AI & Cardio CT Scans

Current cardiac CT scans can identify areas of coronary artery disease

Area of work

Cardiovascular

Imaging modality

CT

Size of data set

54000

image/svg+xml

Project lead

Professor Charalambos Antoniades

Number of NHS partners

3

Unmet need

Caristo has developed a structured learning programme to analyse perivascular fat and predict the likelihood of a vascular event, thus stratifying patients for preventative pharmacotherapy.

Caristo Diagnostics and the University of Oxford have developed a new method for detecting inflammation within the coronary arteries using advanced processing of routine CT scans of the heart. A new measurement, the perivascular Fat Attenuation Index (FAITM) has been developed to capture this information and has been shown to have a striking predictive value for future cardiovascular events. Caristo Diagnostics has recently developed a new algorithm (CaRi-Heart®) which uses AI to analyse features of the arterial wall, the perivascular space (around the arteries supplying blood to the heart) and also fat tissue in the chest. This algorithm further increases the predictive value of FAITM technology

Project aims

In this project, we propose to collect CT imaging data from all NHS trusts participating in NCIMI who undertake Coronary Computed Tomography Angiography (CCTA). The images will then be used for the optimisation and automation of the CaRi-Heart® technology.

If paired patient outcomes data is available, this will be used to train our risk prediction algorithms and make further improvements to FAITM and the CaRi-Heart® platform.

This will allow us to confirm the specificity, utility and predictive value of the optimised algorithms. Finally, we will deploy the CaRi-Heart® technology to NHS trusts participating in NCIMI, our early adopters.

By analysing FAITM and CaRi-Heart® from the CT data from the last 5 years in NCIMI trusts, we will identify vulnerable patients within an otherwise “low risk” group, which will enable a better understanding of the health economic impact of this technology for the NHS and identify patients who may not otherwise have been understood to be at risk. This will build a case for the inclusion of the technology into NICE guidelines and the deployment in the NHS, a goal that is achievable within 3 years.

Objectives

  1. Automate the FAITM tools for cardiovascular risk prediction, using AI applied to cardiac/chest CT data from the NCIMI participating Trusts. Measured by the assumed reduction in “cycle time” and human “touch time” required to analyse a given CCTA once it arrives.
  2. Run a health-economic study on the added value of the FAITM technologies to the NHS and build a case for inclusion in NICE guidelines. 
  3. Return risk-profile data of the study participants to the participating NHS Trusts, free of charge, to enable local deployment of prevention measures to reduce cardiovascular risk in individual patients. 

No articles found