2.3 Data extraction and risk of bias assessment
For all enrolled studies, the following data were extracted by two reviewers in regards of authors, publication years, county, number of patients, anti-PD-L1 antibody, definitions of PD-L1 positivity, IHC cutoffs, anatomic location, clinicopathological and prognostic variables. The risk of bias in the individual studies was appraised using the Reporting Recommendations for Tumour Marker Prognostic (REMARK) guidelines13,which were composed of six dimensions as summarised in Supplemental Fig.S2.