Hidden heart risks: How AI finds what doctors often miss

We’ve all heard the story — a patient feeling unwell goes to the doctor, only to be told their test results were normal. It happens more often than we’d like to admit, as traditional diagnostic methods don’t always catch early signs of disease and leave a gap between patients and the treatment they need. An emerging AI-powered digital health platform is hoping to bridge this gap in cardiology by detecting hidden heart risks.

This AI is a game-changer, according to a study from Imperial College London, as it can spot patients at double the risk of experiencing major adverse cardiac events (MACE), including heart attacks and heart failure.

“Notably, the AI identified at-risk patients who had unremarkable results from traditional diagnostic tools, such as echocardiograms,” said Prof. Nicholas S. Peters, Director of the Health Impact Lab at Imperial, in a press release. “This highlights the technology’s ability to detect hidden risk factors, offering healthcare providers a powerful tool for early intervention and improved patient outcomes.”

The study evaluated more than 1,000 patients who had a single-lead electrocardiogram (ECG) recording performed during routine clinical practice. By applying the detection platform’s AI algorithm to these ECGs, researchers aimed to identify those at elevated risk for major cardiac events and mortality over a two-year period. The goal was to enhance early detection of heart conditions that might be missed by standard diagnostics, ultimately improving patient prognosis through timely intervention.

According to Patrik Bächtiger, PhD., one of the co-leads of this research at Imperial, the AI technology only requires the straightforward use of a single-lead ECG, yet it represents a huge step forward in our ability to stratify cardiac risk at an early stage. More than 800,000 people suffer heart attacks every year in the U.S. alone. How many of these could be prevented with earlier detection?

The technology featured in the study is Eko Health’s SENSORA platform, which combines ECG and heart sound analysis to detect low ejection fraction conditions as well as structural murmurs.

Eko’s AI was trained using vast numbers of ECG recordings and taught to recognize patterns and anomalies associated with reduced cardiac function and elevated risk of adverse events. It leverages this experience when analyzing new ECGs, making it capable of spotting risk factors and indicators easily missed by traditional diagnostic methods.

To make its technology immediately accessible to clinicians, Eko has developed digital stethoscopes powered by its AI. Dozens of healthcare practices have already begun using the stethoscopes, which can be incorporated into routine care with minimal training.

Bringing AI-powered care home

Looking ahead, the impact of this new technology extends well beyond use in hospitals and primary care practices. Eko’s digital stethoscopes don’t require a clinician to operate, opening the door to more effective at-home care and remote patient monitoring. With minimal supervision, patients can use the AI-enabled stethoscopes to transmit data to their healthcare providers from their homes, improving the timeliness of identifying otherwise hidden heart risks.

An additional Imperial study indicated that Eko’s stethoscopes could predict changes in left ventricular ejection fraction (LVEF), which indicates how well a heart is working. By allowing patients to monitor these changes from home, the use of digital stethoscopes can lead to earlier interventions and timely adjustments to patients’ treatment plans with fewer hospital visits, a development the study describes as nothing short of “transformative.”

Another area where AI-powered stethoscopes can have a significant impact is maternal healthcare. The early detection of potential heart conditions can be lifesaving for both mother and child, and pregnant women, especially those in high-risk categories, would benefit from using Eko’s stethoscopes for regular, remote monitoring.

A study published in Nature Medicine found that Eko’s digital stethoscope helped doctors in Nigeria, where more women experience pregnancy-related heart failure than anywhere else, identify double the rate of heart failure compared to a control group receiving standard care.

The study also found that the AI-powered stethoscope was able to identify the possibility of peripartum cardiomyopathy, a specific type of heart failure, 12 times more successfully than traditional screening. This type of heart failure is treatable, but when undiagnosed can worsen to the point of threatening the life of the mother.

The impact of Eko’s stethoscopes on improving healthcare and potentially saving the lives of pregnant women are huge — in the U.S., 1 in 2,000 women and 1 in 700 African American women are affected by peripartum cardiomyopathy, as noted in the study, and by making remote monitoring possible for pregnant women, the AI-powered stethoscopes can improve access to care for patients in remote or underserved areas, reducing disparities in healthcare delivery.

When considering the widespread benefits that tools like Eko’s AI-enabled stethoscopes can bring to patient care, including earlier detection of hidden heart risks and potentially life-threatening conditions, suddenly the distance that has often existed between the patient experience and the ability of healthcare providers to diagnose health issues seems much smaller.