Networks of Cardiovascular Digital Twins (CVD-Net)

Networks of Cardiovascular Digital Twins (CVD-Net): Transforming Healthcare through Personalised Predictive Modelling The Networks of Cardiovascular Digital Twins (CVD-Net) Programme Grant aims to revolutionise healthcare by harnessing the power of digital twin (DT) technology. Patient DTs are virtual replicas that continuously assimilate patient data into sophisticated models to provide personalised predictions and inform clinical decisions. Healthcare, despite its national importance (consuming 12% of GDP, generating £70 billion/year and 240,000 jobs), remains unserved by DT technologies. CVD-Net will build a critical mass of research around patient DTs for healthcare, identify the challenges and opportunities in the clinical setting, and provide a roadmap for NHS implementation. We take the view that we must begin by focussing on a specific clinical use case, and that we need to learn by doing, using real-world data, on clinical timescales and making testable predictions. We propose a flexible Programme structure built around developing a minimum viable DT, then testing, optimising, and evaluating the this over iterative design cycles. We focus on pulmonary arterial hypertension (PAH), a life-threatening cardiovascular disease with high mortality and adverse event rates, as a specific use case to develop a demonstrator NHS DT care pathway. The public and patients are receptive to the idea of DTs with 90% (173/196) agreeing with the statement "I would find a digital twin smartphone app that represents my individual cardiovascular health useful". PAH patients suffer high mortality, frequent clinical worsening events and are served by a limited number of national centres. These high event rates and concentration of patients make it possible (and important) to develop and test the forecasting capabilities of a DT in proof-of-concept studies within CVD-Net. Our objective is to create a comprehensive patient DT that can monitor and forecast disease progression, treatment response, and quality of life for individual patients. The DTs will combine data from hospitals, wearable and implantable sensors, and patient-reported outcomes. To realise DTs at the scale and speed of a clinical service, we propose a novel networking approach, where individual "digital threads" (within a DT) will be 'woven' together to form an interconnected 'digital tapestry' to facilitate shared learning and communication. We will utilise innovative techniques including knowledge graphs, transfer learning, federated learning, and meta-learning to address scalability, variability, uncertainty, and data security challenges. We have brought together a unique interdisciplinary team of engineers, clinicians, computational statisticians, and research engineers to deliver CVD-Net. We will access retrospective and collect prospective data to train, test and validate the network of DTs. We will build the IT infrastructure, and analysis workflows to run a demonstrator DT care pathway within the NHS infrastructure. We will work with patients, clinicians, and stakeholders to assess its usability and added value. Via stakeholder engagement, we will evaluate the feasibility, scalability, and wider adoption potential of networked patient DTs in patient care. By generating robust evidence and understanding patient, clinician, and policy considerations, by completion of CVD-Net, we aim to have moved DTs towards prospective evaluation in a clinical trial. Ultimately, CVD-Net has the potential to transform healthcare by providing personalised predictive modelling, enhancing clinical decision-making, and improving patient outcomes. Its applications will benefit patients, clinicians, policymakers, and the research community, making healthcare more precise and efficient while contributing to the transformation of NHS care.