AbstractBACKGROUND The addition of CAP to maintenance BEV demonstrated statistically significant and clinically relevant improvements in progression-free survival (PFS [primary endpoint]; HR 0.38 [95% CI 0.27–0.55]; log-rank p<0.001) and overall survival (OS [secondary endpoint]; HR 0.43 [95% CI 0.26–0.69]; log-rank p<0.001) in patients (pts) without disease progression (PD) on initial first-line BEV–DOC for HER2-negative mBC in the IMELDA trial. This benefit was achieved despite the smaller than planned sample size due to premature recruitment discontinuation because of regulatory withdrawal of BEV–DOC.METHODS Pts with HER2-negative measurable mBC, ECOG PS <2, and no prior chemotherapy for mBC were eligible. After 3–6 cycles of BEV–DOC, pts without PD were randomized to either BEV alone or BEV–CAP (BEV 15 mg/kg q3w; CAP 1000 mg/m2 bid d1–14 q3w) until PD. PROs (secondary endpoint) were assessed using the EORTC QLQ-C30 completed at screening (before BEV–DOC), at randomization to CAP vs no CAP, then every 3 cycles until PD, and at (but not beyond) PD. Analyses of mean change from randomization were prespecified. A 28-day window around the scheduled timepoints from randomization was applied to maximize the number of questionnaires available for analysis. Exploratory post hoc analyses included mixed-model repeated measures (MMRM; modeling weighted treatment effect from randomization across all available timepoints) and responder analyses using the global health status/QoL subscale. Pts were categorized as having improved (≥10-point increase), stable (change of <10 points), or worsened (≥10-point decrease) scores from randomization [Osoba, 2005].RESULTS Adherence with questionnaire completion was 65–85% for all assessment timepoints during the first year of maintenance therapy. MMRM analysis of the global health status/QoL subscale showed no difference between the treatment arms in change from randomization (least squares mean estimate 0.40 [95% CI –6.07 to 6.87]). Similar results were observed for other subscales, including the diarrhea symptom subscale.No. of pts (%)BEV (N=94)BEV–CAP (N=91)Week 9aN=51N=59Improved15 (29.4)17 (28.8)Stable26 (51.0)34 (57.6)Week 18aN=29N=57Improved11 (37.9)12 (21.1)Stable12 (41.4)30 (52.6)Week 27aN=23N=43Improved7 (30.4)16 (37.2)Stable12 (52.2)20 (46.5)Week 36aN=15N=35Improved4 (26.7)14 (40.0)Stable9 (60.0)17 (48.6)aNo. of patients with completed questionnaires at both randomization and the respective week. Only weeks with ≥10 pts in both arms shown.CONCLUSIONS The IMELDA sample size was smaller than planned but protocol adherence with PRO completion was relatively high. Prespecified change from randomization and exploratory post hoc MMRM analyses of PROs suggest that the clinically meaningful PFS and OS benefit from adding CAP to BEV is achieved while maintaining QoL, with no difference between BEV and BEV–CAP treatments. Responder analyses over time showed improved or stable global health status/QoL scores in the majority of pts at each timepoint in both treatment arms.Citation Format: Dinesh Doval, Saverio Cinieri, Hakan Bozcuk, Jean-Yves Pierga, Kadri Altundag, Xiaojia Wang, Sudeep Gupta, Guillermo Lopez Vivanco, Vineet Gupta, Ewa Chmielowska, Jose Bines, Philippe Montcuquet, Alfred Namour, Emilio Alba, Giorgio Mustacchi, Paulo Cortes, Sabine de Ducla, Ulrich Freudensprung, Lesley Fallowfield, Joseph Gligorov. Exploratory post hoc analyses of patient-reported outcomes (PROs) in the IMELDA randomized phase III trial: Maintenance bevacizumab (BEV) ± capecitabine (CAP) after initial first-line BEV plus docetaxel (DOC) for HER2-negative metastatic breast can [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P2-12-16.