Background: Programmed cell death ligand-1 (PD-L1) expression is a predictive biomarker for anti-PD-1 immunotherapy in non-small cell lung cancer. Different immunohistochemistry (IHC) assays have been developed on histologic material with different cutoffs for positivity. More than one third of the patients are diagnosed on cytology alone. We hypothesized that cytologic cell block material is suitable for PD-L1 analysis.
Materials and methods: Eighty-six paired samples of malignancies from the lung where cytologic cell block and histologic material were available from the same lesion were stained with PD-L1 IHC 28-8pharmDx and PD-L1 IHC 22C3pharmDx. Scorings of like material (cytology or histology) stained with different assays were analyzed in order to evaluate the analytical agreement between assays. Scoring of different materials stained with like assays were analyzed in order to evaluate the agreement between cytology and histology.
Results: A high degree of agreement was found between 28-8pharmDx and 22C3pharmDx, whether applied to histologic or cytologic cell blocks, with Pearson R of 0.95. The Pearson R between 2 rounds of assessment of the same assay on the same type of material was also 0.95. The agreement between histologic and cytologic specimens was high with Pearson R 0.87 to 0.89 and overall agreement between 85% and 95%. There was no bias toward lower prevalence of positivity with cytology than with histology. Disagreement was related to heterogeneity of the histologic tumor sample.
Conclusion: PD-L1 assessment is feasible on cytologic material with the tested assays using cutoffs for positivity similar to those used on histologic material.