PD-L1 expression in bladder cancer: Which scoring algorithm in what tissue?

Abstract Introduction For cisplatin-ineligible patients, approval of first-line immune-checkpoint inhibitor therapy relies on the programmed death ligand 1 (PD-L1) expression assay employed, namely, the combined positive score (CPS) or immune cell (IC) score. This study compares PD-L1 diagnostic scores and positivity in primary and matched metastatic bladder cancer tissue. Methods A total of 108 patients undergoing radical cystectomy for urothelial bladder cancer and lymphatic spread (pN+) were included. PD-L1 expression was compared by immunohistochemistry (IHC) between the primary bladder tumor and associated lymph node metastases using Ventana SP263 anti-PD-L1 antibody. In a subset of cases further IHC was performed with Ventana SP142 and Dako 22C3 antibodies. Second, the PD-L1 scoring algorithms for the CPS and IC score were compared. Correlation of PD-L1 positivity with clinical parameters and tumor stage was assessed. Intra- and intertissue analyses were performed with Pearson's chi square test, McNemar test and Spearman correlation employing IBM SPSS 25. Results PD-L1 expression did not correlate with clinicopathological parameters. The CPS (43.5% vs. 35.6%; P=0.006) and the IC score (28.7% vs. 21.2%; P=0.002) yielded PD-L1 positivity significantly more often in primary BC than in matched lymph node metastasis. Both the CPS (r=0.775; P Conclusion Compared to lymph node analysis, bladder tissue yields more PD-L1 positivity assessed with the CPS and IC scores. This is particularly evident when employing the CPS. The findings highlight that employing both PD-L1 assays may maximize eligibility for first-line checkpoint-inhibitors to treat bladder cancer patients unfit for cisplatin-based chemotherapy.