How is Artificial Intelligence used in cardiology?

Discover the uses of Artificial Intelligence in monitoring and preventing cardiovascular diseases in patients
Thu 08 Oct 2020

Cardiovascular diseases (CVDs) are a leading cause of death globally. According to the World Health Organization, around 17.9 million people lose their lives from CVDs each year, an estimated 31% of all deaths worldwide. Of these deaths, 85% are due to heart attack and stroke. People with cardiovascular disease, or people who are at high cardiovascular risk need early detection and timely prevention, since sometimes all it takes are seconds to prevent a fatality.

Let’s see how Artificial Intelligence (AI) can help patients suffering from cardiovascular diseases and provide better information for physicians.

The current state of Artificial Intelligence in cardiovascular medicine

Why Artificial Intelligence is becoming so popular in healthcare

The volume of data collected in healthcare is increasing dramatically over the past years. Artificial Intelligence is used to interpret complex data from CT scans, MRIs, EHR, biobanks, clinical trials, wearables, clinical sensors, and many more.

Artificial Intelligence can :

  • Deal with highly dimensional data (made up of many variables);

  • Find invisible patterns in this data;

  • Automatically learn from this data to define what is the best course of action (diagnosis, prediction, etc..);

  • Assist practitioners in their work.

Areas of development

In recent years, the technological improvements in Artificial Intelligence have encouraged the deployment of algorithms in prevention, diagnostics, risk stratification, and treatment selection tasks in medicine.

AI algorithms can assist cardiologists in making better diagnostic decisions, monitor patients’ health, predict the risk of a heart disease, and recurrence of a pathological condition.

Artificial Intelligence for pathology diagnosis

Artificial Intelligence can give cardiologists a ‘leg up’ when evaluating hearts for possible conditions. It can enable early detection of a dysfunction and identify individuals at increased risk for future development of a heart disease. For example, a team of researchers at Stanford University is using Artificial Intelligence to detect heart abnormalities through an algorithm that assesses the rate at which the heart pumps blood. The researchers showed that their algorithm performs at the same level as a high-trained cardiologist when it comes to reading echocardiograms, and is able to assess health conditions more rapidly and more thoroughly.

Artificial Intelligence for heart monitoring and self-care

Artificial Intelligence wearable sensors (often called wearable Artificial Intelligence) can help medical professionals monitor and follow patients' health conditions. The sensors can help medical professionals monitor the heart rate, heart rhythm, respiratory rate, walking, sleeping, and other patient’s activities. Artificial Intelligence can help determine a normal baseline for each individual patient, and trigger a clinical alert if/when a patient’s condition is getting worse. This may lead to earlier and easier detection of heart problems and, therefore, ensure patients get the right treatment, saving lives.

For example, Apple has been promoting the watch to doctors and scientists as a serious health tool. Together with a startup called Cardiogram, that uses Artificial Intelligence to analyze wearable’s data, Apple Watch could flag signs of a stroke, and even spot signs of diabetes.

Apart from monitoring patient’s health conditions, wearable devices can further serve as a good platform for building Artificial Intelligence prediction models, focusing on early warning signs of lifestyle diseases including cardiovascular anomalies.

Artificial Intelligence for pathology prediction

Artificial Intelligence holds an immense potential to improve the conventional way of predicting a person’s risk of suffering from a heart disease.

Kantify’s discovery on Atrial Fibrillation prediction

Atrial Fibrillation (AFib) is a quivering or irregular heartbeat (arrhythmia) that can lead to blood clots, stroke, heart failure, and other heart-related complications. It is a pathology that affects 2% of the total world population and is largely un- or underdiagnosed.

Over the past year, the Kantify team has worked together with Universite Libre de Bruxelles (ULB) to tackle the challenge of predicting an upcoming episode of Atrial Fibrillation in individual patients. We have successfully developed an AI-based model that can predict Atrial Fibrillation at individual level, without any history or data about the patient, except for their RR-intervals collected through a Holter monitor. We managed to achieve good results, 30 seconds before the Atrial Fibrillation event occurs. The model can also be used to detect AFib.

This is particularly noteworthy because it was previously not known that Atrial Fibrillation events did show any perceptible marks before their occurrence. “We are still in the prediction phase. We have to make sure that this becomes prevention in order to help patients in a concrete way”, says Segolene Martin, CEO of Kantify, who emphasizes that the algorithm is a world-first.

Google’s model to predict cardiac risk based on eye scans

Google published last year in the Nature journal Biomedical Engineering a surprising discovery. By analyzing images of the back of a patient’s eye, the company’s AI model is able to accurately deduce parameters, including an individual’s age, blood pressure, and whether or not they smoke. This can then be used to predict their risk of suffering a major cardiac event - such as a heart attack - with roughly the same accuracy as current leading methods.

Preparing for a world of fully individualized medicine?

The rise of AI in healthcare opens new doors to individualized diagnoses and treatments. But the road until a world of personalized medicine is still long.