbronchial needle aspiration in the diagnosis of peripheral pulmonary lesions. Chest 116:704, 1999 25. Seltzer SE, Judy PF, Adams DF, et al: Spiral CT of the chest: Comparison of cine and film-based viewing. Radiology 197:73, 1995 26. Siegelman SS, Khouri NF, Leo FP, et al: Solitary pulmonary nodules: CT assessment. Radiology 160:307, 1986 27. Siegelman SS, Khouri NF, Scott WW Jr, et al: Pulmonary hamartoma: CT findings. Radiology 160:313, 1986 28. Siegelman SS, Zerhouni EA, Khouri NF, et al: CT of the solitary pulmonary nodule. AJR Am J Roentgenol 135:1, 1980 29. Swensen SJ, Viggiano RW, Midthun DE, et al: Lung nodule enhancement at CT: Multicenter study. Radiology 214:73, 2000 30. Webb WR: Radiologic evaluation of the solitary pulmonary nodule. AJR Am J Roentgenol 154:701, 1990 31. Woodring JH, Fried AM: Significance of wall thickness in solitary cavities of the lung: A follow-up study. AJR Am J Roentgenology 140:473, 1983 32. Woodring JH, Fried AM, Chuang VP: Solitary cavities of the lung: Diagnostic implications of cavity wall thickness. AJR Am J Roentgenol 135:1269, 1980 33. Yankelevitz DF, Henschke CI, Altorki NK: Cost analysis of competing strategies for evaluating and treating solitary pulmonary nodules. Radiology 201(P): 269, 1996

THE SOLITARY PULMONARY NODULE 34. Yankelevitz DF, Reeves AP, Kostis WJ, et al: Small pulmonary nodules: Volumetrically determined growth rates based on CT evaluation. Radiology 217:251, 2000 35. Zerhouni EA, Stitik FP, Siegelman SS, et al: CT of the

143

pulmonary nodule a cooperative study. Radiology 160:319, 1986 36. Zwirewich CV, Vedal S, Miller RR, et al: Solitary pulmonary nodule: High resolution CT and radiologic-pathologic correlation. Radiology 179:469, 1991 Address reprint requests to Jeffrey S. Klein, MD Department of Radiology Fletcher Allen Health Care 111 Colchester Avenue Burlington, VT 05401

HIGH-RESOLUTION CT OF THE LUNG II

0033–8389/02 $15.00
.00

QUANTITATIVE CT OF THE LUNG Jonathan G. Goldin, MD, PhD

Lung disease is a leading cause of mortality in the United States.2, 69 Chronic lung disease encompasses a spectrum of diffuse lung diseases including emphysema, chronic bronchitis, asthma, respiratory bronchitis, bronchiectasis, pulmonary thromboembolic disease, and interstitial lung disease. CT, currently the best test to assess lung involvement in emphysema and interstitial lung disease, relies on abnormalities being detected when there is sufficient morphologic distortion to result in visually identified changes that occur relatively late in the pathogenesis of most disease processes. There is, however, poor correlation with the conventional measure of lung function and outcome. As a researcher highly respected in aspects of chronic obstructive lung disease recently reflected, ‘‘the simplicity of the lesion-holes in the lung, some (too-small) small airways, and bronchial gland enlargement, belies the functional and pathogenic complexity.’’93 The same limitation likely applies to the reliance of visually evident parenchymal distortions in the CT images of patients with interstitial lung disease. Radiologists, however, have been slow to go beyond the description of structural changes and provide quantitative measures of function. Further, what is needed is a method to distinguish the static and dynamic airway and parenchymal alterations exhibited This work was supported by Grant No. P01 CA5119806 from the National Cancer Institute.

in lung disease and to assess regional variations. High-resolution CT (HRCT) is an established technique for the detailed evaluation of the pulmonary parenchyma and can characterize anatomic details of the lung as small as 200 to 300 mm, which corresponds to approximately the seventh to ninth generations of the airways and lung segments.48, 57, 66, 100 The addition of volume scanning allows data to be acquired through broad regions of interest during different phases of respiration or under different physiologic conditions providing critical insights into the relationships between structure and function. 14, 15, 37 For quantitative image analysis to be useful and applied clinically it needs to be easy to perform, reproducible, observer independent, and offer a valid measurement of disease presence and extent by comparison with physiologic or pathologic criteria. Further imaging protocols need to be tailored to ensure the validity and reproducibility of the quantitative measures. This article gives an overview of quantitative techniques used and their application to various pulmonary diseases. IMAGING PROTOCOLS The best imaging protocol for the quantitation of diffuse lung disease remains unresolved. With rapidly changing CT technology this remains an area of constant evolution. Currently, there is the need to choose, at the

From the Department of Radiological Sciences, University of California at Los Angeles Medical Center, Los Angeles, California

RADIOLOGIC CLINICS OF NORTH AMERICA VOLUME 40 • NUMBER 1 • JANUARY 2002

145

146

GOLDIN

time of image acquisition, between sampling the lungs with high-resolution thin sections at selected intervals and the use of a volume acquisition technique to acquire a three-dimensional CT data set. The advantage of a rapidly acquired volumetric data set is that the images can be acquired in a single standardized breath-hold and the lung is completely sampled allowing the creation of three-dimensional models. Before the advent of subsecond multidetector and spiral CT scanners, thick sections (5 to 10 mm) were necessary to ensure a contiguous acquisition of the entire lung in the same breath-hold. As a consequence, subtle abnormalities in the lung texture and density caused by emphysema, lung fibrosis, or air trapping were difficult to detect in the relatively large voxel sizes. For this reason, thin-section sampling at intervals through the lung was preferentially chosen to allow detection of greater lung detail. With the advent of multidetector scanners and subsecond scanning, however, it is now possible to obtain full-volume datasets in a single breath-hold that can be reconstructed with true thin-section resolution of the entire lung for the purpose of quantitative image analysis (QIA). Standardization of the lung volume to either total lung capacity (TLC), residual volume (RV), or a known lung volume in between the two at the time of scan acquisition is critical to ensure both the validity of quantitative measures and reproducibility of measures on repeated studies. This is most accurately achieved with spirometric gating, in which the CT scan is triggered, and airflow mechanically inhibited, at a predetermined user selected level of breath-hold.40 A similar technique using a pneumotachometer allows the patient to breathe between acquisitions and then on return to a preselected inspiratory level scanning is initiated.39 It is also possible to obtain acceptably good reproducibility of lung volumes without the use of electronically gated CT scanners by the use of an incentive spirometer used during the time of scanning. The availability of commercial apparatus or spirometric gating or triggering is scarce and most quantitative studies are performed without such methods. It is possible with volume acquisition techniques to confirm lung volume reproducibility without the need for spirometric gating,5 but this is only useful when the entire lung volume is acquired in a single breath-hold and not

when several high-resolution images are obtained at intervals through the lungs. In addition to volume triggered scanning, the ability to acquire or reconstruct images at 50- to 100-millisecond intervals, achieving real-time cine´ sequences during the performance of respiratory maneuvers without compromise of image quality by motion, can also be useful.24, 29 Although the necessary technical parameters vary according to the specific application, degree of resolution required, and anatomic region under examination, images of 1- to 3-mm collimation obtained between 300- and 500-millisecond intervals can be acquired as sequences that are triggered or simultaneously monitored by spirometry. ECG-gating may also be used to study the lower lobes free of cardiac motion artifacts.72 Similar results have been reported using fast CT (subsecond helical and multidetector scanners). For most pulmonary functional applications, sequences acquired through the upper, middle, and lower lungs provide an adequate sample of the lungs. By integrating the spirometric and imaging data, changes in lung attenuation for isolated regions of interest can be measured as a function of time, airflow, and lung volume (Fig. 1).

QUANTITATIVE IMAGE ANALYSIS The quantitative information obtained with HRCT may advance our understanding of pulmonary pathophysiology and offer insights into the potential mechanisms involved in the progression of lung disease.* Additional advantages of this technique are its noninvasive nature and ability to offer a comprehensive regional assessment not possible with conventional lung function tests. There has been an explosion in investigations to provide quantitative structural and functional information from digitally acquired image data. These methods include visual quantitation scoring systems; image display (e.g., multiplanar reformations and surface shading for three-dimensional and volume rendering) 53, 71, 82–84 ; anatomic image quantitation (e.g., area and volume of airways and lungs)†; and regional characterization of lung tissue (analyzing attenuation, changes in at*References 1, 27–30, 41, 50, 56, 64, 66–68, 71, 75, 77, 78, 80, 87, 91, 96–98, 101, 104, 107. †References 1, 3, 4, 10, 13, 21, 29, 39, 68, 67, 105, and 106.

QUANTITATIVE CT OF THE LUNG

147

Figure 1. Quantitative image analysis (QIA) is achieved by obtaining CT data at controlled, spirometrically gated lung volumes and transferring the data to automated analysis and display workstations. WS  workstation.

tenuation, and texture patterns in the imaged lung).8, 20, 28, 95, 97, 98 Approaches to CT quantitation may be visual or computer aided. Visual quantitation approaches use various scoring systems that categorize different parenchymal patterns and attempt to quantitate their extent and severity using an assessment of the amount of lung involved. In designing a visual scoring system it is important to score the extent of different CT patterns separately because these different patterns most likely have different functional effects. Such visual scoring CT systems have been developed for emphysema, idiopathic pulmonary fibrosis (IPF), scleroderma lung disease, and chronic pulmonary embolism. In most studies, this final score is calculated either by agreement between two reviewers at a joint reviewing session or by obtaining the mean of the reading scores for the two reviewers. For the most part, they are relatively subjective and lack reproducibility with larger inter-

reader and intrareader variation. Attention has turned to computer-aided detection and quantitative techniques. For computer-aided techniques the anatomic components of the cardiopulmonary system need to be identified and segmented before quantitation. Several approaches and software packages have been developed to assess the lung quantitatively with varying degrees of complexity and overlap in functionality (Fig. 2). These packages consist of varying sets of image segmentation and analysis tools written to answer specific thoracic clinical and research questions.12, 20, 58–60, 82, 97 Image segmentation is achieved using various combinations of threshold- and knowledge-based algorithms to identify the lung and contained airways and blood vessels. For the most part the segmentation process is fully automated but allows manual correction if needed. Three basic types of measurements are

148

GOLDIN

Figure 2. Automated knowledge-based segmentation of lung parenchyma of the entire right lung (A), the central portion (B), and the peripheral region (C).

made on a segmented region of interest: (1) measures relating to the size and shape of the region of interest, (2) attenuation (gray level) statistics, and (3) image texture within the region of interest. Much of the lung consists of intrathoracic gas that, when referenced to specific phases of respiration, can be used to assess parenchyma and airway integrity. An increase in either blood volume or the interstitial, cellular, or fluid content of the lung parenchyma results in increased attenuation of pixels relative to the low attenuation of air. Measures that reflect changes in attenuation are also indirect measures of these parenchymal changes.27, 56, 80 The greater the leftward shift of the lung attenuation curve (e.g., reflecting a greater proportion of pixels with low attenuation), the more extensive the lung destruction caused by emphysema or expiratory airflow obstruction. Conversely, a rightward shift of the curve (to a greater proportion of high attenuation) reflects an increase in either blood flow or the interstitial, cellular, or fluid content of the lung parenchyma. By integrating spirometric and imaging data, changes in lung attenuation in defined regions of interest can be measured as a function of time (attenuation-time curves), airflow, or lung volume. In addition, texture measurements quantify the nature of the local attenuation differences and take into account both the distance and direction of these differences.33, 34, 53, 89, 95, 96, 98 Such measurements as energy and entropy examine the overall homogeneity or heterogeneity of lung attenuation. Such measurements as autocorrelation and covariance may be derived from the second-order histogram of lung density. Fractal analysis may also be applied to the analysis

of the CT patents.81 Total and regional lung volumes can also be calculated accurately (Fig. 3).13, 17 Using computer-aided detection techniques a computer system can be used to establish rules for recognition of image texture features, such as normal parenchyma, bronchus, emphysema, ground-glass attenuation, and fibrosis (Fig. 4). 20 The system can then be tested on a set of test images or regions of interest. In a study by Delorme et al20 this approach resulted in actual recognition of 70% to 80% of patients in 1889 five-pixel by five-pixel test regions. A similar approach, called the adaptive model feature method, has been used by Uppaluri et al95 to classify the images of patients with emphysema and lung fibrosis. These techniques have been applied to distinguish: (1) normal and severely emphysematous patients on either a global basis (entire lung section) or regional basis (portion of lung section) 98 ; (2) healthy nonsmokers, smokers, and smokers with abnormal lung function (chronic obstructive pulmonary disease) on either a global or regional basis97; (3) airway reactivity and treatment response29, 31; (4) other patient populations, such as interstitial lung disease (i.e., IPF and sarcoidosis, asbestosis, and cystic fibrosis)36, 94; and (5) areas of normal parenchyma, vessels, bronchi, emphysema, ground-glass, and lesions with intralobular fibrosis in HRCT images of patients with usual interstitial pneumonia (UIP).20 QIA can also be used as a measure of the efficacy of novel therapeutic techniques because of their ability to detect subtle regional variations not possible with conventional lung function tests.31

QUANTITATIVE CT OF THE LUNG

149

Figure 3. CT measurements of total lung volume (TLV), acquired volumetrically at full inspiration to TLC, correlate extremely well (r  0.98) with TLC measured by body plethysmography.

DISEASE-SPECIFIC APPLICATIONS OF QUANTITATIVE IMAGE ANALYSIS Emphysema Visual Quantitation Chest CT provides an accurate assessment of the distribution and severity of lung destruction and a measure of impaired respiratory mechanics in patients with emphysema.6, 32, 49, 52, 65, 93 Bergin et al6 reported a CT–pathologic correlation study on 32 patients who subsequently underwent surgical resection for lung tumors. In this group of patients, the CT scans were scored using a simple visual scoring system that assessed decreases in lung attenuation and vascular attenuation as measurements of the presence of emphysema. Each lung on each image was scored from 0 to 4 according to the following grading: 0  normal, 1  less than 25% emphysema, 2  25% to 50% emphysema, 3  50% to 75% emphysema, and 4  more than 75% emphysema. The scores were then combined from each lung and by multiplying the number of images obtained of the lungs a final score was calculated. This final score was expressed as a percentage of the maximum possible CT score to account for the different number of images obtained in different patients. A good correlation between the CT emphysema scores and the pathology emphysema scores on the resected lobes was obtained. There was also significant correla-

tion between the diffusing capacity (DLco); spirometric measures of airflow (forced expiratory volume in 1 second [FEV1], FEV1/forced vital capacity [FVC] percent); and the pathologic emphysema scores. This study also showed CT to be a better method of predicting emphysema than pulmonary function tests. Sanders et al86 reported the result of studying 60 male patients with the presence of emphysema with chest radiography, CT, and pulmonary function testing. A visual scoring 1000 800 600 400 200 0 –200 –0.30

–0.20

–0.10

0

First Measure of Correlation Figure 4. Lung texture measures allow for the detection of different lung structural changes as a consequence of parenchymal destruction (in emphysema), air trapping (in asthmatics), and alveolitis and fibrosis (in patients with scleroderma), as compared with normal lung structure (in well-characterized normal patients). Gradations of severity also can be determined by the relationship of the texture measures’ position relative to normal lung texture. Diamond  asthma; square  emphysema; triangle  scleroderma; circle  normal.

150

GOLDIN

system was used similar to that used by Bergin et al.6 They concluded that CT had excellent correlation of pulmonary function testing evidence of emphysema (DLco/alveolar volume [VA], FEV1 percent and FVC percent). CT was shown to be more sensitive than radiography in detecting emphysema. They also identified CT evidence of emphysema in 24 out of 35 patients with no pulmonary function testing evidence for emphysema, suggesting again that CT is a more sensitive measure of mild emphysema than conventional lung function tests. Two studies have compared visual quantitation with QIA on the same image datasets. Eda et al22 used a volume acquisition technique acquiring 3-mm collimation datasets at four levels in the chest. Visual scoring was obtained on inspiratory helical CT images and the mean inspiratory and mean expiratory number was calculated from the same datasets. The mean CT numbers were used to calculate an expiratory to inspiratory (E/I) ratio. Both the visual score and the E/I ratio show good correlations for FEV1, FEV1/FVC, and vital capacity [VC]. The visual score showed better correlation what the DLco percent predicted and the E/I ratio showed better correlation of RV (percent predicted). From this study the investigators concluded that HRCT was useful in assessing emphysema and that the visual scores obtained on the inspiratory helical scans were a better assessment of emphysema extent than the E/I ratio because of the better correlation demonstrated with DLco percent. The E/I ratio was a better assessment of the degree of air trapping present, however, because of the higher correlation with RV percent than visual scores. An increase in CT number on expiration may be caused by either emphysema or air trapping related to coexisting small airway disease. Emphysema is difficult to separate from air trapping on expiratory CT. Bergin et al6 also compared quantitative image assessment with the visual scoring system previously described. They were particularly interested in interscan and inter-reader variation between the two techniques. It was demonstrated that QIA had much less interreader and interscan variation than visual techniques and they recommended QIA as the preferential method. Quantitative Image Analysis Quantitative image analysis of CT images obtained in patients with emphysema has

been studied extensively. In 1988, Mu¨ller et al64 reported the successful correlation of CT image analysis of emphysema with pathologically resected lung specimens using the open, density mask, computer program available on CT scanners. CT scans obtained at total lung capacity using 10-mm collimation (thick-section) contiguous scanning techniques were obtained of the entire thorax in this study. The density mask program calculated the amount of emphysematous lung present, using the fact that normal lung has an attenuation greater than 910 H and the emphysematous lung has an attenuation less than 910 H. The same density masked approach was used by Kinsella et al47 to compare CT image analysis with pulmonary function testing. In this study they calculated mean lung density and the amount of lung less than 910 H in density. There was a good correlation between mean lung density and the volume of emphysema present, and CT-determined total lung volumes and measures of FEV1 percent, FEV1/FVC percent, functional residual capacity percent, RV percent, and TLC percent. A good correlation was also demonstrated between mean lung density and volume of emphysema with measures of DLco percent and DLco/VA percent. The investigators noted that mean lung density was not a good indicator of emphysema because it did not separate cases with less than 5% involvement by emphysema. Knudson et al,49 analyzing expiratory CT scans, reported that they were superior to inspiratory CT scans in quantitating the amount of emphysema present. The density mask approach was also used in the program to calculate the amount of lung with emphysema present on one inspiratory scan and one expiratory scan obtained at either the level of the transverse aorta or at the level of the carina using 10-mm thick-section images. In this study, normal lung attenuation was assumed to be 600 to 900 H. Emphysematous lung was assumed to have an attenuation greater than 900 H. The amount of abnormal lung on the two inspiratory images was averaged and the amount of abnormal lung on the two expiratory images was averaged. The amount of emphysema present on the E/I scans correlated with the measures of pulmonary function including FEV1 percent, FEV1/FVC percent, RV/TLC percent, DLco percent, and DLco/VA percent. They demon-

QUANTITATIVE CT OF THE LUNG

strated better correlation with the expiratory scan data with these parameters of pulmonary function than with the inspiratory data. The expiratory scans also correlated significantly with a measure of elastic recoil of lung (k parameter), whereas the inspiratory scan data did not. Miniati et al62 reported in 1995 the results of evaluating emphysema in 46 patients with chest radiography, qualitative CT, and quantitative CT. These investigators reported a good correlation between visual scoring of chest radiograph and HRCT scans and physiologic measurements of airflow obstruction, increased lung volumes, and impaired diffusing capacity. They also demonstrated that expiratory scans were a better measurement of abnormal lung function than inspiratory scans. Crausman et al,18 in the same year, conducted a study of 14 patients with emphysema and reported excellent correlation between quantitative analysis of expiratory HRCT images of the lungs and measures of pulmonary function, including excise testing. They also reported good correlation with the amount of lung less than 900 H and physiologic measures of lung function, including maximum workload achieved during exercise and maximum oxygen use. Gierada24a in 1997 compared quantitative CT scan analysis, preoperative physiologic assessment, and outcome measures in 46 patients who underwent lung volume reduction surgery for palliation for severe emphysema. Several quantitative CT and preoperative physiologic values correlated with each other and several quantitative CT and outcome measures correlated with each other. The postoperative outcome was better when the mean lung attenuation was greater than 900 H. It was also better when 75% or more of upper lung fell below 900 H (emphysema index) and when more than 25% of the lung fell below 960 H (surveyor emphysema index). Postoperative outcome was better when the ratio of the upper to lower lung emphysema indices was 1.5 or higher. They also noted a better outcome in patients who had more than 1 L of normal lung (defined as lung with an attenuation between 850 H and 701 H). There was a more favorable outcome when the full width at half maximum of attenuation frequency distribution was 80 H or less. There were often twofold to threefold differences in outcome measures between groups that have been stratified using quantitative CT values. The patients who had a large number of low quantitative CT

151

values had a better outcome. There were few correlations between preoperative physiologic values and outcome measures. In addition to quantitation of structural damage at static lung volumes, analysis of the rate of change in Hounsfield units (H) during a FVC maneuver plotted against time (T) can give additional functional information. The plot of H/T curves for normal lungs demonstrates maximum flow and increase in lung attenuation occurs within the first 2 to 4 seconds of forced expiration that parallels the volume time curve. In patients with obstructed airflow, such as severe emphysema, airflow is considerably decreased and there is minimal change in lung attenuation over time, resulting in marked flattening of the H/T curve. Despite similar visually apparent extent of structural damage, caused by emphysema, H/T and lung attenuation change (LAC) curves can show marked differences in airflow and air trapping offering further insights into the pathophysiology of emphysema (Fig. 5). This can also be useful in assessing split lung function in patients with emphysema who have undergone single lung transplantation. A classic early dynamic collapse pattern with late terminal flow is noted on spirometry in these patients. Although at first thought to be attributable to the dynamic collapse of the anastomotic lung, Hounsfield unit versus time curve analysis performed during a forced expiration maneuver demonstrates that the terminal phase is, in fact, most likely caused by the slow flow from the native obstructed lung where flow is seen to occur after complete emptying of the transplant lung.92 Further, the differential airflow of the transplant lung can be assessed easily independently of the native lung in these patients. This can be useful in the detection of early anastomotic complications leading to obstruction (Fig. 6).54 Asthma Airway reactivity, the hallmark of asthma, refers to the ability of the airways to alter reversibly their diameters in response to stimuli. It is well documented that heightened airway reactivity can also contribute significantly to the morbidity and mortality of other airway diseases, including chronic bronchitis and cigarette smoke–induced chronic obstructive pulmonary disease.16, 19, 26, 35, 42–45, 63, 73, 85 Airway reactivity can be evaluated directly

152

GOLDIN

Figure 5. Three patients with similar visually appreciated extent of emphysema (A) demonstrate marked variation in the plot of frequency distribution curve of lung attenuation (B) within the given cross-section region of interest (at TLC and residual volume [RV]). From the curves the extent of parenchymal destruction (below 950 HU) and degree of air trapping (shift of curve between TLC and RV) can be determined. For the plots of attenuation and flow versus time (Hounsfield units/time [HU/T]) (C) the median lung attenuation at each point in time is derived from the anatomically like region of interest (ROI) within an axial section. This offers a regional measure of airflow. As seen, there is a variation between subjects and between lungs in the same patient.

by measuring airflow changes induced by a bronchoconstrictive challenge or evaluated indirectly by measuring reversible airflow obstruction following the administration of a bronchodilator in a patient with airflow obstruction.38, 55 The use of pulmonary function tests to measure airflow obstruction is limited by the fact that these offer a global assessment of the lungs and can neither express the potential regional heterogeneity of bronchial reactivity and airflow obstruction nor localize the distribution or generations involved. The small air-

ways cannot be visualized directly with current radiographic techniques; however, they can be assessed indirectly using HRCT. Obstruction of small airways, even in mild asthma, results in air trapping and regional hyperinflation. Regional hyperinflation can be assessed using functional imaging techniques and postprocessing image analysis measures.25, 29, 58, 66 A number of investigators have used CT to study patients in vivo with asthma and chronic obstructive pulmonary disease. 17, 25, 79, 99 Using spirometrically gated CT, Lamers

QUANTITATIVE CT OF THE LUNG

153

Figure 6. Attenuation and flow versus time curves during a forced expiration in a patient with emphysema following single lung transplant reflect the differential functions of the native and allograft lungs. For the transplant, lung airflow limitation caused by anastomotic site narrowing is detected by the decreased rate of airflow, which is corrected poststent placement. The flow volume loops (FVL) did not show the abnormality or air difference pre- and post-stent, but the patient was markedly less symptomatic. FVL and HU/T curves pre-and post-stent placement. HU/T curves for the transplant lung demonstrated a right shift (
lower flow) although the lung emptied to RV by end expiration. Poststent shift of the HU/T curves of the transplant 1 to the left reflecting faster airway flow. FLVs pre- and post-stent are unchanged.

et al52a compared measurements of lung attenuation obtained at 10% and 90% of VC in patients with chronic bronchitis and emphysema relative to normal subjects. Whereas patients with emphysema exhibited significant decreases in lung attenuation relative to normal subjects at both levels of breath-hold, patients with chronic bronchitis exhibited abnormally decreased lung attenuation only at 10% of VC (P ⬍ 0.001). The mean change in lung attenuation between the two breath-hold maneuvers was substantially greater in normal patients than in either category of patients with chronic obstructive pulmonary disease. There were significant correlations between lung attenuation at 10% VC and FEV1 and DLcosb and between 90% VC and DLcosb. The investigators proposed that decreased lung attenuation in emphysema represents lung destruction and reflects expiratory air trapping in chronic bronchitis. Moreover, they concluded that in patients with airflow obstruction, images obtained at 10% and 90% VC at the same anatomic level may be sufficient to characterize the nature of chronic airflow obstruction. In a prospective study of asymptomatic asth-

matics and nonsmoking normal subjects, Newman et al70 compared lung attenuation on CT and HRCT images obtained through the lung bases and found that the expiratory mean pixel index (percentage of pixels ⬍ 900 H) was significantly higher in 18 nonsmoking asthmatics than in 22 normal control subjects. The mean pixel index correlated significantly with pulmonary function parameters that reflect air trapping and airflow obstruction including FEV1, RV, and functional residual capacity. More recently, in 10 patients with asthma Gevenois et al24 showed that CT density on the inspiratory scans was unaffected by acute air trapping caused by bronchial challenge or by chronic hyperinflation patients with chronic asthma. Goldin et al29 showed significant leftward shifts in frequency distribution of lung parenchymal attenuation values measured by 3mm-thick CT scans of lungs before and after methacholine challenge in 15 mild asthmatic patients. These effects returned to normal after the administration of albuterol. The same effects were not seen in the six control subjects. These studies suggest the lower attenuating areas observed on expiratory HRCT

Figure 7. Images demonstrate right upper lobe matched pairs pre- and post-methacholine challenge at visits two (baseline) and four (post-treatment) in a patient representative of the CFC cohort. The resultant lung attenuation curve (LAC) analysis (in the midzone) demonstrates no significant change in distribution (and therefore no change in air trapping) either pre- or post-methacholine challenge on visit four compared with visit two.

154

QUANTITATIVE CT OF THE LUNG

scans correlate to areas of air trapping rather than emphysema. Okazawa et al72 evaluated six normal patients and six mild to moderate asthmatics before and after methacholine challenge and found that airway luminal narrowing could be quantified both in normal patients and asthmatics. The airway wall area decreased in normal subjects who developed bronchial constriction caused by methacholine and in the asthmatics. In asthmatics this decrease in airway wall area did not occur and it was postulated that this was caused by the more edematous, stiffer airway wall being relatively uncoupled from the elastic recoil of the surrounding lung. In a separate study, Kee et al46 confirmed that CT produced quantified changes in the internal luminal diameter of asthmatic airways provoked by methacholine and albuterol inhalation. Quantitative image analysis of HRCT29 was used to assess the relative efficacy of HFA-BDP and CFC-BDP on small airway hyperreactivity and regional air trapping. A double-blind, randomized, parallel-group pilot study compared the relative efficacy of HFA-134a beclomethasone (Qvar, mass median aerodynamic diameter [MMAD] 0.8 to 1.2 ␮m) with CFC-11/12 beclomethasone (Beclovent, MMAD 3.5 to 4 ␮m) in steroidnaive patients with mild to moderate asthma (PCO2 ⱕ 4 mg/mL). Pretreatment HRCT was performed at RV before and after methacholine challenge test. After 4 weeks of treatment, functional imaging was repeated before methacholine challenge test and after the same concentration of methacholine concentration as the pretreatment scan (N20), or after a 20% or greater fall in FEV 1 (N6 with a higher PCO2, N5 with a lower PCO2). Quantitative assessment of changes in the distribution of lung attenuation was performed. No significant difference was demonstrated between the two treatment groups with respect to improvement in symptoms, spirometry, or methacholine concentration sensitivity, except for a greater decrease in shortness of breath in the HFA-BDP group (P0.035). Before methacholine challenge test at the post-treatment scan (treatment efFVCt), the HFA-BDP group (N14) showed less air trapping in the middle lung zone (less shift in the lowest tenth percentile of the lung attenuation curve at RV) (P0.048). The CFC-BDP group (N15) showed nonsignificant changes in air trapping in the same area (P0.474). The HFA-BDP group showed significantly more

155

improvement in air trapping than the CFCBDP group (P0.044). Given an equal constrictor stimulus (methacholine concentration) HFA-BDP subjects (N1) showed less increase in air trapping in the middle lung zones (less shift in the median of the lung attenuation curve at RV) than CFC-BDP subjects (N9) at the post-treatment scan compared with the pretreatment scan (P0.037). The investigators concluded that HFA-BDP may have greater efficacy in the peripheral airways and that this effect is best assessed using quantitative imaging CT techniques (Figs. 7 and 8.)31 In summary, CT may be useful in asthma either to quantify the extent of air trapping or quantify the degree of airway wall thickening and luminal narrowing and assess novel drug treatment efficacy. These approaches are complimentary. Assessment of airway wall changes requires meticulous techniques to achieve accurate reproducible measurement. Interpretation of airway wall measures is difficult unless one knows the order of the measure bronchus within the bronchial tree. Measurement of air trapping, although indirect, is a relatively simple and useful method for the quantitation of asthma severity. INTERSTITIAL LUNG DISEASE Visual Quantitation Thoracic HRCT has documented use as a noninvasive means of detecting and characterizing interstitial pulmonary parenchymal abnormalities in systemic sclerotic (SSc) lung disease.61, 74, 76 In addition, HRCT is more sensitive to the detection of pulmonary parenchymal changes of interstitial lung disease than projectional chest radiography. 74, 88, 90 HRCT evidence of interstitial lung disease is present in up to 91% of SSc patients,76, 88 a percentage similar to the percentage of patients with SSc who have pulmonary fibrosis on postmortem examination. The closest correlation between pulmonary parenchymal abnormalities identified on HRCT and histology exists in the presence of interstitial lung disease.74 Typical interstitial abnormalities identified on HRCT include thickened interlobular septa, subpleural lines and parenchymal bands, architectural distortion, subpleural cysts, and honeycomb lung formation. Additional findings include subpleural micro-

Figure 8. Images demonstrate right upper lobe matched pairs pre- and post-methacholine challenge at visits two (baseline) and four (post-treatment) in a patient representative of the HFA cohort. The matched premethacholine images (in the midzone) show an apparent decrease in prominence of low-attenuation areas (and therefore air trapping) at baseline, and the resultant LAC curve analysis shows significantly less shift (almost none in this case) postmethacholine at visit four compared with visit two. PTH  Patient H; LAC  lung attenuation curve.

156

QUANTITATIVE CT OF THE LUNG

nodules, small airway ectasia (bronchiectasis and bronchiolectasis), and ground glass opacity.79 In the absence of associated airway changes (e.g., traction bronchiectasis or bronchiolectasis), ground-glass opacification of the lung identified on HRCT corresponds to histologic evidence of alveolar inflammation in patients with chronic diffuse infiltrative lung disease.78 In this setting, ground-glass opacification seems to precede the HRCT appearance of honeycombing, supporting the idea that alveolitis precedes irreversible fibrosis.78 In patients with SSc, ground-glass opacification may be present in combination with varying degrees of interstitial pulmonary fibrosis. In these patients, the implications of groundglass opacification are less well understood. More extensive ground-glass opacification on HRCT is significantly associated with a lower DLco, 76 similar to the association between bronchoalveolar lavage (BAL) evidence of alveolitis and a reduced DLco. A recent study comparing the ability to detect alveolitis by HRCT and by BAL found that nearly 50% of systemic sclerosis patients with a normal HRCT had abnormal cellularity by BAL. 76 This exemplifies the need to detect and quantitate better the presence or absence caused by active alveolitis from irreversible fibrosis in patients with interstitial lung disease. Collins et al16 showed moderate interobserver agreement in scoring pattern type and disease extended patients with IPF. The agreement found was substantially greater than that obtained by scoring chest radiographs. Kazerooni et al, 45 more recently, found that visual scoring of HRCT scans for fibrosis and ground-glass opacity using a 5point scale was associated with moderate to good interobserver agreement (kappa values 0.51 to 0.83). The CT fibrosis score showed moderate correlation (r0.50) with pathologic evidence of fibrosis and slightly less strong correlation (r0.40) with pathologic signs of inflammation. Disappointingly, the CT score for ground-glass attenuation showed relatively weak correlation (r0.26) with pathologic evidence of inflammation. A more detailed scoring system was introduced by Al Jarad et al1a to assess thoracic CT scans in patients with asbestosis. Lung fibrosis and emphysema were scored using a 12-point scoring scale similar to that used in the International Labor Organization classification system for pneumoconiosis. Each of three lung zones were scored separately. No refer-

157

ence standard images were used and the criteria for different categories were not detailed. Despite this, they found good interobserver and intraobserver agreement on scoring the lung zones, with 90% of the fibrosis scores being within two categories of each other. There were moderate correlations between the scores of fibrosis and emphysema and physiologic indices. Morphometry is a point-counting technique developed for the quantitative assessment of histologic features of the lung parenchyma.38 This technique has been modified and applied to the assessment of CT features of bleomycin-induced lung disease in rabbits.38 The extent of reticular-linear opacity seen on CT morphometry correlated well with the extent of lung fibrosis on histology but this was not true for ground-glass opacities. More recently, in the same model, Lynch et al56 showed that reticular opacities were inversely correlated with TLC and with the alveolar-arterial gradient, whereas groundglass opacification did not correlate with any abnormality. CT morphometry is a powerful tool for estimating the extent of different types of disease present on CT and correlating this with the degree of physiologic impairment. It is, however, tedious and difficult to apply to large populations. Visual estimation has as its greatest advantage the simplicity of its approach. The disadvantages are subject to subjectivity and the difficulty in estimating the contribution of the different components of disease (honeycombing, reticular, and linear ground-glass opacity). A further difficulty is in the complex task of integrating the extent of the abnormalities seen on several CT slices and deriving a quantitative measure of the total extent of abnormality on any lung zone or within the lung. Attempts to develop a reference set of images to standardize this type of quantitation may be useful in this regard. Quantitative Image Analysis CT density has also been used to evaluate the extent of diffuse parenchymal lung diseases, such as IPF,36 asbestosis,103 sarcoidosis,25 and radiation fibrosis (Fig. 9).99 Correlation with indices of pulmonary function was achieved using these quantitative methods. In two studies of asbestos workers, Wollmer et al 103 Reuter et al 79 demonstrated that lung density was increased and correlated in-

158

GOLDIN

versely with lung volumes. Chest radiographs and CT scans showed normal or near normal appearance, reinforcing the potential for computer-aided quantitation techniques. Hartley et al36 used the analysis of Hounsfield density histogram to assess patients with IPF. In normal subjects, the Hounsfield histogram is sharply peaked, kurtotic, and skewed to the left by comparison with the gaussian distribution. In IPF, because of increased amounts of fibrotic or inflammatory tissue in the lung, the mean lung attenuation is increased and the histogram becomes less skewed and much less kurtotic. Measures of mean CT density, skewing, and kurtosis in patients with IPF correlate to a degree of impairment as a measure of pulmonary function testing. 36 In another study, Coxson et al 17 showed that the ratio of air to tissue in the lung estimated from CT density values correlated with that found on quantitative histology. The author has used QIA to detect and assess the nature of opacities seen on HRCT of patients with scleroderma and symptoms suggestive of lung disease but normal chest radiographs (Goldin et al, unpublished data). Subjects underwent HRCT sequences in supine position at TLC and RV, respectively, using 1-mm collimation at spaced intervals from apex to lung base. Bronchoalveolar la-

vage (BAL) was performed on the patients with scleroderma within 1 week of CT in two lung regions, selected to include at least one site corresponding to a region of abnormality localized on HRCT. In subjects with no HRCT abnormalities, BAL was performed in the right middle lobe and lingula. Analysis consisted of quantitative assessment of texture measures obtained from the region of the image that best correlated with the site of BAL. The posterior third of the left and right lower lobes was performed on 6 normal, healthy individual volunteers and 19 UIP patients in the supine position at RV. A classification tree7 was built using the texture measures and LAC measures on these 50 lung segments. For this preliminary analysis the measures obtained from the left and right lobes of each subject were assumed to be semi-independent. Cross-validation (leaving one out) was used to determine the sensitivity and specificity of the resulting tree. Three measures (correlation, standard deviation, and sum of squares variance) separated normal subjects from UIP subjects in this data set. The sensitivity of the classifier is 91.7% (11 of 12 normals correctly classified, lower 95% confidence interval [80.5, 100]) and the specificity is 89.5% (34 of 38 UIP subjects correctly classified, lower 95% confidence interval [76.5, 100]). In six patients the QIA detected areas

Figure 9. QIA based on two texture features contrast and SIMC clearly separates bronchoalveolar lavage (BAL) defined alveolitis from normal lung (A). Using two other texture features fifth centile and sum of variance (B) separation of normal, ground-glass and fibrotic parenchyma can be achieved. The overlap between alveolitis and fibrosis in visually apparent ground may account for the overlap seen between these two entities. This QIA offers the potential to separate alveolitis from ground glass if texture features can be trained against correlative pathology rather than visual assessment. N  normal lung; S  scleroderma; G  ground glass; F  fibrosis; SUMVAR  sum of variance; SIMC  second information measure of correlation and contrast.

QUANTITATIVE CT OF THE LUNG

of alveolitis that appeared visually normal indicating the added potential of QIA to assess UIP over classic visual inspection. The first two variables and their splits used in the classification tree are shown in Figure 9A. Two texture measures (second information measure of correlation and contrast) separated BAL-positive from BAL-negative defined acute alveolitis with a sensitivity of 92.3% (12 of 13 BAL-positive regions correctly classified, lower 95% confidence interval [78.6, 100]) and a specificity of 88.9% (eight of nine BAL-negative regions correctly classified, lower 95% confidence interval [75.0, 100]). The first two variables and splits are shown in Figure 9. In four patients there was discrepancy between the BAL cell analysis between the two lobes lavaged. In all four cases QIA correctly detected the presence of abnormality, which was only visually obvious in two patients (see Fig. 9). The Future The advent of rapid volumetric image data acquisition at high spatial resolution, coupled with sophisticated postprocessing technology, provides unique opportunities to explore the relationships between regional pulmonary structure and function. Future developments to extend the practical use of QIA will likely include the development of normative tables for functional imaging parameters independent of traditional pulmonary function tests, reductions in exposure times while preserving image quality, and advanced image processing capability for rapid analysis and display. New contrast agents, such as injectable biodegradable microspheres,105 carry promise to enable real-time acquisition of perfusion and ventilation and cell function. It is exciting to consider that QIA will continue to develop from the realm of research tool into a practical technology for the assessment of patients with early pulmonary conditions. The QIA of CT data may be useful in diffuse lung disease either to detect the nature of airway or parenchymal abnormality or to quantify the functional impairment and assess novel drug treatment efficacy. These approaches are complimentary. QIA requires meticulous imaging techniques to achieve accurate reproducible measurement. Postprocessing software must become more robust and easy to use, but there is a clear trend in this direction. Interpretation of airway wall

159

measures is difficult unless one knows the order of the measure bronchus within the bronchial tree. Measurement of air trapping, although indirect, is a relatively simple and useful method for the quantitation of asthma severity.

SUMMARY Lung disease is a leading cause of morbidity and mortality. HRCT, currently the best test to assess lung involvement in emphysema and interstitial lung disease, relies on abnormalities being detected when there is sufficient morphologic distortion to result in visually identified changes that, for the most part, correlate poorly with conventional lung function tests and outcome. QIA offers a technique to assess both structure and function on a regional and global basis. With the advent of user-friendly software packages, this approach is finding application in clinical practice and in clinical studies of new treatment alternatives for diffuse lung disease

References 1. Adachi S, Kono M, Takemura T: Evaluation of 3D spiral CT bronchoscopy in patients with lung cancer. Radiology 189:264-269, 1993 1a. Al Jarad N, Wilkinson P, Pearson MC, et al: A new high resolution computed tomography scoring system for pulmonary fibrosis, pleural disease, and emphysema in patients with asbestos related disease. Br J Indust Med 49:73–84, 1992 2. American Lung Association: Lung Disease Data. New York, American Lung Association, 1993 3. Archer DC, Coblentz CL, deKemp R, et al: Automated in vivo quantification of emphysema. Radiology 188:835-838, 1993 4. Bae KT, Sloane RM, Gierada DS, et al: Patients with emphysema: Quantitative CT analysis before and after lung volume reduction surgery. Radiology 203:705-714, 1997 5. Becker MD, Berkmen YM, Austin JH, et al: Lung volumes before and after lung volume reduction surgery: Quantitative CT analysis. Am J Respir Crit Care Med 1575 (5 pt 1):1593-1539, 1998 6. Bergin C, Muller N, Nichols DM, et al: The diagnosis of emphysema: A computed tomographic-pathologic correlation. Am Rev Respir Dis 133:541-546, 1986 7. Breiman L, Friedman JH, Olshen RA, et al: Classification and Regression Trees. New York, Chapman and Hall, 1984 8. Brown MS, Goldin JG, McNitt-Gray MF, et al: Automated assessment of split lung function for lung transplant patients. Radiology 205(P):395, 1997 9. Brown MS, Goldin JG, McNitt-Gray MF, et al: Knowledge-based segmentation of thoracic CT im-

160

10. 11. 12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23. 24.

24a.

GOLDIN ages for assessment of split lung function. Med Phys, 27:592–598, 2000 Brown MS, McNitt-Gray MF, Goldin JG, et al: Automated measurement of split and total lung volume from helical CT. Radiology 205(P):224, 1997 Brown MS, McNitt-Gray MF, Goldin JG, et al: Automated measurement of single and total lung volume from CT. J Comput Assist Tomogr 23:632-640, 1999 Brown MS, McNitt-Gray MF, Goldin JG, et al: Knowledge-based method for segmentation and quantitative analysis of lung function from CT. Proceedings of the 1st International Workshop on Computer-Aided Diagnosis. Chicago, September 20–23, 1998 Brown MS, McNitt-Gray MF, Mankovich NJ, et al: Knowledge-based automated technique for measuring total lung volume from CT. Proc SPIE Medical Imaging 2709:63-74, 1996 Brown RH, Mitzner W, Zerhouni E, et al: Direct in vivo visualization of bronchodilation induced by inhalational anesthesia using high-resolution computed tomography. Anesthesiology 78:295-300, 1993 Brown RH, Zerhouni EA, Mitzner W: Visualization of airway obstruction in vivo during pulmonary vascular engorgement and edema. J Appl Physiol 78:1070-1078, 1995 Collins CD, Wells AU, Hansell DM, et al: Observer variation in pattern type and extent of disease in fibrosing alveolitis on thin section computed tomography and chest radiography. Clin Radiol 49:236240, 1994 Coxson HO, Hogg JC, Mayo JR, et al: Quantification of idiopathic pulmonary fibrosis using computed tomography and histology. Am J Respir Crit Care Med 155:1649-1656, 1997 Crausman RS, Ferguson G, Irvin CG, et al: Quantitative chest computed tomography as a means of predicting exercise performance in severe emphysema [published erratum appears in Acad Radiol 2:870, 1995]. Acad Radiol 2:463-469, 1995 Daniloff E, Lynch D, Bartelson B, et al: Observer variation and relationship of computed tomography to severity of beryllium disease. Am J Respir Crit Care Med 155:2047-2056, 1997 Delorme S, Keller-Reichenbecher MA, Zuna I, et al: Usual interstitial pneumonia: Quantitative assessment of high-resolution computed tomography findings by computer-assisted texture-based image analysis. Invest Radiol 32:566-574, 1997 Disler DG, Marr S, Rosenthal DI: Accuracy of volume measurements of computed tomography and magnetic resonance imaging phantoms by threedimensional reconstruction and preliminary clinical application. Invest Radiol 29:739-745, 1994 Eda S, Kubo K, Fujimoto K, et al: The relations between expiratory chest CT using helical CT and pulmonary function tests in emphysema. Am J Respir Crit Care Med 155:1290-1294, 1997 Gail DB, Lenfant CJM: Cells of the lung: Biology and clinical implications. Am Rev Respir Dis 127:366-387, 1983 Gevenois PA, Scillia P, de Maertelaer V, et al: The effects of age, sex, lung size, and hyperinflation on CT lung densitometry. AJR Am J Roentgenol 167:1169-1173, 1996 Gierada DS, Slone RM, Bae KT, et al: Pulmonary emphysema: Comparison of preoperative quantitative CT and physiologic index values with clinical outcome after lung volume reduction surgery. Radiology 205:235–242, 1997

25. Gilman M, Laurens R, Somogyi J, et al: CT attenuation values of lung density in sarcoidosis. J Comput Assist Tomogr 7:407-410, 1983 26. Glenn W, Greene R: Rapid computer-aided radiographic calculation of total lung capacity. Radiology 117:269-273, 1975 27. Goldin JG, Aberle DR: Functional imaging of the airways. J Thorac Imaging 12:29-37, 1997 28. Goldin J, Brown MS, McNitt-Gray MF, et al: Automated assessment of split lung function in post lung transplant evaluation. Proc SPIE 3337:14-23, 1998 29. Goldin JG, McNitt-Gray MF, Sorenson SM, et al: Airway hyperreactivity: Assessment with helical thin-section CT. Radiology 208:321-329, 1998 30. Goldin JG, Szold O, McNitt-Gray MF, et al: Use of EBCT in the evaluation of patient’s post single lung transplantation. Am J Respir Crit Care Med 9:81, 1995 31. Goldin JG, Tashkin DP, Kleerup EC, et al: Comparative effects of HFA- and CFC-beclomethasone diproprionate inhalation on small airways: Assessment using functional helical thin-section computed tomography. J Allergy Clin Immunol 104:S258–S267, 1999 32. Gurney JW, Jones K, Robbins RA, et al: Regional distribution of emphysema: Correlation of HRCT with pulmonary function tests in unselected smokers. Radiology 183:457-463, 1992 33. Haralick K: Statistical and structural approaches to texture. Proc IEEE 67:786-804, 1979 34. Haralick K, Shanmugan I, Dinstein P: Textural features for image classification. IEEE Trans Man Sys Cyber 3:610-621, 1973 35. Harris T, Pratt P, Kilburn K: Total lung capacity measured by roentgenograms. Am J Med 50:756763, 1971 36. Hartley PG, Galvin JR, Hunninghake GW, et al: High-resolution CT-derived measures of lung density are valid indexes of interstitial lung disease. J Appl Physiol 76:271-277, 1994 37. Herold CJ, Brown RH, Mitzner W, et al: Assessment of pulmonary airway reactivity with high-resolution CT. Radiology 181:369-374, 1991 38. Hirose N, Lynch DA, Cherniack RM, et al: Correlation between high resolution computed tomography and tissue morphometry of the lung in bleomycininduced pulmonary fibrosis in the rabbit. Am Rev Respir Dis 147:730-738, 1993 39. Hoffman EA, McLennan G: Assessment of the pulmonary structure-function relationship and clinical outcomes measures: Quantitative volumetric CT of the lung. Acad Radiol 4:758-776, 1997 40. Kalender W, Rienmuller R, Seissler W, et al: Measurement of pulmonary parenchymal attenuation: Use of spirometric gating with quantitative CT. Radiology 175:265-268, 1990 41. Kaneko M, Ehuchi K, Ono R: Three-dimensional display of bronchial images by helical scanning CT. Radiology 177(P):174, 1990 42. Katsuragawa S, Doi K, MacMahon H, et al: Classification of normal and abnormal lungs with interstitial diseases by rule-based method and artificial neural networks. J Digit Imaging 10:108-114, 1997 43. Katsuragawa S, Doi K, MacMahon H: Image feature analysis and computer-aided diagnosis in digital radiography: Classification of normal and abnormal lungs with interstitial disease in chest images. Med Phys 16:38-44, 1989 44. Katsuragawa S, Doi K, MacMahon H, et al: Quanti-

QUANTITATIVE CT OF THE LUNG

45.

46.

47.

48. 49.

50. 51.

52.

52a.

53. 54.

55.

56.

57.

58.

59. 60.

tative computer-aided analysis of lung texture in chest radiographs. Radiographics 10:257-269, 1990 Kazerooni EA, Martinez FJ, Flint A, et al: Thinsection CT obtained at 10-mm increments versus limited three-level thin-section CT for idiopathic pulmonary fibrosis: Correlation with pathologic scoring. AJR Am J Roentgenol 169:977-983, 1997 Kee ST, Fahy JV, Chen DR, et al: High-resolution computed tomography of airway changes after induced bronchoconstriction and bronchodilation in asthmatic volunteers. Acad Radiol 3:389-394, 1996 Kinsella M, Muller NL, Abboud RT, et al: Quantitation of emphysema by computed tomography using a ‘‘density mask’’ program and correlation with pulmonary function tests. Chest 97:315-321, 1990 Klein J, Gamsu G: High-resolution computed tomography of diffuse lung disease. Invest Radiol 24:805-812, 1989 Knudson RJ, Standen JR, Kaltenborn WT, et al: Expiratory computed tomography for assessment of suspected pulmonary emphysema. Chest 99:13571366, 1991 Kramer SS, Hoffman EA: Physiologic imaging of the lung with volumetric high-resolution CT. J Thorac Imaging 10:280-290, 1995 Kuhlman JE, Ney DR, Fishman EK: Two-dimensional and three dimensional imaging of the in vivo lung: Combined spiral computed tomography with multiplanar and volumetric rendering techniques. J Digit Imaging 7:42-47, 1994 Kyung JP, Bergin CJ, Clausen JL: Quantitation of emphysema with three-dimensional CT densitometry: Comparison with two-dimensional analysis, visual emphysema scores, and pulmonary function test results. Radiology 211:541-547, 1999 Lamers JR, Kemerink GJ, Drent M, et al: Reproducibility of spirometrically controlled CT lung densitometry in a clinical setting. Eur Respir J 11:942– 945, 1998 Laws Kl: Textured Image Segmentation. USCIPI Technical Report No 940. University of Southern California, Los Angeles, CA, 1980 Levine MS, Shpiner RB, Martin K, et al: Clinical evaluation of stent placement in patients with persistent dyspnea in post single lung transplant (SLT). Am J Respir Crit Care Med 155:A275, 1997 Lynch D, Hirose N, Cherniack R, et al: Bleomycininduced lung disease in an animal model: Correlation between HRCT-determined abnormalities and lung function. Acad Radiol 4:102-107, 1997 Lynch DA, Newell JD, Tschomper BA, et al: Uncomplicated asthma in adults: Comparison of CT appearance of the lungs in asthmatic and healthy subjects. Radiology 188:829-833, 1993 Mathieson JR, Mayo JR, Staples CA, et al: Chronic diffuse infiltrative lung disease: Comparison of CT and chest radiography. Radiographics 171:111-116, 1989 McNitt-Gray MF, Goldin JG, Johnson TD, et al: Development and testing of image processing methods for the quantitative assessment of airway hyperresponsiveness (AHR) from high-resolution CT images. J Comput Assist Tomogr 21:939-947, 1997 McNitt-Gray MF, Johnson TD, Aberle DR: TRIAD: A thoracic radiology image analysis and display workstation. Med Phys 21:905, 1994 McNitt-Gray MF, Kimme-Smith C, Aberle DR: Calibration of airway area measurements from HRCT images of normal and asthmatic patients. Med Phys 21:926, 1994

161

61. Meziane MA: High-resolution computed tomography scanning in the assessment of interstitial lung disease. J Thorac Imaging 7:13-25, 1992 62. Miniati M, Filippi E, Falaschi F, et al: Radiologic evaluation of emphysema in patients with chronic obstructive pulmonary disease: Chest radiography versus high resolution computed tomography. Am J Respir Crit Care Med 151:1359-1367, 1995 63. Monnier-Cholley L, MacMahon H, Katsuragawa S, et al: Computer-aided diagnosis for detection of interstitial opacities on chest radiographs. AJR Am J Roentgenol 171:1651-1656, 1998 64. Mu¨ller NL, Staples CA, Miller RR, et al: Density masks: An objective method to quantitative emphysema using computed tomography. Chest 94:782787, 1988 65. Mu¨ller NL, Staples CA, Miller RR, et al: ‘‘Density Mask:’’ An objective method to quantitate emphysema using computed tomography. Chest 109:131137, 1996 66. Murata K, Itoh H, Todo G, et al: Centrilobular lesions of the lung: Demonstration by high-resolution CT and pathologic correlation. Radiology 161:641645, 1986 67. Naidich DP: Volumetric scan change perceptions in thoracic imaging. Diagn Imaging 4:70-74, 1993 68. Naidich DP, Webb WR, Mu¨ller NL: Thoracic computed tomography: Current concepts. In Potchen EJ, Grainger RG, Greene R (eds): Pulmonary Radiology. Philadelphia, WB Saunders, 1993, pp 386–404 69. National Center for Health Statistics: Vital Statistics for the United States, Volume 11: Mortality, Part A for Delta Years 1986–1989. Washington, DC, Public Health Service, 1990 70. Newman KG, Lynch DA, Newman LS, et al: Quantitative computed tomography detects air trapping due to asthma. Chest 106:105-109, 1994 71. Ney DR, Kuhlman JC, Hruban RH: Three-dimensional CT-volumetric reconstruction and display of the bronchial tree. Invest Radiol 25:736-742, 1990 72. Okazawa M, Muller N, McNamara AE, et al: Human airway narrowing measured using high resolution computed tomography. Am J Respir Crit Care Med 154:1557-1562, 1996 73. Pappas G, Brodkin C, Sheppard L, et al: The validity of radiographic estimation of total lung capacity in patients with respiratory disease. Chest 114;513520, 1998 74. Pignone A, Matuci-Cerini M, Lombardi A, et al: High resolution computed tomography in systemic sclerosis: Real diagnostic utilities in the assessment of pulmonary involvement and comparison of other modalities of lung investigation. Clin Rheumatol 11:465-472, 1992 75. Reinhardt JM, Hoffman EA: Quantitative pulmonary imaging: Spatial and temporal considerations in high-resolution CT. Acad Radiol 5:539-546, 1998 76. Remy-Jardin M, Giraud F, Remy J, et al: Importance of ground glass attenuation in chronic diffuse infiltrative lung disease: Pathologic-CT correlation. Radiology 189:693-698, 1993 77. Remy-Jardin M, Remy J, Deffontaines C, et al: Assessment of diffuse infiltrative lung disease: Comparison of conventional CT and high-resolution CT. Radiology 181:157-162, 1991 78. Remy-Jardin M, Remy J, Marquette CH, et al: Central airways in asthmatics and control subjects: Evaluation with spirometric-gated spiral CT. Radiology 185(P):131, 1992

162

GOLDIN

79. Reuter M, Holling I, Emde L, et al: Quantitative computed tomography of the lung in the early diagnosis of asbestosis: Initial results in asbestos-exposed navy personnel. Rofo Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr 162:288–295, 1995 80. Rienmu¨ller R, Behr J, Kalender WA, et al: Standardization quantitative high-resolution CT in lung diseases. J Comput Assist Tomogr 15:742-749, 1991 81. Rodriguez LH, Vargas PF, Raff U, et al: Automated discrimination and quantification of idiopathic pulmonary fibrosis from normal lung parenchyma using generalized fractal dimensions in high-resolution computed tomography images. Acad Radiol 2:10-18, 1995 82. Rubin GD, Beaulieu CF, Argiro V, et al: Perspective volume rendering of CT and MR images: Applications for endoscopic imaging. Radiology 199:321330, 1996 83. Rubin GD, Dake MD, Semba CP: Current status of three-dimensional spiral CT scanning for imaging the vasculature. Radiol Clin North Am 33:51-70, 1995 84. Rubin GD, Napel S, Leung A: Volumetric analysis of volume data: Achieving a paradigm shift. Radiology 200:312-317, 1996 85. Sakai F, Gamsu G, Im J-G, Ray C: Pulmonary function abnormalities in patients with CT-determined emphysema. J Comput Assist Tomogr 11:963-968, 1987 86. Sanders C, Nath PH, Bailey WC: Detection of emphysema with computed tomography: Correlation with pulmonary function tests and chest radiography. Invest Radiol 155:1290-1294, 1988 87. Schaefer CM, Prokop M, Zinh C: Spiral CT of the anastamotic complications after lung transplantation. Radiology 189(P):263, 1993 88. Schurawitzki H, Stiglbauer R, Graninger W, et al: Interstitial lung disease in progressive systemic sclerosis: High-resolution CT versus radiography. Radiology 176:755-759, 1990 89. Sonka M, Hlavac V, Boyle R: Image Processing, Analysis and Vision. London, Chapman and Hall Computing, 1993 90. Strickland B, Strickland NH: The value of high definition, narrow section computed tomography in fibrosing alveolitis. Clin Radiol 39:589-594, 1988 91. Swensen SJ, Aughenbaugh GL, Douglas WW, et al: High resolution CT of the lungs: Findings in various pulmonary diseases. AJR Am J Roentgenol 158:971979, 1992 92. Szold O, Levine MS, Goldin JG, et al: Late expiratory plateau patients in post-single lung transplant patients with emphysema. Am J Respir Crit Care Med 151:A85, 1995

93. Thurlbeck WM: Emphysema then and now. Can Respir J 1:21-39, 1994 94. Uppaluri R, Hoffman EA, Sonka M, et al: Computer recognition of regional lung disease patterns. Am J Respir Crit Care Med 160:648-654, 1999 95. Uppaluri R, Hoffman EA, Sonka M, et al: Interstitial lung disease: A quantitative study using the adaptive multiple feature method. Am J Respir Crit Care Med 159:519-525, 1999 96. Uppaluri R, McLennan G, Enright P, et al: Adaptive multiple feature method (AMFM) for the early detection of parenchymal pathology in a smoking population. Proc SPIE 3337:8-13, 1998 97. Uppaluri R, McLennan G, Sonka M, et al: Computer-based objective quantitative assessment of pulmonary parenchyma via x-ray CT. Proc SPIE 3337:377-383, 1998 98. Uppaluri R, Mitsa T, Sonka M, et al: Quantification of pulmonary emphysema from lung computed tomography images. Am J Respir Crit Care Med 156:248-254, 1997 99. Van Dyk J, Hill R: Postirradiation lung density changes as measured by computerized tomography. Int J Radiat Oncol Biol Phys 9:847-852, 1983 100. Webb WR: High-resolution CT of the lung parenchyma. Radiol Clin North Am 27:1085-1097, 1989 101. Webb WR, Stern EJ, Kanth N, et al: Dynamic pulmonary CT: Findings in healthy adult men. Radiology 186:117-124, 1993 102. Weibel ER, Crystal LRG: Structural organization of the pulmonary interstitium. In Crystal RG, West JB, Weibel ED, (eds): The Lung: Scientific Foundations, ed 2. Philadelphia, Lippincott-Raven, 1997, pp 685– 695 103. Wollmer P, Jakobsson K, Albin M, et al: Measurement of lung density by x-ray computed tomography: Relation to lung mechanics in workers exposed to asbestos cement. Chest 91:865-869, 1987 104. Wood SA, Hoford JD, Hoffman EA, et al: A method for measurement of cross sectional area, segment length, and branching angle of airway tree structures in situ. Comput Med Imaging Graph 19:145152, 1995 105. Workman MJ, Tajik JK, Robinson MT, et al: Biodegradable radioplaque microspheres for the evaluation of regional blood flow distribution using x-ray computed tomography. SPIE Med Imaging 216836:344, 1994 106. Wu M, Chang J, Chiang A, et al: Use of quantitative CT to predict postoperative lung function in patients with lung cancer. Radiology 191:257-262, 1994 107. Zagers R, Vrooman HA, Aarts NJM, et al: Quantitative analysis of computed tomography scans of the lungs for the diagnosis of pulmonary emphysema: A validation study of a semiautomated contour detection technique. Invest Radiol 30:552-562, 1995 Address reprint requests to Jonathan G. Goldin, MD, PhD Department of Radiological Sciences UCLA Medical Center 10833 Le Conte Avenue B2-168, CHS Los Angeles, CA 90095–1721 e-mail: [email protected]