A digital twin of a patient’s heart will track how the organ changes over time, predicting when treatment is needed rather than just simulating a few heartbeats. Current patient-specific heart models can guide procedures, but they only capture a snapshot. They cannot forecast how a disease will progress or when a therapy will stop working. This project builds digital twins that update as new imaging data arrives, turning static models into dynamic forecasts. The technical challenge is calibrating these models across large patient groups and efficiently updating them through time. If successful, the approach could transform how clinicians manage heart failure. In three clinical studies, the team will test whether digital twins can identify causes of pump dysfunction, predict which recovered patients can safely stop medication, and forecast when newly diagnosed patients will need advanced therapy. This is the first demonstration of cardiac biomechanics digital twins using real clinical data to answer specific treatment questions. The research is applied and patient-facing. It does not aim to discover fundamental biological mechanisms but to build a practical forecasting tool that could eventually be updated in real time, tracking a patient throughout their life and feeding directly into care decisions.
View original technical description
Modelling and simulation play important roles in designing everything from planes to cars to bridges. However, advances in connectivity and computing now enable models to be linked directly to a specific object or system, creating a "digital twin". Digital twins represent a computational surrogate for a particular object and are updated through time as more information becomes available. However, digital twins are not limited to manufactured objects alone. This project aims to develop digital twins of patients, where a model will track a patient through time. We focus on making digital twins of patients' hearts using detailed imaging data sets over the period of a clinical trial. This is the first step towards models that are updated in real-time, track the patient throughout their life and directly feed back into informing patient care. The digital twin approach builds on patient-specific computer models of the heart that are currently being evaluated to guide procedures in the UK at King's College London and in the US. These models are designed to optimise treatments for a specific patient's pathophysiology but only simulate a small number of heartbeats. Digital twins, which track a patient through time, will forecast disease progression and response to therapy. This represents the next step in simulation guided therapy, where the optimal treatment and, importantly, when to deliver it, will be predicted. This project will address the technical challenges in calibrating computer models of large numbers of patients, how to efficiently update these models through time as more data becomes available, how to analyse images of the heart recorded over the duration of a clinical trial and how to predict complex changes in shape and function of the heart. The approaches will be applied to study three patient groups in three studies. First, we will test if multi-scale cardiac biomechanics models can identify common causes of pump dysfunction in heart failure patients. Second, we will test if digital twins can predict which patients who have recovered from heart failure can stop their heart failure mediation. Thirdly, we will test if digital twin forecasts can be used to predict recovery and pre-empt the need for advanced heart failure therapy in newly diagnosed heart failure patients. This will provide the first demonstration of cardiac biomechanics digital twins using real clinical data to answer important clinical questions.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know