Angiography and blood parameters based prediction of fractional flow reserve in coronary arteries using machine learning

Date

2026-6-26

Author

Kaçar, Mehmet Nazir

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

71
views

38
downloads

Cite This

Coronary artery disease (CAD) is the most common cardiovascular disease with high morbidity, disability, and societal burden. The CAD occurs due to atherosclerotic occlusions of the coronary arteries. Coronary angiography is the gold standard of clinical diagnosis of coronary heart disease. Measurement of fractional flow reserve (FFR) during routine coronary angiography by using a pressure wire to calculate the ratio between coronary pressure distal to a coronary artery stenosis (after the blockage in the vessel) and aortic pressure aids clinicians to make decisions in certain circumstances. However, FFR is an invasive, complicated and expensive process and has to be performed during angiography. Several methods of FFR estimation by angiography without requiring the instrumentation of the coronary artery have been developed. These methods often employ various approaches to fluid dynamics computation (CFD). Complexity of CFD process has led to various assumptions regarding boundary conditions and blood physical properties. Various machine learning (ML) models are being applied to predict FFR using angiography images. The main limitation of angiography-based predictions is its failure to consider the physical properties of the fluid being measured: blood. Because FFR measures blood flow impairment, it is inherently affected by this rheological behaviour. We propose the use of blood parameters with X-ray coronary angiography images to improve prediction, and a novel model that combines blood parameters with angiography videos. We retrospectively collected patient data and experimentally showed that adding blood parameters improves prediction.

Subject Keywords

Angiography, Machine learning, Coronary artery disease, Fractional flow reserve, Blood parameters

URI

https://hdl.handle.net/11511/119785

Collections

Graduate School of Natural and Applied Sciences, Thesis