Identifying and Characterising Response in Clinical Trials: Development and Validation of a Machine Learning Approach in Colorectal Cancer

Researchers developed a machine learning approach combining Virtual Twins method with survLIME to identify patient subgroups who respond differently to treatments in clinical trials. The method achieved 0.77 AUC for identifying treatment responders in colorectal cancer trials, finding genetic mutations, metastasis sites, and ethnicity as key response factors.

• →New machine learning approach combines partly conditional modelling with Virtual Twins method for precision medicine applications.
• →Method achieved 0.77 AUC for identifying fixed responders and improved dynamic responder identification from 0.597 to 0.685 AUC.
• →Applied to colorectal cancer trials, the approach identified genetic mutations, metastasis sites, and ethnicity as important treatment response factors.
• →The approach accommodates dynamic treatment responses while potentially outperforming existing methods for fixed responses.
• →Results from clinical data application were consistent with existing medical literature findings.