Sustained benefits for women with HER2-positive early breast cancer JORGE MADRID BIG GOCCHI PROTOCOLO HERA

Sustained benefits for women with HER2-positive early breast cancer JORGE MADRID BIG GOCCHI PROTOCOLO HERA

The fascinating history of Herceptin 1981 1985 1987 1990 1992 1998 2000 2005 2006 2008 2011 Murine HER2 / neu gene cloned Human HER2 gene cloned mumab 4D5 Phase I IND for rhumab HER2 Paclitaxel + H and H mono US approval Paclitaxel + H and H mono EU approval Adjuvant H approval 1st IA of HERA HERA 2-year follow-up HERA 4-year follow-up and 1-year H vs 2-year H IA Association of HER2 with poor clinical outcome HERA recruitment opens HERA final analysis 1-year H vs 2-year H HER2, human epidermal growth factor receptor 2; H, Herceptin; IA, interim analysis

HERA study design HER2-positive early breast cancer (IHC 3+ and / or FISH+) n=5102 Surgery + (neo)adjuvant chemotherapy + radiotherapy Observation Herceptin q3w x 1 year Herceptin q3w x 2 years Option to cross over to Herceptin (after IA 2005) IHC, immunohistochemistry; FISH, fluorescence in situ hybridisation

End points of the HERA trial Primary end point DFS 1-year Herceptin vs observation 2-year Herceptin vs observation Secondary end points OS, RFS, distant DFS, safety 1-year Herceptin vs observation 2-year Herceptin vs observation compare DFS, OS, RFS, distant DFS and safety 1-year Herceptin vs 2-year Herceptin DFS, disease-free survival; OS, overall survival; RFS, relapse-free survival

HERA 2008 interim analysis: 2-year vs 1-year Herceptin Statistical assumptions HR <0.80 DFS absolute reduction 4.9% 5-year DFS 1-year arm: 70% 5-year DFS 2-year arm: 74.9% p value <0.014 for early release of results 2008 interim analysis 500 events reached June 2008 October 2008 IDMC advises Executive Committee Interim analysis positive Data release at SABCS HR, hazard ratio; IDMC, Independent Data Monitoring Committee Database cleaning Trial continues as planned No data release Final analysis triggered by 725 events (~2011)

HERA: IDMC recommendations October 2008 Do not release information on the 2-year Herceptin arm Continue the 1-year Herceptin vs 2-year Herceptin comparison Release updated information on 1-year Herceptin vs observation No conclusions can be drawn regarding the efficacy of Herceptin therapy for 2 years vs 1 year

HERA study design HER2-positive early breast cancer (IHC 3+ and / or FISH+) n=5102 Surgery + (neo)adjuvant chemotherapy + radiotherapy Observation Herceptin q3w x 1 year Herceptin q3w x 2 years Option to cross over to Herceptin (after IA 2005)

HERA: DFS and OS over time 1 and 2 years follow-up Median follow-up (% follow-up time after selective crossover) DFS benefit No. of DFS events H 1 year vs observation Median follow-up (% follow-up time after selective crossover) OS benefit No. of deaths H 1 year vs observation 2005 1 1 year (0%) 127 vs 220 p<0.0001 2005 1 1 year (0%) 29 vs 37 p=0.26 2006 2 2 years (4.3%) 218 vs 321 p<0.0001 2006 2 2 years (4.1%) 59 vs 90 p=0.0115 0 Favours 1 Favours no 2 Herceptin HR Herceptin 0 Favours 1 Favours no 2 Herceptin HR Herceptin 1 Piccart-Gebhart et al 2005; 2 Smith et al 2007

DFS (ITT): 4-year median follow-up Patients (%) 100 80 60 1-year Herceptin Observation 6.4% 40 20 0 Events 458 369 4-year DFS 72.2 78.6 HR 0.76 95% CI 0.66, 0.87 0 6 12 18 24 30 36 42 48 Months from randomisation p value <0.0001 No. at risk 1698 1703 1564 1619 1440 1552 1363 1485 1297 1414 1240 1352 1180 1280 992 1020 712 854

OS (ITT): 4-year median follow-up Patients (%) 100 80 Observation 1-year Herceptin 1.6% 60 40 20 0 Events 213 182 4-year DFS 87.7 89.3 HR 0.85 95% CI 0.70, 1.04 0 6 12 18 24 30 36 42 48 Months from randomisation p value 0.1087 No. at risk 1698 1703 1642 1660 1601 1640 1556 1615 1519 1577 1471 1524 1398 1447 1175 1149 828 953

HERA: DFS and OS over time Median follow-up (% follow-up time after selective crossover) DFS benefit No. of DFS events H 1 year vs observation Median follow-up (% follow-up time after selective crossover) OS benefit No. of deaths H 1 year vs observation 2005 1 1 year (0%) 127 vs 220 p<0.0001 2005 1 1 year (0%) 29 vs 37 p=0.26 2006 2 2 years (4.3%) 218 vs 321 p<0.0001 2006 2 2 years (4.1%) 59 vs 90 p=0.0115 2008 4 years (33.8%) 369 vs 458 p<0.0001 2008 4 years (30.9%) 182 vs 213 p=0.1087 0 Favours 1 Favours no 2 Herceptin HR Herceptin 0 Favours 1 Favours no 2 Herceptin HR Herceptin 1 Piccart-Gebhart et al 2005; 2 Smith et al 2007

Specific question 1 Crossover to Herceptin of 52% of the patients originally allocated to observation disrupted the randomised comparison between 1-year Herceptin and observation Question: To what extent might crossover have biased the ITT analysis?

Flow chart of observation patients: by status on 16 May 2005 1698 patients originally randomised to observation 344 patients 1354 patients 16 May DFS event or lost to follow-up 2005 alive and disease-free 198 alive post DFS event 469 patients remained on observation 344 patients ineligible for crossover

Time to selective crossover by calendar date (n=885) Proportion 0.6 Switched to Herceptin No. at risk Observation 0.5 0.4 0.3 0.2 0.1 0.0 16 May 2005 22 Aug 2005 28 Nov 2005 6 Mar 2006 12 Jun 2006 18 Sep 2006 25 Dec 2006 1354 1193 596 209 116 71 30 Randomisation to 1st dose Diagnosis to 1st dose Follow-up from 1st dose Median time (range), months 22.8 (<1-52.7) 30.9 (9.1-58.3) 29.1 (0.8-34.5)

Baseline characteristics of observation patients alive and disease free on 16 May 2005 Compared to patients who did not selectively cross over to Herceptin, those who did were more likely to: be younger have received anthracyclines and anthracyclines plus taxanes be diagnosed with node-positive disease have hormone receptor-positive tumours

Specific question 2 885 of 1354 patients (65%) in the observation group who were alive and disease free on May 16 2005 crossed over and received Herceptin Questions: What was the course of disease in the subgroups of observation patients who did or did not cross over to active therapy? Is there any effect of the late introduction of Herceptin?

Landmark of 16 May 2005 The landmark analysis considers only patients who were alive and disease free on 16 May 2005

DFS (landmark analysis): Herceptin vs observation Patients alive and disease free (%) 100 80 60 40 20 Observation: Alive and disease free on 16 May 2005 Herceptin: Alive and disease free on 16 May 2005 No. at risk 0 0 6 12 18 24 30 36 42 48 Months from randomisation 1354 1353 1339 1316 1278 1239 1180 992 712 1481 1480 1473 1447 1399 1351 1280 1020 854

DFS (landmark analysis): observation (alive, no DFS event), selective crossover and no crossover Patients alive and disease free (%) 100 80 60 40 No. at risk 20 Observation: Alive and disease free on 16 May 2005 Selective crossover 0 No crossover 0 6 12 18 24 30 36 42 48 Months from randomisation 1354 1353 1339 1316 1278 1239 1180 992 712 885 885 884 878 870 851 822 690 480 469 468 455 438 408 388 358 302 232

OS (landmark analysis): Herceptin vs observation Patients alive and disease free (%) 100 80 60 40 No. at risk 20 0 0 1354 1481 Observation: Alive and disease free on 16 May 2005 Herceptin: Alive and disease free on 16 May 2005 6 12 18 24 30 36 42 48 Months from randomisation 1354 1481 1350 1481 1344 1474 1332 1461 1316 1438 1270 1378 1065 1094 759 910

OS (landmark analysis): crossover vs no-crossover Patients alive (%) 100 80 60 40 No. at risk 20 0 0 Observation: Alive and disease free on 16 May 2005 Selective crossover No crossover 6 12 18 24 30 36 42 48 Months from randomisation 1354 1354 1350 1344 1332 1316 1270 1065 759 885 885 885 883 881 875 852 718 499 469 469 465 461 451 441 418 347 260

Cardiac safety: safety analysis population a No. patients (%) Cardiac death Severe CHF (NYHA III and IV) Symptomatic CHF (II, III and IV) Confirmed significant LVEF drop Herceptin discontinued due to cardiac problems Observation a n=1719 1 (0.1) 0 (0.0) 3 (0.2) 13 (0.8) 1-year Herceptin n=1682 0 (0.0) 13 (0.8) 33 (2.0) 62 (3.7) 87 (5.2) a Patients who crossed over are censored from the date of starting Herceptin treatment CHF, congestive heart failure; NYHA, New York Heart Association; LVEF, left ventricular ejection fraction

Cardiac safety: observation group Cardiac death Severe CHF (NYHA III and IV) Symptomatic CHF (II, III and IV) Confirmed significant LVEF drop Herceptin discontinued due to cardiac problems No crossover after 16 May 05 n=469 0 (0.0) 0 (0.0) 1 (0.2) 5 a (1.1) Crossover n=885 0 (0.0) 0 (0.0) 9 (1.0) 26 (2.9) 43 (4.9) a For 3 of the patients, the LVEF drop occurred between 16 May 05 and the date of the patient decision and may have influenced the patient decision

HERA 4-year follow-up data: summary (1) The updated analysis at 4 years was limited to 1-year Herceptin vs observation as recommended by IDMC Extensive selective crossover of observation patients to active therapy biased the ITT comparison Landmark analysis of observation patients who were disease free on 16 May 2005 explored the effects of later introduction of Herceptin Lack of randomisation limits the interpretation of the landmark analysis different outcome due to drug effect or patient characteristics?

HERA 4-year follow-up data: summary (2) In HERA, the DFS benefit associated with Herceptin is maintained at 4-year median follow-up 50% of patients in the observation arm crossed over to Herceptin treatment, therefore the OS benefit is no longer statistically significant Patients crossing over at a later date appear to benefit from 1 year of Herceptin

HERA: conclusions and next steps 4-year follow-up data support the hypothesis that the risk of relapse in HER2-positive early breast cancer persists over time Prolonged exposure to the Herceptin antibody may improve efficacy This is being tested in the comparison of the 1-year and 2-year groups in the HERA study

Additional studies demonstrate consistent DFS benefit for Herceptin HERA CTx H 1 year B-31 / N9831 AC PH BCIRG 006 AC TH TCarboH DFS benefit Median follow-up, years 4 3 3 3 NOAH CTx / H H 1 year 3 FinHer a VH / TH CEF b PACS-04 a CTx H 1 year n=231 n=528 5 4 0 Favours 1 Favours no 2 Herceptin HR Herceptin a Based on small subgroups of patients with HER2-positive breast cancer; Gianni et al 2008; b DDFS; CTx, chemotherapy; AC, doxorubicin, cyclophosphamide; Gianni et al 2009; Joensuu et al 2009; P, paclitaxel; T, docetaxel; Carbo, carboplatin; V, vinorelbine; Slamon et al 2006; Perez et al 2007; CEF, cyclophosphamide, epirubicin, 5-fluorouracil Smith et al 2007; Spielmann et al 2007

1-year Herceptin treatment consistently reduces the risk of death by one-third OS benefit Median follow-up, years HERA CTx H 1 year 4 B-31 / N9831 AC PH 3 BCIRG 006 AC TH 3 TCarboH 3 FinHer VH / TH CEF n=231 5 0 Favours 1 Favours no 2 Herceptin HR Herceptin Gianni et al 2009; Joensuu et al 2009; Slamon et al 2006; Perez et al 2007; Smith et al 2007

HER2-positive breast cancer: outstanding questions Concurrent or sequential Herceptin therapy? Herceptin efficacy in lower-risk patients? Optimal treatment duration? Translational research? New combinations? ALTTO (Herceptin + lapatinib) BETH (Herceptin + Avastin)

ALTTO: Phase III randomised open-label trial comparing adjuvant lapatinib +/ Herceptin HER2-positive early breast cancer (n=8000) Surgery and completion of (neo)adjuvant anthracycline-based chemotherapy No taxane Concurrent taxanes e for 12 weeks H a q3w for 52 weeks L b qd for 52 weeks H c qw for 12 weeks L d qd + H c q3w for 52 weeks H a q3w for 52 weeks L b qd for 52 weeks H c qw for 12 weeks L d qd + H c q3w for 52 weeks 6-week washout 6-week washout L b qd for 34 weeks L b qd for 34 weeks a Herceptin 8 mg/kg iv loading dose followed by 6 mg/kg q3w; b Lapatinib 1500 mg; c Herceptin 4 mg/kg iv loading dose followed by 2 mg/kg qw; d Lapatinib 1000 mg; e Paclitaxel 80 mg/m 2 qw or docetaxel q3w

BETH: Phase III randomised trial comparing Herceptin-containing adjuvant regimens +/ Avastin Resected node-positive or high-risk node-negative HER2-positive early breast cancer n~3600 (maximum) a CTx b + Herceptin CTx b + Herceptin + Avastin Herceptin continued q3w until 1 year total Herceptin + Avastin continued q3w until 1 year total Primary endpoint: DFS Secondary endpoints: OS; RFS; distant recurrence-free interval; safety; biomarker analysis a Actual recruitment to date ~300; b Docetaxel 75 mg/m 2 q3w + carboplatin AUC 6 q3w or docetaxel 100 mg/m 2 FEC; CTx, chemotherapy; q3w, three weekly

Conclusions 1 year of Herceptin remains the most appropriate and evidence-based approach for patients with HER2-positive early breast cancer Studies are ongoing to address optimal treatment duration Development of new anti-her2 treatment regimens will lead to greater patient benefits