Background and objective: Analysis of programmed death ligand-1 (PD-L1) in tumour samples is necessary to identify candidates for anti-PD-L1/PD-L1 therapy. Because PD-L1 is evaluated by immunohistochemistry (IHC), an adequate amount of tumour tissue is a prerequisite for PD-L1 testing. To examine whether pleural fluid might be an alternative to biopsy/resection specimens for IHC evaluation of PD-L1 in patients with non-small cell lung carcinoma (NSCLC), we compared PD-L1 by IHC between histological specimens and matched pleural fluid.
Methods: A retrospective cohort study of patients with NSCLC who underwent core biopsy of a lung mass/surgical resection with PD-L1 IHC and had a pleural fluid cell block (CB) available for PD-L1 staining was conducted. PD-L1 was categorized as negative (PD-L1 in <1% of tumour cells), moderately positive (PD-L1 in ≥1% to <50%), strongly positive (PD-L1 ≥ 50) or inadequate for PD-L1 testing (<100 tumour cells in the CB). Weighted Cohen's kappa was calculated to evaluate the agreement between PD-L1 on biopsy/resection specimen and pleural fluid for variables with more than two categories.
Results: Of the 115 patients included in this study, 82 (71.3%) had at least 100 tumour cells and were included in the analysis. Of these, 80 (97.6%) had adenocarcinoma. For PD-L1 of histological specimens versus pleural fluid categorized as negative, moderately positive or strongly positive, the weighted kappa statistic was 0.76 (95% CI: 0.64-0.88), and the concordance was 0.78 (95% CI: 0.68-0.86).
Conclusion: Correlation and concordance are high between PD-L1 in histological specimens and matched pleural fluid. Evaluation of PD-L1 in pleural fluid should be considered in patients unable to undergo histological biopsies.
Keywords: cytology; malignant; pleural effusion; programmed death ligand-1.
© 2019 Asian Pacific Society of Respirology.