UC San Diego AI Model Predicts Cancer Treatment Responses

Researchers at the University of California, San Diego have unveiled MutationProjector, an artificial‑intelligence system that translates a tumor’s genetic profile into predictions of how the cancer will respond to therapy. Trained on genomic data from more than 30,000 tumors spanning ten solid‑cancer types, the model generates a compact representation of the tumor’s biological state, enabling it to identify both known and novel biomarkers linked to treatment outcomes. In validation studies across independent patient cohorts—including bladder, lung, and melanoma cancers—the system matched or exceeded existing prediction methods for immunotherapy and chemotherapy response, addressing the current limitation that only about 8 % of cases are matched to an FDA‑approved therapy based on genetics.

MutationProjector also provides interpretable insights into the molecular pathways driving the predictions, which could help clinicians refine biomarker selection and treatment strategies. The research was published in Cancer Discovery and the team plans to expand the model to additional cancer types and international genomic datasets.