INTRODUCTION: While BCG remains standard of care adjuvant treatment for high-risk nonmuscle invasive bladder cancer (NMIBC), predicting response to BCG therapy is difficult. PD-L1 expression in the tumor microenvironment is a biomarker which may impact the efficacy of immunotherapy. The purpose of this study is to determine the predicative ability of PD-L1 for BCG-unresponsiveness in BCG-naïve high-risk NMIBC. MATERIALS AND METHODS: We performed a retrospective review of high-risk NMIBC patients treated with BCG in the United States (108) and Israel (124) from 2008 to 2014. PD-L1 status of BCG-naïve specimens determined with Dako 22c3 assay (CPS >
0) at initial transurethral resection was correlated with outcomes, primarily BCG-unresponsiveness. The impact of BCG exposure on PD-L1 expression was analyzed in patients who had sequential tumor sampling. RESULTS: Of 232 patients, 48 (21.6%) were BCG-unresponsive. PD-L1 expression was positive in 20 (9.2%) and in 5 (5.4%), 15 (14.3%), and 0 of Ta, T1, and CIS, respectively. In the US cohort, BCG-unresponsiveness occurred in 2 (14.3%) of PD-L1 positive and 36 (42.9%) of PD-L1 negative patients (P = 0.042). Tumors with baseline PD-L1 positivity were associated with a lower rate of BCG-unresponsiveness (OR 0.14
95%CI 0.03-0.76). The predictive value of PD-L1 status in determining BCG-unresponsiveness was poor (AUC 0.57, 95%CI 0.46-0.69). Changes from PD-L1 negative to positive status after BCG exposure were observed, indicating a possible mechanism of resistance to BCG. CONCLUSIONS: While PD-L1 status was an imperfect overall biomarker for BCG unresponsiveness, there was an association between initial PD-L1 expression and BCG response. Additionally, there was an upregulation of PD-L1 expression in BCG unresponsive specimens. These findings generate a hypothesis-generating discussion regarding the tumor immune-microenvironment and its response to BCG. Further investigation is necessary to better understand this interaction and its impact on tumor biomarker profiling and patient selection for combination therapy.