1135-Drug Response Models for Personalised Cancer Therapy

Researchers have developed a machine learning methodology to predict how individual cancer patients will respond to various medications by using functional drug screening rather than relying solely on genetic data. By testing a small panel of "probing" drugs on patient-derived cell cultures, the system analyzes historical data from other cell lines to accurately forecast the effectiveness of a much larger library of treatments. This approach proved highly successful across multiple datasets, including FDA-approved drugs and clinical biopsies, often identifying potent therapies that significantly reduced cell viability. The study suggests that these predictive models offer a faster, more cost-effective path toward personalized medicine by bypassing the biological complexity that often limits genomic methods. Ultimately, this framework supports the creation of tailored drug cocktails designed to improve patient outcomes and overcome treatment resistance.

References:

• Abdel-Rehim A, Orhobor O, Griffiths G, Soldatova L, King RD. Establishing predictive machine learning models for drug responses in patient derived cell culture. NPJ Precis Oncol. 2025 Jun 13;9(1):180. doi: 10.1038/s41698-025-00937-2. PMID: 40514399; PMCID: PMC12166088.