Introduction
The digital twins concept stems from engineering and manufacturing. It consists in the use of a real-time virtual counterpart of the physical entity, such a plane or bridge, is used for monitoring. Due to the two-way data interaction between the real entity and the virtual replica, the digital twin accurately reflects changes over time. In healthcare, digital twins could be used to monitor whole patients, organs, or processes, yet this technology is not widely employed. This project aims to develop digital twins for monitoring patients with pulmonary arterial hypertension (PAH), a life-threatening cardiovascular disease.
Explaining the science
Digital twins of the cardiovascular system are virtual copies of the human heart and circulatory system, created using advanced computational models derived from physics principles and medical data. These twins aim to simulate physiological behaviour of the cardiovascular system, including the heart’s pumping action and blood flow dynamics. In this context model parameters can include patient-specific physiological properties of the heart and vessels, such as cardiac contractility, blood volume and compliance, properties which are difficult or impossible to monitor remotely. By constantly updating model using data from implanted monitors, as well as imaging data, this project will investigate if digital twins can be used to better forecast the patients’ wellbeing and the response to treatment.
Project aims
The project will assess whether the use of digital twins for the cardiovascular system in NHS patient care for PAH is feasible, scalable, and affordable. The project goals include:
• Creating a sensitivity analysis pipeline based on model emulators to inform model selection based on available data.
• Investigating different cardiovascular system models and their ability to capture the pathological changes caused by PAH
Applications
The purpose of this project is to complement the data available to clinicians by unlocking insight into the mechanistic changes as a result of pathology. Potential applications include:
• Treatment planning: manipulating model parameters to simulate the effects of different strategies.
• Monitoring and planning: tracking parameter evolution over time to assess the effectiveness of interventions and/or drug response and predict future outcomes and complications.