Heart failure is a serious condition that affects millions of people worldwide. It happens when the heart becomes too weak or stiff to pump enough blood through the body.
This can lead to symptoms such as breathlessness, swelling, tiredness, dizziness, and difficulty performing normal daily activities.
Although modern medicine has improved treatment for heart failure, early diagnosis remains a major challenge in many parts of the world.
This problem is especially severe in low-income countries where hospitals may not have enough advanced equipment or trained heart specialists. Many people are diagnosed only after the disease has already caused serious damage.
Now, a new study suggests that artificial intelligence could help solve part of this problem using one of the simplest and cheapest heart tests available.
Researchers from UT Southwestern Medical Center and collaborators in Kenya found that artificial intelligence combined with standard electrocardiograms, called ECGs, can successfully detect hidden heart weakness in patients.
The findings were published in JAMA Cardiology.
An ECG is a common medical test that records the electrical activity of the heart using small sensors attached to the skin. The test is quick, inexpensive, and already widely used in clinics around the world.
Normally, ECGs alone cannot always clearly identify early heart pumping problems. However, researchers trained an artificial intelligence algorithm to recognize subtle hidden patterns linked to heart dysfunction.
The scientists focused on a condition called left ventricular systolic dysfunction, or LVSD.
The left ventricle is the heart’s main pumping chamber. When it becomes weak and cannot pump blood properly, patients are at higher risk of developing full heart failure later.
Doctors usually diagnose LVSD using echocardiography, an ultrasound imaging test that creates moving pictures of the heart.
Echocardiograms are considered the gold standard for evaluating heart pumping ability. However, the machines are expensive, and trained specialists are needed to perform and interpret the scans.
Many healthcare systems in sub-Saharan Africa and other developing regions do not have enough access to these services.
To explore whether AI-enhanced ECGs could help, researchers studied nearly 6,000 patients receiving routine healthcare at eight medical facilities in Kenya.
Every participant received AI-ECG screening. A subgroup of 1,444 participants also underwent echocardiograms so researchers could compare the AI findings with standard imaging results.
The AI system performed surprisingly well.
The algorithm correctly identified more than 95% of patients who actually had LVSD.
The test also showed strong specificity, correctly identifying many people who did not have the condition.
Perhaps most importantly, the AI-ECG achieved a negative predictive value of more than 99%. This means that when the test suggested someone did not have LVSD, it was almost always correct.
This could be extremely useful in healthcare systems with limited resources because it allows doctors to quickly rule out serious heart dysfunction in many patients without needing expensive scans.
Researchers also found that patients with positive AI-ECG results often showed other signs of unhealthy heart remodeling, including thickened heart muscle and problems with heart relaxation.
These findings suggest the AI system may help detect multiple forms of hidden heart disease.
The study was led by Dr. Ambarish Pandey from UT Southwestern Medical Center together with researchers from Kenya, including Dr. Bernard Samia from M.P. Shah Hospital and the Kenya Cardiac Society.
The researchers believe AI-ECG screening could become a practical and scalable way to improve heart disease detection in underserved regions.
Heart failure rates are rising rapidly in many developing countries. Patients in sub-Saharan Africa often develop heart disease at younger ages and face worse outcomes compared with patients in wealthier countries.
Limited access to early diagnosis is one major reason.
The researchers say combining inexpensive ECG testing with artificial intelligence could help identify high-risk patients earlier, allowing treatment to begin before severe heart failure develops.
The study also demonstrates how AI technology may help reduce healthcare inequalities around the world.
Instead of relying only on expensive machines available in large hospitals, clinics with basic ECG equipment may eventually be able to use AI software to screen large numbers of patients quickly and affordably.
Still, researchers emphasize that AI systems are not perfect and should be used alongside medical professionals rather than replacing doctors.
More studies will be needed to confirm how well AI-ECG screening works in different populations and healthcare systems.
However, the findings offer hope that simple AI tools could dramatically improve heart disease screening in parts of the world where advanced medical testing is difficult to access.
Overall, the study highlights how combining artificial intelligence with existing low-cost technology may help doctors identify heart problems earlier, improve treatment access, and potentially save many lives worldwide.
If you care about heart health, please read studies about how eating eggs can help reduce heart disease risk, and Vitamin K2 could help reduce heart disease risk.
For more information about heart health, please see recent studies about how to remove plaques that cause heart attacks, and results showing a new way to prevent heart attacks, strokes.
Source: UT Southwestern Medical Center.
