Abstract

Background The detection of programmed death-ligand 1 (PD-L1) expression can enrich for patients who respond to anti-programmed cell death 1 (PD-1)/PD-L1 therapies. Though, for most laboratories, the cost of PD-L1 22C3 pharmDx is prohibitive for widespread use, whereas the laboratory-developed test (LDT) PD-L1 E1L3N antibody clone is widely available and inexpensive. This study aims to explore the analytical performance of E1L3N on the Dako Autostainer Link-48 platform and further evaluate the concordance of E1L3N and 22C3 expression in non-small cell lung cancer (NSCLC) biopsy samples. Methods A total of 171 NSCLC biopsy samples were utilized in this study. Cases with less than 100 tumor cells were excluded. Serial sections of representative blocks were used for immunohistochemistry (IHC) staining. The staining protocol was performed according to the standard PD-L1 IHC 22C3 pharmDx package. PD-L1 staining on the tumor cell membrane was detected by immunofluorescence. Results At a 1% cutoff value, PD-L1 was positive in 46.2% of patients using clone 22C3 and 42.1% of patients using E1L3N assays. At a 50% cutoff value, PD-L1 was positive in 16.4% of patients using clone 22C3 and 15.2% of the patients using E1L3N assays. Cohen’s kappa was used to evaluate the concordance of the PD-L1 expression between clone 22C3 and E1L3N. The kappa values were 0.893 [95% confidence interval (CI): 0.826–1] at the 1% cutoff and 0.868 (95% CI: 0.764–1) at the 50% cutoff. An evaluation of the intraclass correlation coefficients (ICCs) between the antibodies was used to quantify the interassay variability for PD-L1 expression in tumor cells. ICCs showed high concordance between the two antibodies (0.955, 95% CI: 0.939–0.967). Cohen’s kappa was also used to assess the consistency of the PD-L1 evaluation between two pathologists.