Active Heart, Stroke & Blood Lungs & Breathing

Connectivity to Enable Accurate Patient-Specific Cardiovascular Assessments

In plain English

AI plain-English summary

A pilot project will refine a non-invasive 3D heart model to improve how doctors diagnose blocked arteries from standard X-ray images. Cardiovascular disease kills one in three people globally and costs the NHS £10 billion each year. The current gold-standard diagnostic test, fractional flow reserve, is invasive, risky, and expensive. A non-invasive alternative called virtual FFR (VIRTUheart) already exists, but it becomes inaccurate when arteries are twisted or poorly visible on angiograms. This project tackles those errors by improving the 3D reconstruction of blood vessels from 2D images, simulating how dye moves through the arteries, and accounting for blood flow into tiny side branches. If successful, the work could make virtual FFR reliable enough to replace many invasive procedures, cutting costs and expanding access to accurate diagnosis. The techniques developed here—modelling uncertainty, pulsatile flow, and contrast dynamics—could also transfer to other fields such as lung, brain, or cancer imaging. The team expects their findings to inform national clinical guidelines and help set regulatory standards for computational diagnostics, moving the technology toward commercial use.

View original technical description
Context Cardiovascular diseases (CVD) are the leading cause of death globally, responsible for around 1 in 3 deaths annually. The economic burden is significant, costing the NHS approximately £10 billion and the wider economy £25 billion each year. Fractional flow reserve (FFR) is the current gold standard for the assessment of coronary artery disease (CAD), the most common CVD. But it is invasive, time-consuming, expensive and risky. Accurate, non-invasive diagnostic tools such as virtual FFR (vFFR - VIRTUheart) and absolute coronary blood flow can improve patient outcomes and comfort, pathway efficiency and reduce costs. Challenges to be Addressed State-of-the-art diagnostic models, including VIRTUheart, face limitations in clinical deployment due to reconstruction instability (Objectives 1-2) but provide an ideal foundation for more comprehensive diagnostic metrics based upon improved and expanded physiological content (Objectives 3-5). Complex cases, characterised by very twisted vessels and inadequate contrast in 2D angiograms inter alia result in vFFR inaccuracies. These same inaccuracies obstruct the shift from the relative metric of vFFR to absolute flow for CAD assessment. Meanwhile, an ageing population with increasing prevalence of CAD is escalating the demand for rapid, efficient, scalable, and deployable diagnostic solutions. Objectives and Methods This pilot project will strengthen collaborations between computing engineer and mathematician Xu, consultant cardiologists Gunn and Morris, and medical physicist Halliday. It will also enable Xu to gain cardiology insight and deliver novel engineering and physical sciences (EPS) research from a clinical user perspective, addressing EPSRC health technologies’ ‘transforming prediction and early diagnosis’ challenge. We aim to refine and extend the current patient-specific VIRTUheart models with more accurate captures of anatomy and embedded physiology, for effective and efficient CAD assessment and treatment planning. The team seek to: Enhance tool geometry reconstruction and quantify uncertainty: Utilise digital and 3D-printed phantoms to validate and minimise errors in 3D artery reconstruction from 2D angiographic images. Quantify uncertainty in the models to produce predicted metrics with confidence bounds. Simulate contrast medium dynamics: Understand the impact of radio-opaque dye transport on angiography, using convection-diffusion models to predict angiographic image accuracy. Incorporate sequestration flow: Develop novel models to account for blood flow into unresolved side-branch vessels, modelled as a porous wall flux in 3D. Verify absolute distal flow as an alternative CAD diagnostic metric: Validate blood flow simulations against in vivo measurements to establish and verify the diagnostic reliability of absolute flow. Introduce pulsatility: Incorporate pulsatile blood flow effects into current VIRTUheart models. Develop and validate pulsatile flow boundary conditions using digital phantoms and quantify the diagnostic significance of temporal complex flow dynamics. Potential Applications and Benefits Multiple benefits stem from this project: (i) enhanced accuracy in detecting and assessing CAD, leading to better patient outcomes; (ii) more accurate vFFR calculations will reduce the need for invasive procedures and impacting costs, and increase accessibility and patient participation; (iii) advanced understanding of coronary anatomy and haemodynamics will support new treatment approaches and personalised interventions accounting for patient characteristics. The novel EPS research and techniques developed in this pilot project could be applied to other computational medicine areas modelling physiological flow, e.g. pulmonology, neurology, oncology and orthopaedics. We expect the project findings to inform clinical practice and national guidelines on the use of vFFR and establish regulatory standards for the accuracy and reliability of medical imaging and computational diagnostics - a step towards commercialisation.

View the original record at the funder ↗

Researchers

Ian Halliday (Co-Investigator)Julian Gunn (Co-Investigator)Paul Morris (Co-Investigator)Xu Xu (Principal Investigator)

Related Research

Grants with similar aims, by meaning.

Non-invasive assessment and management of coronary heart disease - a translational, data driven approach
Optimising a computational model of the physiology of coronary artery disease: time to make it personal
Hidden haemodynamics: A Physics-InfOrmed, real-time recoNstruction framEwork for haEmodynamic virtual pRototyping and clinical support (PIONEER)
Computational Modelling to Improve the Assessment and Management of Patients With Coronary Artery Disease
Personalised 3D modelling of the right ventricle as an approach to ARVC

Original classification

Research and Innovation

Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.