Why does the fight against heart disease need AI?

Lately, we discussed the topic of the development of predictive algorithms in cardiovascular diseases The ability to predict the presentation or progression of a disease does not necessarily have to be associated only with the latest technology (although custom medical algorithms development probably won’t hurt). It is a possibility that is somehow inherent in our organism. Today, we’d like to focus on a very specific predictor. It helps us combat the world’s leading cause of death: cardiovascular disease. It’s called coronary artery calcification (CAC). And it needs AI.

Calcium Score: 1st sign of coronary artery atherosclerosis?

The largest group among cardiovascular diseases is coronary artery disease (CAD). CAD is a clinical manifestation of coronary artery atherosclerosis. In this condition, plaque builds up inside the walls of coronary arteries, causing artery narrowing. Narrowing reduces blood flow to the heart and increases the risk of heart attack and other cardiovascular events. A higher calcium score indicates a greater amount of calcified plaque buildup in the arteries and a higher risk of cardiovascular problems. Though symptoms of coronary atherosclerosis typically emerge later in life, its silent precursor, coronary artery calcification, can be detected much earlier with computed tomography (CT) scans.

Coronary calcium CT scan prognostic value

The calcium score CT scan, also known as coronary artery calcium scoring (CAC scoring, CACS) is a diagnostic method. It originated in the 1980s with pioneering research aimed at visualizing and quantifying coronary artery calcifications. The 2010s witnessed increasing recognition of the CAC score’s clinical utility, supported by numerous studies demonstrating its potential for risk stratification. Thanks to the prognostic value, both American and European medical guidelines now recommend using CAC scans in certain situations. It’s worth mentioning that US guidelines are more enthusiastic about CAC scans, considering them a helpful tool for making treatment decisions in people at intermediate risk of heart disease. [1]

a figure presenting the role of Coronary Artery Calcium Score in primary prevention according to US Guidelines

Utilizing CAC According to US Guidelines. Image source.

The calcium score algorithm we developed

Several methodologies for calculating coronary calcification have been established in clinical practice (Agatston method, volume of calcium, and calcium mass). However, clinicians typically perform them manually or semi-automatically. And this is laborious work.

Spotting and assigning calcium buildup in heart arteries is a pain. Moreover, relying just on CT scans without contrast can be tricky. That’s why we developed a custom AI algorithm to analyze CT scans of hearts with calcium buildup. This AI can not only find the calcified plaque. Additionally, it figures out which artery it’s in, even if it’s spread across multiple arteries, even into the main aorta.

To find out more, read our latest posts where we described our way to automate coronary artery calcium scoring with AI.

Example of segmentation and labelling of calcified plaques performed by the algorithm developed by Graylight Imaging
Coronary artery calcification: power of zero?

A calcium CT scan, a non-invasive and painless imaging technique, gives a unique ability to detect and quantify calcium buildup in the coronary arteries. However, we cannot consider it a sole remedy for heart disease.

The “power of zero” in CAC refers to the strong association between a zero CAC score (meaning no detectable calcified plaque in coronary arteries) and a very low risk of future cardiovascular events. Transitioning from a zero CAC score to a positive one marks a crucial turning point in heart health. This shift indicates the presence of coronary artery calcium, the underlying cause of many heart problems. From this point, CAC tends to increase exponentially, so the treatment should be introduced.

Thereupon, one would assume that zero CACS (CAC score) practically excludes coronary atherosclerosis and heart attack risk. But here comes a plot twist. Zero does not necessarily mean zero risk.

Before the calcification of coronary artery plaque

As technology advances, medicine gains new weapons against the main global killer as well as new predictive methods. As plenty of research demonstrated, we shall measure coronary artery calcification. But at the same time, we cannot underestimate its earlier variant: the non-calcified plaque. Clinical research proved that the prevalence of non-calcified plaque is not insignificant: also, among those persons with calcium score of zero [2].

Identifying asymptomatic people with plaques that haven’t calcified yet (however create risk of CAD) is truly pivotal to guiding decision-making for primary prevention now. It is estimated that over 70% of sudden cardiac events happen to asymptomatic patients. The problem is the non-calcified plaque is almost undetectable manually in clinical practice.

Heart disease battle depends on precise detection and management of all plaque buildup in coronary arteries. This is where medical technology steps in, offering custom algorithms development and AI-powered analysis of medical images. We believe that this might be the place where technological advancements will refine our risk assessment strategies. We must ensure everyone receives the appropriate preventative measures, even those with seemingly low cardiovascular disease risk scores.