Feasibility study of an ex vivo functional precision medicine platform, Optim.AI™, in guiding treatment for gynecological cancers

Background: Gynecological cancers, like ovarian and endometrial, face therapeutic challenges due to their molecular heterogeneity and high rates of relapse following standard platinum-based chemotherapy. Access to genetic information, such as BRCA mutation and homologous recombination deficiency (HRD) status, can predict potential responders to PARP inhibitors. However, these actionable mutations are amenable for a subset of patients only, underscoring the need for complementary functional precision medicine approaches. This ex vivo screening strategy could potentially support patient treatment management. Optim.AI™, a combinatorial functional precision medicine platform, has previously validated clinical utility for hematological cancers and sarcoma. In this feasibility study, we explored the application of Optim.AI™ on ovarian and endometrial cancers. Methods: Tumor cells were isolated from tissue samples from both biopsies and resections. Short-term patient-derived organoids were formed before combinatorial treatment with 12 drugs containing both FDA-approved chemotherapy and targeted drugs. Cell viability was quantified post-drug treatment for Optim.AI™ analysis, ranking all possible top combinatorial therapies for report generation. Retrospective concordance analysis was carried out after clinical responses were collected. Results: Based on the samples received, minimum tissue mass of 0.0835g and 0.368g for ovarian and endometrial samples respectively yielded sufficient cells to proceed with Optim.AI™ testing. Reports were successfully generated for 94% of these samples, with a mean turnaround time of seven working days. Z' factor, a statistical, quality measure for high-throughput screening, was demonstrated to be more than 0.5 for all reports generated, indicative of very good assays. Across the eight ovarian cancer reports generated, Gemcitabine-based combinations were among the most frequently top-ranked treatments. Notably, for HRD-negative patients, Optim.AI™ predictions commonly suggested increased sensitivity to gemcitabine paired with doxorubicin and paclitaxel Preliminary retrospective concordance analysis highlights Optim.AI™'s ability to predict response, where general correlation was observed between lower NCV and higher chemotherapy response score. Conclusion: This study showcases Optim.AI™ as a viable clinical-decision support platform for aiding treatment management for gynecological cancers, particularly for patients who do not harbor actionable mutations. Prospective clinical concordance analysis of Optim.AI™-guided treatments would further validate its clinical utility in these cancers and provide potential precision medicine insights for HRD-negative patients.