Understanding the need for digital twins’ data in patient advocacy and forecasting oncology

Abstract

Digital twins are made of a real-world component where data is measured and a virtual component where those measurements are used to parameterize computational models. There is growing interest in applying digital twins-based approaches to optimize personalized treatment plans and improve health outcomes. The integration of artificial intelligence is critical in this process, as it enables the development of sophisticated disease models that can accurately predict patient response to therapeutic interventions. There is a unique and equally important application of AI to the real-world component of a digital twin when it is applied to medical interventions. The patient can only be treated once, and therefore, we must turn to the experience and outcomes of previously treated patients for validation and optimization of the computational predictions. The physical component of a digital twins instead must utilize a compilation of available data from previously treated cancer patients whose characteristics (genetics, tumor type, lifestyle, etc.) closely parallel those of a newly diagnosed cancer patient for the purpose of predicting outcomes, stratifying treatment options, predicting responses to treatment and/or adverse events. These tasks include the development of robust data collection methods, ensuring data availability, creating precise and dependable models, and establishing ethical guidelines for the use and sharing of data. To successfully implement digital twin technology in clinical care, it is crucial to gather data that accurately reflects the variety of diseases and the diversity of the population.