Programmed death ligand 1 protein expression, histological tumour differentiation and intratumoural heterogeneity in
Vladimír Tancoš1, Marián Grendár2, Anna Farkašová3
1Department of Pathological Anatomy, Comenius University in Bratislava, Jessenius Faculty of Medicine and University Hospital in Martin, Slovakia.
Abstract:
Intratumoural heterogeneity of pulmonary adenocarcinoma challenges the accurate interpretation of programmed death ligand 1 (PD-L1) immunohistochemistry, which is the only validated predictive marker for successful anti-PD-1/PD-L1 immunotherapy. The aim of this study was to determine whether PD-L1 expression is related to adenocarcinoma histological differentiation in a retrospective analysis of tumour biopsies with intratumoural histological heterogeneity. Adenocarcinomas with high intratumoural heterogeneity were categorised as 'mixed adenocarcinomas'. PD-L1 expression was determined immunohistochemically using tumour proportion scores (TPS). In 'mixed adenocarcinomas' PD-L1 scores were assessed across tumour areas with specific histological patterns. Comparisons were performed between histologically distinct differentiated tumours and/or histological areas. Poorly differentiated adenocarcinomas, represented by predominantly solid or micropapillary histological patterns, showed significantly higher expression of PD-L1 than other subtypes (p<0.001). Differentiation of intra-adenocarcinoma components was inversely correlated with PD-L1 expression: there were more PD-L1 positive cells in poorly differentiated areas than less differentiated (p<0.001), or than well differentiated areas (p<0.001), and in less differentiated more than well differentiated areas (p=0.001). In conclusion, PD-L1 expression is associated with poorly differentiated morphology in adenocarcinomas with intratumoural histological heterogeneity. Consequently, a TPS approach may not account for the contribution of more aggressive tumour components with higher levels of PD-L1 expression in within the tumour. Performing spectral analyses of PD-L1 expression across tumours is likely to be more accurate.


