Role of health digital twins in oncology drug development - a primer

Abstract

Health Digital Twins (HDTs) have attracted increasing interest as a potential tool for improving both patient care and clinical research, particularly in oncology where clinical trials remain slow, costly, and operationally complex. This primer introduces the concept of digital twins in a form accessible to professionals involved in oncology clinical research and outlines the main components of health digital twins, describes how these systems are generated, and highlights their potential relevance to oncology trials. HDTs are personalized, dynamically updated, predictive, and decision−informing computational representations that integrate clinical records, laboratory measurements, imaging, pathology, and molecular data. The generation of an HDT begins with the integration of multiple sources of information to create a digital representation that reflects the patient’s disease state and can be updated as new information becomes available. These systems may be built using data−driven AI models, mechanistic physiological models, or hybrid approaches that combine both. Emerging HDT−based approaches have been proposed as a means of addressing several of the structural limitations of oncology trials by enabling in−silico experimentation, improved biological stratification of patients, and more efficient use of longitudinal real−world and historical data. In early−phase studies, HDTs can support dose selection, safety evaluation, and cohort selection, while in later phases they may contribute to optimized trial designs, synthetic control arms, and adaptive trial methodologies. These applications may reduce patient enrolment time, improve efficiency, and support simulation−driven decision−making. Despite this promise, the application of HDTs remains associated with technical, ethical, regulatory, and data−governance challenges, including data quality, model validity, bias, and uncertainty. Regulatory agencies currently view HDTs as complementary tools within model−informed drug development rather than as replacements for conventional clinical evidence. In conclusion, HDTs have the potential to transform clinical trials by enabling personalized treatments, optimizing trial design, and facilitating predictive analytics, although their applications remain emerging and require further validation, governance frameworks, and regulatory alignment to support widespread adoption.