Cumulative Exposure to Bevacizumab (BV) After Disease Progression (PD) Correlates with Survival in Non-Small Cell Lung Cancer (NSCLC): A Time-Dependent Analysis of the Aries Observational Cohort Study

ABSTRACT Background Duration of BV use appears to contribute to treatment (tx) efficacy in some cancers (eg, CRC, ovarian). A phase 3 mCRC trial met its 1° endpoint of improved postprogression overall survival (ppOS) when continuing BV with chemotherapy (CT) into 2nd-line (2L) tx. A phase 2 study showed a trend for OS benefit when adding BV to CT in BV-naive 2L NSCLC (HR 0.71; 95% CI 0.41–1.21) (Herbst JCO 2007). This analysis evaluated whether BV exposure (exp) after PD correlates with ppOS in NSCLC. Methods ARIES 1st-line (1L) BV-treated NSCLC patients (pts) who survived 1st PD were included. ppOS equaled the time from 1st PD to any-cause death. BV exp, over follow-up, equaled the cumulative days of BV from 1st PD. A time-dependent Cox regression model was fitted to assess the effect of cumulative BV exp on ppOS, while controlling for potential time-dependent and -fixed confounders. A landmark sensitivity analysis, also adjusting for confounders, compared ppOS in pts treated with BV beyond PD (BBP) and pts treated otherwise (No BBP) ≤30 days after PD. Results As of 09/2011, of 1967 enrolled 1L pts, 1461 (74%) had 1st PD. Characteristics (n = 1461) were: 48% had 1st PD within 6 mos, 52% male, 31% ≥70 y, 13% never smokers, and 13% ECOG status ≥2. The median ppOS for all pts with 1st PD was 6 mos (95% CI 5.6–6.7). Among pts with any BV tx, the mean cumulative BV exp was 116 days (range, 2–1140). Across follow-up, the HRs for ppOS decreased by ∼4% for each additional 21-day interval of cumulative exp (Table). Cumulative BV duration was associated with improved ppOS (P  Conclusions This analysis suggests that cumulative exp to BV after 1st PD may correspond with incremental increases in ppOS for NSCLC pts, and supports the conduct of the phase 3 AvaALL trial. Cycle a No. of pts who received cumulative cycles of BV continuously b Hazard ratio (HR) for ppOS 95% confidence interval (CI) 1 229 0.959 0.941–0.978 2 159 0.920 0.885–0.956 3 118 0.883 0.833–0.935 4 84 0.847 0.784–0.914 5 72 0.812 0.738–0.894 6 55 0.779 0.694–0.874 7 41 0.747 0.653–0.855 8 32 0.717 0.614–0.836 a A cycle is calculated as 21 days of cumulative exposure after PD. b Eg, n = 55 patients were continuously dosed through approximately day 126 (cycle 6) after PD. Disclosure T.J. Lynch: Dr. Lynch is on the Infinity Pharmaceuticals board of directors, consults for Merck, Boehringer-Ingleheim and Astex, and holds a patent on EGFR testing from Partners Healthcare. M. Jahanzeb: Dr. Jahanzeb is a consultant/advisor for Genentech. D.R. Spigel: Dr. Spigel is an unpaid advisor/consultant to Genentech. A. Wozniak: Dr. Wozniak has received honoraria and research support from Genentech and research funding from Eli Lilly. L. Leon: Dr. Leon is an employee of and holds stock options in Roche/Genentech. S. Fish: Dr. Fish is an employee of and holds stock options in Roche/Genentech. E.D. Flick: Dr. Flick is an employee of and holds stock options in Roche/Genentech. D. Dalal: Dr. Dalal is an employee of and holds stock options in Roche/Genentech. M.P. Kosty: Dr. Kosty has participated in the Genentech Speakers' Bureau.