Kol Slide Kit Atac 100

ATAC Trial 100-month median follow-up data

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Introduction ● The primary aim of treatment in EBC is preventing recurrence ● The benefit of tamoxifen on tumour recurrence and mortality is well-established ● Nevertheless, recurrence rates are >2% per year long term, and >30% of women develop recurrent disease within 15 years1 ● Clinical trials have confirmed that AIs reduce recurrence rates further and offer tolerability advantages compared with tamoxifen during adjuvant treatment2–4 ● For AIs, several questions remain unanswered – including the extent to which treatment benefits continue after completion of therapy 1

AIs, aromatase inhibitors

EBCTCG overview. Lancet 2005; 365: 1687-1717 2 ATAC Trialists’ Group. Lancet 2005; 365: 60-62 3 Coates et al. J Clin Oncol 2007; 25: 486-92 4 Coombes et al. Lancet 2007; 369: 559-70

ATAC trial design Postmenopausal women with invasive breast cancer (n = 9366) Surgery ± radiotherapy ± chemotherapy Randomisation 1:1:1 for 5 years Anastrozole (n = 3125)

Tamoxifen (n = 3116)

ITT population n = 3125 Safety population n = 3092 HR+ subpopulation n = 2618

ITT population n = 3116 Safety population n = 3094 HR+ subpopulation n = 2598

ITT, intent-to-treat; HR+, hormone receptor-positive

Discontinued following Combination initialn=3125 analysis as no efficacy or tolerability benefit compared with tamoxifen arm

‘Arimidex’, Tamoxifen, Alone or in Combination (ATAC) trial ● The ATAC trial dataset is robust and mature1–3 ● Anastrozole significantly prolongs disease-free survival and time to recurrence compared with tamoxifen1–3 ● Anastrozole is better tolerated than tamoxifen2 ● 68-month median follow-up suggests that the efficacy benefits observed extend for at least one year beyond treatment completion3

The ATAC Trialists’ Group. Lancet 2002; 359: 2131-39 The ATAC Trialists’ Group. Lancet Oncol 2006; 7: 633-43 3 ATAC Trialists’ Group. Lancet 2005; 365: 60-62 1

2

Efficacy analyses ● Efficacy analyses were assessed in both the ITT population and the more clinically relevant HR+ subgroup (84% of all patients) – Kaplan-Meier prevalence curves are shown for the HR+ population only

● Primary endpoint: – Disease-free survival

● Secondary endpoints: – Time to recurrence – Time to distant recurrence – Contralateral breast cancer – Death: all causes – Death after recurrence – Death without recurrence

Safety analyses ● All adverse events occurring on treatment were recorded ● After treatment*, all fractures and serious adverse events continued to be recorded blindly up to the time of recurrence or death

*14 days after treatment termination

ATAC additional follow-up (all monotherapy patients) March 2007 (n = 6241)

March 2004 (n = 6241)

Increment n (%)

Median follow-up (months to death)

100

68

32 (47)

Women years’ follow-up (to death)

46202

33465

12737 (38)

Total events

1704

1226

478 (39)

Time to recurrence

1183

898

285 (32)

Time to distant recurrence

911

699

212 (30)

Contralateral breast cancer

148

94

54 (57)

Death − all causes

1253

831

422 (51)

Death after recurrence

732

503

229 (46)

Death without recurrence

521

328

193 (59)

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53 AstraZeneca, data on file

Summary of endpoints (A vs T) (all patients) Endpoint

No. events

Hazard ratio

95% CI

p-value

Disease-free survival

817 v 887

0.90

(0.82-0.99)

0.025

Time to recurrence

538 v 645

0.81

(0.73-0.91) 0.0004

Time to distant recurrence

424 v 487

0.86

(0.75-0.98)

0.022

Contralateral breast cancer

61 v 87

0.68

(0.49-0.94)

0.02

Death − all causes

629 v 624

1.00

(0.89-1.12)

0.99

Death after recurrence

350 v 382

0.91

(0.79-1.05)

0.2

Death without recurrence

279 v 242

1.12

(0.94-1.33)

0.2

A, anastrozole; T, tamoxifen; CI, confidence interval The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Summary of endpoints (A vs T) (HR+ patients) Endpoint

No. events

Hazard ratio

95% CI

p-value

Disease-free survival

618 v 702

0.85

(0.76-0.94)

0.003

Time to recurrence

391 v 494

0.76

(0.67-0.87) 0.0001

Time to distant recurrence

305 v 357

0.84

(0.72-0.97)

0.022

Contralateral breast cancer

50 v 80

0.60

(0.42-0.85)

0.004

Death − all causes

472 v 477

0.97

(0.86-1.11)

0.7

Death after recurrence

245 v 269

0.90

(0.75-1.07)

0.2

Death without recurrence

227 v 208

1.05

(0.87-1.26)

0.6

A, anastrozole; T, tamoxifen; HR+, hormone receptor-positive; CI, confidence interval The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Efficacy endpoints for all patients and HR+ patients All patients (ITT) HR+ patients

Favours anastrozole (A)

Favours tamoxifen (T)

Hazard ratio Hazard ratio All HR+ patients patients

Disease-free survival

0.90

0.85

Time to recurrence

0.81

0.76

Time to distant recurrence

0.86

0.84

Contralateral breast cancer

0.68

0.60

Death − all causes

1.00

0.97

Death after recurrence

0.91

0.90

Death without recurrence

1.12

1.05

0.2

0.4

0.6

0.8 1.0 1.2 1.5

2.0

Hazard ratio (A / T) and 95% CI The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Disease-free survival HR+ patients Patients 30 (%) 25

HR

95% CI

HR+ 0.85 (0.76, 0.94)

20

Tamoxifen (T) Anastrozole (A)

29.9%

p-value 0.003

30 25

25.8% 16.4%

15

20 15

10

10

13.9%

5

5 Absolute difference 2.5%

0 0

1

2

3 4 5 6 Follow-up time (years)

4.1%

7

8

At risk: A 2618 2541 2453 2361 2278 2159 1995 1801 1492 T 2598 2516 2400 2306 2196 2075 1896 1711 1396 HR, hazard ratio; CI, confidence interval

0

9 608 547

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Time to recurrence HR+ patients Patients 30 (%) 25

HR

95% CI

30

p-value

HR+ 0.76 (0.67, 0.87) 0.0001

20

Tamoxifen (T) Anastrozole (A)

15

21.8%

25 20

12.5%

17.0%

10

15 10

9.7%

5

5 4.8%

0 0

1

2

3 4 5 6 Follow-up time (years)

7

8

At risk: A 2618 2541 2453 2361 2278 2159 1995 1801 1492 T 2598 2516 2400 2306 2196 2075 1896 1711 1396

0

9 608 547

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Time to recurrence HR+ patients Patients 30 (%) 25

HR

95% CI

30

p-value

HR+ 0.76 (0.67, 0.87) 0.0001

20

Tamoxifen (T) Anastrozole (A)

15

21.8%

25 20

12.5%

17.0%

10

15 10

9.7%

5

5

Absolute difference 2.8%

0 0

1

2

3 4 5 6 Follow-up time (years)

4.8%

7

8

At risk: A 2618 2541 2453 2361 2278 2159 1995 1801 1492 T 2598 2516 2400 2306 2196 2075 1896 1711 1396

0

9 608 547

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

How does the absolute benefit compare with other trials? ● 5 years tamoxifen vs 2 years tamoxifen – an absolute reduction of 2.7% (11.6% on 5 years tamoxifen vs 14.3% on 2 years)2

● Anthracycline-based vs CMF chemotherapy – For recurrence, breast cancer mortality and overall mortality, the absolute difference is about 3% at 5 years and 4% at 10 years3

● Absolute difference in recurrence in the HR+ population in ATAC at 9 years is 4.8% in favour of anastrozole1

ATAC Trialists’ Group, Lancet 2005; 365: 60-62 EBCTCG overview, Lancet 1998; 351: 1451-1467 3 EBCTCG overview: Lancet 2005; 365: 1687-1717 1

2

Time to recurrence: smoothed hazard estimates HR+ patients

Annual 4.0 hazard rates 3.0 (%)

4.0 3.0

2.0

2.0

1.0

1.0

Tamoxifen (T) Anastrozole (A)

0.0

0.0 0

1

2

3 4 5 6 Follow-up time (years)

7

8

9

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

TTR: Carryover effect in post-treatment period ● Recurrence rates continued to be lower with anastrozole after treatment completion ● In the HR+ subgroup, the absolute difference in recurrence increased from 2.8% at 5 years to 4.8% at 9 years ● Tamoxifen has a known carryover effect in years 5-9, of about two-thirds the size of that during active treatment1 ● There is a statistically significantly larger carryover effect for anastrozole (HR=0.75, 95% CI 0.61-0.94, p=0.01)

TTR, time to recurrence

1

EBCTCG overview. Lancet 2005; 365: 1687-1717

Time to distant recurrence HR+ patients Patients 30 (%) 25

HR

95% CI

HR+ 0.84 (0.72, 0.97)

20

p-value

30

0.022

25 20

Tamoxifen (T) Anastrozole (A)

15

15.6% 9.1%

10

13.2%

7.8%

5 0

1

2

3 4 5 6 Follow-up time (years)

10 5

Absolute difference 1.3%

0

15

2.4%

7

8

At risk: A 2618 2551 2470 2393 2320 2201 2042 1854 1536 T 2598 2533 2440 2363 2263 2151 1982 1809 1484

0

9 636 591

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Time to distant recurrence: smoothed hazard estimates HR+ patients

Annual 3.0 hazard rates (%) 2.0

3.0

2.0

1.0

1.0 Tamoxifen (T) Anastrozole (A)

0.0

0.0 0

1

2

3 4 5 6 Follow-up time (years)

7

8

9

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Time to distant recurrence ● At this 100-month median follow-up, a 30% increase in the number of distant recurrences was seen overall since the last analysis at 68 months (911 vs 699 events) ● Smoothed hazard plots clearly demonstrate that lower recurrence rates are maintained with anastrozole, even after treatment has been completed ● Distant recurrence rates continued to diverge with increasing follow-up time, being 1.3% lower for anastrozole at Year 5 and 2.4% lower at Year 9

Contralateral breast cancer HR+ patients Patients (%)

5

HR

95% CI

HR+ 0.60 (0.42, 0.85)

4

5

p-value 0.004

4.2%

Tamoxifen (T) Anastrozole (A)

3

4 3

2.5% 1.8%

2

2

1.0%

1

1

AD 0.8%

0 0

1

2

3 4 5 6 Follow-up time (years)

1.7%

7

8

At risk: A 2618 2541 2453 2361 2278 2159 1995 1801 1493 T 2598 2516 2400 2306 2196 2075 1896 1711 1396 AD, absolute difference

0

9 608 547

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Death: all causes HR+ patients Patients 30 (%) 25

HR

95% CI

HR+ 0.97 (0.86, 1.11)

20

p-value

30

0.70

25 20

Tamoxifen (T) Anastrozole (A)

15

15

10

10

5

5

0

0 0

1

2

3 4 5 6 Follow-up time (years)

7

8

At risk: A 2618 2567 2511 2445 2389 2274 2102 1911 1586 T 2598 2549 2504 2432 2339 2227 2068 1888 1551

9 659 620

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53 AstraZeneca, data on file

Overall survival ● There was no difference in overall survival between groups in either the ITT or HR+ populations ● Mean age for this analysis is 72 years so there is increased risk of serious co-morbidities ● Deaths from other causes formed a major component of overall survival – In the HR+ population, 48% and 44% of total deaths in the anastrozole and tamoxifen arms, respectively, were non-breast cancer deaths – As treatment only impacts breast-specific deaths, the true treatment effect is diluted by the high proportion of deaths from other causes

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Competing causes of mortality ● A recent study investigated non-breast cancer deaths in women aged ≥50 with node-negative, oestrogen receptor-positive primary breast cancer treated with adjuvant tamoxifen ● For at least 10 years after diagnosis, the large majority of deaths were unrelated to breast cancer

Hanrahan et al. J Clin Oncol 2007; 25: 4952-60

Competing causes of mortality Age ≥50 years and ER+ Probability 0.4

Breast-cancer deaths Other deaths

0.3 0.2 0.1 0.0 0

ER+, oestrogen receptor-positive

5 10 Time from diagnosis (years)

15

Hanrahan et al. J Clin Oncol 2007; 25: 4952-60

Death after recurrence HR+ patients Patients 30 (%) 25

HR

95% CI

HR+ 0.90 (0.75, 1.07)

20

p-value

30

0.20

25 20

Tamoxifen (T) Anastrozole (A)

15

15

10

10

5

5

0

0 0

1

2

3 4 5 6 Follow-up time (years)

7

8

At risk: A 2618 2567 2511 2445 2389 2274 2102 1911 1586 T 2598 2549 2504 2432 2339 2227 2068 1888 1551

9 659 620

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53 AstraZeneca, data on file

Death after recurrence ● The endpoint of death after recurrence* also includes those patients who recurred and then died of other causes ● Therefore there is some dilution in assessing the true mortality from breast cancer ● Anastrozole reduces the risk of death following recurrence by 10% – Although currently not statistically significant (p=0.2), there is a 90% chance that the time to death following recurrence is longer with anastrozole

*Previously labelled ‘Breast cancer deaths’ in previous ATAC reports and elsewhere

Death without recurrence HR+ patients Patients 30 (%) 25

HR

95% CI

HR+ 1.05 (0.87, 1.26)

20

p-value

30

0.60

25 20

Tamoxifen (T) Anastrozole (A)

15

15

10

10

5

5

0

0 0

1

2

3 4 5 6 Follow-up time (years)

7

8

At risk: A 2618 2567 2511 2445 2389 2274 2102 1911 1586 T 2598 2549 2504 2432 2339 2227 2068 1888 1551

9 659 620

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53 AstraZeneca, data on file

Analysis of time to recurrence (A vs T) according to baseline and treatment characteristics in the HR+ population Favours Anastrozole

Nodal status Positive Negative Unknown Tumour size <2cm ≥2cm Receptor status ER+ / PgR+ ER+ / PgRPrevious chemotherapy Yes No Age <65 years ≥65 years All patients 0.2

Favours Tamoxifen

Hazard ratio (95% CI) 0.84 (0.70-1.00) 0.68 (0.55-0.84) 0.48 (0.23-1.00) 0.81 (0.66-0.98) 0.74 (0.62-0.88) 0.87 (0.74-1.02) 0.42 (0.31-0.58) 0.89 (0.70-1.13) 0.71 (0.61-0.83) 0.76 (0.63-0.91) 0.77 (0.63-0.93) 0.76 (0.67-0.87)

0.4

0.6

0.8 1.0 1.2 1.5

2.0

Hazard ratio (A / T) and 95% CI A, anastrozole; T, tamoxifen

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Subgroup analysis of time to recurrence ● The benefit of anastrozole for time to recurrence was seen in all subgroups regardless of baseline and treatment characteristics ● There was no significant heterogeneity across treatment subgroups except for: – the small subgroup ER+/PgR- (19% of ER+) for which the benefit in favour of anastrozole was larger than for the ER+/PgR+ subgroup (p<0.001 for heterogeneity between these subgroups) • This difference was not observed in BIG1-98 or for the subset of patients in ATAC from whom tissue was collected fro translational research1,2

Viale et al. J Clin Oncol 2007; 25: 3846-52 Dowsett & Allred. Breast Cancer Res Treat 2007; 100 (Suppl 1)

Safety and tolerability

Predefined adverse events at any time on treatment or any severity

Hot flushes Nausea and vomiting Fatigue / tiredness (asthenia) Mood disturbances Musculo-skeletal disorders Vaginal bleeding Vaginal discharge Ischaemic cardiovascular disease Ischaemic cerebrovascular event Venous thromboembolic events Deep venous thromboembolic events Cataracts Carpal tunnel syndrome† †

Anastrozole (N = 3092)

Tamoxifen (N = 3094)

1102 (35.6) 394 (12.7) 578 (18.7) 599 (19.4) 1104 (35.7) 167 (5.4) 110 (3.6) 130 (4.2) 64 (2.1) 87 (2.8) 48 (1.6) 189 (6.1) 79 (2.5)

1263 (40.8) 358 (12.4) 544 (17.6) 555 (17.9) 915 (29.6) 319 (10.3) 409 (13.2) 106 (3.4) 91 (2.9) 141 (4.6) 75 (2.4) 218 (7.0) 22 (0.7)

included as a non-predefined adverse event of interest ATAC Trialists’ Group. Lancet Oncology 2008; 9:45-53, Webextra table

Serious adverse events: on and off treatment (number and annual rate – safety population) On treatment

Off treatment

Anastrozole

Tamoxifen

Anastrozole

Tamoxifen

12781

12331

9351

9448

1125 (9.12)

356 (3.81)

339 (3.59)

153 (1.20)

284 (2.30)

49 (0.52)

57 (0.60)

Endometrial cancer

4 (0.03)

12 (0.10)

1 (0.01)

12 (0.13)

Myocardial infarction

34 (0.27)

33 (0.27)

26 (0.28)

28 (0.30)

Cerebrovascular accident

20 (0.16)

34 (0.28)

22 (0.24)

20 (0.21)

Fracture episodes*

375 (2.93)

234 (1.90)

146 (1.56)

143 (1.51)

Women years’ follow-up

All serious adverse events 1054 (8.25) Treatment-related serious adverse events†

as judged by the investigator *one or more fractures on the same day based on adverse events and serious adverse-event reports The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53 †

Treatment-related serious adverse events ● The incidence of treatment-related serious adverse events remained lower with anastrozole for the entire follow-up period – statistically significantly lower during treatment and similar to tamoxifen after treatment completion

● In particular, the incidence of endometrial cancer with anastrozole was much lower both on- and off-treatment

New primary cancers at non-breast cancer sites prior to recurrence No. patients (%) New primary cancers

Anastrozole (n = 3092)

Tamoxifen (n = 3094)

TOTAL

292 (9.4)

288 (9.3)

Head and neck

12 (0.4)

5 (0.2)

Upper gastrointestinal

8 (0.3)

6 (0.2)

Colorectal

56 (1.8)

36 (1.2)

Lung

42 (1.4)

24 (0.8)

Skin (non-melanoma)

96 (3.1)

107 (3.5)

Melanoma

8 (0.3)

18 (0.6)

Ovary

12 (0.4)

26 (0.8)

Endometrium*

5 (0.2)

24 (0.8)

Cervix

2 (0.1)

6 (0.2)

Kidney / bladder

17 (0.5)

15 (0.5)

Leukaemia / lymphoma / myeloma

22 (0.7)

19 (0.6)

Other

37 (1.2)

32 (1.0)

*includes uterine cancers not specified as cervix The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

New primary cancers ● The only difference in new primary-cancer rates that was statistically significant* was the lower number of endometrial cancers seen with anastrozole ● As tamoxifen is known to increase the incidence of endometrial cancer,1 this difference is not surprising, and could be a result of either protection from lowered oestrogen levels or the increase associated with tamoxifen, or both ● Numerical differences were observed with lower ovarian cancer and melanoma rates with anastrozole, and lower colorectal and lung cancer rates with tamoxifen, none statistically significant* ● Further investigation is needed to assess if these are just due to random variations or not Fisher et al. J Natl Cancer Inst 1994; 86: 527-37

1

*after a Bonferroni correction for multiple comparisons

Deaths according to treatment group (ITT population)

No. patients (%)

Cause of death

Anastrozole (n = 3125)

Tamoxifen (n = 3116)

Total deaths

629 (20)

624 (20)

Deaths after recurrence

350 (11)

382 (12)

Deaths without recurrence

279 (9)

242 (8)

Cardiovascular

67 (2)

66 (2)

Cerebrovascular

25 (1)

29 (1)

Second primary non-breast cancer

84 (3)

60 (2)

Other

103 (3)

87 (3)

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Cause of death by recurrence status ● There were 732 deaths after recurrence compared with 911 distant recurrences (and 1183 recurrences at any site) ● Further follow-up is needed to assess breastcancer mortality rate when more events are accrued ● As all types of recurrence have important implications for long-term survival, future analysis is awaited with interest

Fractures (occurring at any time before recurrence) Fractures before recurrence

A T N=3092 N=3094 (%) (%) Odds ratio

A vs T 95% CI

p-value

Patients with one or more fracture episodes1

421 (13.6)

311 (10.1)

1.41

1.21-1.65 <0.0001

Hip1

49 (1.6)

42 (1.4)

1.17

0.75-1.82

0.46

Spine2

60 (1.9)

37 (1.2)

1.64

1.08-2.48

0.02

Wrist / colles2

94 (3.0)

83 (2.7)

1.14

0.84-1.54

0.4

All other sites2

270 (8.7)

191 (6.2)

1.46

1.20-1.77

0.0001

A, anastrozole; T, tamoxifen 1

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53 2 AstraZeneca, data on file

Fracture episode rates throughout the study Annual fracture episode rates (%)

4

Tamoxifen (T) Anastrozole (A)

3

2

1

0 0

1

At risk: A 2984 T 2976

2 2859 2824

3 4 5 6 7 Time since randomisation (years) 2745 2699

2640 2572

2496 2419

2306 2208

2077 2000

8

9

1713 1645

702 659

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Fracture episode rates ● The increase in fracture episode rate with anastrozole appears to only be associated with the active treatment period and does not continue after treatment completion ● In the post-treatment follow‑up period, the fracture episode rate for anastrozole and tamoxifen was similar (IRR=1.03, non-significant)

IRR, incidence rate ratio

Conclusions ● At 100-month median follow-up, data show anastrozole is: –

significantly superior to tamoxifen at preventing recurrence beyond the completion of treatment

– significantly superior to tamoxifen at preventing both distant recurrence and contralateral breast cancer

● The absolute difference in recurrence rates continues to increase after treatment completion: – HR+ population: 2.8% at 5 years increased to 4.8% after 9 years – HR = 0.75; p=0.01

● Anastrozole is well-tolerated long term: – No excess fracture rate seen after treatment completion – Risk of fracture on treatment needs to be taken into consideration and can be managed according to established guidelines – No new morbidity or mortality increases seen after treatment completion

Conclusions: benefits continue and detrimental effects decline after treatment cessation Time to recurrence Patients 30 (%)

Fracture episode rates

Tamoxifen (T) Anastrozole (A)

25

21.8%

Annual 4 fracture episode rates (%) 3

Tamoxifen (T) Anastrozole (A)

20 15

12.5%

2

17.0%

10 1

9.7% 5 Absolute 2.8% difference

0 0

1

2

3

4

4.8% 0 5

6

Follow-up time (years)

7

8

9

0

1

2

3

4

5

6

7

8

9

Time since randomisation (years)

The ATAC Trialists’ Group. Lancet Oncol 2008; 9:45-53

Conclusions ● Statistically significant evidence for a larger carryover effect with anastrozole compared with tamoxifen – Starting treatment with anastrozole has a significant impact on outcomes for at least 9 years after diagnosis

● Anastrozole as initial adjuvant therapy should be considered standard of care for postmenopausal women with hormone-sensitive early breast cancer