Background
Patients with homologous-recombination-deficient (HRD) pancreatic cancer respond to platinum chemotherapy and PARP inhibition, but durable benefit is limited. POLAR asked whether adding immune checkpoint blockade to PARP inhibition could deepen and extend that benefit as a chemotherapy-free maintenance strategy.
Design
POLAR is an investigator-initiated phase 2 trial evaluating maintenance pembrolizumab + olaparib in biomarker-stratified metastatic pancreatic cancer. Cohort A enrolled patients with BRCA1/2 or PALB2 alterations.
Results
In cohort A (n=33), which enrolled BRCA1/2- or PALB2-altered tumors, median PFS was 8.3 months, median OS was 28 months, and estimated 3-year OS was 44%. The formal primary efficacy threshold was not met, so these findings should be interpreted as a biomarker-rich signal rather than proof of superiority over another maintenance strategy.
The canonical BRCA1/2/PALB2 cohort had stronger activity than the non-core HRD cohort. This distinction cautions against treating every DNA-repair alteration — for example, an ATM alteration in isolation — as biologically equivalent to BRCA2 or PALB2 for PARP-inhibitor sensitivity.
Translational correlates
POLAR profiled ctDNA, tumor immune cells, genomic scars, DNA-repair alterations, and neoantigens. Falling or clearing ctDNA, greater tumor-infiltrating lymphocytes, and frameshift-indel neoantigen enrichment were associated with durable clinical benefit.
These associations support the hypothesis that a subset of DNA-repair-deficient pancreatic cancers is immunologically distinct. They are exploratory biomarkers, not yet validated patient-selection tests.