In Europe, two bevacizumab-containing regimens (bevacizumab/capecitabine and bevacizumab plus paclitaxel) are approved by regulatory authorities based on superior progression-free survival (PFS) demonstrated with both of these regimens versus chemotherapy alone (Milleret al., NEJM 2007; Robert etal., JCO 2010). However, until now, the relative efficacy of these two regimens in terms of overall survival (OS) has been unclear.
Final OS results of the phase III TURANDOT study (NCT00600340) comparing these regimens were reported at the European Cancer Congress (ECC 2015) taking place in Vienna, Austria, f25 - 29 September, 2015. The trial demonstrated non-inferior OS with bevacizumab/capecitabine compared with bevacizumab/paclitaxel in the primary analysis (stratified analysis in the per-protocol (PP) population). However, results of unstratified sensitivity analyses were inconsistent with the primary results.
Primary objective met but sensitivity analyses showed inconsistent results
Thomas Brodowicz, Medical University Vienna in Vienna, Austria presented findings from the open-label randomised phase III TURANDOT trial during the Proffered Paper Session on advanced breast cancer, held on 26 September 2015. The trial was designed to demonstrate non-inferior OS with first-line bevacizumab/capecitabine versus bevacizumab/paclitaxel in patients with locally recurrent/metastatic breast cancer (LR/mBC).
TURANDOT enrolled 564 patients from 51 centres in 12 countries. Eligible patients had HER2-negative LR/mBC and had received no prior chemotherapy for advanced disease. Patients were randomised to receive either bevacizumab/paclitaxel (intravenous bevacizumab at 10 mg/kg on days 1 and 15 plus intravenous paclitaxel at 90 mg/m2 on days 1, 8 and 15, repeated every 4 weeks; n=285) or bevacizumab/capecitabine (intravenous bevacizumab at 15 mg/kg on day 1 plus oral capecitabine at 1000 mg/m2 twice daily on days 1–14, repeated every 3 weeks; n=279) until disease progression or unacceptable toxic effects. The stratification factors were country, oestrogen/progesterone receptor status and menopausal status.
The primary objective of TURANDOT was to demonstrate non-inferior OS with first-line bevacizumab/capecitabine versus bevacizumab/paclitaxel in the PP population by rejecting the null hypothesis of inferiority (hazard ratio [HR] ≥1.33) using a stratified Cox proportional hazard model. Sensitivity analyses were performed in the intent to treat (ITT) population and also using an unstratified Cox model in both populations. Subgroup analyses in clinically relevant populations were prespecified.
Previously reported results of the prespecified interim analysis were inconclusive and non-inferior OS could not be demonstrated. The stratified HR was 1.04 (97.5% repeated confidence interval (CI) ‒∞ to 1.69, crossing the boundary for non-inferiority) (Langet al., Lancet Oncol 2013). However, progression-free survival (PFS; secondary endpoint) was significantly superior with bevacizumab/paclitaxel compared with bevacizumab/capecitabine.
Professor Brodowicz reported the primary final OS analysis (stratified analysis, PP population), which was performed after deaths in 183 (69%) of 266 patients in the bevacizumab/paclitaxel arm and 201 (76%) of 265 patients in the bevacizumab/capecitabine arm. The stratified HR was 1.02 (97.5% repeated CI: ‒∞ to 1.26; p=0.007), meeting the criteria for non-inferiority. Therefore the primary objective of the TURANDOT trial was met.
Median OS was 30.2 months in the bevacizumab/paclitaxel arm versus 26.1 months in the bevacizumab/capecitabine arm.
Results of the stratified ITT analysis were consistent with the PP results; however, results from stratified and unstratified analyses were inconsistent. In the unstratified analysis in the PP population, the HR was 1.13 (97.5% repeated CI: ‒∞ to 1.39).
The PFS benefit with bevacizumab/paclitaxel over bevacizumab/capecitabine was sustained in this final analysis, as was the superior objective response rate.
Hypothesis-generating subgroup findings
Subgroup analyses in several clinically-relevant subgroups, including analyses according to hormone receptor status, ECOG performance status, age, presence of visceral metastases and prior anthracycline and/or taxane therapy, showed no difference in OS between the treatment arms, consistent with the primary analysis in the overall population. However, in the subgroups of premenopausal females and patients with a body surface area ≥1.8 m2, OS appeared to be more favourable with bevacizumab/paclitaxel than with bevacizumab/capecitabine. Further exploratory analyses suggested significant interactions between these factors and treatment.
Professor Brodowicz explained that the marked heterogeneity in these subgroup analyses might be used when considering treatment options, but requires confirmation in further studies.
Dr Fatima Cardoso who discussed the study results said that TURANDOT was a multi-centric, open-label, randomised, non-inferiority phase III study with good statistical design. Bevacizumab was part of both arms, so studying the role of bevacizumab in advanced breast cancer was not a goal of the study.
Dr Cardoso agreed with the study conclusions and meeting the primary endpoint. However, regarding sub-group analysis she cautioned that only 95 pre-menopausal patients were included. This study is important because it shows that capecitabine is as good as paclitaxel in HER-2 negative disease, strengthening the ABC Guidelines regarding preferred regimens for first-line chemotherapy.
The study authors noted that although the TURANDOT trial met its primary objective in the primary final OS analysis, there was inconsistency between the stratified and unstratified analyses. The authors plan to explore this inconsistency, as well as the heterogeneity between subgroups. Professor Brodowicz concluded that based on findings of the TURANDOT trial, selection of a particular bevacizumab-containing regimen should take into account each individual patient’s treatment priorities.
1800 Final results for overall survival (OS), the primary endpoint of the CECOG.
TURANDOT prospective randomised trial evaluating bevacizumab–paclitaxel (BEV–PAC) vs BEV–capecitabine (CAP) for HER2-negative locally recurrent/metastatic breast cancer (LR/mBC).
The study was sponsored by the Central European Cooperative Oncology Group (CECOG) and funded by Roche.