Sensitivity analysis and publication bias
To further explore the sources of heterogeneity, a sensitivity analysis was conducted by excluding each study. The results are presented in Figure 5. No study had a significant effect on the pooled results, which suggested that although heterogeneity exists, the summary outcomes of the meta-analysis are stable.
Egger’s regression test was employed to detect publication bias. As shown in Figure 6, the ORR, DCR, mPFS, mOS and incidence of grade 3 or higher AEs were p=0.427 (95% CI: -4.25 to 8.79), p=0.571 (95% CI: -18.22 to 11.61), p=0.193 (95% CI: -1.03 to 4.25), p=0.264 (95% CI: -0.23 to 6.43) and p=0.258 (95% CI: -20.34 to 6.85), respectively. No publication bias was identified for any of the outcomes. In brief, the results of this meta-analysis were reliable.