ARTIFICIAL intelligence (AI) is emerging as crucial component of health care to help improve performance and efficiency of physicians. AI includes deep-learning algorithms, machine learning, computer-aided detection (CAD) systems, and convolutional neural networks.

In medical imaging, and more particularly cardiovascular imaging, AI has become very instrumental in efficiently managing imaging work lists, enabling structured reporting, auto detecting injuries and diseases, and pulling in relevant prior exams and patient data.

In clinical cardiology practice AI helps automate tasks and measurements in imaging and in reporting systems, guides novice diagnostic imaging operators to improve imaging accuracy, and can risk stratify patients.

AI and electrocardiograms (ECG)

An electrocardiogram (ECG) is a primary cardiac diagnostic test that measures the electrical activity in the heart to identify overall heart function, heart rhythm abnormalities, and areas of poor blood circulation and prior heart attack. A standard 12-lead ECG breaks the 3D structure of the heart into 12 zones, each showing the electrical activity in that specific area of the heart. This helps pinpoint areas where there are issues with heart blood vessel blockage or electrical performance. Many home-use heart monitors and consumer grade ECG machines use fewer leads and so are less specific as to cardiac conditions or location of abnormal heart rhythms but they can show areas of concern that may warrant further diagnostic testing or treatment. With AI we are able to expand the utility of the ECG, even when obtained with machines with otherwise limited capability.

AI can be applied to electrocardiogram (ECG) data captured by smart watches for example to identify patients with weak heart pumps, according to a new analysis published in Nature Medicine. These findings were reported in a recent study led by cardiologists from Mayo Clinic exploring ECG data from nearly 2,500 patients. The mean patient age was 53 years old and 56 per cent of the patients were women.

The group applied a 12-lead algorithm for poor heart function to the ECG data and found that AI was able to accurately detect poor heart function in nearly 90 per cent of cases when compared with a cardiac ultrasound. Currently, we diagnose heart dysfunction — a weak heart pump — through an echocardiogram (cardiac ultrasound), CT scan or an MRI but these are expensive, time-consuming and often not routinely accessible, especially in patients in rural areas. Therefore, the ability to diagnose a weak heart pump remotely, from an ECG, could lead to a significant improvement in the care of heart patients as it could allow for a timelier detection of potentially life-threatening disease at massive scale and in a more cost-effective way.

With the advent of consumer grade wearable medical devices the opportunity to extend care to patients remotely is profound. Increasing the capability to accumulate data from wearable consumer electronics and provide analytic capabilities to prevent disease or improve health remotely in the manner demonstrated by this study can transform and expand access to health care.

AI in atrial fibrillation

Atrial fibrillation (Afib) is a common electrical disturbance of the heart manifesting with irregular heart rhythm. Afib is commonly seen in patients with long-standing hypertension and in older persons. There is accumulating evidence that women may be more at risk for Afib than men. Afib over time can cause the clotting of blood inside the heart and may lead to strokes. The economic cost of the complications of Afib on individuals and society can be very significant. Using AI to screen patients for atrial fibrillation can detect new cases that may have otherwise been missed, according to new research published in The Lancet. The application of this work hopefully can play a major role in preventing life-threatening strokes.

Traditional screening programmes select patients based on age (65 or older) or the presence of conditions such as high blood pressure. While these approaches seem reasonable because advanced age is one of the most important risk factors for Afib, it is not feasible to repeatedly conduct intensive heart monitoring in millions of older adults globally.

The investigators aimed to see if AI could provide clinicians with another screening option for identifying Afib patients. In the study involving over 2,000 patients the researchers demonstrated that AI-guided screening was linked to more detected cases of Afib among both high- and low-risk patients, and confirmed that an AI algorithm can select a subgroup of older adults who might benefit from more intensive monitoring. This is particularly valuable in low-resource environments where undetected and untreated Afib can lead to catastrophic consequences and impose additional drain on an already burdened health system.

Use of AI in chest pain evaluation

In the past 25 years cardiac computed tomography (CT) has emerged as a primary cardiovascular imaging modality and has been recommended as a viable tool for chest pain assessment in the hands of experienced and appropriately trained operators. CT plays a key role in evaluation for structural heart disease — especially for heart valves, repair of congenital defects, and left atrial appendage occlusion (LAAO) for both pre-procedure planning and procedural guidance in patients with Afib.

CT calcium scoring is also now used routinely for risk assessment for coronary artery disease and to determine which patients could benefit from the use of cholesterol-lowering and stabilising medications. Coronary CT angiography (CCTA) is used for non-invasive atomical assessment of the heart arteries to determine plaque burden and to identify areas of blockage that may cause ischemia and heart attacks. Additional use of contrast CT perfusion or fractional flow reserve CT (FFR-CT) can offer physiological information on the function of the heart.

Applying advanced AI algorithms for CCTA has been shown to provide additional benefits in patients with stable chest pain and suspected coronary artery disease (CAD), according to a late-breaking clinical trial presented at the recently concluded American Heart Association Scientific Sessions held in Chicago in early November 2022.

The Prospective Randomised Trial of the Optimal Evaluation of Cardiac Symptoms and Revascularization (PRECISE) Study reported data on more than 2,000 adult patients who presented with stable chest pain and suspected CAD.

Dr Ernest Madu, MD, FACC and Dr Paul Edwards, MD, FACC are consultant cardiologists for Heart Institute of the Caribbean (HIC) and HIC Heart Hospital. HIC is the regional centre of excellence for cardiovascular care in the English-speaking Caribbean and has pioneered a transformation in the way cardiovascular care is delivered in the region. HIC Heart Hospital is registered by the Ministry of Health and Wellness and is the only heart hospital in Jamaica. Correspondence to info@caribbeanheart.com or call 876-906-2107