A hundred patients with pulmonary arterial hypertension will wear sensors that feed data into personalised computer models of their hearts, updated in real time as their condition changes. This matters because pulmonary arterial hypertension—a rare but devastating condition causing fainting and breathlessness—is currently diagnosed through an invasive procedure, and doctors have no reliable way to predict how individual patients will respond to treatment. Existing heart models are built for short-term planning, not for tracking disease over months or years. The project will integrate physiological models of sleep and exercise into these "digital twins," testing whether daily activity patterns improve the accuracy of predictions about disease progression and treatment outcomes. If successful, this approach could replace one-size-fits-all treatment forecasts with individualised, continuously updated predictions. For patients, that might mean earlier warnings of deterioration and more precisely timed interventions. For clinicians, it offers a non-invasive way to monitor disease without repeated hospital visits. The work is fundamentally about building a new kind of predictive tool—one that learns from a patient’s everyday life, not just from snapshots taken in a clinic.
View original technical description
Clinical decisions are often informed by hospital measurements, but patients spend their lives in the community. The wider use of wearables and implantable sensors has empowered individuals to track their health and provides new measurements to diagnose, select therapies, and monitor responses. However, these data sets are often complex, and linking them to underlying pathologies remains challenging. Patient digital twins are computational models that evolve with time by assimilating personalised data to track disease progression for real-life patients. Cardiac digital twins simulate how heart cells, tissues, and blood flow work together. Studies have focused on procedural planning and predicting short-term treatment responses; they do not track the patient over time and are not updated as new data becomes available. There is a growing interest in long-term heart modelling, especially how it remodels in response to disease or treatment. Pulmonary hypertension (PH) is defined clinically as a mean pulmonary artery pressure > 20 mmHg. Pulmonary arterial hypertension (PAH, Group I PH) affects 15 to 55 per million, poorly impacting their quality of life with fainting and breathing difficulty. Current prediction methods are not measurable in many patients, and diagnosing PH is invasive. My PhD project will carry out a long-term prospective study on 100+ PAH patients, creating cardiac digital twins that are updated when new sensor data are available. It is hypothesised that integrating physiological models of sleep and exercise into cardiac digital twins will significantly improve the prediction accuracy of disease progression and treatment outcomes in patients compared to models without sleep and exercise. These simulations will enable more precise forecasting targeted to an individual rather than a patient population. This will allow definitive and continuous bidirectional data exchange between the real heart and digital twin to truly inform disease progression.
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