Incidental Thyroid Nodule: Patterns Of Diagnosis And Rate Of Malignancy

Dissections

OBSERVATIONAL 27 April 2009

Evidence-based Medicine for Surgeons Incidental thyroid nodule: patterns of diagnosis and rate of malignancy

Authors: Jin J, Wilhelm SM, McHenry CR Journal: The American Journal of Surgery 2009; 197: 320–324 Centre: Case Western Reserve University, School of Medicine, Cleveland, Ohio, United States

BACKGROUND

A thyroid "incidentaloma" is defined as an unsuspected, focal thyroid lesion discovered by a radiographic imaging modality or at the time of a nonthyroid neck surgery in a patient without history of thyroid disease. It has been suggested that thyroid incidentalomas have a higher rate of thyroid malignancy than the previously reported rate of 5% for palpable thyroid nodules Authors' claim(s): “...an incidental thyroid nodule is associated with a high rate of malignancy.”

RESEARCH QUESTION Population Patients referred for evaluation of incidentally detected thyroid nodules during work up for metastatic disease or investigation of other medical problems. Indicator variable Thyroid nodule detected radiological imaging study.

IN

SUMMARY Thyroid incidentaloma

Work up for metastatic disease

Investigations for other medical problems

88

62

Number of incidentalomas on

a

Outcome variable Finding of histologically confirmed thyroid cancer in the nodules. Comparison Historical controls. Cancer in the two subsets that constituted the study.

Mean size of nodule

2.1 cm (0.4 - 8 cm)

Number submitted to FNAC

125 (82%)

FNAC - "benign"

62

FNAC - "indeterminate"

49

FNAC - "malignant"

9

FNAC - "non-diagnostic"

5

Number submitted to thyroidectomy

65

Thyroid cancer on HPE

21

Microcarcinoma

6

Prevalence of thyroid cancer

THE

15/150 (10%)

BOTTOM LINE

As the authors themselves admit, there is a very large selection bias. The majority of patients (88 of 150) were known to have a pre-existing malignancy and the thyroid nodules were detected on imaging studies carried out for metastatic disease. The remainder were also patients with other medical complaints who were being worked up. This cannot be seen in any way as a representative sample of a normal population. The sample population is highly skewed. There is no attempt to construct a fair control. Despite that, the prevalence of thyroid cancer was 10% (excluding the microcarcinomata). Considering the small sample size and the skewed sample, this is not likely to be significantly higher than the quoted 5% in the literature. Interestingly, the mean size of the nodules was 2.1 cm: large enough to be clinically palpable. Are they really "incidentalomas" if this large? The danger in this paper lies in its being used as the rational for ordering imaging studies in a thoughtless fashion.

EBM-O-METER Evidence level

Overall rating

Bias levels

Double blind RCT

Sampling

Randomized controlled trial (RCT) Prospective cohort study - not randomized Case controlled study Case series - retrospective

Trash Life's too short for this

Swiss cheese Full of holes

Safe Holds water



Newsworthy “Just do it”

Comparison Measurement

l | Novel l | Feasible l Ethical l | Resource saving l

Interesting

The devil is in the details (more on the paper) ...

© Dr Arjun Rajagopalan



SAMPLING Sample type Simple random Stratified random Cluster Consecutive Convenience

Inclusion criteria Patients referred to an endocrine surgical unit with a thyroid nodule detected during imaging for metastatic disease or other medical conditions 

Exclusion criteria

Final score card

Previous history of thyroid cancer, previous thyroid surgery, or a known history of nodular thyroid disease 

Study

Judgmental

Target

?

Accessible

?

Intended

150

Drop outs

25 (18%)

Study



= Reasonable |

?

= Arguable |



125

= Questionable

Sampling bias: The study was carried out on patients who were undergoing work up for metastatic disease (88) or other medical problems (62). This is not representative of a normal population of individuals. The sample is highly biased. Twenty five patients (18%) of the referrals underwent no evaluation.

COMPARISON Randomized

Case-control

Non-random

Historical

None

Controls - details Allocation details

All patients underwent US examination of the thyroid gland. Fine-needle aspiration (FNA) biopsy was performed for all incidentally discovered thyroid nodules with abnormal sonographic features. Thyroidectomy was recommended for all patients with a malignant, indeterminate, or persistently nondiagnostic cytologic result.

Comparability

Patients were analyzed as two groups: those whose nodules were discovered on imaging for metastatic disease (88) and those who were studied as part of other medical problems (62). There is no mention of the basic features of either group or of the comparability.

Disparity

The disparity between the two groups and between the whole sample and a general population, though not stated, is obvious and large.

Comparison bias: The study is deeply flawed in terms of the comparison. Both groups are non-representative. There is no baseline, disease-free, presumably normal population to serve as a control. All comparisons are against historical controls.

MEASUREMENT Measurement error Blinding

Scoring

N

Protocols

?

Training

Y

Observer error Gold std.

Device error Device suited to task

Repetition

Device used

1.Ultrasound assessment of nodule

Y

N

Y

N

N

N

N

2.FNAC of suspicious nodules

Y

N

Y

N

Y

N

N

3.HPE of thyroidectomy tissue

Y

-

Y

N

N

-

-

Measurement bias: Histopathological examination of the excised thyroid is the only gold standard for confirmation of the presence of malignancy in the nodule. Only 65 of the 150 patients had such confirmation. Admittedly, it is impossible to submit all patients for thyroidectomy and some sort of screening process is required. This makes the need for a good control group even more essential - something that this study lacks in entirety.

© Dr Arjun Rajagopalan