Results of biomarker analysis from phase II trial of nivolumab in refractory biliary tract cancer

338 Background: Our previous phase II study demonstrated that nivolumab provides modest but durable clinical efficacy in patients (pts) with refractory biliary tract cancer (BTC), suggesting the significant clinical benefit of nivolumab in selected pts and identification of potential predictive biomarkers. Here, we report outcomes from the biomarker analysis. Methods: Pre-treatment tumor samples were obtained and assessed for immunohistochemistry (IHC) with PD-1 and PDL-1 antibodies, mRNA sequencing (RNAseq), and whole exome sequencing (WES). Clinical efficacy was correlated with the molecular analysis. Prior cisplatin exposure was reviewed to evaluate the impact of cisplatin on the tumor mutational burden (TMB) and clinical outcome. Results: Among 46 pts who had tumor response evaluation, 31 pts received cisplatin prior to nivolumab, and 15 were cisplatin-naive. Pre-treatment tumor samples were assessed for IHC (n = 42), RNAseq (n = 11), and WES (n = 11) based on tissue availability. There was no statistically significant correlation between prior cisplatin exposure and clinical outcome. Among 42 available tumor samples, 18 (43%) expressed PD-L1 positivity ( > 1%), which was associated with a statistically significant prolonged progression free survival (PFS). Median PFS was 10.4 months for PD-L1 positive vs 2.4 months for PD-L1 negative (HR 0.23, 95% CI 0.10-0.51; P < 0.001), while there was no statistically significant correlation between PD-1 expressing tumor infiltrating lymphocytes and clinical outcome. There was also no statistically significant difference in TMB of cisplatin-exposed (n = 8) vs. cisplatin-naïve tumor samples (n = 3) (9162, SD = 150 vs. 8898, SD = 806, p = 0.62). In comparing the group with prolonged disease control of at least 16 weeks (n = 4) to those with rapid disease progression (n = 7) who had mRNA sequencing performed, levels of AC005609.1, FAT3, TMEM151A, ADARB2, FAM153A were all significantly upregulated (p < 0.01), while levels of CLCA1, MUC2, IGHV3-43, SWORA6, and CRISP3 were all significantly downregulated (p < 0.01). Conclusions: Prior cisplatin exposure did not lead to statistically significant differences in clinical outcome or TMB in advanced BTC pts treated with nivolumab. However, PD-L1 expression > 1% correlated with improved PFS, and variations in level of certain mRNA sequences were noted when comparing pts who had rapid disease progression vs. prolonged disease control with nivolumab. This suggests that BTC pts with positive PD-L1 expression and a particular mRNA profile may have a favorable response to nivolumab. Further studies are needed to confirm these findings. Clinical trial information: NCT02829918.