Abstract
Purpose Conflicting results are reported for maintenance treatment with bevacizumab during chemotherapy-free intervals (CFI) in metastatic colorectal cancer after induction chemotherapy. Patients and Methods In this open-label, phase III, randomized controlled trial, we compared the tumor control duration (TCD) observed with bevacizumab maintenance and with no treatment (observation) during CFI subsequent to induction chemotherapy with 12 cycles of fluorouracil, leucovorin, and irinotecan plus bevacizumab. After disease progression, the induction regimen was repeated for eight cycles, followed by a new CFI. Results From March 2010 to July 2013, 491 patients were randomly assigned. Disease progression or death occurred during induction chemotherapy in 85 patients (17%); 261 patients (53%) had at least one reinduction, 107 (22%) had two reinductions, and 56 (11%) had three or more reinductions. The median TCD was 15 months in both groups; the median progression-free survival (PFS) from randomization was 9.2 and 8.9 months in the maintenance group and observation groups, respectively. The TCD observed in both groups was higher compared with the TCD hypotheses of the trial. The median overall survival (OS) was 21.7 and 22.0 months in the maintenance and observation groups, respectively. In the per-protocol population, defined as patients with at least one reinduction after the first progression, the median duration of the first CFI was 4.3 months in both arms; the median TCD was 17.8 and 23.3 months ( P = .339), the median PFS was 9.9 and 9.5 months, and the median OS was 27.6 and 28.5 months in the maintenance and observation groups, respectively. Multivariable analysis revealed that female gender, WHO performance status ≥ 2, and unresected primary tumors were associated with a shorter TCD. Conclusion Bevacizumab maintenance monotherapy did not improve TCD, CFI duration, PFS, or OS.