Cytologic-histologic correlation of programmed death-ligand 1 immunohistochemistry in lung carcinomas.

ABSTRACT

BACKGROUND: Programmed cell death protein 1 inhibitors increasingly are being used to treat patients with advanced lung carcinomas. Programmed death-ligand 1 (PD-L1) immunohistochemistry (IHC) in tumor cells (TCs) and tumor-infiltrating immune cells (ICs) is used to select patients for programmed cell death protein 1 inhibition, but few studies have evaluated PD-L1 IHC in cytology specimens. The objective of the current study was to compare PD-L1 IHC in cytology cell blocks and matched surgical specimens. METHODS: A total of 56 cytology specimens obtained between 2013 and 2016 with matching surgical specimens were stained with anti-PD-L1. Membranous positivity was scored as a percentage of the TCs and ICs by 2 pathologists. Results were compared between cytology and surgical specimens, and interobserver concordance was assessed. RESULTS: The average PD-L1 positivity rate was 28% in TCs and 5% in ICs in surgical specimens (standard deviations of 37% and 7%, respectively), and 21% in TCs and 8% in ICs in cytology specimens (standard deviations of 33% and 15%, respectively). Interobserver concordance was high for TCs in surgical and cytology specimens (intraclass correlation coefficients of 0.96 and 0.96, respectively), and was moderate for ICs in surgical and cytology specimens (intraclass correlation coefficients of 0.47 and 0.67, respectively). There was moderate to high correlation between PD-L1 positivity in TCs between surgical and cytology specimens (Spearman correlation coefficient [Spearman r], 0.69), particularly among fine-needle aspiration specimens (Spearman r, 0.78), but not between PD-L1 positivity in ICs in surgical and cytology specimens (Spearman r, 0.14), including among fine-needle aspiration specimens (Spearman r, 0.23). CONCLUSIONS: Tumor PD-L1 IHC positivity in cytology specimens appears to correlate strongly with results obtained from matching surgical specimens. PD-L1 IHC in ICs within cytology specimens does not reflect results in matched surgical specimens and should not be used in clinical decision making. Cancer Cytopathol 2018;126253-63. © 2018 American Cancer Society.