Coronary artery disease, occurring due to a combination of genetic and non-genetic factors, can cause serious damage to heart and human health. How these factors contribute to the occurrence of the disease is being investigated using statistical models. Through artificial intelligence, more precise roles of these factors and new factors are being explored to find applications to prevent the disease and intervene early in high-risk populations.

A heart beats throughout a person’s life, but coronary artery disease (CAD) — which ranks as the top cause of human death in the world — can cut a lifespan short. CAD is a category of heart disease that results from a combination of multiple factors, including genetics, personal lifestyle, and social environment. In most cases, the disease develops silently but strikes suddenly, causing severe chest pain, with or without breathing difficulty. Early filtering of CAD high-risk groups can not only reduce the pain of individuals but also decrease public health expenses. What if early prevention for high-risk populations became widely available? My research, directed by Professor Sarah A. Gagliano Taliun at the Montreal Heart Institute’s computational genetics laboratory, focuses on leveraging artificial intelligence for comprehensive and accurate prediction of CAD using genetic and non-genetic factors to provide metrics for early prevention and intervention.

To better understand our research objective, let’s first talk about the heartbeat and CAD. A heartbeat comes from the rhythmic contraction and relaxation of the muscles located in the inner wall of the heart, which pumps blood with abundant oxygen from the lungs and other nutrition from the body’s other organs to the whole body and transports waste from the whole body to the lungs, liver, and other organs through the blood vessels. Waste is finally excreted by exhaled air, urine, and feces. Like all human organs, to do its job, the heart also needs oxygen and other nutrition, which are delivered through blood vessels inside the muscles of the heart (called coronary arteries). If something goes wrong with the coronary arteries, such as blocking due to the buildup of fatty deposits and scar tissue usually over a long period of time, normal circulation will be restricted. This effect on circulation leads to insufficient oxygen and nutrition supply to the heart and chest pain: the clinical manifestation of CAD. If this stage is prolonged, it will cause irreversible damage to the heart muscles and eventually death.

In most cases, CAD doesn’t happen suddenly: there is a gradual buildup that eventually results in the blocking. But once it shows up, the disease is very dangerous. Without timely treatment, death is not a rare outcome. The World Health Organization reported that ischemic heart disease, one kind of CAD, is the leading cause of death globally, accounting for 16% of the world’s total deaths in 2019. Great progress has been made in the treatment of CAD, which mainly focuses on surgeries to restore blood supply to the injured muscles and maintain circulation. In some cases, to counteract the side effects of certain operations, patients are required to take medication for the rest of their lives, which significantly decreases quality of life.

The incidence of CAD in certain populations is much higher than the average level; women seem less likely to suffer from CAD. Smoking, lack of physical exercise, and diet preference also influence the disease development. Long working hours or life in a stressful social environment are also factors that cannot be ignored. Many genetic factors have been discovered that strongly relate to CAD, meaning if a person carries one or several genes that increase the risk of developing CAD, the person is more likely to develop and suffer from CAD in the future. Non-genetic factors, like the amount a person has smoked over a period of time and the level of certain chemical molecules in blood, have been used to measure the impact of lifestyle and social environment on CAD risk.

To quantify the contribution of the genetic and non-genetic factors to CAD, statistical models have been developed. In many of the models, a “score” is computed according to a linear equation to predict a person’s overall risk of developing CAD, where each factor has a positive or negative scale. Linear models play important roles in predicting incidence of CAD in individuals. In many cases, the result is approximate but satisfactory.

In recent years, artificial intelligence has proven to be an extremely powerful tool to take on prediction, classification, and other tasks. One of its powers comes from non-linear modelling to solve real problems, including the prediction of CAD. Our research is to use genetic and non-genetic factors that may affect the development of CAD in individuals to predict at-risk individuals, and we will compare which factors differ between males and females to allow for more personalized treatment options and earlier intervention.