Digital twins in medicine: between simulating the future and clinical uncertainty

Abstract

Digital twins are individualized, dynamic, and predictive virtual representations of a patient, organ, or biological system, updated with data from their real-world counterpart. Their development could transform medicine by integrating clinical information, imaging, biomarkers, genomic data, wearable devices, and environmental variables, thereby enabling disease trajectories to be simulated and therapeutic responses to be explored before interventions are applied to the patient. In cardiology, their potential applications include phenotyping, risk prediction, electrophysiology, and treatment selection. However, most models remain at an early stage and have limited prospective validation. Their clinical adoption will require evidence of accuracy, clinical benefit, generalizability, and equity, together with safeguards for interoperability, privacy, transparency, and appropriate uncertainty quantification. Digital twins should not replace physicians, but rather complement clinical judgment and provide a virtual experimentation environment to improve real-world decisions without dehumanizing medicine.