Assessment of heterogeneity and reporting biases
We assumed that there would be clinical heterogeneity in the studies, including different ages of the study population and differences in immunotherapy protocol. We assessed heterogeneity between studies using the I2 test with a value greater than 50% representing substantial heterogeneity. We estimated a pooled RR using a Mantel-Haenzel fixed-effect model if I2 test ≤50% or a random-effect model if I2 test>50%. Funnel plot was used to assess potential publication bias.