Real-time Virtual Fractional Flow Reserve Assessment via Deep Learning

Complete

24/05/2018

South East

Data

Awarded amount: £238,762

Grant scheme: NET

Institution: King's College London

Principle Investigator: Dr Jack Lee

---

This project seeks to combine AI and computational methods to develop a non-invasive method for measuring pressure within blood vessels. These measurements are important in determining whether patients with coronary artery disease would benefit from a stenting procedure.

The current methods by which the measurement is made are invasive and the uptake of them is very low as a result. In order to develop a novel, non-invasive protocol, the team aimed to construct 3D images of coronary blood vessels, estimate pressures within the vessels and then evaluate these measurements using traditional techniques in the cathlab. This will allow an AI system to be trained to estimate pressures in blood vessels based on angiogram images.

Unfortunately, this project was heavily impacted by the pandemic and the clinical aspects of the project were unable to be fulfilled. However, the outcome that was achieved represents the first development of a an automated AI system for the assessment of coronary artery pressure.

This will ultimately advance diagnostic methods for CAD and produce better outcomes at reduced spend. A number of smaller concurrent projects are currently in motion to continue development of the technology, and a follow up project for clinical translation is in preparation. This project has also resulted in a high number of publications with more to come over the coming months.