RÉSUMÉ
INTRODUCTION: Controversy regarding the prognostic and/or predictive role of PD-1 and PD-L1 expression for upper tract urothelial carcinoma (UTUC) could partly be explained by inconsistencies in the immunohistochemistry (IHC) methodology. Objective is to standardize the methodology for routine evaluation of PD-1 and PD-L1 expression in UTUC patients. MATERIALS AND METHODS: Twenty-two cases treated with radical nephroureterectomy between 1996 and 2015 at 11 French hospitals were randomly selected to compare different methodologies for evaluation of PD-1 and PD-L1 expression. IHC was carried out on whole tissue sections and 0.6 mm- or 2 mm-core tissue micro-arrays (TMAs) using PD-1 NAT105 and PD-L1 28.8 or E1L3N on both tumor cells and tumor-infiltrating immune cells (TILs). Results obtained with whole tissue sections (WTS) were compared to those obtained with 0.6 mm- and 2 mm-core TMAs. Concordance was evaluated using Kappa coefficient. RESULTS: For evaluation of PD-1 and PD-L1 expression, the best concordance with WTS was observed using the PD-1 NAT105 and PD-L1 28.8 antibody on 2 mm-core TMAs, with 5% cut off for positivity on TILs and tumor cells, respectively (Kappa = 0.8). CONCLUSIONS: The most accurate methodology for routine evaluation of PD-1 and PD-L1 expression in UTUC may be based on 2 mm-core TMAs using NAT105 and 28.8 antibodies with a 5% cut off for positivity on TILs and tumor cells, respectively.
