Limitations
Only five RCTs and two RCTs referring to PD-1/PD-L1 antibodies combining with other therapy we collected limited the analysis. All the RCTs were open-label trials, which might influence compliance and detection bias. We also included some single-arm trials, whereas some had small numbers of participants, especially in the trials using combination therapy. Publication bias of PFS, subgroups analyses of OS with clinical characteristics (female, smoke current/former, never smoke, ECOG=0, and ECOG≥1) some types of any grade AEs of PD-1/PD-L1 inhibitors (e.g., vomit, weight loss, dermatitis acneiform, leukopenia, neuropathy peripheral and hypothyroidism), grade ≥3 AEs of PD-1/PD-L1 inhibitors (e.g., neutrophil count decreased, decreased appetite, and leukopenia), any grade AEs of PD-1 inhibitors (e.g., vomit, weight loss, alopecia, neutropenia, nausea, and diarrhea), and grade ≥3 AEs of PD-1 inhibitors (neutrophil count decreased) could not be detected because of the limited number of available studies. However, the occurrence of any grade hypothyroidism in PD-1/PD-L1 inhibitors was much higher than standard of care, which could not be overlooked.