Lapatinib versus trastuzumab in combination with neoadjuvant anthracycline-taxane-based chemotherapy (GeparQuinto, GBG 44): a randomised phase 3 trial

Articles

Lapatinib versus trastuzumab in combination with neoadjuvant anthracycline-taxane-based chemotherapy (GeparQuinto, GBG 44): a randomised phase 3 trial Michael Untch, Sibylle Loibl, Joachim Bischoff, Holger Eidtmann, Manfred Kaufmann, Jens-Uwe Blohmer, Jörn Hilfrich, Dirk Strumberg, Peter A Fasching, Rolf Kreienberg, Hans Tesch, Claus Hanusch, Bernd Gerber, Mahdi Rezai, Christian Jackisch, Jens Huober, Thorsten Kühn, Valentina Nekljudova, Gunter von Minckwitz, for the German Breast Group (GBG) and the Arbeitsgemeinschaft Gynäkologische Onkologie-Breast (AGO-B) Study Group

Summary Background We compared the efficacy and safety of the addition of lapatinib versus trastuzumab to anthracyclinetaxane-based neoadjuvant chemotherapy. Methods In the GeparQuinto randomised phase 3 trial, patients with untreated HER2-positive operable or locally advanced breast cancer were enrolled between Nov 7, 2007, and July 9, 2010. Patients were eligible if their tumours were classified as cT3/4a-d, or hormone receptor (HR)-negative, HR-positive with clinically node-positive and cT2 disease (cT2 cN+), or HR-positive and pathologically node-positive in the sentinel lymph node for those with cT1 disease (cT1 pNSLN+). Patients were randomly assigned in a 1:1 ratio to receive neoadjuvant treatment with four cycles of EC (epirubicin [90 mg/m² intravenously] plus cyclophosphamide [600 mg/m² intravenously], every 3 weeks), and four cycles of docetaxel (100 mg/m² intravenously every 3 weeks) with either trastuzumab (6 mg/kg intravenously, with a starting loading dose of 8 mg/kg, for eight cycles, every 3 weeks) or lapatinib (1000–1250 mg per day orally) throughout all cycles before surgery. Randomisation was done by dynamic allocation with the minimisation method of Pocock and patients were stratified by participating site, HR status, and extent of disease (cT1–3 cN0–2 vs T4 or N3). The primary endpoint was pathological complete response (defined as ypT0 and ypN0) and was analysed in all patients who received at least one cycle of EC. Participants and investigators were not masked to treatment assignment. Pathologists in centres assessing surgery outcomes were masked to group assignment. This trial is registered with ClinicalTrials.gov, number NCT00567554. Findings Of 620 eligible patients, 309 were randomly assigned to chemotherapy with trastuzumab (ECH-TH group) and 311 to chemotherapy with lapatinib (ECL-TL group). Two patients in the ECH-TH group and three patients in the ECL-TL group did not start treatment because of withdrawal of consent or immediate surgery. 93 (30·3%) of 307 patients in the ECH-TH group and 70 (22·7%) of 308 patients in the ECL-TL group had a pathological complete response (odds ratio [OR] 0·68 [95%CI 0·47–0·97]; p=0·04). Chemotherapy with trastuzumab was associated with more oedema (119 [39·1%] vs 88 [28·7%]) and dyspnoea (90 [29·6%] vs 66 [21·4%]), and ECL-TL with more diarrhoea (231 [75·0%] vs 144 [47·4%]) and skin rash (169 [54·9%] vs 97 [31·9%]). 43 (14·0%) patients discontinued in the ECHTH group and 102 (33·1%) in the ECL-TL group. 70 serious adverse events were reported in the ECH-TH group and 87 in the ECL-TL group. Interpretation This direct comparison of trastuzumab and lapatinib showed that pathological complete response rate with chemotherapy and lapatinib was significantly lower than that with chemotherapy and trastuzumab. Unless longterm outcome data show different results, lapatinib should not be used outside of clinical trials as single anti-HER2treatment in combination with neoadjuvant chemotherapy. Funding GlaxoSmithKline, Roche, and Sanofi-Aventis.

Introduction The use of neoadjuvant chemotherapy for the treatment of patients with primary breast cancer has increased during the past decade. The aim of neoadjuvant chemotherapy in routine practice is to improve operability of the breast tumour; however, within clinical trials, it is considered as an in-vivo test for chemosensitivity to new agents and treatments, with the aim to precede and anticipate the results from large adjuvant trials.1 www.thelancet.com/oncology Vol 13 February 2012

In the NeOAdjuvant Herceptin (NOAH) study,2 HER2positive patients with locally advanced or inflammatory breast cancer were randomly assigned to a neoadjuvant chemotherapy with or without trastuzumab. Trastuzumab significantly improved pathological complete response rates and 3-year event-free survival.2 Results from the Taxol Epirubicin Cyclophosphamide Herceptin NeOadjuvant (TECHNO) study3 showed that those 39% of patients who achieved a pathological complete response with epirubicin plus cyclophosphamide (EC)

Lancet Oncol 2012; 13: 135–44 Published Online January 17, 2012 DOI:10.1016/S14702045(11)70397-7 See Comment page 112 See Lancet Online/Articles DOI:10.1016/S01406736(11)61847-3 See Lancet Online/Comment DOI:10.1016/S01406736(12)60068-3 Helios-Klinikum, Berlin-Buch, Berlin, Germany (Prof M Untch MD); Headquarters, German Breast Group, Neu-Isenburg, Germany (S Loibl MD, V Nekljudova PhD, Prof G von Minckwitz MD); Universitäts-Frauenklinik, Magdeburg, Germany (J Bischoff MD); Universitäts-Frauenklinik, Kiel, Germany (H Eidtmann MD); Universitäts-Frauenklinik, Frankfurt am Main, Germany (Prof M Kaufmann MD); St Gertrauden Krankenhaus, Berlin, Germany (Prof J-U Blohmer MD) Henriettenstift, Hannover, Germany (Prof J Hilfrich MD); Marienhospital Herne, Universitätsklinikum, Bochum, Germany (Prof D Strumberg MD); Department of Gynaecology and Obstetrics, University Hospital Erlangen, Friedrich-Alexander-University Erlangen-Nuremberg, Germany and Comprehensive Cancer Centre Erlangen-Nuremberg, Erlangen, Germany (Prof P A Fasching MD); Universitäts-Frauenklinik, Ulm, Germany (Prof R Kreienberg MD); Onkologie Bethanien, Frankfurt, Germany (Prof H Tesch MD); Klinikum zum Roten Kreuz, München, Germany (C Hanusch MD);

135

Articles

Universitäts-Frauenklinik, Rostock, Germany (Prof B Gerber MD); Luisenkrankenhaus, Düsseldorf, Germany (M Rezai MD); Klinikum Offenbach, Offenbach am Main, Germany (Prof C Jackisch MD); Kantonsspital, St Gallen, Switzerland (Prof J Huober MD); and Klinikum Esslingen, Esslingen, Germany (Prof T Kühn MD) Correspondence to: Prof Gunter Von Minckwitz, German Breast Group, c/o GBG Forschungs GmbH, Martin-Behaim-Straße, 63263 Neu-Isenburg, Germany gunter.vonminckwitz@ germanbreastgroup.de

followed by paclitaxel plus trastuzumab had a significantly longer 3-year disease-free and overall survival than did patients who received the same treatment, but who had no pathological complete response.3 The GeparQuattro study4 included 450 patients with HER2-positive tumours treated with EC followed by docetaxel and trastuzumab. A pathological complete response rate of 31·7% was reported whereas a group of 1050 patients with HER2negative tumours treated by the same chemotherapy but without trastuzumab achieved only a pathological complete response rate of 15·7%. Cardiac event rates were low in these trials, despite simultaneous treatment with trastuzumab with anthracyclines.2,4 Lapatinib is an oral dual tyrosine kinase inhibitor that targets EGFR and HER2. Lapatinib in combination with capecitabine in heavily pretreated patients with HER2-positive, metastatic breast cancer resulted in a prolonged time to progression and suggested an improved overall survival compared with those treated with capecitabine alone.5 Other studies have shown improved

620 patients completed eligibility criteria

309 randomised to ECH-TH

311 randomised to ECL-TL

3 did not start ECL-TL because of withdrawal or immediate surgery

2 did not start ECH-TH because of withdrawal or immediate surgery

308 started ECL-TL

307 started ECH-TH

32 discontinued ECH-TH 13 adverse events during T 7 investigators’ decision 8 patient’s decision 3 tumour progression 1 unknown

51 discontinued ECL-TL 19 adverse events during T 9 investigators’ decision 18 patient’s decision 2 tumour progression 3 unknown

11 discontinued H only 0 adverse events during EC 0 adverse events during T 2 investigators’ decision 2 patient’s decision 2 other 0 unknown 5 missing

51 discontinued L only 5 adverse events during EC 15 adverse events during T 8 investigators’ decision 4 patient’s decision 0 other 18 unknown 1 missing

264 completed ECH-TH

206 completed ECL-TL

307 had surgery

308 had surgery

267 analysed per protocol*

172 analysed per protocol†

Figure 1: Flow of patients throughout the study EC=epirubicin plus cyclophosphamide. H=trastuzumab. T=docetaxel. L=lapatinib. *Three patients with tumour progression who discontinued in the ECH-TH group were included in the per-protocol analysis. No patients who received all cycles of ECH-TH were excluded from the per-protocol analysis. †Two patients with tumour progression who discontinued in the ECL-TL group were included in the per-protocol analysis. 36 patients who discontinued in the ECL-TL group were excluded from the per-protocol analysis because they did not receive the full dose.

136

outcome when lapatinib was given to paclitaxel or aromatase inhibitors.6,7 Furthermore, patients treated with the combination of lapatinib plus trastuzumab have shown longer overall survival than those given lapatinib alone in patients with heavily pretreated metastatic breast cancer.8 These data provide the rationale for lapatinib to be tested in the non-metastatic early breast cancer setting. The GeparQuinto phase 3 study randomly assigned patients with primary breast cancer to neoadjuvant treatment within three groups (for those with HER2positive disease, patients were treated with chemotherapy plus trastuzumab vs chemotherapy plus lapatinib; for the HER2-negative group, chemotherapy plus or minus bevacizumab was given; and for the HER2-negative, nonresponding patients, chemotherapy plus or minus everolimus was given). Here, we report on the HER2positive group.

Methods Patients Female patients with previously untreated unilateral or bilateral primary invasive breast carcinoma were enrolled in the study. Diagnosis of breast cancer had to be confirmed histologically by core biopsy. HER2 status of the tumour had to be positive by HercepTest (Dako, Clostrup, Denmark; score three times positive) or in-situ hybridisation (ratio ≥2·0) by the local pathologist. Tumour lesions were required to have a palpable size of 2 cm or more, or a sonographical size of 1 cm or more in maximum diameter, and had to be measurable in two dimensions, preferably by sonography. In case of inflammatory disease, the clinical extent of inflammation was used as the measurable lesion. Patients with locally advanced tumour stages cT4 or cT3, hormone receptor (HR)-negative tumours, or HR-positive tumours with clinically positive axillary nodes (cN+; for cT2) or pNSLN+ for cT1 disease were eligible. Further relevant inclusion criteria were normal cardiac function (left ventricular ejection fraction ≥55%) and no evidence of distant disease; further exclusion criteria included known or suspected cardiac disease, previous thromboembolic event, known haemorrhagic diathesis or coagulopathy, disease significantly affecting gastrointestinal function, major surgery within the last 28 days or anticipation of the need for major surgery during study treatment, and concurrent treatment with other anti-cancer or investigational agents. All patients provided written informed consent and the protocol was reviewed by all responsible local ethics committees and competent authorities. The conduct of the trial was supervised by an independent data monitoring committee.

Randomisation and masking Patients were randomly assigned to receive trastuzumab (ECH-TH group) or lapatinib (ECL-TL group) in addition to EC followed by docetaxel. Central randomisation was done by dynamic allocation with the www.thelancet.com/oncology Vol 13 February 2012

Articles

minimisation method of Pocock9 in a 1:1 ratio. Patient allocations were stratified by participating site, HR status (negative vs positive), and extent of disease (cT1–3 cN0–2 vs T4 or N3). The study statistician (VN) prepared the database and randomisation program. Patients with complete baseline documentation were filed in the randomisation database at the German Breast Group (GBG). The patient number and treatment group was assigned by the program and communicated to the participating site. Participants and investigators were not masked to treatment assignment, pathologists in centres assessing surgery outcome were masked to treatment assignment. A central blinded review of pathology reports was done.

Procedures All patients were scheduled to receive four cycles of EC (epirubicin 90 mg/m², cyclophosphamide 600 mg/m², day 1, every 3 weeks) followed by four cycles of docetaxel (100 mg/m², day 1, every 3 weeks). Patients were randomly assigned to receive trastuzumab 6 mg/kg intravenously, every 3 weeks, starting with a loading dose of 8 mg/kg intravenously on day 1 of the first EC cycle or lapatinib 1250 mg per day starting on day 1 of the first cycle of EC until day 21 of the fourth cycle of docetaxel concomitantly to all chemotherapy cycles treatment. The first 30 patients randomly assigned to lapatinib received only 1000 mg per day during the first EC cycle and first cycle of docetaxel, and dose was escalated to 1250 mg per day for subsequent cycles in case of adequate tolerability (run-in phase).10 Dose of lapatinib was reduced to 1000 mg per day to improve tolerability for all subsequent cycles after a protocol amendment. This was implemented after 210 patients were accrued into the lapatinib group of the study. Patients completed post-surgery trastuzumab treatment for 1 year in both treatment groups. No further post-surgical chemotherapy regimen was recommended by the protocol. Pegfilgrastim was given with lapatinib as primary prophylaxis for febrile neutropenia and with trastuzumab as secondary prophylaxis. Loperamide was prescribed as hands-on-medication and patients receiving lapatinib were informed to use it immediately after the first onset of diarrhoea. In the case of tumour progression during chemotherapy, study treatment was discontinued and further treatment was up to the investigator. No crossover for the anti-HER2 agents was recommended. Patients had to undergo surgery within 21–35 days after last chemotherapy infusion. Sentinel node biopsy was allowed before registration or at the time of definitive surgery, or both. This procedure was allowed instead of axillary clearance in patients with no involvement of the lymph nodes. We assessed haematological and biochemical variables on a weekly basis and examined the target lesion and regional lymph nodes by palpation at every cycle. Breast www.thelancet.com/oncology Vol 13 February 2012

ECH-TH (n=307)

ECL-TL (n=308)

Age (years) <30

6 (2%)

9 (3%)

30–39

38 (12%)

45 (15%)

40–49

109 (36%)

97 (31%)

50–59

99 (32%)

103 (33%)

60–69

50 (16%)

48 (16%)

≥70 Median (range)

5 (2%)

6 (2%)

50 (25–74)

50 (21–73)

cT4a–c

14 (5%)

14 (5%)

cT4d

42 (14%)

41 (13%)

cT1–3

251 (82%)

253 (82%)

Clinical tumour stage

By palpation cT1

47 (16%)

35 (12%)

cT2

130 (45%)

136 (48%)

cT3

56 (19%)

59 (21%)

Missing

18

23

By sonography cT1

51 (17%)

45 (15%)

cT2

180 (59%)

184 (61%)

cT3

17 (6%)

18 (6%)

Missing

3

6

Stage Operable Locally advanced (cT4 or cN3) Missing

256 (83%)

254 (82%)

51 (17%)

54 (18%)

0

0

Clinical nodal status cN0

93 (31%)

93 (31%)

cN1

209 (69%)

208 (69%)

Missing

5

7

Sentinel node biopsy done No

223 (73%)

222 (72%)

Yes

83 (27%)

85 (28%)

Missing

1

1

Tumour type Ductal invasive

271 (89%)

Lobular invasive

11 (4%)

6 (2%)

Other

24 (8%)

23 (7%)

Missing

1

278 (91%)

1

Tumour grade I

5 (2%)

3 (1%)

II

159 (52%)

152 (49%)

III

141 (46%)

152 (49%)

Missing

2

1

Hormone-receptor status Negative

137 (45%)

137 (44%)

Positive

170 (55%)

171 (56%)

226 (74%)

217 (70%)

Focality Unifocal Multifocal

53 (17%)

55 (18%)

Multicentric

28 (9%)

36 (12%)

Table 1: Patients characteristics at baseline

137

Articles

ultrasound was repeated after every second cycle and ultrasound and mammography was done before breast surgery. We repeated cardiac ultrasound after four cycles of therapy and before surgery. The local pathologist assessed the pathological response of the breast tumour and infiltration of regional lymphnodes using a modified regression grading system11 (grade 5: no microscopic evidence of residual viable tumour cells [invasive or non-invasive] in breast and nodes [pathological complete response]; grade 4: no residual tumour in breast tissue, but involved nodes; ECH-TH (n=307)

ECL-TL (n=308)

No

214 (69·7%)*

238 (77·3%)*

··

··

Yes

93 (30·3%)

70 (22·7%)

··

··

95% CI for ypT0, ypN0

25·2–35·8

18·2–27·8

··

ypT0, ypN0 (primary endpopint)

Difference between arms (% [95% CI]) 7·6 (0·6 to 14·5)

ypT0, ypN0/+

8·2 (1·0 to 15·4)

p value 0·04

·· 0·03

No

202 (65·8%)

228 (74·0%)

··

··

Yes

105 (34·2%)

80 (26·0%)

··

··

95% CI for ypT0, ypN0/+

28·9–39·8

21·2–31·3

ypT0/is, ypN0

·· 14·4 (6·9 to 22·0)

·· <0·0001

No

170 (55·4%)

215 (69·8%)

··

··

Yes

137 (44·6%)

93 (30·2%)

··

··

39·0–50·4

25·1–35·7

··

95% CI ypT0/is, ypN0/+

14·1 (6·4 to 22·0)

·· 0·001

No

154 (50·2%)

198 (64·3%)

··

··

Yes

153 (49·8%)

110 (35·7%)

··

··

44·1–55·6

30·4–41·3

95% CI for ypT0/is, ypN0/+ Regression grade (RG)

·· NA

·· 0·02

RG 0

16 ( 5·2%)

20 ( 6·5%)

··

··

RG 1

97 (31·6%)

128 (41·6%)

··

··

RG 2

41 (13·4%)

50 (16·2%)

··

··

RG 3

48 (15·6%)

30 (9·7%)

··

··

RG 4

12 ( 3·9%)

10 (3·2%)

··

··

RG 5

93 (30·3%)

70 (22·7%)

Clinical response

·· –0·1 (–4·8 to 4·7)

·· 1·00

CR

100 (33·0%)

84 (28·5%)

··

··

Statistical analysis

PR

173 (57·1%)

182 (61·7%)

··

··

Overall (CR or PR)

273 (90·1%)

266 (90·2%)

··

··

95% CI for overall

86·2–93·2

86·2–93·3

··

··

No change

24 (7·9%)

23 (7·8%)

··

··

Progressive disease

6 (2·0%)

6 (2·0%)

··

··

Missing

4

··

··

All patients receiving at least one cycle of EC were included in the efficacy and safety analyses. Treatment groups were compared by continuity corrected twosided Pearson’s χ² test (for efficacy parameters) and Fisher’s exact test. 95% CIs are provided for the efficacy endpoints. Patients with missing response data were counted as having no response. Sensitivity analyses were done of the primary endpoint in patients who received all cycles of chemotherapy and at least 90% of the planned dose of trastuzumab or lapatinib (except those who discontinued chemotherapy due to progression). A multivariable logistic regression without stepwise selection was used to adjust for the baseline factors. Univariable logistic regression was used in subgroup analyses and a Breslow-Day test for interaction12 to test the homogeneity of odds ratios (ORs) across subgroups. We assessed the duration of

13

Breast-conserving surgery

4·9 (–3·1 to 13·0)

0·27

No

102 (36·4%)

115 (41·4%)

··

··

Yes

178 (63·6%)

163 (58·6%)

··

··

Type of surgery unknown

27

30

··

··

95% CI for breast-conserving surgery

57·6–69·2

52·6–64·5

··

··

Data are number (%) unless otherwise specified. BCS=breast-conserving surgery. CR=complete response. PR=partial response. NA=not applicable. *11 patients in each arm had no information on pathological response but were counted as having no response. Table 2: Efficacy of treatment, according to histological, surgical, and clinical outcome

138

grade 3: only residual non-invasive tumour in breast tissue; grade 2: focal invasive tumour measuring ≤5 mm; and grade 0–1 for all remaining scenarios). If new lesions were detected, the response was graded as 0–1. Clinical complete response was defined as no evidence of disease in the breast by ultrasound, or, if ultrasound was not possible, by mammography or physical examination. A partial response was defined as a reduction in the product of the two largest perpendicular diameters of the primary tumour by 50% or more; progressive disease was defined as an increase in tumour size by 25% or more or the presence of a new lesion. All remaining scenarios were classified as no change. Breast-conserving surgery was considered if tumorectomy, segmentectomy, or quadrantectomy was the final surgical procedure. Toxic effects were graded with the National Cancer Institute Common Terminology Criteria (NCI-CTC) version 3.0. The primary aim of the study for the HER2-positive primary breast cancer group in GeparQuinto was to compare the pathological complete response rates (defined as ypT0 and ypN0) after treatment with trastuzumab or lapatinib, given concomitantly with neoadjuvant chemotherapy. Secondary outcomes included toxic effects, compliance, response rates of the tumour and axillary nodes by physical examination and imaging tests (sonography, mammography, or MRI) before surgery and pathological complete response according to different definitions (ypT0 ypN0/+, ypT0/is ypN0; ypT0/is ypN0/+), and the breast conservation rate. Efficacy was assessed for predefined subgroups as tumour stage (cT1–3 cN0–2 vs cT4 or cN3), HR status (HRpositive [oestrogen-receptor or progesterone receptor, or both, ≥10%] vs HR-negative [both receptors <10%]), and response by best appropriate imaging method to the first four cycles of treatment (complete vs partial vs no change). Long-term follow-up will be reported later and the biomarker analysis has not yet started.

www.thelancet.com/oncology Vol 13 February 2012

Articles

unreduced anti-HER2 treatment with the Kaplan-Meier method and compared treatment groups with the logrank test. A tail-oriented subpopulation treatment effect pattern plot (STEPP)-analysis13 was used to explore the relation between total dose of lapatinib and pathological complete response rate. All statistical analyses were done with SAS version 9.2. A pathological complete response of 26% was assumed in the ECH-TH group (based on the GeparQuattro study4) and expected a pathological complete response of 37% in the ECL-TL group (OR 1·67). According to a two-sided continuity corrected Pearson’s χ² test with α=0·05 and β=0·20, a sample size of 594 patients was needed. An interim safety analysis of the first 30 patients having received at least two cycles of EC was done to optimise the supportive treatment.10 This study is registered with ClinicalTrials.gov, number NCT00567554.

The overall clinical response before surgery did not differ between groups (table 2). A non-significant numerical difference in the rate of patients who had breast conserving surgery was noted in favour of the ECH-TH group (table 2). Figure 2 shows the effect of lapatinib versus trastuzumab on pathological complete response rates within clinically relevant subgroups. The effects in subgroups and the overall effect did not differ. When analysing the group of patients who received treatment as planned, 82 (30·7%) of 267 patients in the ECH-TH group and 43 (25·0%) of 172 patients in the ECL-TL group (p=0·24) had a pathological complete response, whereas in patients with less treatment than planned, 11 (27·5%) of 40 patients in the ECH-TH group Category*

Univariable Odds ratio (95% CI)

Multivariable p value

Odds ratio (95% CI) p value

Arm

ECH-TH vs ECL-TL

0·68 (0·47–0·97)

0·03

0·68 (0·47–0·99)

0·04

Role of the funding source

Age (years)

<40 vs ≥40

1·49 (0·88–2·53)

0·14

1·49 (0·87–2·56)

0·15

GBG Forschungs GmbH was the legal sponsor of the study. The study design and the protocol were written by GvM, MU, and the members of the neoadjuvant subboard of GBG and AGO B and were reviewed by the pharmaceutical funders. These funders had no role in the collection, analysis, or interpretation of the data. Only the study statistician (VN) had access to the raw data. The report was first drafted by GvM and reviewed by all authors and the funders. The corresponding author and MU had final responsibility for the decision to submit for publication.

cT

cT1–3 vs cT4

1·15 (0·73–1·82)

0·54

1·05 (0·64–1·70)

0·85

cN

Negative vs positive

0·78 (0·53–1·15)

0·21

0·69 (0·46–1·03)

0·07

Histological type Ductal/others vs lobular 1·53 (0·56–4·19)

0·41

1·63 (0·57–4·67)

0·36

Grading

G1/2 vs G3

1·47 (1·02–2·11)

0·04

1·35 (0·92–1·99)

0·12

HR status

Negative vs positive

0·49 (0·34–0·70)

<0·0001

0·52 (0·35–0·76)

0·001

Results From Nov 7, 2007, to July 9, 2010, 620 patients were enrolled at 126 centres in Germany and one centre in Switzerland into the HER2-positive group of the GeparQuinto study (figure 1). Of these patients, 309 were randomly assigned to the ECH-TH group and 311 to the ECL-TL group. Baseline patient’s disease characteristics were well balanced between groups (table 1). The median tumour size was 40 mm (range 10–300) by palpation in both groups. Two patients in the ECH-TH group and three patients in the ECL-TL group did not start chemotherapy because of withdrawal of consent or immediate surgery. As predefined in the protocol, these patients were excluded from the analysis. 93 (30·3%) of 307 patients in the ECH-TH group and 70 (22·7%) of 308 patients in the ECL-TL group had a pathological complete response (ypT0, ypN0; OR 0·68 95% CI 0·47–0·97; p=0·04; tables 2, 3). This inferior response rate for ECL-TL was also noted when other definitions of pathological complete responses were used. Fewer tumour regressions were observed in the ECL-TL group than in the ECH-TH group (p=0·02). The results remained unchanged when corrected for relevant baseline characteristics in a multivariable model (table 3). www.thelancet.com/oncology Vol 13 February 2012

ECH-TH=epirubicin, cyclophosphamide, trastuzumab–docetaxel, trastuzumab. ECL-TL=epirubicin, cyclophosphamide, lapatinib–docetaxel, lapatinib. HR=hormone receptor. The multivariable analysis shows the treatment effect adjusted by these baseline parameters. *First group is the reference.

Table 3: Odds ratios for pathological complete response according to subgroups

Subgroup

Odds ratio (95% CI)

Age (years) <40 ≥40 T stage cT1–3 cT4 N stage cN0 cN+ Stage Operable Locally advanced (cT4 or cN3) Histological type Ductal or other Lobular Grade 1 or 2 3 HR Negative Positive Overall 0·2

Test for interaction 0·41

0·46 (0·17–1·26) 0·73 (0·49–1·07) 0·87 0·67 (0·45–0·997) 0·72 (0·32–1·65) 0·63 0·60 (0·32–1·13) 0·72 (0·46–1·13) 0·42 0·63 (0·42–0·94) 0·92 (0·40–2·15) 0·18 0·65 (0·45–0·94) 2·67 (0·33–21·3) 0·45 0·78 (0·46–1·32) 0·59 (0·36–0·97) 0·82 (0·50–1·36) 0·53 (0·31–0·91) 0·68 (0·47–0·97) 0·4

0·6

0·8 1·0

Higher pCR with trastuzumab

0·40

2·0 Higher pCR with lapatinib

Figure 2: Odds ratios for achieving a pCR according to subgroups pCR=pathological complete response. HR=hormone receptor. The analyses for stage and HR status were preplanned and stratified.

139

Articles

250

500

Theoretical daily dose of lapatinib (mg) 750

1000

1250

60

Pathological complete response (%)

50

40

30

20

10

0 20 000

40 000

60 000

80 000

100 000

120 000

140 000

160 000

180 000

200 000

220 000

Mean cumulative total dose of lapatinib (mg)

Figure 3: STEPP analysis on the correlation of mean cumulative total dose of lapatinib with pathological complete response rate Daily lapatinib doses of 700 mg, 1000 mg, and 1250 mg correspond to cumulative total doses of 117 600 mg, 168 000 mg, and 210 000 mg, respectively. Blue lines are the 95% CIs for the data (red line).

See Online for appendix

140

and 27 (19·9%) of 136 patients in the ECL-TL group had a pathological complete response (p=0·41). A STEPP analysis on mean cumulative total dose of lapatinib showed low pathological complete response rates with a daily dose of less than 700 mg, but constant rates for daily doses of 700–1250 mg (figure 3). The planned chemotherapy cycles were completed in 275 (90%) of patients in the ECH-TH group and 257 (83%) in the ECL-TL group. Additionally, 11 (4%) of patients stopped trastuzumab and 51 (17%) stopped lapatinib although chemotherapy was continued (figure 1). This different adherence occurred mainly during docetaxel cycles after four cycles of EC-H or EC-L. Chemotherapy dose delays were necessary in 120 (39%) of 307 patients in the ECH-TH group and 116 (38%) of 308 patients in the ECL-TL group (p=0·74). Dose reductions of the cytotoxic agents were necessary in 43 (14%) of 307 patients in the ECH-TH group and 61 (20%) of 308 patients in the ECLTL group (p=0·07) and of the targeted agent in three (1%) and 99 (32%) patients in each group, respectively (p<0·0001). These dose reductions of lapatinib occurred during EC in 67 (22%) patients and during docetaxel in 38 (12%) patients. Median duration of anti-HER2treatment without dose reductions was significantly longer with trastuzumab (24 weeks) than with lapatinib (21 weeks; p<0·0001), despite a significant increase in treatment duration without dose reductions of lapatinib after implementation of the protocol amendment (24 weeks vs 21 weeks, p=0·0005; appendix). Almost all

patients showed toxic effects of grades 3–4. ECL-TL treatment showed more frequent non-haematological toxic effects (grade 3–4), especially diarrhoea and skin rash, whereas ECH-TH was associated with more oedema and dyspnoea (table 4). A non-significant excess of congestive heart failures was noted in the ECL-TL group (one vs seven). 70 serious adverse events were reported in the ECH-TH group and 87 in the ECL-TL group. No patients died during the treatment phase of this study.

Discussion This neoadjuvant phase 3 study compared two anti-HER2 targeted agents. Pathological complete response rates were significantly lower with lapatinib treatment than with trastuzumab, irrespective of the definitions of pathological complete response that were used. These results confirm the efficacy of a neoadjuvant regimen containing trastuzumab.2–4,14,15 Lower pathological complete response rates in the ECLTL group might be explained by a lower capability of tyrosine-kinase-inhibitor lapatinib to block the HER2 pathway compared with the antibody, trastuzumab. Trastuzumab might have additional anti-tumour efficacy by inducing an immune response via antibody-derived cellular cytotoxicity.16 Furthermore, lower drug exposure could be another explanation for lower activity of lapatinib. Lapatinib, but also the simultaneous chemotherapy, was frequently discontinued, or the doses reduced. Because of toxic effects, a protocol amendment www.thelancet.com/oncology Vol 13 February 2012

Articles

ECH-TH (n=307)

ECL-TL (n=308)

p value

Anaemia

ECH-TH (n=307)

ECL-TL (n=308)

p value

(Continued from previous column)

1–4

292 (96·4%)

288 (94·4%)

0·33

Hand-foot syndrome

3–4

6 (2·0%)

9 (3·0%)

0·60

1–3

Leucopenia

117 (38·5%)

134 (43·5%)

0·22

11 (3·6%)

15 (4·9%)

0·55

97 (31·9%)

169 (54·9%)

<0·0001

2 (0·7%)

22 (7·1%)

<0·0001

147 (48·4%)

128 (41·6%)

0·10

2 (0·7%)

3 (1·0%)

1·0

3

1–4

288 (95%)

286 (93·5%)

0·49

Skin rash (acneiform)

3–4

208 (68·6%)

189 (61·8%)

0·09

1–3

Neutropenia

3

1–4

269 (90·3%)

250 (85·0%)

0·06

Nail changes

3–4

237 (79·5%)

222 (75·5%)

0·28

1–3

Febrile neutropenia

3

Any

22 (7·2%)

30 (9·7%)

0·31

1–4

175 (57·6%)

155 (50·3%)

0·07

1–4

91 (30·0%)

101 (33·1%)

0·43

3–4

6 (2·0%)

7 (2·3%)

1·0

3–4

3 (1·0%)

9 (3·0%)

0·14

Fever 1–4

64 (21·1%)

51 (16·6%)

0·18

1–4

14 (4·7%)

18 (6·2%)

0·47

3–4

10 (3·3%)

7 (2·3%)

0·47

3–4

3 (1·0%)

1 (0·3%)

0·62

Thromboembolic events

Thrombopenia

Bilirubin

Alanine aminotransferase

Sensory neuropathy

1–4

16 (5·3%)

13 (4·2%)

0·57

1–4

178 (59·3%)

173 (58·6%)

0·87

3–4

6 (2·0%)

7 (2·3%)

1·0

3–4

7 (2·3%)

4 (1·4%)

0·54

Bleeding 54 (17·8%)

46 (14·9%)

Aspartate aminotransferase

1–4

0·38

1–4

138 (46·0%)

138 (46·8%)

0·87

3–4

3–4

2 (0·7%)

2 (0·7%)

1·0

Dyspnoea 1–4

90 (29·6%)

66 (21·4%)

0·02

1–4

50 (16·7%)

37 (12·8%)

0·20

3–4

4 (1·3%)

4 (1·3%)

1·0

3–4

2 (0·7%)

1 (0·3%)

1·0

Arterial hypertension 1–4

25 (8·2%)

24 (7·8%)

0·88

1–4

221 (72·7%)

221 (71·8%)

0·86

3–4

1 (0·3%)

3 (1·0%)

0·62

3–4

14 (4·6%)

11 (3·6%)

0·55

Cardiovascular disorders (except congestive heart failure) 1–4

33 (10·9%)

23 (7·5%)

0·16

1–4

104 (34·2%)

125 (40·6%)

0·11

3–4

4 (1·3%)

2 (0·6%)

0·45

3–4

6 (2·0%)

5 (1·6%)

0·77

Congestive heart failure 1–4

1 (0·3%)

7 (2·3%)

0·07

1–4

144 (47·4%)

231 (75·0%)

<0·0001

3–4

0

2 (0·3%)

0·5

3–4

8 (2·6%)

36 (11·7%)

<0·0001

LVEF <50% and >10 points drop from baseline

1–4

221 (72·7%)

230 (74·9%)

0·58

Other non-haematological adverse events

3–4

12 (3·9%)

18 (5·9%)

0·35

1–4

255 (83·1%)

255 (82·8%)

1·0

3–4

41 (13·4%)

52 (16·9%)

0·26

Alkaline phosphatase

Nausea

Vomiting

Diarrhoea

Mucositis

NA

Allergic reactions 1–4

48 (15·8%)

52 (16·9%)

0·74

3–4

3 (1·0%)

5 (1·6%)

0·72

1–4

119 (39·1%)

88 (28·7%)

3–4

5 (1·6%)

4 (1·3%)

0·75

1–4

244 (80·3%)

244 (79·2%)

0·76

3–4

21 (6·9%)

30 (9·7%)

0·24

Oedema 0·006

0

0

4 (1·4%)

1 (0·4%)

NA

0·43

Data are n (%) unless otherwise specified. LVEF=Left ventricular ejection fraction. NA=not applicable. *Grades of maximum severity per patient are based on the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI-CTC) Version 3.0, except congestive heart failure, which was based on New York Heart Association (NYHA) classification.

Table 4: Clinically relevant adverse events by maximum grade*

Fatigue

(Continues in next column)

was made, in which the dose of lapatinib was decreased from 1250 mg to 1000 mg per day for all the patients who were subsequently enrolled. However, when patients without dose reductions were compared, a non-significant www.thelancet.com/oncology Vol 13 February 2012

difference in the pathological complete rates with comparable ORs remained in favour of trastuzumab. Additionally, the STEPP analysis showed no dose– response relation for daily lapatinib doses between 700 mg and 1250 mg (figure 3). Despite the large CIs, this STEPP analysis also does not support the hypothesis that differences in the numbers of patients with dose 141

Articles

Panel: Research in context Systematic review At the time we planned our study, no trial reporting on lapatinib for the treatment of early breast cancer existed (information confirmed also by the manufacturer). Several authors of this report were also involved in the planning of the NeoALTTO17 trial that addresses a similar question on the use of lapatinib. Interpretation As shown in an overview of studies in table 5, four studies17,18,19,20 on lapatinib as part of neoadjuvant treatment for breast cancer have been reported since our trial started. The GeparQuinto trial is the only phase 3 study providing level 1 evidence for the inferiority of lapatinib compared with trastuzumab. All other phase 2 studies,17,19,20 except the CHER-LOB study,18 support this finding with at least numerically lower pathological complete response rates for the lapatinib-containing treatment than for trastuzumab-containing treatments.

reductions resulted in the differing pathological complete response rates noted in the two treatment groups. So far, no head-to-head comparison of these two antiHER2 agents has been available to allow a direct comparison of their efficacy and safety (panel). Concomitantly to the GeparQuinto study, results from the neoadjuvant NeoALTTO study17 were reported; however, the study design differed from GeparQuinto’s in duration of chemotherapy (12 vs 24 weeks), use of anthracycline (adjuvant vs neoadjuvant), choice of taxane (paclitaxel vs docetaxel), a window of anti-HER2 treatment without

Chemotherapy backbone

chemotherapy (yes vs no), dose of lapatinib (1500 mg daily vs 1250–1000 mg daily), and sample size per arm (≤154 patients vs ≤311 patients). Treatment of lapatinib plus lapatinib-paclitaxel (L-PL) suggested a lower pathological complete response rate (ypT0/is, ypN0) than was noted with trastuzumab plus trastuzumab-paclitaxel (H-PH), but the highest pathological complete response rate was achieved when both the HER2 inhibitors were used together with paclitaxel (HL-PHL group). The CHER-LOB study,18 which used a similar three-drug approach, but a weekly paclitaxel plus FEC chemotherapy schedule, reported similar early results for the dual antiHER2 blockade. The pathological complete response rate (ypT0/is, ypN0) of the trastuzumab-containing group of GeparQuinto seemed to be higher than the rate noted in NeoALTTO17 (44·0% vs 27·6%). Apart from potential variations in the study population, longer duration of chemotherapy and use of anthracyclines in combination with trastuzumab might explain the higher rates in GeparQuinto. In fact, the NeoSphere study21 reported a pathological complete response rate as low as 21·5% with four cycles of docetaxel plus trastuzumab. By contrast, the 36-week regimen used in the NOAH study2 achieved a pCR rate of 38·0% similar to that seen in the GeparQuinto study. Dual blockade of HER2—eg, by trastuzumab plus lapatinib—together with a long duration and anthracycline-containing chemotherapy might reach even higher pathological complete response rates. This hypothesis was suggested by a small phase 2 trial19 including 78 evaluable patients that reported a pathological response rate of 74·0% (table 5). Neoadjuvant chemotherapy aims to maximise the chance for a pathological complete response. We

GeparQuinto GEICAM 2006-1420

NeoALTTO17

CHER-LOB18 Holmes19

NeoSphere21

NOAH2

P

P-FEC

T

AP-P-CMF

EC-T

EC-T

Duration (weeks)

24

24

12+6

24

FEC-P 24+2

12

30

Anthracycline containing

Yes

Yes

No

Yes

Yes

No

Yes 117

Arm with trastuzumab alone Number

307

50

145

36

26

107

Pathological complete response* (n)

137

24

40

9

nk

23

nk

48·0%

27·6%

25·0%

54·0%

21·5%

38·0%

Percentage

44·6%

Arms with lapatinib alone 308

52

150

39

29

na

na

Pathological complete response* (n)

Number

93

12

30

10

nk

na

na

Percentage

30·2%

24·0%

20·0%

26·3%

29·0%

na

na

na

na

HL

HL

HL

HPt

na

na

na

145

45

23

107

na

Arms with two anti-HER2-directed agents Agents Number Pathological complete response* (n)

na

na

68

21

nk

42

na

Percentage

na

na

46·9%

46·7%

74·0%

39·3%

na

A=doxorubicin. C=cyclophosphamide. T=docetaxel. E=epirubicin. F=fluorouracil. H=trastuzumab. L=lapatinib. M=methotrexate. P=paclitaxel. Pt=pertuzumab. na=not applicable. nk=not known. *Pathological complete response rate defined as ypT0/is, ypN0.

Table 5: Comparison of trials of trastuzumab alone, lapatinib alone, and two anti-HER2-directed agents given simultaneously with short or long duration chemotherapy, with or without anthracyclines

142

www.thelancet.com/oncology Vol 13 February 2012

Articles

therefore intended to expose the tumour for as long as possible to the synergistic effect of chemotherapy and trastuzumab before surgery. Simultaneous use with anthracyclines was therefore necessary. An increased cardiac toxicity in unselected patients not monitored for cardiac function was reported in a pivotal study in metastatic breast cancer.22 However, recent studies in metastatic disease, as well as neoadjuvant studies, with stringent cardiac inclusion criteria and close cardiac monitoring, have reported only low cardiac toxicity rates.2,4,23 The GeparQuinto study now provides further evidence on the feasibility of this approach. Recently, these trial results have led to an extension of the label to use trastuzumab simultaneously with anthracyclinecontaining neoadjuvant chemotherapy.24 The GeparQuinto study has several strengths, such as the use of an established, neoadjuvant chemotherapy regimen containing anthracyclines and taxanes, the large sample size, and the central verification of pathology reports by a trained reviewer. No central review of the surgical specimen was done. We used an algorithm to assess patient eligibility that is concordant with our national treatment guidelines for neoadjuvant chemotherapy. Neoadjuvant chemotherapy has a long tradition in Germany, and the eligibility criteria of this study were broader than in other trials from European countries, or in the USA. We used a definition for pathological complete response that best discriminates patients with favourable and unfavorable long-term outcome.25 We also provided information on the degree of tumour regression, although this score provides only limited prognostic information in addition to the ypT, ypN stage.25 Practical, logistical, and financial reasons did not allow us to complete central testing for HER2 status before study entry. The absence of a third arm investigating the combination of trastuzumab and lapatinib, as well as the unblinded assessment of efficacy endpoints are potential weaknesses. The run-in phase provided early tolerability data on the combination of lapatinib with epirubicin, which would otherwise have needed a separate phase 1–2 study, and would have postponed this phase 3 study for at least 2 years. However, the run-in phase did not provide information early enough to indicate an immediate dose reduction of lapatinib. On the basis of these findings, lapatinib should not be used outside of clinical trials as single anti-HER2treatment in combination with neoadjuvant chemotherapy. Because all patients in the ECL-TL group received trastuzumab for 1 year after surgery, survival data will provide information on a prolonged sequential treatment with two anti-HER2-directed treatments. Currently, the highest pathological complete response rates can be achieved when anti-HER2-directed agents are combined with a 6-month duration of anthracyclinetaxane-containing chemotherapy. www.thelancet.com/oncology Vol 13 February 2012

Contributors MU, SL, HE, J-UB, JHi, CH, BG, CJ, JHu, VN, and GvM had the idea for and designed the study. MU, SL, JB, HE, MK, J-UB, JHi, DS, PAF, RK, HT, CH, BG, MR, CJ, JHu, TK, and GvM provided study materials and obtained data. MU, HE, SL, VN, and GvM analysed and interpreted data. MU, SL, VN, and GvM wrote the report. All authors reviewed the report. Conflicts of interest BG received honoraria from AstraZeneca and GlaxoSmithKline. GvM received honoraria, consulting fees, and research grants from Roche, GlaxoSmithKline, and Sanofi-Aventis. HE received travel grants from Roche and GlaxoSmithKline and received payment for board membership from Roche. HT received honoraria from AstraZeneca, GlaxoSmithKline, Roche, Sanofi-Aventis, and Novartis. JB received consulting fees and research funding from GlaxoSmithKline. JHu received consulting fees and honoraria from Roche, Sanofi-Aventis, and GlaxoSmithKline and a research grant from GlaxoSmithKline. J-UB received honoraria from GlaxoSmithKline, Roche, and Sanofi-Aventis. MK received honoraria from AstraZeneca, GlaxoSmithKline, Novartis, Pfizer, Roche, and Sanofi-Aventis. PAF received honoraria, consulting fees, and research grants from Novartis. RK received payment for board membership from AstraZeneca and GlaxoSmithKline. SL received honoraria from Roche. TK received honoraria from Roche, GlaxoSmithKline, Novartis, and AstraZeneca. All other authors declare that they have no conflicts of interest. Acknowledgments The conduct of the study was funded by an unrestricted financial grant from GlaxoSmithKline, Roche, and Sanofi-Aventis, Germany, and lapatinib was provided free of charge. We thank Sabine Kleinefeld for support in data management, Andrea Maisch for medical site support, Konstantin Reißmüller and Mathias Uhlig for clinical project management, and the members of the independent data-monitoring committee, Anthony Howell (chair), Monica Castiglione, Renate Haidinger, Dirk Hasenclever, and Dieter Hölzer. Results from this trial were presented in part at the 33rd San Antonio Breast Cancer Symposium, Dec 8–12, 2010, San Antonio, TX, USA. References 1 Kaufmann M, von Minckwitz G, Bear HD, et al. Recommendations from an international expert panel on the use of neoadjuvant (primary) systemic treatment of operable breast cancer: new perspectives 2006. Ann Oncol 2007; 18: 1927–34. 2 Gianni L, Eiermann W, Semiglazov V, et al. Neoadjuvant chemotherapy with trastuzumab followed by adjuvant trastuzumab versus neoadjuvant chemotherapy alone, in patients with HER2-positive locally advanced breast cancer (the NOAH trial): a randomised controlled superiority trial with a parallel HER2-negative cohort. Lancet 2010; 375: 377–84. 3 Untch M, Fasching PA, Konecny GE, et al. Pathologic complete response after neoadjuvant chemotherapy plus trastuzumab predicts favorable survival in human epidermal growth factor receptor 2-overexpressing breast cancer: results from the TECHNO trial of the AGO and GBG Study Groups. J Clin Oncol 2011; 29: 3351–57. 4 Untch M, Rezai M, Loibl S, et al. Neoadjuvant treatment with trastuzumab in HER2-positive breast cancer: results from the GeparQuattro study. J Clin Oncol 2010; 28: 2024–31. 5 Cameron D, Casey M, Oliva C, Newstat B, Imwalle B, Geyer CE. Lapatinib plus capecitabine in women with HER-2-positive advanced breast cancer: final survival analysis of a phase III randomized trial. Oncologist 2010; 15: 924–34. 6 Di Leo A, Gomez HL, Aziz Z, et al. Phase III, double-blind, randomized study comparing lapatinib plus paclitaxel with placebo plus paclitaxel as first-line treatment for metastatic breast cancer. J Clin Oncol 2008; 26: 5544–52. 7 Johnston S, Pippen J Jr, Pivot X, et al. Lapatinib combined with letrozole versus letrozole and placebo as first-line therapy for postmenopausal hormone receptor-positive metastatic breast cancer. J Clin Oncol 2009; 27: 5538–46. 8 Blackwell KL, Burstein HJ, Storniolo AM, et al. Randomized study of lapatinib alone or in combination with trastuzumab in women with ErbB2-positive, trastuzumab-refractory metastatic breast cancer. J Clin Oncol 2010; 28: 1124–30.

143

Articles

9

10

11

12

13 14

15

16

17

144

Pocock SJ, Simon R. Sequential treatment assignment with balancing for prognostic factors in the controlled clinical trial. Biometrics 1975; 31: 102–15, von Minckwitz G, Eidtmann H, Loibl S, et al. Integrating bevacizumab, everolimus, and lapatinib into current neoadjuvant chemotherapy regimen for primary breast cancer. Safety results of the GeparQuinto trial. Ann Oncol 2011; 22: 301–06. Sinn HP, Schmid H, Junkermann H, et al. Histologic regression of breast cancer after primary (neoadjuvant) chemotherapy. Geburtshilfe Frauenheilkd 1994; 54: 552–58. [in German]. Breslow NE, Day NE. Statistical methods in cancer research, volume 1: the analysis of case-control studies. IARC Scientific Publications, No. 32. Lyon, France: International Agency for Research on Cancer, 1980. Bonetti M, Gelber RD. Patterns of treatment effects in subsets of patients in clinical trials. Biostatistics 2004; 5: 465–81. Buzdar AU, Ibrahim NK, Francis D, et al. Significantly higher pathologic complete remission rate after neoadjuvant therapy with trastuzumab, paclitaxel, and epirubicin chemotherapy: results of a randomized trial in human epidermal growth factor receptor 2-positive operable breast cancer. J Clin Oncol 2005; 23: 3676–85. Buzdar AU, Valero V, Ibrahim NK, et al. Neoadjuvant therapy with paclitaxel followed by 5-fluorouracil, epirubicin, and cyclophosphamide chemotherapy and concurrent trastuzumab in human epidermal growth factor receptor 2-positive operable breast cancer: an update of the initial randomized study population and data of additional patients treated with the same regimen. Clin Cancer Res 2007; 13: 228–33. Barok M, Isola J, Palyi-Krekk Z, et al. Trastuzumab causes antibody-dependent cellular cytotoxicity-mediated growth inhibition of submacroscopic JIMT-1 breast cancer xenografts despite intrinsic drug resistance. Mol Cancer Ther 2007; 6: 2065–72. Baselga J, Bradbury I, Eidtmann H, et al, on behalf of the NeoALTTO Study Team. Lapatinib with trastuzumab for HER2-positive early breast cancer (NeoALTTO): a randomised, open-label, multicentre, phase 3 trial. Lancet 2012; published online Jan 17. DOI:10.1016/S0140-6736(11)61847-3.

18

19

20

21

22

23

24

25

Guarneri V, Frassoldati A, Bottini A, et al. Preoperative chemotherapy plus trastuzumab, lapatinib or both in HER2 positive operable breast cancer: results of the randomized phase II CHER-LOB study. J Clin Oncol (in press). Holmes FA, Nagarwala YM, Espina VA, et al. Correlation of molecular effects and pathologic complete response to preoperative lapatinib and trastuzumab, separately and combined prior to neoadjuvant breast cancer chemotherapy. Proc Am Soc Clin Oncol 2011; 29 (suppl): abstr 506. Alba E, Albanell J, de la Haba J, et al. Lapatinib vs. trastuzumab in combination with standard EC-D chemotherapy in the neoadjuvant treatment of HER2+ patients. Results from the GEICAM 2006-14 phase II randomized trial. Cancer Res 2011; 71 (24 suppl): abstr PD07–04. Gianni L, Pienkowski T, Im Y-H, et al. Efficacy and safety of neoadjuvant pertuzumab and trastuzumab in women with locally advanced, inflammatory, or early HER2-positive breast cancer (NeoSphere): a randomised multicentre, open-label, phase 2 trial. Lancet Oncol 2012; 13: 25–32. Slamon DJ, Leyland-Jones B, Shak S, et al. Use of chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer that overexpresses HER2. N Engl J Med 2001; 344: 783–92. Untch M, Muscholl M, Tjulandin S, et al. First-line trastuzumab plus epirubicin and cyclophosphamide therapy in patients with human epidermal growth factor receptor 2-positive metastatic breast cancer: cardiac safety and efficacy data from the herceptin, cyclophosphamide, and epirubicin (HERCULES) trial. J Clin Oncol 2010; 28: 1473–80. European Medicines Agency. Herceptin—trastuzumab. http://www. ema.europa.eu/docs/en_GB/document_library/Summary_of_ opinion/human/000278/WC500117821.pdf (accessed Nov 25, 2011). von Minckwitz G, Untch M, Blohmer JU, et al. Definition and impact of pathological complete response on prognosis after neoadjuvant chemotherapy in various intrinsic breast cancer subtypes. J Clin Oncol (in press).

www.thelancet.com/oncology Vol 13 February 2012