Post-progression therapies and outcomes for patients with advanced, BRCA-related pancreatic cancer after receipt of PARP inhibitors: A national retrospective cohort study.

4156 Background: Maintenance therapy with poly-ADP ribose polymerase (PARP) inhibitors (PARPi) can delay progression of advanced platinum-sensitive pancreatic cancer in patients with mutations in BRCA1, BRCA2or PALB2. However, there are limited data to guide treatment decisions and outcomes following progression with PARPi. Thus, we sought to compare outcomes of post-progression platinum and non-platinum therapies for patients with pancreatic cancer after receipt of PARPi using a nationwide cohort. Methods: This study used the nationwide Flatiron Health electronic health record (EHR)-derived de-identified database. During the study period, the de-identified longitudinal patient-level data originated from approximately 280 cancer clinics. Patients were eligible for inclusion if they received maintenance olaparib for advanced pancreatic cancer and subsequently received post-progression chemotherapy. Post-progression chemotherapy was characterized as platinum (P) or non-platinum (NP). Demographics, tumor stage, and lab values were compared using the Chi-squared test or Mann Whitney U test. Overall survival (OS) and time to treatment discontinuation (TTD) were evaluated using the Kaplan Meier method with an index date of initiation of post-progression therapy. Given the limited anticipated sample size, no formal comparisons were made between treatment groups. Cox proportional hazards models were constructed to identify associations between baseline clinical factors and outcomes. Results: Of 11,179 patients were included in the database, 26 met inclusion criteria for this analysis after progressing on olaparib and subsequently receiving post-progression chemotherapy. Most patients were white (81%) and non-Hispanic (77%). Twelve (46%) received post-progression P, while 14 (54%) received post progression NP. No significant differences were noted in baseline characteristics between groups. Among patients on P, the most common regimens were gemcitabine and cisplatin (50% n=6), FOLFIRINOX (17%, n=2), and FOLFOX (17%, n=2). Of those on NP, the most common regimens were gemcitabine and nab-paclitaxel (50%, n=7) and FOLFIRI (29%, n=4). The median OS for post progression therapies was 4.8 months for P and 4.1 months for NP group. The median TTD for these post progression therapies was 0.9 months for P and 1.5 months for NP. No baseline clinical factors were associated with time-to-event in Cox regression analyses. Conclusions: In this analysis, patients who progressed on olaparib rapidly progressed on post-progression P and NP and had limited survival. Further work will be needed to identify mechanisms of resistance to PARPi and to optimize post-progression treatments. This work adds to a growing body of literature supporting an individualized approach to post-progression therapy decisions for these patients.