Articles

2-year follow-up of trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer: a randomised controlled trial
Ian Smith, Marion Procter, Richard D Gelber, Sbastien Guillaume, Andrea Feyereislova, Mitch Dowsett, Aron Goldhirsch, Michael Untch, Gabriella Mariani, Jose Baselga, Manfred Kaufmann, David Cameron, Richard Bell, Jonas Bergh, Robert Coleman, Andrew Wardley, Nadia Harbeck, Roberto I Lopez, Peter Mallmann, Karen Gelmon, Nicholas Wilcken, Erik Wist, Pedro Snchez Rovira, Martine J Piccart-Gebhart, for the HERA study team

Lancet 2007; 369: 2936 See Comment page 3 Royal Marsden Hospital, London, UK (Prof I Smith MD, Prof M Dowsett PhD); Institute of Cancer Research, London (Prof I Smith, M Dowsett); Frontier Science (Scotland), Kingussie, UK (M Procter MSci); Department of Biostatistics and Computational Biology, DanaFarber Cancer Institute, Boston, MA, USA (Prof R D Gelber PhD); Breast Data Centre (S Guillaume MSc), and Department of Medicine (Prof M J Piccart-Gebhart MD), Jules Bordet Institute, Universit Libre de Bruxelles, Brussels, Belgium; Roche, Basel, Switzerland (A Feyereislova MD); Department of Medicine, European Institute of Oncology, Milan, Italy (Prof A Goldhirsch MD); Oncology Institute of Southern Switzerland, Bellinzona, Switzerland (A Goldhirsch); Klinik fr Frauenheilkunde und Geburtshilfe, LudwigMaximilians-Universitt Mnchen, Grosshadern, Munich, Germany (Prof M Untch MD); Istituto Nazionale per lo Studio e la Cura dei Tumori, Milan, Italy (G Mariani MD); Medical Oncology Service, Vall dHebron University Hospital, Barcelona, Spain (Prof J Baselga MD); Department of Obstetrics and Gynaecology, Johann Wolfgang Goethe-Univeritat, Franfurt, Germany (Prof M Kaufmann MD); Department of Oncology, Western General Hospital, Edinburgh, UK (D Cameron MD); Andrew Love Cancer Centre, Geelong Hospital, Geelong, Australia (Prof R Bell MD); Department of Oncology, Karolinska Institute and University Hospital, Stockholm, Sweden (Prof J Bergh MD); Cancer Research Centre, Academic Unit of Clinical Oncology, Weston Park Hospital, Sheeld, UK (Prof R Coleman MD); Medical Oncology, Christie Hospital, Manchester, UK (A Wardley MD); Frauenklinik der Technischen Universitat Munchen, Klinikum rechts der Isar, Munich, Germany (Prof N Harbeck MD); Instituto Oncologico Nacional Panama, Panama City, Panama (R I Lopez MD); Department of Obstetrics and Gynaecology, University Hospital of Cologne, Cologne, Germany (Prof P Mallmann MD);

Summary
Background Trastuzumaba humanised monoclonal antibody against HER2has been shown to improve diseasefree survival after chemotherapy in women with HER2-positive early breast cancer. We investigated the drugs eect on overall survival after a median follow-up of 2 years in the Herceptin Adjuvant (HERA) study. Methods HERA is an international multicentre randomised trial that compared 1 or 2 years of trastuzumab treatment with observation alone after standard neoadjuvant or adjuvant chemotherapy in women with HER2-positive node positive or high-risk node negative breast cancer. 5102 women participated in the trial; we analysed data from 1703 women who had been randomised for treatment with trastuzumab for 1 year and 1698 women from the control group, with median follow-up of 235 months (range 048 months). The primary endpoint of the trial was diseasefree survival. Here, we assess overall survival, a secondary endpoint. Analyses were done on an intent-to-treat basis. This trial is registered with the European Clinical Trials Database, number 200500238511. Findings 97 (57%) patients randomised to observation alone and 58 (34%) patients randomised to 1 year of treatment with trastuzumab were lost to follow-up. 172 women stopped trastuzumab prematurely. 59 deaths were reported for trastuzumab and 90 in the control group. The unadjusted hazard ratio (HR) for the risk of death with trastuzumab compared with observation alone was 066 (95% CI 047091; p=00115). 218 disease-free survival events were reported with trastuzumab compared with 321 in the control group. The unadjusted HR for the risk of an event with trastuzumab compared with observation alone was 064 (054076; p<00001). Interpretation Our results show that 1 year of treatment with trastuzumab after adjuvant chemotherapy has a signicant overall survival benet after a median follow-up of 2 years. The emergence of this benet after only 2 years reinforces the importance of trastuzumab in the treatment of women with HER2-positive early breast cancer.

Introduction
Trastuzumab (Herceptin; Roche, Basel, Switzerland) is a humanised monoclonal antibody that is targeted against the extracellular domain of the HER2 transmembrane growth factor receptor.1 Amplication of the HER2 gene and overexpression of the receptor occurs in around 1525% of women with early breast cancer, and is associated with an aggressive disease course.2,3 Trastuzumab has been shown to be of overall survival benet to women with HER2-positive metastatic breast cancer administered alone4,5 or in combination with chemotherapy.6,7 The Herceptin Adjuvant (HERA) trial (Breast International Group 0101) is one of several large trials designed to test the ecacy of trastuzumab in the adjuvant (ie, postsurgery) treatment of women with HER2-positive early breast cancer. Results of a rst planned interim analysis with a median 1-year follow-up showed that trastuzumab given every 3 weeks for 1 year after adjuvant or neoadjuvant (ie, presurgery) chemotherapy achieved a signicant improvement in disease-free survival compared with women treated with adjuvant chemotherapy alone, with a hazard ratio (HR) of 054.8
www.thelancet.com Vol 369 January 6, 2007

The combined analysis of two similar North American trials (North Central Cancer Treatment Group Trial N9831 and National Surgical Adjuvant Breast and Bowel Project B-31) has also shown a signicant improvement in disease-free survival for trastuzumab given concurrently with four courses of paclitaxel either every week or every 3 weeks after a combination of doxorubicin and cyclophosphamide and continued for 1 year compared with the same chemotherapy schedule alone.9 A fourth adjuvant trastuzumab trial, known as BCIRG 006, has also shown much the same disease-free survival benet when trastuzumab is given either with docetaxel after doxorubicin and cyclophosphamide or with docetaxel and carboplatin.10 Finally, a fth, much smaller, trial has also shown an improvement in disease-free survival after only 9 weeks of trastuzumab given at the start of treatment concurrently with adjuvant chemotherapy.11 The magnitude of the benet, with a reduction in the early risk of recurrence of around 50% in all these trials, has led to the widespread use of trastuzumab as adjuvant therapy. However, this use has been criticised by some

29

Articles

British Columbia Cancer Agency, Vancouver, Canada (Prof K Gelmon MD); Department of Medical Oncology and Palliative Care, Westmead Hospital, Westmead, Australia (N Wilcken PhD); Department of Oncology, Ullevl University Hospital, Oslo, Norway (Prof E Wist PhD); and S Oncologa Mdica, Complejo Hospitalario de Jan, Jan, Spain (P Snchez Rovira MD) Correspondence to: Prof Ian Smith, Breast Unit, Royal Marsden Hospital, Fulham Road, London SW3 6JJ, UK [email protected]

because the ndings arise from interim analyses, followup is short, and signicant overall survival benet has not been shown in any stand alone trial.12 Our aim is to report an analysis of overall survival, together with an updated assessment of disease-free survival, in the HERA trial with a median 2-year follow-up.

Methods
Patients
The study design, eligibility criteria, treatment schedules, monitoring, and statistical analysis plan have been described in detail elsewhere.8 Briey, the HERA trial is an international intergroup open-label phase III randomised trial involving women with centrally conrmed HER2-positive (immunohistochemistry score 3 or uorescence in-situ hybridisation positive) early stage invasive breast cancer who had completed local regional therapy and a minimum of four courses of predened standard adjuvant or neoadjuvant chemo-

therapy. Eligibility criteria included node-positive disease or node-negative disease if the pathological tumour size was larger than 1 cm. Women with locally advanced disease including inammatory breast cancers were excluded. Women with a left ventricular ejection fraction (LVEF) of less than 55% after completion of chemotherapy and radiotherapy, congestive cardiac failure, or other major cardiac problems8 were excluded. The ethics review boards at all the participating institutions approved the study protocol and all patients gave written informed consent.

Procedures
After local regional therapy (surgery with or without radiotherapy), patients were randomly assigned to one of three groups: observation only, 8 mg/kg trastuzumab given intravenously by 90 minute infusion as a loading dose followed by 6 mg/kg every 3 weeks for 1 year, or the same schedule of trastuzumab for 2 years (not reported here). Randomisation was done within 7 weeks from day 1 of the last chemotherapy cycle or 6 weeks from the end of radiotherapy or denitive surgery, whichever was the last. A minimisation procedure was used with stratication according to region of the world, age, nodal status, title of chemotherapy, and hormone receptor status together with intention to use endocrine therapy. When a signicant disease-free survival in favour of 1 years treatment with trastuzumab over observation alone emerged with a median follow-up of 1 year,8 a protocol amendment was made after recruitment had been completed (except for the last ve patients) to allow women in the observation group the option of switching to trastuzumab, irrespective of the interval since randomisation. Women who opted to switch were also given the further choice of a secondary randomisation to 1 year versus 2 years of treatment with trastuzumab. The primary endpoint of the trial was disease-free survival (dened as time from randomisation to the rst occurrence of any of the following events: recurrence of breast cancer at any site; the development of ipsilateral or contralateral breast cancer including ductal carcinoma in situ but not lobular carcinoma in situ; second nonbreast malignant disease other than basal cell or squamous cell carcinoma of the skin or carcinoma in situ of the cervix; or death from any cause without documentation of a cancer-related event), and the eect of trastuzumab on this endpoint has been described previously.8 Overall survival was a secondary endpoint; other secondary ecacy endpoints included time to recurrence, time to distant recurrence, and safety, including cardiac safety. Criteria for interrupting or stopping trastuzumab therapy on the basis of cardiotoxicity or other side-eects have been previously described.8 Severe congestive heart failure, which does not include death from cardiac causes, was dened as New York Heart Association grade III or IV functional class
www.thelancet.com Vol 369 January 6, 2007

Diagnosis Local determination of HER2-positive invasive breast cancer

Primary treatment Surgery and adjuvant or neoadjuvant chemotherapy or both, with or without radiation therapy

Conrmation of HER2-positive breast cancer (IHC 3+ or FISH+ by central review laboratory) and LVEF55% after primary treatment

Randomisation

2 years of treatment with trastuzumab Initial dose 8 mg/kg, maintenance dose 6 mg/kg every 3 weeks for 2 years 1701 patients

1 year of treatment with trastuzumab Initial dose 8 mg/kg, maintenance dose 6 mg/kg every 3 weeks for 1 year 1703 patients

Observation alone 1698 patients

2 year median follow-up (range 048 months)

2 year median follow-up (range 049 months)

2 year median follow-up (range 048 months)

No further data released by the independent data monitoring committee Patients continue treatment and follow-up as planned

218 disease-free survival events, 59 deaths Patients continue treatment and follow-up as planned

321 disease-free survival events, 90 deaths 861 patients chosen to start trastuzumab

Figure 1: Trial prole FISH+=uorescence in-situ hybridisation positive. IHC 3+=immunohistochemistry score 3.

30

Articles

conrmed by a cardiologist and a decrease in LVEF of at least 10% below baseline and to less than 50%. Symptomatic congestive heart failure was dened as symptomatic congestive heart failure conrmed by a cardiologist and LVEF less than 50% and a decrease in LVEF of at least 10% from baseline. A signicant LVEF drop was dened as a decrease in LVEF of 10 points or more from baseline to a level below 50 points. A conrmed signicant LVEF drop was dened as an asymptomatic (NYHA class I) or mildly symptomatic (NYHA class II) signicant drop in LVEF which is also conrmed on repeat LVEF assessment about 3 weeks after the rst documented drop, or as identied by the cardiac advisory board review.

led the writing of the paper with input from the HERA executive committee, which includes a Roche representative who was not allowed to inuence the paper in any way other than as approved by the executive committee. All authors had access to all the data. The trials steering committee had nal responsibility to submit the manuscript for publication.

Results
5102 women were recruited between December, 2001, and June, 2005, including 1698 in the observation group and 1703 assigned to receive 1 year of treatment with trastuzumab (gure 1). The baseline characteristics of patient tumours and treatment, updated from the
Observation group (n=1698) Age (years) <35 3549 5059 60 Prior (neo)adjuvant chemotherapy No anthracyclines Anthracyclines, no taxanes Anthracyclines and taxanes Menopausal status* Premenopausal Uncertain Postmenopausal Hormone receptor status Negative Positive Nodal status Not assessed (neoadjuvant chemotherapy) Negative 13 4 178 (10%) 555 (33%) 490 (29%) 474 (28%) 194 (11%) 544 (32%) 486 (29%) 479 (28%) 843 (50%) 855 (50%) 843 (50%) 860 (50%) 234 (14%) 692 (41%) 770 (45%) 257 (15%) 681 (40%) 765 (45%) 101 (6%) 1156 (68%) 441 (26%) 101 (6%) 1154 (68%) 448 (26%) 126 (7%) 752 (44%) 548 (32%) 272 (16%) 127 (7%%) 756 (44%) 548 (32%) 272 (16%) 1-year trastuzumab group (n=1703)

Statistical analysis
A target accrual of 4482 patients was planned to identify a 23% reduction in the risk of a disease-free survival event with 80% power with a two-sided level of signicance of 25% for both of the pairwise comparisons: 2 years of treatment with trastuzumab versus observation alone and 1 year of treatment with trastuzumab versus observation alone. 951 endpoint events were required for the nal analysis. One interim ecacy analysis was planned after 475 events, the results of which, reviewed by the independent data monitoring committee in April, 2005, resulted in the initial HERA trial publication.8 The ecacy analyses were done on an intention-totreat basis. tests for categorical data and log-rank tests for time-to-event endpoints provided two-sided p values. Kaplan-Meier curves were calculated. Cox proportional hazards regression analysis was used to estimate hazard ratios and 95% CI. Statistical analyses were done with SAS version 8. In addition to the main intent-to-treat analysis, an analysis that censored women at the time of switching to trastuzumab has also been done to compensate for a potential eect of delayed administration of trastuzumab. The independent data monitoring committee continues to review data about deaths, compliance, and safety every 6 months. On the basis of a review in March, 2006, they recommended that overall survival results for observation alone versus treatment with trastuzumab for 1 year with a median follow-up of 2 years should be made public. Results for the group of patients treated with trastuzumab for 2 years remain blinded because the comparison with the group treated for 1 year continues to mature. This trial is registered with the European Clinical Trials Database, number 200500238511.

Data are number (%). Percentages have been rounded. *Status at randomisation; in the observation group, one patient with unknown menopausal status at randomisation and one patient with missing menopausal status. One patient with missing nodal status in the observation group.

Table 1: Baseline patient characteristics

Observation group (n=1698) Number of events Deaths Distant event Central nervous system Locoregional event Contralateral breast cancer Second non-breast malignant disease 321 (19%) 90 (5%) 233 (14%) 22 (1%) 68 (4%) 9 (05%) 8 (05%)

1-year trastuzumab group (n=1703) 218 (13%) 59 (3%) 152 (9%) 26 (2%) 45 (3%) 7 (04%) 6 (04%)

Role of the funding source

The trial was sponsored and funded by Roche. The collection, analysis, and interpretation of the data were done entirely independently, under the auspices of the Breast International Group. The corresponding author
www.thelancet.com Vol 369 January 6, 2007

Data are number of events (%). Percentages have been rounded.

Table 2: Site of rst disease-free survival event (ITT analysis)

31

Articles

original publication, are shown in table 1. 97 (57%) patients randomly assigned to observation alone and 58 (34%) patients randomly assigned to 1 year of treatment with trastuzumab were lost to follow-up. As of May 15, 2006, 861 women in the observation group had switched to trastuzumab. In the February, 2006, analysis that censored patients at the time of moving to trastuzumab, the median time that had elapsed from the point of switching was 26 months. 705 patients originally randomly assigned to observation alone were censored for disease-free survival and overall survival at the date of switching treatment. Results from the censored analysis are much the same as those from the intention-to-treat analysis.
A 100

80 Disease-free survival (%) 1 years treatment with trastuzumab Observation alone

63%

60

40

20

0 Numbers at risk Trastuzumab Observation 1703 1698 1591 1535 1434 1330 1127 984 742 639 383 334 140 127

B 100 27% 80 Overall survival (%) 1 years treatment with trastuzumab 60 Observation alone

40

20

0 0 6 12 18 24 30 36 Months from randomisation Numbers at risk Trastuzumab Observation 1703 1698 1627 1608 1498 1453 1190 1097 794 711 407 366 146 139

Figure 2: Kaplan-Meier estimates of disease-free survival and overall survival (A) Disease-free survival for 1 year of trastuzumab vs observation with a median follow-up of 2 years. (B) Overall survival for 1 year of trastuzumab vs observation with a median follow-up of 2 years.

After a median follow-up of 235 months (range 048 months), 539 disease-free survival events had been recorded in the two groups. Table 2 shows data about the site of rst disease-free survival events. The unadjusted HR for the risk of an event in the trastuzumab group compared with the observation group was 064 (95% CI 054076; p<00001 by the log rank test), which corresponds to an absolute disease-free survival benet of 63% (806% vs 743%) at 3 years (gure 2A). The HR for the disease-free survival benet by censored analysis was 063 (053075; p<00001). 149 deaths occurred in the two groups; more deaths occurred in the observation group than in the trastuzumab group (table 2). The unadjusted HR for the risk of death in the trastuzumab group compared with observation alone was 066 (047091; p=00115 by the log rank test); which corresponds with an absolute overall survival benet of 27% (924% vs 897%) at 3 years (gure 2B). The HR for overall survival by censored analysis was 063 (045087; p=00051). More distant metastases were reported in the observation group than in the group receiving trastuzumab (table 2); the HR for time to a distant recurrence in the trastuzumab group compared with the observation group was 060 (049073; p<00001 by log rank test). These results correspond to an absolute time to distant recurrence event-free survival benet of 63% at 3 years (857% vs 794%, p<00001). There was no evidence of substantial heterogeneity in the relative treatment eect on disease-free survival between subgroups, and there was no evidence of any subgroup in which trastuzumab was seen to be less ecacious than observation alone (gure 3). All CIs overlap the overall result. The annualised hazard rates for disease-free survival for both groups from the point of randomisation are shown in gure 4. An increase in risk during the rst year for the observation group was seen, although this risk falls in the second year and beyond. Trastuzumab suppresses this increased early risk. Although the HR between the two groups decreases beyond the second year, there continues to be less chance of a disease-free survival event with trastuzumab compared with observation alone at all points up to 3 years. There were more episodes of at least one grade 3 or grade 4 adverse event and of serious adverse events with trastuzumab than in the observation group (p<00001; table 3). The only grade 3 or grade 4 adverse event experienced by ve or more patients in the observation group was hypertension (n=5). The grade 3 or grade 4 adverse events experienced by ve or more patients in the trastuzumab group were hypertension (12), depression (8), diarrhoea (7), congestive cardiac failure (7), vomiting (6), arthralgia (6), cardiac failure (5), hot ush (5), headache (5), and back pain (5). There was a higher incidence of fatal adverse events in the trastuzumab group than in the observation group
www.thelancet.com Vol 369 January 6, 2007

32

Articles

Subgroup (number of patients) Region of the world Europe, Canada, South Africa, Australia, New Zealand (2438) Asia Pacic, Japan (405) Eastern Europe (369) Central and South America (189) Age at randomisation <35 years (253) 3549 years (1508) 5059 years (1096) 60 years (544) Menopausal status at randomisation Premenopausal (491) Uncertain (1373) Postmenopausal (1535) Nodal status Not assessed (neoadjuvant chemotherapy) (372) Negative (1099) 13 positive nodes (976) 4 positive nodes (953) Pathological tumour size Any (neoadjuvant chemotherapy) (372) 02 cm (1351) >25 cm (1482) >5 cm (171) Hormone receptor status ER-negativePgR-negative (1627) ER-negativePgR-positive (172) ER-positivePgR-negative (460) ER-positivePgR-positive (984) Histological grade 3poorly dierentiated (2047) 2moderately dierentiated (1111) Surgery for primary tumour Breast-conserving procedure (1432) Mastectomy (1968) Previous radiotherapy Yes (2606) No (795) Type of (neo)adjuvant chemotherapy No anthracyclines (202) Anthracyclines, no taxanes (2310) Anthracyclines and taxanes (889) All patients (3401) 00 05 HR 10 15

Number of events trastuzumab vs observation 161 vs 235 21 vs 37 23 vs 36 13 vs 13 19 vs 31 89 vs 150 71 vs 97 39 vs 43 43 vs 49 70 vs 135 105 vs 137 39 vs 50 34 vs 58 50 vs 80 95 vs 132 39 vs 50 61 vs 95 97 vs 150 20 vs 25 126 vs 190 12 vs 12 26 vs 39 46 vs 61 157 vs 201 47 vs 97 77 vs 121 141 vs 200 183 vs 265 35 vs 56 12 vs 15 132 vs 221 74 vs 85 218 vs 321

HR (95% CI)

066 (054081) 053 (031090) 054 (032091) 098 (045211) 057 (032101) 054 (042070) 071 (052097) 091 (059141) 080 (053121) 048 (036064) 075 (058097) 066 (043100) 059 (039091) 061 (043 087) 064 (049083) 066 (043100) 065 (047090) 055 (043071) 114 (063206) 063 (050078) 077 (034174) 082 (050134) 063 (043093) 073 (059090) 046 (033065) 059 (044079) 068 (055084) 064 (053077) 064 (042098) 076 (035162) 057 (046071) 080 (059110) 064 (054076)

Figure 3: Exploratory disease-free survival subgroup analysis for 1 year of trastuzumab vs observation

(p=01601; table 3). However, the nature of the fatal adverse events did not seem to have any causal relation with trastuzumab. Of the 172 women who stopped trastuzumab prematurely, 115 (68%) stopped because of safety issues, 43 (25%) because of refusal, and 14 (08%) because of other reasons. As described previously,13 there was one cardiac death in the observation group and none with trastuzumab. Severe congestive heart failure occurred in more women on trastuzumab than in the observation group (p<00001); likewise, symptomatic congestive heart failure (including severe congestive heart failure) occurred in more patients on trastuzumab than in those in the observation group
www.thelancet.com Vol 369 January 6, 2007

(table 4). Furthermore, a conrmed signicant LVEF drop occurred in more women on trastuzumab than in those in the observation group (p<00001; table 4). 72 (43%) women discontinued trastuzumab because of cardiac problems. A more detailed account of cardiotoxicity in this trial is being prepared.

Discussion
Our results indicate that trastuzumab shows a signicant overall survival benet in early breast cancer over observation alone after chemotherapy. Such a survival benet after only 2 years of follow-up is unusual in breast cancer trials. For example, adjuvant chemotherapy is of
33

Articles

major importance in the treatment of breast cancer and has achieved a long-term survival benet of 38% for younger women,14 but the key early trial for such treatment had not begun to show a substantial survival dierence after only 2 years of follow-up,15 nor have any subsequent chemotherapy trials. In the modern era, the same is true for the aromatase inhibitors, widely deemed to be a major development in adjuvant endocrine therapy.1618 Only tamoxifenthe most successful treatment ever developed
020

Observation 1 years treatment with trastuzumab

015

010

005

000 16 712 1318 1924 2530 3136 Months from randomisation

Figure 4: Annualised disease-free survival hazard rates for 1 year of trastuzumab vs observation Observation group (n=1442)* Patients with one or more grade 3 or 4 adverse event 88 (6%) Patients with one or more serious adverse event Fatal adverse event Treatment withdrawals 97 (7%) 3 (02%) NA 1-year trastuzumab p value group (n=1688)* 190 (11%) 156 (9%) 9 (05%) 172 (10%) <00001 00103 01601 ..

Data are number of events (%). Percentages have been rounded. NA=not applicable. *271 patients with at least one adverse event after moving to trastuzumab are not included. Treatment group vs observation group. Cardiac failure, suicide, unknown cause of death. Cerebral haemorrhage, cerebrovascular accident, sudden death, appendicitis, intestinal obstruction, unknown cause of death after a road accident, pulmonary carcinomatous lympangitis, two unknown. The intestinal obstruction occurred after a second occurrence of non-breast malignant disease.

Table 3: Adverse events Observation (n=1708) Cardiac death Severe congestive heart failure (NYHA III and IV) Symptomatic congestive heart failure Conrmed signicant LVEF drop Trastuzumab discontinued due to cardiac problems 1 (01%) 0 2 (01%) 9 (05%) NA 1-year trastuzumab p value* (n=1678) 0 10 (06%) 36 (2%) 51 (3%) 72 (4%) 1000 <00001 <00001 <00001 <00001

Data are number of events (%). Percentages have been rounded. NA=not applicable. *Treatment group vs observation group. Not including cardiac death. Including severe congestive heart failure but not including cardiac death. Asymptomatic or mildly symptomatic.

Table 4: Cardiac safety

for breast cancershowed a similar survival benet in such a short period.19 Thus, the early evidence of an overall survival benet in the HERA trial after only 2 years of follow-up reinforces the importance of trastuzumab in the adjuvant treatment of women with HER2-positive early breast cancer.20 That this survival dierence will not be sustained with further follow-up is a possibility; however, there are few precedents for this in work already published on adjuvant therapies for early breast cancer. Furthermore, after the publication of disease-free survival results from HERA with a median follow-up of 12 months,8 we estimated that the chance of losing statistical signicance with longer follow-up was less than 20%. This estimate was calculated on the basis of division of HRs for disease-free survival into a xed component already observed and a random component, assuming that the risk for the observation group fell to the level seen in the trastuzumab group and follow-up was continued for a further 4 years. Because more events have occurred in the observation group than in the trastuzumab group during the subsequent 12 months, and the HR has fallen for both arms beyond 2 years, the chance of loss of signicance remains below 20%. The same arguments pertain for overall survival. The one confounding issue is the recent cross-over of patients in the control group to trastuzumab; it could be that, with longer follow-up, a signicant overall survival benet might be maintained in the censored analysis, but not the intention-to-treat analysis because of this crossover. The NSABP B-31 trial compared treatment with trastuzumab for 1 year starting concurrently with four courses of paclitaxel after a combination of doxorubicin and cyclophosphamide with the same chemotherapy schedule alone.9 When data from this trial were combined with results from two similar groups in the NCCTG 9831 trial, a signicant overall survival benet was noted after a median follow-up of 2 years and with an HR of 067 in favour of trastuzumab,9 which is much the same as that recorded here. The consistently high reductions in the risk of breast cancer recurrence of around 50% seen in these two trials, together with that noted in a fourth large adjuvant trastuzumab trial, BCIRG 006,10 suggest that these results will probably translate into stand-alone survival gains in the near future. Our exploratory subgroup analysis suggests that all subgroups of women seem to benet from trastuzumab. In particular, there is so far no signicant dierence in ecacy between women with node-positive and nodenegative disease, nor between those receiving adjuvant compared with neoadjuvant chemotherapy. Whether dierent patterns of treatment eect emerge for some subgroups with further follow-up remains to be seen. That the risk of cardiotoxicity remains low is encouraging, and the absence of any substantial evidence of an increase in cumulative cardiotoxicity beyond 1 year has also been shown in the B-31 trial.21 The overall risk of severe congestive heart failure reported here is lower than
www.thelancet.com Vol 369 January 6, 2007

34

Hazard rates

Articles

that reported in the B-31 trial (06% vs 41%),21 as is the proportion of women who stopped treatment because of cardiac problems (43% and, of evaluable trastuzumab patients, 156%). These ndings could be a result of the longer time interval between stopping chemotherapy with anthracyclines and starting trastuzumab in the HERA trial than in B-31, or the requirement of a postchemotherapy LVEF of 55% or more before enrolment in HERA, or both. In this context, one should note that the incidence of cardiotoxicity in the anthracycline followed by trastuzumab group of the BCIRG 006 trial was much the same as that in B-31, but substantially lower in the trastuzumab group that did not receive anthracycline.10 Thus far, the only data for cardiotoxicity associated with trastuzumab continued for more than a year is in metastatic breast cancer: in one recent study involving 218 patients with a median treatment duration of 21 months and a median follow-up of 326 months, 49 (28%) experienced a cardiac event and 19 (109%) grade 3 cardiotoxicity.22 However, most patients improved after withdrawal of trastuzumab and appropriate treatment, and there was only one cardiac death. One should note that these patients would often be less t than those in HERA and other adjuvant trials because of their metastatic disease, and some would have received a greater cumulative dose of anthracycline chemotherapy. Two major questions remain for adjuvant trastuzumab. The rst is whether trastuzumab started concurrently with taxane chemotherapy (as in the USA trials) is better than trastuzumab starting sequentially after completion of chemotherapy (as here). The NCCTG N9831 trial addresses this issuethis trial includes a third group given sequential trastuzumab. Preliminary data suggest that sequential treatment might be less eective than concurrent treatment,23 but this was an unplanned comparison with low statistical power, and longer follow-up is needed for conrmation. In this context, one should note that the median time from diagnosis of breast cancer to starting trastuzumab in the HERA trial was 85 months, which could have aected ecacy or resulted in some patients with very high-risk disease relapsing before the opportunity arose to enter the trial. The second issue concerns the duration of trastuzumab treatment. A third group in the HERA trial, in which patients are treated with trastuzumab for 2 years, addresses this issue. That trastuzumab suppresses the increased early hazard rate seen in the observation group during the rst year, but has less eect thereafter, might be of relevance (gure 4). This nding supports our decision to compare 2 years of treatment with trastuzumab with 1 year of treatment. However, one should also note that a small Finnish trial involving 232 women has reported a diseasefree survival benet with 9 weeks of trastuzumab treatment given concurrently with chemotherapy at the start of treatment.11 The HR from this trial (042, 95% CI 021083; p=001) is within the same range as those recorded in trials of trastuzumab treatment for 1 year, but
www.thelancet.com Vol 369 January 6, 2007

the small number of women involved, together with the wide CI, require conrmatory data. The data presented here conrm earlier reports that trastuzumab, a rationally designed biological therapy that is targeted against a specic amplied gene (HER2) and its over-expressed receptor protein, is of benet to women with HER2-positive breast cancer when given after completion of adjuvant chemotherapy. The survival benet that has emerged over such a short period emphasises the potential of this approach and underlines the importance of developing further specic targeted therapies in breast and other cancers.
Contributors M Untch was a member of the executive committee and participated in the conduct of the trial and the preparation of the manuscript. M Kaufmann participated in the HERA study with the German GABG-Study Group, of which he is chairman and participated in the conduct of, and recruitment to, the HERA trial as a key investigator. R Bell participated in the concept, management, and conduct of the HERA trial as a member of the executive committee, steering committee, and as an investigator. R I Lopez participated in the conduct of, and recruitment to the HERA trial as a key investigator. M Procter was involved in the statistical analysis. P Mallmann participated in the conduct of, and recruitment to the HERA trial as a key investigator. K Gelmon participated as an investigator, as the representative of the National Cancer Institute of CanadaClinical Trials Group (NCICCTG), and on the trial committee, and contributed to the nal version of the manuscript. E Wist was principal investigator in Norway. J Bergh took part in the study planning, was a member of the steering committee, and made comments on a draft version of the manuscript. G Mariani was principal investigator for the Michelangelo group. M J Piccart-Gebhart participated in the design, conduct, and analysis of the trial as chief investigator and chairman of the steering committee. P Snchez Rovira took part in the conduct of recruitment to the HERA trial as a key investigator. I Smith took part in the planning and execution of the HERA trial, and wrote the rst and nal drafts of this manuscript. N Wilcken was a member of the steering committee and took part in the writing of the manuscript. A Goldhirsch took part in the study design, conduct, and interpretation of the results. R Coleman took part as an investigator, was leader of the participating EORTC collaborative group, and was a member of the HERA steering committee. A Wardley participated in the conduct of, and recruitment to the HERA trial as a key investigator and helped write the manuscript. S Guillaume was the intergroup and monitoring coordinator. N Harbeck was a member of the German HERA steering committee, and took part in the design, revision, and amendment of the study design. She was a local principal investigator, involved in patient recruitment. J Baselga took part in the design, conduct, patient accrual, and interpretation of the data as a member of the executive committee and as an investigator. M Dowsett was a member of the executive committee and co-chair of the translational research committee. D Cameron was an investigator in the HERA trial, and was a member of the steering and executive committees. As the executive committee member representative for all Breast International Group collaborative groups, he was involved in the day-to-day running of the trial. R D Gelber took part in the design, conduct, analysis, and reporting of the trial in his role as senior statistician. He directed statistical analyses and contributed to the writing of the manuscript. A Feyereislova was a coordinator and reviewed the published data. All authors saw and approved the nal version of the manuscript. Conict of interest statement M Untch has received speakers honoraria from Roche. R Bell has served as an adviser to Roche. In the past 2 years, K Gelmon has been on advisory boards and has received honoraria from Roche, sano aventis, BristolMyers Squibb, AstraZeneca, Novartis, Pzer, Lilly, Amgen, Genetech, and GlaxoSmithKline. She has also received research grants from Roche. J Berghs research group has received research support from Roche, and J Bergh has taken part in advisory boards for Roche. M J Piccart-Gebhart has served on an advisory board on Aromasin, has received consulting fees from GlaxoSmithKline, and an unrestricted educational grant from Roche

35

Articles

on behalf of the Breast International Group. I Smith has received honoraria from Roche for lectures and attendance at advisory boards. N Wilcken has received honoraria from Roche for educational presentations. A Goldhirsch has received honoraria and travel expenses from Roche. A Wardley has received honoraria from Roche for speaking engagements including the use of adjuvant trastuzumab. He has also received travel grants from Roche and worked and done a small amount of advisory work for Roche. N Harbeck has received speakers honoraria from Roche and received the investigators fee from the German Breast Group. J Baselga is a member of the Roche herceptin advisory board. M Dowsett has received fees for attending advisory board meetings and giving ad-hoc lectures for Roche. D Cameron has received research funding from Roche (separate to that for the conduct of this trial), and has also received honoraria and consultancy fees from Roche. R D Gelber declares that Roche provided nancial support for the Breast European Adjuvant Study Team (BrEAST), which in turn provided him with partial salary support. A Feyereislova is an employee of Roche. R Coleman, M Kaufmann, S Guillaume, M Procter, P Mallmann, E Wist, G Mariani, P Sanchez Rovira, and R I Lopez declare that they have no conict of interest. Acknowledgments We thank Donna Saery for her invaluable skill and expertise in the preparation of this manuscript. References 1 Gschwind A, Fischer OM, Ullrich A. The discovery of receptor tyrosine kinases: targets for cancer therapy. Nat Rev Cancer 2004; 4: 36170. 2 Slamon DJ, Clark GM, Wong SG, et al. Human breast cancer: correlation of relapse and survival with amplication of the HER-2/ neu oncogene. Science 1987; 235: 17782. 3 Slamon DJ, Godolphin W, Jones LA, et al. Studies of the HER-2/ neu proto-oncogene in human breast and ovarian cancer. Science 1989; 244: 70712. 4 Vogel CL, Cobleigh MA, Tripathy D, et al. Ecacy and safety of trastuzumab as a single agent in rst-line treatment of HER2overexpressing metastatic breast cancer. J Clin Oncol 2002; 20: 71926. 5 Baselga J, Carbonell X, Castaneda-Soto NJ, et al. Phase II study of ecacy, safety, and pharmacokinetics of trastuzumab monotherapy administered on a 3-weekly schedule. J Clin Oncol 2005; 23: 216271. 6 Slamon DJ, Leyland-Jones B, Shak S, et al. Concurrent administration of anti-HER2 monoclonal antibody and rst-line chemotherapy for HER2-overexpressing metastatic breast cancer. A phase III, multinational, randomised controlled trial. N Engl J Med 2001; 344: 78392. 7 Marty M, Cognetti F, Maraninchi D, et al. Randomised phase II trial of the ecacy and safety of trastuzumab combined with docetaxel in patients with human epidermal growth factor receptor 2-positive metastatic breast cancer administered as rst-line treatment: the M77001 Study Group. J Clin Oncol 2005; 23: 426574. 8 Piccart-Gebhart MJ, Procter M, Leyland-Jones B, et al. Trastuzumab after adjuvant chemotherapy in HER2-positive breast cancer. N Engl J Med 2005; 353: 165972. 9 Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus adjuvant chemotherapy for operable HER2-positive breast cancer. N Engl J Med 2005; 353: 167384.

10

11

12 13

14

15

16

17

18

19

20 21

22

23

Slamon D, Eiermann W, Robert N, et al. Phase III randomised trial comparing doxorubicin and cyclophosphamide followed by docetaxel (AC T) with doxorubicin and cyclophosphamide followed by docetaxel and trastuzumab (AC TH) with docetaxel, carboplatin and trastuzumab (TCH) in HER2 positive early breast cancer patients: BCIRG 006 study. 28th Annual San Antonio Breast Cancer Symposium; San Antonio, TX, USA; Dec 710, 2005. Abstract 1. Joensuu H, Kellokumpu-Lehtinen PL, Bono P, et al. Adjuvant docetaxel or vinorelbine with or without trastuzumab for breast cancer. N Engl J Med 2006; 354: 80920. The Lancet. Herceptin and early breast cancer: a moment for caution. Lancet 2005; 366: 1673. HERA Study Team. Trastuzumab (H: Herceptin) following adjuvant chemotherapy (CT) signicantly improves disease-free survival (DFS) in early breast cancer (BC) with HER2 overexpression: the HERA Trial. 28th Annual San Antonio Breast Cancer Symposium; San Antonio, TX, USA; Dec 710, 2005. Abstract 11. Early Breast Cancer Trialists Collaborative Group (EBCTCG). Eects of chemotherapy and hormonal therapy for early breast cancer on recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005; 365: 1687717. Bonadonna G, Brusamolino E, Valagussa P, et al. Combination chemotherapy as an adjuvant treatment in operable breast cancer. N Engl J Med 1976; 294: 40510. Baum M, Budzar AU, Cuzick J, et al. Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early breast cancer: rst results of the ATAC randomised trial. Lancet 2002; 359: 213139. Coombes RC, Hall E, Gibson LJ, et al. A randomized trial of exemestane after two or three years of tamoxifen therapy in postmenopausal women with primary breast cancer. N Engl J Med 2004; 350: 108192. Thurlimann B, Keshaviah A, Coates AS, et al. A comparison of letrozole and tamoxifen in postmenopausal women with early breast cancer. N Engl J Med 2005; 353: 274757. Baum M, Brinkley DM, Dossett JA, et al. Improved survival among patients treated with adjuvant tamoxifen after mastectomy for early breast cancer. Lancet 1983; 2: 450. Hortobagyi GN. Trastuzumab in the treatment of breast cancer. N Engl J Med 2005; 353: 173436. Tan-Chiu E, Yothers G, Romond E, et al. Assessment of cardiac dysfunction in a randomized trial comparing doxorubicin and cyclophosphamide followed by paclitaxel, with or without trastuzumab as adjuvant therapy in node-positive, human epidermal growth factor receptor 2-overexpressing breast cancer. NSABP B-31. J Clin Oncol 2005; 23: 781119. Guarneri V, Lenihan DJ, Valero V, et al. Long-term cardiac tolerability of trastuzumab in metastatic breast cancer: the MD Anderson Cancer Center experience. J Clin Oncol 2006; 24: 410715. Perez EA, Suman VJ, Davidson N, et al. Advances in monoclonal therapy for breast cancer: further analysis of NCCTG N9831. 41st Annual Meeting of the American Society of Clinical Oncology; Orlando, FL; May 16, 2005.

36

www.thelancet.com Vol 369 January 6, 2007