PD-L1 expression in gastric cancer: interchangeability of 22C3 and 28-8 pharmDx assays for responses to immunotherapy

Abstract

Recent clinical trials have shown the promising therapeutic effects of pembrolizumab and nivolumab in patients with advanced gastric cancer. Currently, the programmed death ligand-1 (PD-L1) 22C3 pharmDx assay is the only companion diagnostic assay for assessing the safety and effectiveness of pembrolizumab. The purpose of this study was to compare 22C3 pharmDx and 28-8 pharmDx, a complementary diagnostic assay for nivolumab, in gastric cancer. In this study, 22C3 and 28-8 pharmDx assays were performed on the same formalin-fixed, paraffin-embedded tissue blocks of gastric adenocarcinoma clinical samples (n = 55). The concordance rate was evaluated using combined positive score (CPS) cutoffs of 1, 10, and 50. PD-L1 positivity with CPS >= 1 was 45.5% using the 22C3 pharmDx assay and 49.1% using the 28-8 pharmDx assay. At a CPS cutoff of 1, the overall percentage agreement was 96.4%. The positive and negative percentage agreements were 93.3% and 100%, respectively. All cases positive for PD-L1 using the 22C3 pharmDx assay were also positive using the 28-8 pharmDx assay. At a CPS cutoff of 10, the overall percentage agreement was 96.4%. At a CPS cutoff of 50, the two assays exhibited 100% concordance. Nonspecific cytoplasmic staining in the background tissues and tumor cells was often observed in the 28-8 pharmDx assay. When the results of the two assays were matched for response to immunotherapy, the overall response rate was higher in patients with a PD-L1 CPS >= 1 than in PD-L1-negative patients (22C3 pharmDx, P = 0.001; 28-8 pharmDx, P = 0.002). In conclusion, PD-L1 22C3 and 28-8 pharmDx assays were highly comparable at CPS cutoffs of 1, 10, and 50 in gastric cancer. These results provide evidence for the potential interchangeability of the two PD-L1 assays in gastric cancer.