Abstract B042: Pharmacophenotypic analysis of advanced pancreatic cancer patient-derived tumoroid models uncovers spatiotemporal drug resistance dynamics associated with single-agent and combination therapies
Jarle Bruun, Peter W. Eide, Christer A. Andreassen, Nicolas Pasquier, Tine N. Alver, Christina Biörserud, Sara Blomström, Åke A. Sandberg, Elena RangelovaAbstract
Background:
Despite decades of intensive research, pancreatic cancer remains one of the deadliest malignancies, with only modest improvements in patient survival. Recently, important clinical breakthroughs have been achieved with inhibitors targeting the active RAS(ON) state, doubling the median overall survival of patients with metastatic pancreatic cancer. However, therapeutic resistance inevitably emerges, leading to disease progression and mortality. There is therefore an urgent need, and an unprecedented opportunity, to develop rational combination strategies capable of delaying or overcoming treatment resistance.
Methods and materials:
Pharmacophenotypic profiles were generated for 35 clinically approved and investigational agents across a panel of 20 advanced pancreatic cancer patient-derived tumoroid (PDT) models using the clinical-grade, high-content imaging-based drug screening platform iCAN. Based on these findings, rational follow-up combination studies were performed with longitudinal monitoring over 14 days to investigate resistance dynamics and the emergence of drug-tolerant persister cells. An AI-powered computer vision algorithm (iCANdy) was used to detect, track, and quantify 11 distinct morphological classes associated with cancer cell growth states and drug response phenotypes, enabling spatiotemporal characterization of resistance-associated adaptations.
Results:
Individual pancreatic cancer PDTs exhibited unique baseline morphological signatures and substantial intra-model phenotypic heterogeneity in response to treatment. Drug screening revealed extensive inter-patient variability in sensitivity and resistance to standard-of-care, off-label, and investigational therapies, including the KRAS G12D inhibitor MRTX1133, the RAS(ON) inhibitors daraxonrasib (panRAS) and zoldonrasib (KRAS G12D), and the PRMT5 inhibitors MRTX1719 and vopimetostat. Distinct adaptive phenotypic responses associated with specific drugs and drug classes were observed, particularly involving transitions between cystic and invasive growth states. Drug-tolerant persister populations emerged following exposure to most single agents at clinically relevant concentrations. A focused set of clinically relevant combinations, including paclitaxel, gemcitabine, 5-fluorouracil, daraxonrasib, and zoldonrasib, demonstrated additive or synergistic activity across subsets of PDT models. Notably, combinations involving daraxonrasib with vopimetostat or MRTX1719 induced pronounced cytotoxic effects in a subset of models and were associated with distinct morphological remodeling events.
Conclusions:
Spatiotemporal pharmacophenotypic profiling reveals profound intra-patient and inter-patient heterogeneity in drug response and resistance evolution in pancreatic cancer. These findings support the use of longitudinal phenotypic analysis to identify resistance-associated cellular states and to guide the development of rational combination therapies capable of improving the durability of response to RAS-targeted treatment strategies.
Citation Format:
Jarle Bruun, Peter W. Eide, Christer A. Andreassen, Nicolas Pasquier, Tine N. Alver, Christina Biörserud, Sara Blomström, Åke A. Sandberg, Elena Rangelova. Pharmacophenotypic analysis of advanced pancreatic cancer patient-derived tumoroid models uncovers spatiotemporal drug resistance dynamics associated with single-agent and combination therapies [abstract]. In: Proceedings of the AACR Conference on Pancreatic Cancer: New Frontiers in Biology and Therapeutic Development; 2026 Sep 25-28; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(18_Suppl_2):Abstract nr B042.