Adaptive Responses to PARP Inhibition Predict Response to Olaparib and Durvalumab: Multi-omic Analysis of Serial Biopsies in the AMTEC Trial

Abstract In syngeneic murine breast cancer models, poly ADP-ribose polymerase inhibitor (PARPi) and anti-PD-L1 combinations induce deep, sustained responses independent of BRCA1/2 mutation status. We therefore investigated this combination in the AMTEC clinical trial, in which a one-month olaparib run-in was followed by combined olaparib and durvalumab in participants with BRCA1/2 wild-type metastatic triple negative breast cancer. To characterize adaptive responses to olaparib monotherapy, paired biopsies taken before and during PARPi lead-in were deeply characterized by DNA, RNA, and protein multi-omic analyses, including spatially-resolved single cell proteomics for tumor and immune contexture. We identified multiple potential tumor-intrinsic and microenvironmental biomarkers from pre-treatment and on-olaparib biopsies that robustly predicted participant response to combined olaparib and durvalumab. Notably, the on-olaparib biopsy provided the greatest information content, suggesting that adaptation of malignant cells and the tumor ecosystem to PARPi can serve as a predictor of potential benefit from combined PARPi and anti-PD-L1 therapy.