Different anti-PD-1 and anti-PD-L1 antibodies bind different epitopes. However, whether the results from the SP142 and 22C3 immunochemistry (IHC) assays can be interchanged to determine patient eligibility for immunotherapy remains largely unknown. Histologic sections from 135 tumor samples were probed with both 22C3 and SP142 antibodies. The concordance of PD-L1 expression determined by the two assays was assessed. Additionally, we evaluated the association of PD-L1 expression detected by different assays with clinicopathological features and prognosis. In total, 105 (77.78%) of 135 samples evaluated by the 22C3-IHC platform produced the same results with the SP142-IHC platform (Kappa value: 0.481, p < 0.001). In addition, 69 (51.11%) of 135 samples evaluated by the SP142-IHC platform produced the same results with the 22C3-IHC platform (Kappa value: 0.324, p < 0.001). PD-L1 expression based on the 22C3-IHC assay was significantly correlated with smoking status, whereas that based on the SP142-IHC assay was correlated with smoking status, sex, and histology. Compared to the SP142-IHC assay, the 22C3-IHC assay usually resulted in an underestimation of PD-L1 expression in tumor cells and immune cells. Thus, the results from the two assays cannot be interchanged. Our data also suggest that the use of different reagents may account for inconsistencies in the literature regarding the association between PD-L1 expression and clinicopathological features.