Background: Over 20% of biliary tract cancers (BTCs) carry mutations in homologous recombination DNA damage repair (HR-DDR) pathways. The PARP inhibitor olaparib (O) blocks tumor DNA repair and may induce response in BTCs with such mutations. Furthermore, HR-DDR mutations may engender neoantigens that improve response to immune checkpoint inhibitors. We propose that treatment with O and the anti-PD-1 monoclonal antibody pembrolizumab (P) will produce a durable anti-tumor response to BTC, especially in patients with HR-DDR mutations. Methods: Consent for participation was obtained. Eligibility criteria included age > 18 years with an ECOG score of ≤1 and a histologic diagnosis of advanced or metastatic BTC with progression of disease (PD) on prior first-line therapy. Treatment featured oral O (300 mg twice daily) plus intravenous P (200 mg every 3 weeks). Response was measured radiographically using RECIST 1.1 guidelines. The primary endpoint was overall response rate (ORR), with the alternative hypothesis that O + P would improve ORR compared to historical controls (versus the null hypothesis of no improvement). Type I and II error rates were set at 5% and 15%, respectively. Simon’s optimal two-stage design was used. Planned sample size was 13 patients in the first stage. Continuation to the second stage would only occur if ≥3 patients demonstrated response. The second stage was planned to include 20 additional patients (N = 33 total), which was based on the number needed to demonstrate an improvement in ORR from 17.5% in historical controls to at least 35.0%. Results: Of 21 eligible patients, 14 were accrued between June 2020 and March 2022, and 13 were evaluable for efficacy. Patients had a median age of 63 years, were primarily female (71%), Caucasian (50%), and with metastatic disease (93%). Patients received a mean of 6.5 cycles of therapy. Best treatment response included partial response (N = 2), stable disease (N = 5), and PD (N = 6). The ORR was 15.3% (95% CI = 0.02-0.45). Of the partial responders, 1 patient achieved a duration of response of 8 months (mos), and 1 has ongoing response at 24 mos. Median (med) time to progression was 7.7 mos (95% CI = 1.2-9.3), med progression free survival was 5.5 mos (95% CI = 1.2-7.7), and med overall survival was 11.9 mos (95% CI = 5.5-15.4). Most patients (64%) developed at least one treatment-related adverse event (trAE). Grade 3 trAEs were experienced in 5 patients (36%) and included anemia (N = 3), diarrhea (N = 1), and transaminitis (N = 1). Conclusions: While O + P has acceptable safety and toxicity, the study did not achieve significant improvement in the primary endpoint. However, the two patients who demonstrated durable treatment response were found to have HR-DDR tumor mutations (including RAD51, ATM, and BRCA2), necessitating further research into the efficacy of O + P for patients with similar tumor genomes. Clinical trial information: NCT04306367.