Improved understanding of the molecular mechanisms has led to identification of checkpoint signalling and development of checkpoint inhibitors in the treatment of many cancers, including lung cancer. To be able to select the patients who benefit most from checkpoint inhibitors, predictive biomarkers are needed. Currently, the only predictive biomarker that has been approved in clinical use is PD-L1, the ligand of the inhibitory T-cell checkpoint PD-1. The use of PD-L1 as a predictive biomarker is confounded by multiple unresolved issues, from testing issues (e.g., cut-off values for positivity) to clinical use (e.g., the response to anti-PD-1 and anti-PD-L1 antibodies in patients without any expression of PD-L1). Even more open questions exist in the evaluation of PD-L1 as a prognostic biomarker. In the future, we expect that an improved understanding of immune system, tumor microenvironment, mechanism of action of immunotherapeutic drugs, and PD-L1 testing methods will elucidate the value of PD-L1 as a prognostic and predictive biomarker in detail.