Cisplatin plus gemcitabine versus paclitaxel plus gemcitabine as first-line therapy for metastatic triple-negative breast cancer (CBCSG006): a randomised, open-label, multicentre, phase 3 trial

Articles

Cisplatin plus gemcitabine versus paclitaxel plus gemcitabine as first-line therapy for metastatic triple-negative breast cancer (CBCSG006): a randomised, open-label, multicentre, phase 3 trial Xi-Chun Hu*, Jian Zhang*, Bing-He Xu, Li Cai, Joseph Ragaz, Zhong-Hua Wang, Bi-Yun Wang, Yue-E Teng, Zhong-Sheng Tong, Yue-Yin Pan, Yong-Mei Yin, Chang-Ping Wu, Ze-Fei Jiang, Xiao-Jia Wang, Gu-Yin Lou, Dong-Geng Liu, Ji-Feng Feng, Jian-Feng Luo, Kang Sun, Ya-Jia Gu, Jiong Wu, Zhi-Min Shao

Summary Lancet Oncol 2015; 16: 436–46 Published Online March 18, 2015 http://dx.doi.org/10.1016/ S1470-2045(15)70064-1 See Comment page 357 *Contributed equally Department of Medical Oncology (Prof X-C Hu MD, J Zhang MD, Z-H Wang MD, B-Y Wang MD), Department of Radiology (Prof Y-J Gu MD), Department of Breast Surgery (Prof J Wu MD, Prof Z-M Shao MD), Fudan University Shanghai Cancer Centre, Collaborative Innovation Centre for Cancer Medicine, Shanghai, China; Department of Oncology, Shanghai Medical College, (Prof X-C Hu, J Zhang, Z-H Wang, B-Y Wang, Prof Y-J Gu, Prof J Wu, Prof Z-M Shao); Department of Biostatistics, School of Public Health, Fudan University, Shanghai, China (J-F Luo PhD); Cancer Institute and Hospital, Chinese Academy of Medical Sciences, Beijing, China (Prof B-H Xu MD); Department of Medical Oncology, Harbin Medical University Cancer Hospital, Harbin, China (Prof L Cai MD); Faculty of Medicine, School of Population and Public Health, University of British Columbia, Vancouver, BC, Canada (Prof J Ragaz FRCP); Department of Medical Oncology, The First Hospital of China Medical University, Shenyang, China (Prof Y-E Teng MD); Tianjin Medical University Cancer Institute and Hospital, Tianjin, China (Prof Z-S Tong MD); Department of Medical Oncology, The First Hospital, Anhui Medical University, Hefei, China (Prof Y-Y Pan MD); Department of Oncology, The First Affiliated Hospital of

436

Background Platinum chemotherapy has a role in the treatment of metastatic triple-negative breast cancer but its full potential has probably not yet been reached. We assessed whether a cisplatin plus gemcitabine regimen was noninferior to or superior to paclitaxel plus gemcitabine as first-line therapy for patients with metastatic triple-negative breast cancer. Methods For this open-label, randomised, phase 3, hybrid-designed trial undertaken at 12 institutions or hospitals in China, we included Chinese patients aged 18–70 years with previously untreated, histologically confirmed metastatic triple-negative breast cancer, and an ECOG performance status of 0–1. These patients were randomly assigned (1:1) to receive either cisplatin plus gemcitabine (cisplatin 75 mg/m² on day 1 and gemcitabine 1250 mg/m² on days 1 and 8) or paclitaxel plus gemcitabine (paclitaxel 175 mg/m² on day 1 and gemcitabine 1250 mg/m² on days 1 and 8) given intravenously every 3 weeks for a maximum of eight cycles. Randomisation was done centrally via an interactive web response system using block randomisation with a size of eight, with no stratification factors. Patients and investigator were aware of group assignments. The primary endpoint was progression-free survival and analyses were based on all patients who received at least one dose of assigned treatment. The margin used to establish non-inferiority was 1·2. If non-inferiority of cisplatin plus gemcitabine compared with paclitaxel plus gemcitabine was achieved, we would then test for superiority. The trial is registered with ClinicalTrials.gov, number NCT01287624. Findings From Jan 14, 2011, to Nov 14, 2013, 240 patients were assessed for eligibility and randomly assigned to treatment (120 in the cisplatin plus gemcitabine group and 120 in the paclitaxel plus gemcitabine group). 236 patients received at least one dose of assigned chemotherapy and were included in the modified intention-to-treat analysis (118 per group). After a median follow-up of 16·3 months (IQR 14·4–26·8) in the cisplatin plus gemcitabine group and 15·9 months (10·7–25·4) in the paclitaxel plus gemcitabine group, the hazard ratio for progression-free survival was 0·692 (95% CI 0·523–0·915; pnon-inferiority<0·0001, psuperiority=0·009, thus cisplatin plus gemcitabine was both non-inferior to and superior to paclitaxel plus gemcitabine. Median progression-free survival was 7·73 months (95% CI 6·16–9·30) in the cisplatin plus gemcitabine group and 6·47 months (5·76–7·18) in the paclitaxel plus gemcitabine group. Grade 3 or 4 adverse events that differed significantly between the two groups included nausea (eight [7%] vs one [<1%]), vomiting (13 [11%] vs one [<1%]), musculoskeletal pain (none vs ten [8%]), anaemia (39 [33%] vs six [5%]), and thrombocytopenia (38 [32%] vs three [3%]), for the cisplatin plus gemcitabine compared with the paclitaxel plus gemcitabine groups, respectively. In addition, patients in the cisplatin plus gemcitabine group had significantly fewer events of grade 1–4 alopecia (12 [10%] vs 42 [36%]) and peripheral neuropathy (27 [23%] vs 60 [51%]), but more grade 1–4 anorexia (33 [28%] vs 10 [8%]), constipation (29 [25%] vs 11 [9%]), hypomagnesaemia (27 [23%] vs five [4%]), and hypokalaemia (10 [8%] vs two [2%]). Serious drug-related adverse events were seen in three patients in the paclitaxel plus gemcitabine group (interstitial pneumonia, anaphylaxis, and severe neutropenia) and four in the cisplatin plus gemcitabine group (pathological bone fracture, thrombocytopenia with subcutaneous haemorrhage, severe anaemia, and cardiogenic syncope). There were no treatment-related deaths. Interpretation Cisplatin plus gemcitabine could be an alternative or even the preferred first-line chemotherapy strategy for patients with metastatic triple-negative breast cancer. Funding Shanghai Natural Science Foundation.

Introduction Triple-negative breast cancer is pathologically defined as oestrogen receptor (ER)-negative, progesterone receptor (PR)-negative, and HER2-negative disease,

and accounts for 12–17% of all breast cancers.1 Despite general improvements in the management of breast cancer, triple-negative breast cancer represents a continuing challenge because, when compared with www.thelancet.com/oncology Vol 16 April 2015

Articles

other subtypes, it is associated with a higher frequency of recurrence, shorter disease-free survival, and poorer overall survival, despite similar therapeutic approaches to other breast cancers. The median distant disease-free interval for relapsed triple-negative breast cancer is about 1–2 years, and the median survival for metastatic triple-negative breast cancer is about 1 year.2,3 Because hormone receptors or HER2 receptors are not expressed in triple-negative breast cancer, cytotoxic chemotherapy is the main systemic treatment, with no dedicated biological agents available. Specifically, targeted approaches such as DNA repair agents, PARP inhibitors, EGFR inhibitors, antiangiogenic agents, or checkpoint kinase 1 (Chk1) inhibitors (with or without chemotherapy) have not produced substantial improvements in triplenegative breast cancer outcomes.4,5 The molecular signature of triple-negative breast cancer generally overlaps with basal-like breast cancer, with concordance of 70–90%.6 Additionally, the reported7 prevalence of BRCA mutations in unselected patients with triple-negative breast cancer is 11·2%, several times higher than that in the wider population with breast cancer. Both BRCA-associated breast cancer and sporadic triple-negative or basal-like breast cancer have characteristics consistent with abnormal DNA repair and genome-wide instability,8,9 which lend support to the use of DNA-damaging compounds such as platinums. During the past several years, several studies have investigated platinum-based therapy in the neoadjuvant or metastatic triple-negative breast cancer setting. Results from two randomised phase 2 studies, GeparSixto10 and CALGB 40603,11 showed that the addition of carboplatin to neoadjuvant therapy for triple-negative breast cancer significantly increases the proportion of patients who achieved a pathological complete response. However, no other randomised trials were reported either before or during this study, in the metastatic triple-negative breast cancer setting comparing platinum with other standard chemotherapeutic agents such as anthracyclines or taxanes. Compared with carboplatin, cisplatin has shown superior efficacy in the neoadjuvant setting of locally advanced triple-negative breast cancer, with a higher proportion of patients achieving a pathological complete response and a significant improvement in overall survival.12 Also in the metastatic setting, cisplatin has resulted in more responses than carboplatin.13 In 2010, our group reported a phase 2 study of gemcitabine and cisplatin as first-line combination therapy in patients with metastatic triple-negative breast cancer.14 Thus, based on the better control of triple-negative breast cancer in both the neoadjuvant and metastatic settings with cisplatin, and strong preclinical evidence for a synergistic effect of cisplatin and gemcitabine15,16 in addition to the result of our phase 2 study,14 we initiated the Chinese Breast Cancer Study Group (CBCSG) 006 trial, to investigate whether a cisplatin plus gemcitabine www.thelancet.com/oncology Vol 16 April 2015

regimen given as first-line treatment in metastatic triplenegative breast cancer would be non-inferior or even superior to a standard regimen of paclitaxel plus gemcitabine.17

Methods Study design and participants We did this prospective, open-label, multicentre, randomised, phase 3 trial at 12 institutions or hospitals in China (certified by the Chinese Food and Drug Administration) within the Chinese Breast Cancer Study Group (appendix). The protocol of this study is available online. Eligible patients were Chinese patients with breast cancer aged 18–70 years who had triple-negative breast cancer histologically confirmed at the primary tumour, with clinical, imaging, histological or cytological evidence of metastatic (stage IV) disease. ER, PR, and HER2 status were identified locally at each participating centre by immunohistochemistry of patient tumour sections. The immunohistochemical cutoff for ERnegative and PR-negative status was 10% or less staining in the nuclei. HER2-negative status was a score of 0 or 1 by immunohistochemistry, or the absence of HER2 amplification (ratio <2·2) by fluorescence in-situ hybridisation analysis. The establishment of ER, PR, and HER2 status was done preferentially with tissue from recurrent or metastatic lesions, but primary breast cancer tissue was also acceptable. The main inclusion criteria were metastatic triplenegative breast cancer with no previous chemotherapy for metastatic disease, at least one extracranial measurable lesion by MRI or CT according to Response Evaluation Criteria in Solid Tumors (RECIST, version 1.1), and Eastern Cooperative Oncology Group performance status of 0–1. Previous use of taxanes in the adjuvant or neoadjuvant setting was acceptable if relapse occurred at least 6 months after completion of the primary locoregional and neoadjuvant or adjuvant therapies. Main exclusion criteria were patients with non-triple-negative breast cancer, child-bearing potential but unwillingness to use adequate contraception, symptomatic or unstable CNS metastases, life expectancy of less than 3 months, participation in other trials within 4 weeks before enrolment, reduced haematological, hepatic, or renal functions, congestive heart failure, any concurrent medical disorder that could potentially increase the risk of toxic effects, or other invasive malignant diseases within the past 5 years except excised basal cell skin carcinoma and cervical carcinoma in situ. Independent ethics committees of the participating centres approved the study protocol. We did the study in accordance with the Declaration of Helsinki. All patients provided written informed consent.

Nanjing Medical University, Nanjing, China (Prof Y-M Yin MD); Department of Oncology, The Third Affiliated Hospital of Soochow University, Changzhou, China (Prof C-P Wu MD); Beijing 307 Hospital of PLA, Beijing, China (Prof Z-F Jiang MD); Department of Medical Oncology, Zhejiang Cancer Hospital, Hangzhou, China (Prof X-J Wang MD); Breast Cancer Centre, Ruijin Hospital, Shanghai Jiaotong University School of Medicine, Shanghai, China (G-Y Lou MD); Cancer Centre, Sun Yat-sen University, Guangzhou, China (D-G Liu MD); Jiangsu Cancer Hospital, Nanjing, China (Prof J-F Feng MD); Biostatistics, Incyte Corporation, Wilmington DE, USA (K Sun PhD) Correspondence to: Prof Xi-Chun Hu, Department of Oncology, Shanghai Medical College, Fudan University, Shanghai 200032, China [email protected] See Online for appendix For the study protocol see http://www.cbcs.cn/uploadimg/ protocol%20of%20CBCSG006. pdf

Randomisation and masking We randomly assigned eligible patients (1:1) to receive either cisplatin plus gemcitabine or paclitaxel plus 437

Articles

gemcitabine. Randomisation was done centrally via a block randomisation of size eight, with no stratification factors, via an interactive web-response system. Investigators or research coordinators sent the random assignment forms by fax to the Clinical Research Coordination Office in Fudan University Shanghai Cancer Centre (Shanghai, China). After checking the inclusion criteria, the study coordinator sent the allocated treatment back to the investigator by fax. Patients, investigators, and outcome assessors were aware of treatment group assignment.

Procedures We gave either cisplatin plus gemcitabine (cisplatin 75 mg/m² on day 1; gemcitabine 1250 mg/m² on days 1 and 8) or paclitaxel plus gemcitabine (paclitaxel 175 mg/m² on day 1; gemcitabine 1250 mg/m² on days 1 and 8) intravenously every 3 weeks for a maximum of eight cycles, or until disease progression or intolerable toxic

240 patients randomised

120 assigned to paclitaxel plus gemcitabine

120 assigned to cisplatin plus gemcitabine

2 did not receive allocated intervention 1 did not meet inclusion criteria 1 withdrew consent before intervention

118 received allocated intervention

2 did not receive allocated intervention 1 did not meet inclusion criteria 1 withdrew consent before intervention

118 received allocated intervention

74 discontinued intervention 44 had disease progression 6 had an adverse event 1 died 18 at patient’s decision 4 at physician’s decision 1 unknown

Figure 1: Trial profile

438

77 discontinued intervention 34 had disease progression 15 had an adverse event 0 died 24 at patient’s decision 2 at physician’s decision 2 unknown

40 completed eight cycles 4 still receiving treatment

39 completed eight cycles 2 still receiving treatment

118 included in modified intention-to-treat analysis 107 included in per-protocol analysis 118 included in safety analysis

118 included in modified intention-to-treat analysis 105 included in per-protocol analysis 118 included in safety analysis

effects developed. Before the evidence of protocol-defined disease progression, use of any other anticancer drugs was not recommended. To prevent cisplatin-induced renal toxic effects, we gave hydration treatment for 3 days before initiating the cisplatin plus gemcitabine regimen. Because aprepitant and fosaprepitant were not commercially available in China until 2014, we gave antiemetic measures, such as 5-HT3 receptor antagonists, dexamethasone, and promethazine, to prevent cisplatin-related acute or delayed nausea or vomiting. We also gave standard premedications with dexamethasone and histamine antagonists before paclitaxel to prevent hypersensitivity reactions. Prophylactic treatment with granulocytecolony stimulating factor was not allowed. Febrile neutropenia was managed according to institutional treatment guidelines in China. Concurrent treatment with bisphosphonates was allowed. The following recommendations for chemotherapy dose reductions were applied: in patients who had grade 4 haematological, grade 3 or 4 non-haematological, or other protocol-specified toxic effects, gemcitabine, cisplatin, and paclitaxel treatments were interrupted. If the toxic effects resolved to a grade lower than 2, the dose of gemcitabine and cisplatin was restarted at 75% of the dose at the first appearance of the toxic effects and at 50% of the starting dose at the second appearance. For paclitaxel, up to two dose reductions were allowed at 20% reduction of the previous dose. If whole blood count was low, day 8 gemcitabine could be given at a reduced dose or postponed for up to 7 days to allow recovery, otherwise it was discontinued. We removed patients from the study if their disease progressed, protocol-specified unacceptable side-effects occurred, the allocated treatment was delayed for more than 14 days, or if patients withdrew consent. Treatment was permanently discontinued if more than two dose modifications were needed. Patients would discontinue a drug in the allocated combination if a specific severe adverse event was judged to be related to that particular drug (eg, grade ≥2 pneumonitis or haemolytic uraemic syndrome for gemcitabine; creatinine clearance <30 mL/min for cisplatin; or anaphylaxis for paclitaxel). Tumour response was assessed according to RECIST version 1.1 by a team of local investigators and when needed, with independent central assessment, every two cycles until disease progression. After disease progression, information about survival status and further therapy was obtained systematically every 3 months. Adverse events were recorded at each treatment visit, at each follow-up visit, and at the end-of-study visit. We graded these using the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.0.

Outcomes The primary endpoint was progression-free survival, as determined by local investigators and supported with www.thelancet.com/oncology Vol 16 April 2015

Articles

independent central assessment when necessary. A central review of all progression events, however, was not done. Progression-free survival was defined as the time from the date of randomisation to progression or death from any cause, whichever occurred first. We analysed the primary endpoint in patients who received at least one dose of the study treatment (modified intention-to-treat population). Secondary endpoints were overall survival (defined as the time from randomisation to death from any cause), the proportion of patients who had an objective response (defined as a best response of complete or partial response from the start of treatment until disease progression or death, in accordance with RECIST 1.1 criteria), and safety (numbers and proportions of patients in each treatment group who had any adverse event, any adverse event leading to dose reductions, any adverse event leading to discontinuation, any serious adverse event, any treatmentrelated serious adverse event, and any adverse event leading to death). We did planned exploratory subgroup analyses to investigate the association between baseline characteristics and treatment efficacy.

Paclitaxel plus gemcitabine (n=118) Age (years)

48 (43–55)

Cisplatin plus gemcitabine (n=118) 47 (42–57)

<40

19 (16%)

21 (18%)

≥40

99 (84%)

97 (82%)

ECOG performance status 0

44 (37%)

37 (31%)

1

74 (63%)

81 (69%)

Menopausal status Premenopausal

64 (54%)

55 (47%)

Postmenopausal

54 (46%)

63 (53%)

Time between breast surgery and recurrent disease 6 (5%)

13 (11%)

<1 year

20 (17%)

23 (19%)

≥1 year

92 (78%)

82 (69%)

1

44 (37%)

36 (31%)

2

36 (31%)

36 (31%)

≥3

38 (32%)

46 (39%)

Lung

70 (59%)

64 (54%)

Liver

30 (25%)

30 (25%)

Statistical analysis

Brain

5 (4%)

5 (4%)

From our previous phase 2 trial on cisplatin plus gemcitabine,14 we expected a marginal superiority of the cisplatin plus gemcitabine regimen compared with paclitaxel plus gemcitabine. Therefore, we used a hybrid design as described by Freidlin and colleagues18 to compare the efficacy between the two groups. In this design, the non-inferiority margin (δN) and the modest (not large but sufficient in risk reduction) clinical benefit margin (δM) are specified in terms of the hazard ratio (HR)–ie, the ratio of the hazard of progression or death with cisplatin plus gemcitabine versus paclitaxel plus gemcitabine. Our previous data14 estimated that patients receiving the cisplatin plus gemcitabine regimen would have a 20% reduction in hazard compared with the paclitaxel plus gemcitabine regimen, with a median progression-free survival of 6·2 versus 5·0 months. Thus, for a hybrid design with 80% power at a 0·025 one-sided significance level, the non-inferiority margin was 1·2 and the clinical benefit margin was 0·806; therefore with a 2 year accrual plus 1 year follow-up, we calculated that we would need 215 patients. Accounting for a 10% dropout rate, the final number of patients needed was 236, with 118 patients per group. According to the formula by Freidlin and colleagues,18 at least 199 progression-free survival events were needed for the first non-inferiority analysis. If noninferiority based on the 1·2 margin was achieved, then the prespecified superiority hypothesis would be tested under the null hypothesis that there was no difference in the progression-free survival between the two groups. A formal statistical superiority conclusion of cisplatin plus gemcitabine over paclitaxel plus gemcitabine is based on whether the 95% CI of the HR excludes 1. Without statistical significance of superiority, a result of

Bone

23 (19%)

33 (28%)

www.thelancet.com/oncology Vol 16 April 2015

Primary metastatic

Number of metastatic organ sites

Metastatic sites

Contralateral breast Lymph nodes Pleura

4 (3%)

6 (5%)

71 (60%)

73 (62%)

9 (8%)

13 (11%)

Chest wall or skin

23 (19%)

29 (25%)

Visceral disease

88 (75%)

79 (67%)

116 (98%)

118 (100%)

2 (2%)

0 (0%)

Anthracycline

97 (82%)

98 (83%)

Taxane

77 (65%)

75 (64%)

Anthracycline and taxane

69 (58%)

70 (59%)

Hormonal receptor status Both ER and PR less than <1% positivity Others* Neoadjuvant or adjuvant chemotherapy

Data are median (IQR) or n (%) unless otherwise indicated. *Others included ER-negative and PR 1–9% positive; ER 1–9% positive and PR-negative; and ER 1–9% positive and PR 1–9% positive. ECOG=Eastern Cooperative Oncology Group. ER=oestrogen receptor. PR=progesterone receptor.

Table 1: Baseline patient characteristics

Paclitaxel plus gemcitabine (n=118)

Cisplatin plus gemcitabine (n=118)

648

654

Cycles Total Median (IQR)

6 (4–8)

6 (4–8)

Relative dose intensity Gemcitabine

775·2 mg/m² per week (756·7–833·3), 93%

757·6 mg/m² per week (730·5–831·3), 91%

Paclitaxel/cisplatin

57·3 mg/m² per week (55·5–58·3), 98%

23·6 mg/m² per week (23·4–24·9), 94%

Data are median (IQR), %.

Table 2: Treatment exposure

439

Articles

Before disease progression

≥1 chemotherapy agent

After disease progression

Paclitaxel plus gemcitabine (n=118)

Cisplatin plus gemcitabine (n=118)

p value

Paclitaxel plus gemcitabine (n=118)

Cisplatin plus gemcitabine (n=118)

p value

9 (8%)

11 (9%)

0·640

56 (47%)

51 (43%)

0·513

Vinorelbine

4 (3%)

0 (0%)

0·130

26 (22%)

20 (17%)

0·324

Capecitabine

4 (3%)

5 (4%)

1·000

18 (15%)

18 (15%)

1·000

Oxaliplatin

1 (<1%)

0 (0%)

1·000

9 (8%)

13 (11%)

0·370

Other fluoropyrimidine derivatives

1 (<1%)

0 (0%)

1·000

6 (5%)

12 (10%)

0·141

Gemcitabine

1 (<1%)

3 (3%)

0·614

1 (<1%)

2 (2%)

1·000

Anthracyclines

0 (0%)

0 (0%)

4 (3%)

1 (<1%)

0·366

Taxanes

0 (0%)

3 (3%)

0·006

Cyclophosphamide

0 (0%)

0 (0%)

Cisplatin

2 (2%)

4 (3%)

Bevacizumab

0 (0%)

1 (<1%)

·· 0·245

2 (2%)

12 (10%)

4 (3%)

1 (<1%)

0·366

0·679

21 (18%)

1 (<1%)

<0·0001

1·000

2 (2%)

4 (3%)

0·679

··

Other types of agents used in one or two patients per group included carboplatin, mitomycin, and vinblastine.

Table 3: Post-study chemotherapy before and after disease progression

non-inferiority would be concluded. A multiple-comparison adjustment was not needed to test the two hypotheses.19 We did a per-protocol analysis in patients who had no major protocol violations and who received at least six cycles of allocated intervention, and in patients discontinued treatment before completion of six cycles of treatment because of disease progression or death. For the safety analysis, we assessed data for all patients who received at least one dose of study treatment. We assessed progression-free survival and overall survival, with 95% CI for median time to event, with the Kaplan-Meier method. We estimated the HRs and corresponding 95% CIs with the Cox proportional hazard regression model. Objective responses were analysed with Pearson’s χ² test (or Fisher’s exact test when appropriate). We calculated the dose intensity as the ratio of the total dose divided by the total treatment duration. We used Epidata (version 3.1) and SPSS (version 19.0) for all statistical analyses. The trial is registered with ClinicalTrials.gov, number NCT01287624.

Role of the funding source The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results Between Jan 14, 2011 to Nov 14, 2013, we enrolled and randomly assigned 240 patients to treatment. Four did not receive allocated intervention (two in each group) because they either did not meet inclusion criteria or they withdrew consent. Thus, 236 patients with metastatic triple-negative breast cancer (median age 47 years, IQR 42–56) received at least one dose of assigned chemotherapy (118 assigned to cisplatin plus gemcitabine and 118 to paclitaxel plus 440

gemcitabine) and were included in the modified intentionto-treat population20 for the efficacy and safety analysis (figure 1). Overall, 21 patients in the paclitaxel plus gemcitabine group and 19 in the cisplatin plus gemcitabine group were histologically confirmed as triple-negative on the metastatic tumour tissue, and the rest were assessed on the primary lesion. Baseline characteristics were relatively well balanced between treatment groups (table 1). Most patients were aged 40 years or older and had visceral disease across both groups. More than two-thirds of patients had recurrent disease 1 year or more after the primary breast surgery and most had received adjuvant or neoadjuvant chemotherapy. Patients from both groups had a median of six treatment cycles (table 2). Relative dose intensities are shown in table 2. Allocated treatments were discontinued because of disease progression in 44 (37%) of 118 patients in the paclitaxel plus gemcitabine group versus 34 (29%) of 118 patients in the cisplatin plus gemcitabine group (figure 1); 40 (34%) of 118 patients in the paclitaxel plus gemcitabine group and 39 (33%) of 118 patients in the cisplatin plus gemcitabine group completed the allocated eight cycles. Dose reductions were needed for 36 (30%) of 118 patients in the gemcitabine plus paclitaxel group and 47 (40%) of 118 patients in the gemcitabine plus cisplatin group. Similar proportions (<10%) of patients in both groups received other anticancer drugs before disease progression (table 3), mainly because of consent withdrawal and ethical consideration of breast cancer-related symptoms. After the documentation of progression, almost half of the patients (56 [47%] of 118 patients in the paclitaxel plus gemcitabine group and 51 [43%] of 118 in the cisplatin plus gemcitabine group) received second-line or third-line chemotherapy regimens, with 21 (18%) of 118 patients in the paclitaxel plus gemcitabine group given the cisplatin-based regimen and 12 (10%) of 118 in the cisplatin www.thelancet.com/oncology Vol 16 April 2015

Articles

Discussion Our findings suggest that the platinum-based chemotherapy regimen cisplatin plus gemcitabine given as firstline therapy for patients with metastatic triple-negative breast cancer was both non-inferior to, and superior to the standard paclitaxel plus gemcitabine regimen in terms of progression-free survival. Both regimens were well tolerated with expected toxicity profiles. To our knowledge, our study is the first randomised trial to www.thelancet.com/oncology Vol 16 April 2015

A Cisplatin plus gemcitabine Paclitaxel plus gemcitabine

Progression-free survival (%)

100 80

HR 0·692 (95% CI 0·523–0·915); psuperiority=0·009

60 40 20 0

0 Number at risk Cisplatin plus 118 gemcitabine Paclitaxel plus 118 gemcitabine

3

6

9

12

15

18

21

24

27

30

33

36

102

71

42

18

11

6

3

2

2

1

0

0

93

56

23

10

4

3

1

1

1

0

0

0

B

100

HR 0·902 (95% CI 0·605–1·344); plog-rank=0·611

80 Overall survival (%)

plus gemcitabine group given the taxane-based regimen as salvage therapy (table 3). The data cutoff for progression-free survival analysis was March 27, 2014, with a total of 201 events, 97 in the cisplatin plus gemcitabine group and 104 in the paclitaxel plus gemcitabine group. The median follow-up periods were 16·3 months (IQR 14·4–26·8) in the cisplatin plus gemcitabine group and 15·9 months (10·7–25·4) in the paclitaxel plus gemcitabine group. Median progression-free survival, in the modified intention-to-treat population, was 7·73 months (95% CI 6·16–9·30) for cisplatin plus gemcitabine and 6·47 months (5·76–7·18) for paclitaxel plus gemcitabine (figure 2). With regard to the study primary endpoints, the data not only rule out non-inferiority, but also support a superiority of cisplatin plus gemcitabine compared with paclitaxel plus gemcitabine based on the 95% CI of the HR excluding 1 (figure 2 and table 4). Similar results were seen in the per-protocol population (table 4). Exploratory analyses of progression-free survival across the main clinical subgroups showed HRs that consistently favoured the cisplatin plus gemcitabine group (figure 3), objective responses in both the modified intention-totreat and per-protocol populations were also significantly more common in the cisplatin plus gemcitabine than the paclitaxel plus gemcitabine group (table 4). Overall survival data are immature, with 49 deaths (42%) from any cause in the 118 patients in the paclitaxel plus gemcitabine group and 48 deaths (41%) in the 118 patients in the cisplatin plus gemcitabine group (figure 2). All 236 patients in the safety population had at least one adverse event. Neutropenia was the most frequent adverse event of grade 3 or more, with similar incidences in both groups. Grade 3 and 4 adverse events were significantly higher in the cisplatin plus gemcitabine group compared with the paclitaxel plus gemcitabine group for nausea, vomiting, anaemia, and thrombocytopenia, whereas events of grade 3 and 4 musculoskeletal pain were higher in the paclitaxel plus gemcitabine group (table 5). Additionally, patients in the cisplatin plus gemcitabine group had significantly fewer events of grade 1–4 alopecia and peripheral neuropathy, but more grade 1–4 anorexia, constipation, hypomagnesaemia, and hypokalaemia (table 5). Serious drug-related adverse events were seen in three patients in the paclitaxel plus gemcitabine group and four in the cisplatin plus gemcitabine group (table 6). No treatment-related deaths were reported.

60 40 20 0 0

Number at risk Cisplatin plus 118 gemcitabine Paclitaxel plus 118 gemcitabine

3

6

9

12

15

18

21

24

27

30

33

36

Time from random assignment (months) 116

99

85

60

43

33

22

15

11

6

2

0

115

99

76

57

42

27

15

13

9

4

2

0

Figure 2: Kaplan-Meier plot of progression-free survival (A) and overall survival (B) Data for the modified intention-to-treat population. HR=hazard ratio.

compare these two regimens in the metastatic triplenegative breast cancer setting (panel). Our results were lent support by preclinical data,21,22 two randomised neoadjuvant clinical studies,10,11 and one randomised phase 2 trial in the metastatic setting,23 which all provide some level of evidence for platinum activity in triplenegative breast cancer. Consistent with these results, the platinum-based cisplatin plus gemcitabine regimen in our randomised study yielded a significantly higher number of objective responses compared with the paclitaxel plus gemcitabine. Although small, the improvement of median progression-free survival noted in our study is meaningful and has potential cost implications because cisplatin is substantially lower in price than taxanes in China. The improved progressionfree survival in the cisplatin plus gemcitabine group was also longer than the 4·6 months seen in the control group treated with first-line gemcitabine and carboplatin,24 and also approached that seen with chemotherapy plus bevacizumab25 in the same setting. 441

Articles

Modified intention-to-treat population Paclitaxel plus gemcitabine (n=118)

Cisplatin plus gemcitabine (n=118)

Per-protocol population HR (95% CI)

p value

Paclitaxel plus gemcitabine (n=107)

HR (95% CI)

Cisplatin plus gemcitabine (n=105)

p value

Response Complete response

4 (3%)

8 (7%)

··

··

4 (4%)

8 (8%)

··

··

Partial response

54 (46%)

68 (58%)

··

··

54 (50%)

63 (60%)

··

··

Stable disease

41 (35%)

26 (22%)

··

··

33 (31%)

23 (22%)

··

··

Progressive disease

16 (14%)

10 (8%)

··

··

16 (15%)

10 (10%)

··

··

3 (3%)

6 (5%)

··

··

0 (0%)

1 (1%)

··

58 (49%)

76 (64%)

··

58 (54%)

71 (68%)

··

104 (88%)

97 (82%)

95 (89%)

87 (83%)

Missing data or not assessable* Overall response

0·018

·· 0·045

Progression-free survival Number of events Median progression-free survival, months (95% CI)

6·47 (5·76–7·18)

··

7·73 (6·16–9·30)

··

0·692 (0·523–0·915)

pnon-inferiority <0·0001, psuperiority =0·009

6·47 (5·64–7·30)

··

8·00 (6·63–9·37)

··

0·681 (0·507–0·913)

pnon-inferiority <0·0001, psuperiority =0·009

Overall survival Number of events

49 (42%)

Median overall survival, months (95% CI)

48 (41%)

··

··

··

··

42 (39%)

0·902 (0·605–1·344) plog-rank =0·611

43 (41%)

··

··

··

0·921 (0·601–1·412)

plog-rank =0·707

HR=hazard ratio.*Tumour assessment data were missing or not assessable for response because of consent withdrawal or receiving other anticancer drugs before the first assessment in the modified intentionto-treat and per-protocol populations. Overall survival data (months) are too immature to report here.

Table 4: Efficacy of study treatment

HR (95% CI)

psuperiority

n

PFS (months) Cisplatin plus Paclitaxel plus gemcitabine gemcitabine

All Age, years <40 ≥40 ECOG performance score 0 1 Menopausal status Premenopausal Postmenopausal Time between breast surgery and recurrent disease Primary metastatic <1 year ≥1 year Number of metastatic organ site 1 2 ≥3 Visceral Yes No Previous treatment with anthracycline Yes No Previous treatment with taxane Yes No

0·1

0·2

0·3

0·4

0·5 0·6 0·7 δM 0·9 1·0 1 δN

Favours cisplatin plus gemcitabine

1·5

0·69 (0·52–0·92)

0·009

236

7·73

6·47

0·41 (0·20–0·81) 0·77 (0·57–1·05)

0·011 0·097

40 196

8·77 7·50

5·60 6·63

0·65 (0·40–1·05) 0·70 (0·49–0·99)

0·079 0·042

81 155

8·70 6·70

7·87 5·70

0·72 (0·48–1·07) 0·65 (0·43–0·98)

0·101 0·037

119 117

6·60 8·90

6·47 6·13

0·33 (0·10–1·06) 0·98 (0·54–1·78) 0·64 (0·46–0·90)

0·062 0·942 0·009

18 46 172

6·97 4·77 8·70

2·47 4·20 6·63

0·54 (0·32–0·90) 0·66 (0·39–1·11) 0·78 (0·50–1·23)

0·017 0·118 0·287

80 72 84

9·67 8·77 5·90

7·13 6·07 4·27

0·75 (0·54–1·04) 0·61 (0·35–1·05)

0·087 0·075

167 69

7·00 8·00

6·07 6·47

0·67 (0·49–0·91) 0·86 (0·43–1·72)

0·009 0·666

195 41

8·00 6·97

6·50 5·37

0·65 (0·45–0·93) 0·77 (0·49–1·23)

0·017 0·279

152 84

8·00 7·57

6·50 5·87

2·0

Favours paclitaxel plus gemcitabine

Figure 3: Subgroup analysis of progression-free survival Data are medians for modified intention-to-treat population in clinically relevant subgroups. δN (non-inferiority margin)=1·2. δM=0·806. HR=hazard ratio. ECOG=Eastern cooperative oncology group.

442

www.thelancet.com/oncology Vol 16 April 2015

Articles

Paclitaxel plus gemcitabine (n=118)

Cisplatin plus gemcitabine (n=118)

Grade 1–2

Grade 3

Grade 1–2

Grade 3

Leucopenia

53 (45%)

46 (39%)

9 (8%)

50 (42%)

56 (47%)

1 (<1%)

Neutropenia

34 (29%)

42 (36%)

27 (23%)

40 (34%)

52 (44%)

15 (13%)

Grade 4

Grade 4

Haematological

3 (3%)

0

Anaemia

Febrile neutropenia

72 (61%)

··

6 (5%)

0

65 (55%)

··

30 (25%)

3 (3%)

9 (8%)

0

Thrombocytopenia

36 (31%)

2 (2%)

1 (<1%)

36 (31%)

22 (19%)

16 (14%)

Non-haematological Alopecia

42 (36%)

··

··

12 (10%)

··

Nausea

56 (47%)

1 (<1%)

··

85 (72%)

8 (7%)

Vomiting

39 (33%)

1 (<1%)

0

73 (62%)

12 (10%)

Musculoskeletal pain

26 (22%)

Anorexia Stomatitis

9 (8%)

·· 1 (<1%)

··

11 (9%)

0

··

1 (<1%)

0

31 (26%)

2 (2%)

0

1 (<1%)

0

7 (6%)

0

0

Constipation

11 (9%)

0

0

29 (25%)

0

0

Fatigue or asthenia

22 (19%)

2 (2%)

··

33 (28%)

1 (<1%)

··

3 (3%)

2 (2%)

0

2 (2%)

0

0

Dyspnoea

6 (5%)

10 (8%)

··

Pyrexia

15 (13%)

0

0

17 (14%)

1 (<1%)

0

Peripheral neuropathy

58 (49%)

2 (2%)

0

27 (23%)

0

0

Rash

21 (18%)

0

0

12 (10%)

0

0

2 (2%)

0

0

4 (3%)

1 (<1%)

0

Hearing toxicity Pain (other than musculoskeletal pain)

10 (8%)

0

··

13 (11%)

1 (<1%)

··

3 (3%)

0

0

1 (<1%)

1 (<1%)

0

Anaphylaxis

··

2 (2%)

1 (<1%)

··

0

0

Infection

2 (2%)

1 (<1%)

0

0

0

0

Oedema in limbs

7 (6%)

1 (<1%)

··

5 (4%)

0

··

Pneumonitis

Laboratory-assessed items Increased ALT

19 (16%)

1 (<1%)

0

10 (8%)

0

0

Increased AST

13 (11%)

0

0

6 (5%)

0

0

Increased bilirubin

6 (5%)

0

0

2 (2%)

1 (<1%)

0

Increased creatinine

0

0

0

0

0

0

1 (<1%)

0

30 (25%)

0

0

Hyperglycaemia

31 (26%)

Hypomagnesaemia

5 (4%)

0

0

26 (22%)

1 (<1%)

0

Hyponatraemia

2 (2%)

0

0

2 (2%)

1 (<1%)

0

Hypokalaemia

2 (2%)

0

0

6 (5%)

4 (3%)

0

Hypercalcaemia

0

0

0

0

0

1 (<1%)

1 (<1%)

0

19 (16%)

0

0

0

0

1 (<1%)

4 (3%)

0

Hypocalcaemia Hypophosphataemia

17 (14%) 7 (6%)

ALT=alanine aminotransferase. AST=aspartate aminotransferase. ··=not applicable. Data are n (%) from the safety population. There were no grade 5 drug-related adverse events.

Table 5: Drug-related adverse events

There was little difference in the number of deaths between groups at this analysis, although the data are somewhat immature at this timepoint. Furthermore, patients were treated with a range of other drugs after progression, and although the assessment of chemotherapy outcomes after progression was not protocol-specified, there was crossover between the two groups postprogression. As a result, the effect of the new regimen on overall survival might not be accurately assessed. After our trial was designed and initiated, the KCSGBR07-02 trial26 showed that HER2-negative metastatic breast cancer would benefit from maintenance therapy in www.thelancet.com/oncology Vol 16 April 2015

terms of both progression-free survival and overall survival. However, even with the publication of results from this trial, most guidelines for breast oncology discourage maintenance strategy with chemotherapy for stage IV breast cancer, because of the toxic effects in a setting without evidence for curability. Additionally, maintenance chemotherapy has many unresolved issues, such as the choice of combination versus monotherapy, oral versus intravenous delivery, and others. The tolerability profile of cisplatin plus gemcitabine and paclitaxel plus gemcitabine regimens was as expected in our study, and when compared with the scientific literature, 443

Articles

Grade 3

Grade 4

Action taken

Cisplatin plus gemcitabine group (n=118) Pathological bone fracture

1

0

Treatment continuation

Thrombocytopenia with subcutaneous haemorrhage

0

1

Dose reduction

Severe anaemia

0

1

Treatment continuation

Cardiogenic syncope

1

0

Treatment discontinuation

Paclitaxel plus gemcitabine group (n=118) Interstitial pneumonia

1

0

Treatment discontinuation

Anaphylaxis

0

1

Treatment discontinuation

Severe neutropenia

0

1

Treatment continuation

Table 6: Serious adverse events

Panel: Research in context Systematic review We searched PubMed for published studies, scientific meeting abstracts for any preliminary results, and ClinicalTrials.gov for any additional open or closed phase 2 or 3 trials using search terms such as “cisplatin”, “carboplatin”, “platinum(s)”, “triple-negative”, “metastatic breast cancer”, and “chemotherapy”. To our knowledge, only one such phase 3 study (TNT, NCT00532727) has been initiated to compare carboplatin with docetaxel in metastatic triple-negative breast cancer. Whether a platinum-based combination regimen is non-inferior or even superior to a more established taxane-based combination regimen as first-line treatment in metastatic triple-negative breast cancer is still unknown. Based on the relevant results obtained, the institutional review board reviewed the appropriateness and ethical and scientific aspects of the study, on which to base the approval of the study. Interpretation Our study has shown both the non-inferiority and superiority, in terms of progressionfree survival, of a novel cisplatin plus gemcitabine regimen compared with a more established standard regimen of paclitaxel plus gemcitabine. Both regimens were well tolerated with different toxicity profiles. These findings suggest that cisplatin plus gemcitabine could be an alternative or even the preferred first-line chemotherapy option for patients with metastatic triple-negative breast cancer.

no new safety concerns were identified. The more frequent grade 3 or 4 adverse events in the cisplatin plus gemcitabine group compared with the paclitaxel plus gemcitabine group were mainly nausea, vomiting, anaemia, and thrombocytopenia, all generally self-limiting and well managed. Patients in the paclitaxel plus gemcitabine group had a higher incidence of musculoskeletal pain, peripheral neuropathy, and alopecia. Several other randomised clinical trials have attempted to establish the role of platinum-based chemotherapies as first-line therapy in the treatment of metastatic triplenegative breast cancer. A randomised phase 2 study with 53 patients by Fan and colleagues23 suggested that platinum-containing regimens might be more effective and better tolerated than regimens that do not contain platinum. Another randomised phase 3 study (TNT)27 compared carboplatin monotherapy with a docetaxel monotherapy in unselected metastatic or recurrent locally advanced triple-negative breast cancer or 444

BRCA1-positive or BRCA2-positive breast cancer. TNT showed similar efficacy between carboplatin and docetaxel in terms of objective responses, progressionfree survival, and overall survival. Although the results of TNT might be less relevant to the population of patients with triple-negative breast cancer, because the dose of docetaxel at 100 mg/m² could be too toxic in strictly palliative settings, and cisplatin might be more effective than carboplatin,12,13 the TNT study nevertheless introduced platinum salts as a potentially important class of agents for metastatic triple-negative breast cancer. Although recommendations for asymptomatic stage IV breast cancer suggest the use of sequential monotherapies in patients with more aggressive breast cancer subtypes (ie, metastatic triple-negative breast cancer), combination chemotherapy is regarded as a superior option. Responses to gemcitabine monotherapy in phase 2 studies have been as low as zero and this drug is therefore not recommended as monotherapy; nevertheless, the addition of gemcitabine to taxanes or platinum agents is reasonable based on potential synergistic effects.17 The rather short median progression-free survival of 3·1 months (and median overall survival of only 12·4 months) after carboplatin given as a monotherapy in the TNT study supports the conclusion that combination chemotherapy in this setting is to be recommended. Molecular biology differences in triple-negative breast cancer might exist across different socioeconomic groups or between people of different ethnic origins, and could contribute to variable response to many therapeutic regimens including platinum-based agents. However, to our knowledge, no evidence exists for significant differences of molecular subtyping or frequency of gene mutations in triple-negative breast cancer (or breast cancer in general) between Asian and non-Asian patients. However, evidence for the universal application of our cisplatin plus gemcitabine regimen can come only from validation trials in other ethnic populations. Other data suggest that chemotherapy outcomes seem to be affected by predictive molecular markers such as germline BRCA1 or BRCA2 mutation status. For example, the subgroup analysis of the breast cancer TNT study27 showed significantly greater proportions of objective responses and better progression-free survival for carboplatin compared with docetaxel restricted to BRCA1 or BRCA2 subsets, with no significance noted for the whole population. These results support the potential of biomarkers such as BRCA1 or BRCA2 to correctly select chemotherapy choices in metastatic triple-negative breast cancer.28 Other tumour-based assays that detect cumulative DNA damage secondary to underlying DNA repair defects, such as the homologous recombination deficiency assay, also have the potential to identify non-BRCA1 or BRCA2 mutation carriers with BRCA-like breast cancer, who might respond to DNA repair-targeted treatment strategies, such as platinum agents.29,30 www.thelancet.com/oncology Vol 16 April 2015

Articles

Because the dichotomised myriad homologous recombination deficiency score in the TNT study did not select for sensitivity to carboplatin over docetaxel,27 the further development of such assays for BRCA, as part of the future prospective clinical trials undertaken on metastatic lesions, would be ideal. These studies could eventually determine the predictive value of these assays to successfully select candidates with metastatic triple-negative breast cancer for or against a specific class of new agents, including platinum-based regimens. Our study has several limitations. First, the latest modification of the College of American Pathologists guidelines with different hormone receptor positivity cutoff values was reported after our trial had been designed and might have led to a different definition of triple-negative breast cancer in our study, and thus to a potential bias. However, only two patients in the paclitaxel plus gemcitabine group and none in the cisplatin plus gemcitabine group did not meet the new definition criteria of triple-negative breast cancer. Second, when our study was designed in 2010, paclitaxel plus gemcitabine with paclitaxel given every 3 weeks was still the schedule recommended by most guidelines, including the National Cancer Comprehensive Network guideline group. Since 2010, several studies have suggested that paclitaxel given once per week is more effective than the regimen given once every 3 weeks used in our study. However, the effect of paclitaxel once per week in combination with gemcitabine has not been fully assessed, and although this might become the desired schedule eventually, it will need future testing in the next generation of trials investigating gemcitabine plus paclitaxel. Third, because this was an investigatorinitiated study, we could not do a totally central assessment of response and progression-free survival because of financial limitations. Finally, central reading of pathology slides and the more detailed identification of triple-negative breast cancer pathological changes were not mandatory in this study. For example, application of the PAM50 intrinsic subtype signature to subclassify the triple-negative breast cancer groups as basal-like and non-basal-like, or differentiated according to BRCA1 or BRCA2 status could be of benefit, to better identify the subpopulation of metastatic triple-negative breast cancer patients who might or might not benefit from the novel regimen. In conclusion, compared with the more established paclitaxel plus gemcitabine regimen, our cisplatin plus gemcitabine regimen could be an alternative or even the preferred first-line chemotherapy strategy for patients with metastatic triple-negative breast cancer. Future work could include a validation of our cisplatin plus gemcitabine regimen in non-Asian populations and the identification of predictive markers for treatment with platinum-based regimens for patients with metastatic triple-negative breast cancer. www.thelancet.com/oncology Vol 16 April 2015

Contributors X-CH was the principal investigator. X-CH, Z-HW, B-YW, JW, and Z-MS conceived the idea for the study. X-CH and JZ obtained the data. X-CH, JZ, J-FL, and KS did the statistical analysis. X-CH, B-HX, LC, Y-ET, Z-ST, Y-YP, Y-MY, C-PW, Z-FJ, X-JW, G-YL, D-GL, and J-FF were the principal investigators at the centres with the highest recruitment. X-CH, JZ, JR, and Y-JG analysed and interpreted the data and wrote the report. Declaration of interests We declare no competing interests. Acknowledgments This study was funded by Shanghai Natural Science Foundation, grant number 12ZR1406300. Gemcitabine was provided by Eli Lilly. We thank Boris Freidlin for his help with the statistical analysis and interpretation. We thank all the patients, their families, and the institutions involved in this study. References 1 Foulkes WD, Smith IE, Reis-Filho JS. Triple-negative breast cancer. N Engl J Med 2010; 363: 1938–48. 2 Dent R, Trudeau M, Pritchard KI, et al. Triple-negative breast cancer: clinical features and patterns of recurrence. Clin Cancer Res 2007; 13: 4429–34. 3 Kassam F, Enright K, Dent R, et al. Survival outcomes for patients with metastatic triple-negative breast cancer: implications for clinical practice and trial design. Clin Breast Cancer 2009; 9: 29–33. 4 Clark O, Botrel TE, Paladini L, Ferreira MB. Targeted therapy in triple-negative metastatic breast cancer: a systematic review and meta-analysis. Core Evid 2014; 9: 1–11. 5 Ma CX, Ellis MJ, Petroni GR, et al. A phase II study of UCN-01 in combination with irinotecan in patients with metastatic triple negative breast cancer. Breast Cancer Res Treat 2013; 137: 483–92. 6 Gelmon K, Dent R, Mackey JR, Laing K, McLeod D, Verma S. Targeting triple-negative breast cancer: optimising therapeutic outcomes. Ann Oncol 2012; 23: 2223–34. 7 Couch FJ, Hart SN, Sharma P, et al. Inherited mutations in 17 breast cancer susceptibility genes among a large triple-negative breast cancer cohort unselected for family history of breast cancer. J Clin Oncol 2014; published online Dec 1. DOI: 10.1200/JCO.2014.57.1414. 8 Andre F, Job B, Dessen P, et al. Molecular characterization of breast cancer with high-resolution oligonucleotide comparative genomic hybridization array. Clin Cancer Res 2009; 15: 441–51. 9 Bergamaschi A, Kim YH, Wang P, et al. Distinct patterns of DNA copy number alteration are associated with different clinicopathological features and gene-expression subtypes of breast cancer. Genes Chromos Cancer 2006; 45: 1033–40. 10 von Minckwitz G, Schneeweiss A, Loibl S, et al. Neoadjuvant carboplatin in patients with triple-negative and HER2-positive early breast cancer (GeparSixto; GBG 66): a randomised phase 2 trial. Lancet Oncol 2014; 15: 747–56. 11 Sikov WM, Berry DA, Perou CM, et al. Impact of the addition of carboplatin and/or bevacizumab to neoadjuvant once-per-week paclitaxel followed by dose-dense doxorubicin and cyclophosphamide on pathologic complete response rates in stage II to III triple-negative breast cancer: CALGB 40603 (Alliance). J Clin Oncol 2015; 33: 13–21. 12 Hurley J, Reis IM, Rodgers SE, et al. The use of neoadjuvant platinum-based chemotherapy in locally advanced breast cancer that is triple negative: retrospective analysis of 144 patients. Breast Cancer Res Treat 2013; 138: 783–94. 13 Isakoff SJ, Goss PE, Mayer EL, et al. TBCRC009: a multicenter phase II study of cisplatin or carboplatin for metastatic triplenegative breast cancer and evaluation of p63/p73 as a biomarker of response. Proc Am Soc Clin Oncol 2011; 29 (suppl): abstr 1025. 14 Wang Z, Hu X, Chen L, et al. Efficacy of gemcitabine and cisplatin (GP) as first-line combination therapy in patients with triplenegative metastatic breast cancer: preliminary results report of a phase II trial. Proc Am Soc Clin Oncol 2010; 28 (suppl 15): abstr 1100. 15 Bergman AM, Ruiz VHV, Veerman G, Kuiper CM, Peters GJ. Synergistic interaction between cisplatin and gemcitabine in vitro. Clin Cancer Res 1996; 2: 521–30. 16 van Moorsel CJ, Kroep JR, Pinedo HM, et al. Pharmacokinetic schedule finding study of the combination of gemcitabine and cisplatin in patients with solid tumors. Ann Oncol 1999; 10: 441–48.

445

Articles

17

18

19

20

21

22

23

24

446

Albain KS, Nag SM, Calderillo-Ruiz G, et al. Gemcitabine plus paclitaxel versus paclitaxel monotherapy in patients with metastatic breast cancer and prior anthracycline treatment. J Clin Oncol 2008; 26: 3950–57. Freidlin B, Korn EL, George SL, Gray R. Randomized clinical trial design for assessing noninferiority when superiority is expected. J Clin Oncol 2007; 25: 5019–23. Koyama T, Westfall PH. Decision-theoretic views on simultaneous testing of superiority and noninferiority. J Biopharm Stat 2005; 15: 943–55. Lewis JA. Statistical principles for clinical trials (ICH E9): an introductory note on an international guideline. Stat Med 1999; 18: 1903–42. Tassone P, Tagliaferri P, Perricelli A, et al. BRCA1 expression modulates chemosensitivity of BRCA1-defective HCC1937 human breast cancer cells. Br J Cancer 2003; 88: 1285–91. Leong CO, Vidnovic N, DeYoung MP, Sgroi D, Ellisen LW. The p63/ p73 network mediates chemosensitivity to cisplatin in a biologically defined subset of primary breast cancers. J Clin Invest 2007; 117: 1370–80. Fan Y, Xu BH, Yuan P, et al. Docetaxel-cisplatin might be superior to docetaxel-capecitabine in the first-line treatment of metastatic triple-negative breast cancer. Ann Oncol 2013; 24: 1219–25. O’Shaughnessy J, Schwartzberg L, Danso MA, et al. Phase III study of iniparib plus gemcitabine and carboplatin versus gemcitabine and carboplatin in patients with metastatic triple-negative breast cancer. J Clin Oncol 2014; 32: 3840–47.

25

26

27

28 29

30

Miles DW, Dieras V, Cortes J, Duenne AA, Yi J, O’Shaughnessy J. First-line bevacizumab in combination with chemotherapy for HER2-negative metastatic breast cancer: pooled and subgroup analyses of data from 2447 patients. Ann Oncol 2013; 24: 2773–80. Park YH, Jung KH, Im SA, et al. Phase III, multicenter, randomized trial of maintenance chemotherapy versus observation in patients with metastatic breast cancer after achieving disease control with six cycles of gemcitabine plus paclitaxel as first-line chemotherapy: KCSG-BR07-02. J Clin Oncol 2013; 31: 1732–39. Tutt A, Ellis P, Kilburn L, et al. The TNT trial: a randomized phase III trial of carboplatin (C) compared with docetaxel (D) for patients with metastatic or recurrent locally advanced triple negative or BRCA1/2 breast cancer (CRUK/07/012). San Antonio Breast Cancer Symposium, 2014: abstr S3-01. Peshkin BN, Alabek ML, Isaacs C. BRCA1/2 mutations and triple negative breast cancers. Breast Dis 2010; 32: 25–33. Lips EH, Mulder L, Hannemann J, et al. Indicators of homologous recombination deficiency in breast cancer and association with response to neoadjuvant chemotherapy. Ann Oncol 2011; 22: 870–76. Telli ML, Jensen KC, Abkevich V, et al. Homologous recombination deficiency (HRD) score predicts pathologic response following neoadjuvant platinum-based therapy in triple-negative and BRCA1/2 mutation-associated breast cancer. San Antonio Breast Cancer Symposium, 2012: abstr PD09-04.

www.thelancet.com/oncology Vol 16 April 2015