A Bayesian network meta-analysis on systemic therapy for previously treated gastric cancer.

Abstract

We conducted a systemic literature review of randomized controlled trials (RCTs) to identify phase III RCTs on salvage treatment of advanced gastric cancer (AGC) and performed a Bayesian network meta-analysis with random-effects model. The overall survival (OS) was the primary outcome of interest. A total of 20 randomized phase III trials were selected. For the second-line treatment, olaparib plus paclitaxel had the highest surface under the cumulative ranking curve value (90.5%), followed by paclitaxel plus ramucirumab (88.4%) and pembrolizumab (86.5%), indicating that these treatments could be the most effective regimens in terms of OS. Nivolumab, chemotherapy, and apatinib showed significant OS benefit compared with best supportive care for the third-line treatment. In conclusion, pembrolizumab may be the most preferable regimen as a second-line treatment for patients with PD-L1-expressing AGC, while paclitaxel-based combinations are recommended for PD-L1-negative AGC. Nivolumab might be the most preferable third-line treatment.