Topological Data Analysis Offers New Means of Assessing Coronary Atherosclerosis

A recently published narrative review in Mayo Clinic Proceedings: Digital Health discusses topological data analysis (TDA) as a new and effective method of evaluating coronary atherosclerosis. As  a chronic disorder due to the accumulation of plaques in the coronary arteries, coronary atherosclerosis  is a major cause of heart disease and death. The authors performed a literature review using PubMed, Scopus, and Google Scholar with search terms including topological data analysis and coronary atherosclerosis.

Older imaging technologies, such as coronary artery calcium (CAC) scoring and CT coronary angiography (CTCA), are incapable of the detailed imaging TDA can provide.

TDA adds a new dimension by considering shape and structure of data instead of focusing on the usual metrics of imaging. TDA can reveal hidden structure and correlations because it can identify structures such as loops, holes, and other features related to significant clinical phenomena. This feature of TDA is particularly valuable because of the heterogeneous makeup of atherosclerotic lesions.

The capability of TDA to uncover specific features at varying spatial scales by a method called persistent homology is one of the primary strengths of TDA. This assists in distinguishing between stable results and noise and provides more solid risk prediction. In contrast to most deep learning models, TDA is also more transparent and interpretable and can therefore be more readily integrated into clinical workflows.

The review cites the possibility of TDA in the detection of noncalcified plaques, which are challenging to measure by means of conventional CAC scoring. Such plaques are now being identified as high risk yet underdiagnosed. TDA allows a more subtle evaluation with the recording of complex geometry of such lesions.

In addition to plaque assessment, TDA can stratify data from patients into groups based on similarities in plaque behavior and probable clinical course. Clinicians can visualize coronary data, using mathematical algorithms and distance-based filtration, making it more personalized to the treatment plan.

The review indicates that clinicians might develop a better perception of plaque biology by combining TDA with existing imaging tools such as CTCA and CAC. This may improve risk stratification and lead to better diagnosis and more effective treatment plans.

Although it is still a developing area, TDA has the potential to change the nature of evaluation and management of coronary atherosclerosis. Its clinical usefulness still requires validation, but initial reports suggest its applicability in the future of precision cardiovascular care.