Abstract Background: Combined anti-PD-1 (pembrolizumab) and PARP (olaparib) inhibition shows promising clinical activity in homologous recombination deficient (HRD) pancreatic cancer (PC), yet the drivers of resistance remain unexplored. Methods: Tumor single-cell RNA and TCR sequencing were performed on 66 longitudinal fresh tumor samples from 44 participants (pts) with metastatic PC treated with pembrolizumab and olaparib in the POLAR trial (NCT04666740) (31 baseline, 22 on-treatment at 2 months and 13 post-progression tumor biopsies). Bulk TCR sequencing was conducted in 32 pts on peripheral blood samples collected at 0, 6, 18, and 30 weeks. Data were integrated with IMPACT-HRD profiling, neoantigen prediction and histopathological analysis. Pts were enrolled across three cohorts: cohort A with canonical HRD mutations (BRCA1/2, PALB2) ; cohort B with non-core HRD gene alterations; and cohort C with homologous recombination proficient (HRP) tumors. This design enabled dissection of molecular mechanisms across HRD and non-HRD contexts. Results: A total of 263, 780 tumor and stromal cells were successfully profiled. Primary resistant tumors were enriched with basal-like phenotype and myofibroblastic CAFs (myCAFs), regardless of HRD status. HRP tumors showed increased activation of RAS/MAPK and iron metabolism pathways. Histopathological analysis performed before treatment showed reduced CD8+ T cell infiltration in de novo resistant tumors, along with a dense stromal architecture indicative of fibroblast-mediated immune exclusion. Rapid progression (3. 5 months) was linked to immune evasion characterized by MHC-I downregulation and adipose-like CAF enrichment. In contrast, responders (6 months PFS) demonstrated early activation of cytotoxic T cells, along with helper T cells showing upregulated MHC-I expression at 2 months. Secondary resistant samples from responders exhibited a phenotypic shift in tumor cells from classical to proliferative basal-like, enrichment of inflammatory CAFs (iCAFs), CD8+T cells senescence and polarization of the myeloid compartment toward foam-like M2-macrophages and toward N2 neutrophils. Peripheral blood TCRseq identified clonotypes that were expanded in the blood and infiltrated the tumor in a subset of 6 HRD patients, all long-term survivors (18 months), including 4 with ongoing responses 30 months. In tumor tissue, 95% of these clonotypes corresponded to CD8+ T cells exhibiting strong cytotoxic activity (PRF1hiGMZBhiGNLYhi) and increased TCR expression. Conclusions: Resistance to combined PARP and PD-1 blockade in PC arises from both tumor-intrinsic features and microenvironmental barriers, including CAF-mediated immune exclusion and myeloid polarization, and is agnostic to HRD status. Durable responses are associated with both systemic expansion and effective tumor infiltration of cytotoxic T cells. These findings support combinatorial strategies targeting stroma and tumor plasticity and highlight TCR tracking as a biomarker of durable benefit. Citation Format: Marc Hilmi, Jin Park, Wilson Mckerrow, Shigeaki Umeda, Catherine O'Connor, Yuval Elhanati, Elias-Ramzey Karnoub, Roshan Sharma, Kevin Soares, Zeynep Tarcan, Nuray Tezcan, Olca Basturk, Nicolas Lecomte, Joshua Schoenfeld, Nadeem Riaz, Vinod Balachandran, Dana Pe'er, Benjamin Greenbaum, Eileen M O’Reilly, Christine Iacobuzio-Donahue, Wungki Park. Tumor cell plasticity, stromal rewiring, and clonal T cell expansion define response and resistance to combined PARP and PD-1 blockade (POLAR) in pancreatic cancer abstract. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pancreatic Cancer Research—Emerging Science Driving Transformative Solutions; Boston, MA; 2025 Sep 28-Oct 1; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2025;85 (18Suppl₃): Abstract nr A125.
Hilmi et al. (2025) studied this question.