Machine learning models to predict refractory responses in ovarian cancer

Project summary

Despite decades of research, there is currently no way to predict platinum-refractory (who do not respond to platinum-based therapy) high-grade serous ovarian cancer (HGSOC) prior to treatment. Our recent study (Cell 2023, PMID: 37541199) identified and validated a proteomic signature capable of predicting refractory tumors before chemotherapy. We also found that mechanisms of refractoriness vary across HGSOC subtypes. Building on this, we propose a precision medicine approach that leverages deep learning and subtype-specific biology to improve prediction. With clinical validation using higher throughput assays, these signatures can predict refractory HGSOC patients, and eventually could guide alternative therapies through clinical trials.