Abstract

We developed a digital twin framework for guiding therapy decisions for individual patients. Given that cancer is an eco-evolutionary system that adapts to evade therapy, static treatment regimens eventually fail in many advanced cancers. A personalized, proactive, and dynamic approach to therapy has the potential to improve outcomes. To explore clinical feasibility of implementing such a framework, we developed a clinical trial (NCT04343365), the Evolutionary Tumor Board (ETB), which uses eco-evolutionary theory and predictive modeling to assist clinical decision making for individual cancer patients (n=24, ongoing). The framework relies on detailed data curation and imaging measurements for each patient, and mathematical modeling incorporates multi-lesion tumor growth, treatment response, and evolution of resistance. Digital twin models are calibrated by historical datasets of similar patients, clinical trial data, and the patient’s longitudinal data. A “Phase i trial” approach accounts for prediction uncertainty, and information is delivered via a software interface. The ETB has provided novel and useful decision support for the patients and oncologists. At the same time, our efforts show that there are both challenges and opportunities in using digital twins to predict and personalize therapy, particularly in the context of real-time clinical care.