Background: The randomised open-label phase III XELAVIRI trial failed to demonstrate non-inferiority of the sequential application of fluoropyrimidine plus bevacizumab followed by additional irinotecan at first progression (Arm A) versus initial combination of all agents (Arm B) for untreated metastatic colorectal cancer in the initial analysis of time-to-failure-of-strategy (TFS, 90% confidence boundary of 0.8). Here, we evaluate efficacy in the full analysis set (FAS), the per-protocol set, in addition to age-related and molecular subgroups.
Methods: Median TFS, overall survival (OS) and progression-free survival (PFS) were estimated by Kaplan-Meier method and log-rank test. Cox regression models assessed hazard ratios (HRs) and confidence intervals (CIs) (TFS: 90%; OS, PFS: 95%).
Results: Of 421 patients, 390 (92.6%), 391 (92.9%) and 357 (84.8%) events for TFS, OS and PFS were observed in the FAS with a median follow-up of 54.2 months (Arm A) versus 52.9 months (Arm B). Non-inferiority of sequential treatment for TFS was missed in the FAS (HR 0.93; 90% CI, 0.79-1.10; P = 0.482) and not shown in the per-protocol set (HR 0.93; 90% CI, 0.75-1.13, P = 0.433). Formal non-inferiority for TFS was observed for patients older than 70 years (HR 1.06; 90% CI, 0.80-1.41; P = 0.670) and patients with RAS mutant tumours (HR 1.12; 90% CI, 0.87-1.43; P = 0.465). In RAS/BRAF wild-type tumours, combination treatment was significantly superior to sequential therapy in all end-points.
Conclusions: In the overall population, XELAVIRI just missed to demonstrate the non-inferiority of sequential compared to combination therapy for TFS. However, the non-inferiority of sequential treatment was observed in elderly patients and RAS mutant tumours.
Trial registration: Trial registration ID (clinicaltrials.gov) NCT01249638.
Keywords: Bevacizumab; Elderly; Final survival analysis; Metastatic colorectal cancer; RAS.
Copyright © 2022 Elsevier Ltd. All rights reserved.