Impact of hepatitis B virus (HBV) infection on efficacy and safety in the KEYNOTE-966 study of pembrolizumab (pembro) plus gemcitabine and cisplatin (gem/cis) for advanced biliary tract cancer (BTC).

4097 Background: In KEYNOTE-966, pembro + gem/cis significantly increased overall survival compared with placebo + gem/cis alone as first-line therapy for patients (pts) with advanced BTC (hazard ratio HR = 0.83, 95% confidence interval CI 0.72-0.95; P = 0.0034). Chronic HBV infection is a known risk factor for BTC and may impact the effectiveness of immunotherapy in patients with BTC. An exploratory analysis showed low incidence of HBV reactivation and no cases of HBV-associated hepatitis in HBV-positive pts from KEYNOTE-966. Here, we present a post hoc analysis from KEYNOTE-966 to assess whether patient HBV infection status is associated with differences in efficacy and safety outcomes. Methods: Pts were randomized 1:1 to pembro 200 mg IV or placebo IV Q3W for ≤35 cycles plus gem 1000 mg/m 2 IV until disease progression and cis 25 mg/m 2 IV for ≤8 cycles. HBV status was determined by anti-HBc (total and IgM), HBV DNA, and HBsAg. Pts with chronic HBV infection (HBsAg positive and/or detectable HBV DNA) and clinically resolved HBV infection (HBsAg negative and anti-HBc positive, and undetectable HBV DNA) at baseline were eligible and comprised the HBV-positive subgroup. Pts with chronic HBV infection needed to have begun antiviral therapy ≥4 weeks before, and HBV DNA <100 IU/mL prior to, starting study treatment. The data cutoff date for this analysis was November 14, 2023. Results: 1069 pts were enrolled in KEYNOTE-966. 329 (30.8%) were HBV positive (164 from pembro arm vs 165 from placebo arm) and 732 (68.5%) were HBV negative (366 from pembro arm vs 366 placebo arm); 8 had unknown HBV status at baseline. More pts in the HBV-positive subgroup were male (60.8% vs 47.4% in the HBV-negative group), ≥65 years old (53.2% vs 44.5%), and from Asia (76.9% vs 31.8%). In the HBV-positive subgroup, median OS was 12.3 mo (95% CI 9.5-14.5) for the pembro arm vs 10.9 mo (95% CI 9.7-13.2) for the placebo arm (HR = 0.87, 95% CI 0.69-1.10). For the HBV-negative subgroup, median OS was 12.8 mo (95% CI 11.4-13.9) for the pembro arm vs 10.7 mo (95% CI 9.4-11.7) for the placebo arm (HR = 0.85, 95% CI 0.73-0.99). Grade 3-5 treatment-related AE rates in the pembro arm vs the placebo arm were 69% vs 73% of pts in the HBV-positive subgroup and 72% vs 68% in the HBV-negative subgroup, respectively. Rates of discontinuation of any drug due to a treatment-related AE in the pembro arm vs the placebo arm were 14% vs 15% in the HBV-positive subgroup and 22% vs 16% in the HBV-negative subgroup. Conclusions: Efficacy and safety outcomes remained consistent between HBV-positive and HBV-negative subgroups from KEYNOTE-966. These data further support pembro + gem/cis as first-line therapy for advanced BTC regardless of HBV infection status. Clinical trial information: NCT04246177 .