(1032-A) Clinical evaluation of functional combinatorial precision medicine platform to predict treatment outcomes and enhance combination therapy design in soft tissue sarcomas

Abstract: Soft tissue sarcomas (STS) are a highly heterogeneous collection of tumors that arise from primitive mesenchymal cells. Due to a reliance on cytotoxic chemotherapy as standard of care, survival outcomes for advanced STS patients remain poor. While efforts to understand sarcomagenesis have revealed potential biomarkers, the diversity of STS subtypes and interpatient tumour heterogeneity has led to the lack of effective molecular targeted therapies available. Given the uniquely individualized nature of STS, we hypothesize that the application of an ex vivo drug sensitivity platform, Quadratic Phenotypic Optimization Platform (QPOP), towards primary STS patient samples can improve identification of patient-specific drug combinations. Using a 12-drug set comprising STS standard of care, FDA approved drugs and promising investigational drugs, QPOP ranks and compares all possible therapeutic combinations using a predesigned array of 155 test combinations. We first evaluated clinical concordance of QPOP to predict treatment outcomes in an STS cohort by comparing patient sample results with either parallel treatment initiated at time of sample collection or physician choice to use QPOP-guided treatment. Clinical concordance analysis showed QPOP had a total predictive value (TPV) of 76.9% and an AUCROC of 85.7%. Exceptional responders to guided treatment with pazopanib or eribulin were observed. We also explored QPOP’s combination therapy ranking function to identify novel combinations that may benefit larger groups of STS patients when appropriately guided. Across a total of 45 patient samples collected, QPOP identified a novel BRD4 inhibitor-based combination as most frequently top-ranked, outperforming standard of care ifosfamide and doxorubicin. Single-drug and combination dose-response assays performed in a panel of established patient lines supported its synergistic interaction, accompanied by increased PARP and caspase-3 cleavage with immunoblotting, and increased apoptotic response with Annexin V assay. Mechanistically, this synergy results in cell cycle arrest and induction of programmed cell death pathways, as revealed by molecular profiling on representative patient and cell lines. We further showed that co-treatment synergistically repressed oncogenic MYC and activated the p27kip1 tumor suppressor. Using an STS xenograft mouse model, we demonstrated this combination to be more effective at impairing tumor progression than single agents alone. Taken together, these results showcased the potential for QPOP to predict clinical response in solid cancers, such as STS, and provided preliminary clinical evidence for the prioritization of a novel drug pair in addressing STS beyond known clinicopathologic parameters.