AI-guided stethoscope helps detect cardiomyopathy during pregnancy

A screening programme that incorporates artificial intelligence (AI)-guided digital stethoscope improves clinicians’ ability to detect peripartum cardiomyopathy in the Nigerian SPEC-AI trial.

Nigeria has a high maternal mortality, with one in 96 pregnancies attributed to peripartum cardiomyopathy, according to study author Dr Demilade Adedinsewo from the Mayo Clinic, Jacksonville, Florida, US. She added that Non-Hispanic Black women are disproportionately affected by maternal death, with a threefold higher risk than White women.

 “This portable, digital stethoscope [Eko Health] can take a single-lead EKG and a phonocardiogram,” she said during her presentation at AHA 2023. “The embedded AI can predict the likelihood of a left ventricular systolic dysfunction, with the results displayed on a phone or tablet in real time.”

Future applications

Adedinsewo said that the AI-guided stethoscope can be used in both clinical and nonclinical settings. “This simple intervention has the potential to improve the quality of obstetric care by reducing delays in the diagnosis of this life-threatening but treatable maternal condition.”

Dr Alexander Sandhu from the Stanford University, California, US, meanwhile commented that clinicians really need portable and scalable AI tools to solve patients’ health problems. “This digital stethoscope is a phenomenal example of that.”

He was surprised that 4 percent of otherwise healthy young pregnant women in the intervention arm had reduced ejection fraction, which speaks to the importance of improving the diagnosis in this space.

All- Nigerian cohort

The study involved 1,232 Black women (mean age 31 years) from Nigeria, who were pregnant or had delivered within the past year. They were enrolled between August 2022 and September 2023 at six teaching hospitals in the country. [AHA 2023, abstract LBS.08]

Women randomly assigned to the intervention group had an ECG plus the AI-guided stethoscope, whereas those in the control group received usual care plus an ECG. Adedinsewo explained that both groups received the ECG to validate the stethoscope in an all-Black, out-of-US population as the device was developed and validated in a largely White population in the US.

Left ventricular systolic dysfunction

Detection of left ventricular systolic dysfunction (<50 percent) on echocardiography occurred in 4.1 percent of patients in the intervention group and 1.8 percent in the control group (odds ratio [OR], 2.3, 95 percent confidence interval [CI], 1.1-4.8). The results were similar across all subgroups regardless of region, age, and pregnancy status at baseline.

The AI-guided group also had a higher rate of cardiomyopathy, with left ventricular ejection fraction <50 percent (OR, 2.05, 95 percent CI, 1.03-4.32) compared with the control group.

Sensitivity and specificity

Overall, the digital stethoscope had an area under the curve of 0.93 (95 percent CI, 0.88-0.99) for detection of LVEF <50 percent and 0.98 for detection of LVEF <40 percent (95 percent CI, 0.97-1.00). Sensitivity was 92 percent. Specificity was 80 percent.

Study discussant Dr Marco Perez from Stanford University Hospital, California, US said the study raises the question of who benefits the most from this screening technique. He added that genetic studies are needed to understand why peripartum cardiomyopathy is more prevalent in Nigeria than in other countries.

As a parting shot, Adedinsewo said the AI-guided stethoscope could broaden opportunities for screening in resource-limited settings. However, the algorithm that uses the AI technology was created for the study alone and is not currently embedded in the stethoscope, she added.