Abstract
Medical digital twins (MDTs) are virtual representations of patients that simulate the biological, physiological, and clinical processes of individuals to enable personalized medicine. With the increasing complexity of omics data, particularly multiomics, there is a growing need for advanced computational frameworks to interpret these data effectively. Foundation models (FMs), large-scale machine learning models pretrained on diverse data types, have recently emerged as powerful tools for improving data interpretability and decision-making in precision medicine. This review discusses the integration of FMs into MDT systems, particularly their role in enhancing the interpretability of multiomics data. We examine current challenges, recent advancements, and future opportunities in leveraging FMs for multiomics analysis in MDTs, with a focus on their application in precision medicine.