Study finds AI-enabled stethoscope falls short in diagnosing heart conditions in cats and dogs

An AI-enabled digital stethoscope may have significant limitations when used to identify heart murmurs and abnormal heart rhythms in cats and dogs, according to a new study from North Carolina State University.

Researchers found that the device performed moderately well at detecting heart murmurs in dogs but missed most murmurs in cats. It also produced unreliable classifications of heart rhythms in dogs. The findings suggest that the technology may be more useful as an aid to veterinary examination than as a standalone diagnostic tool.

The prospective study, published online August 5 in the Journal of the American Veterinary Medical Association, evaluated 105 companion animals—54 dogs and 51 cats. Each animal underwent cardiac auscultation by a board-certified veterinary cardiologist, a cardiology resident and a fourth-year veterinary student. Their findings were compared with those produced by the AI-enabled stethoscope.¹

The device tested in the study was designed for human use, rather than specifically for veterinary patients. The researchers said this distinction may be important because cardiac sounds and rhythms can differ substantially between species.

“However, the diagnostic AI for these stethoscopes is trained on human data, not dog or cat data,” said Kursten Pierce, DVM, DACVIM (Cardiology), FACVIM (Interventional Cardiology), an assistant professor of clinical sciences and board-certified cardiologist at NC State's College of Veterinary Medicine²

AI performs differently in dogs and cats

Among the 54 dogs examined, clinicians identified heart murmurs in 38 animals, or 70%. The AI software correctly detected 33 of those murmurs, for a sensitivity of approximately 87%. Fourth-year veterinary students performed at the same level as the device in identifying canine murmurs.

The results were less consistent when the device assessed heart rhythms. Clinicians identified arrhythmias in 24 dogs, including 6 cases of atrial fibrillation. The AI system correctly identified all 6 dogs with atrial fibrillation but also incorrectly classified 22 additional dogs as having the condition.

The device did not classify any of the dogs as free of an arrhythmia, according to the study. That means its automated rhythm assessment produced false-positive findings even among dogs whose rhythms were considered normal by the veterinary clinicians.¹

The device's performance was substantially worse in cats. Veterinary clinicians identified murmurs in 22 of the 51 cats, or 43%. The AI system detected murmurs in only 2 cats, missing 20 of the murmurs identified by the veterinary experts. Fourth-year veterinary students detected nearly two-thirds of the murmurs in affected cats.

“The most clinically concerning result to me was cats,” said Jake Johnson, DVM, a cardiology resident at NC State and first author of the study. “This device essentially couldn’t find a murmur, so a tool vets lean on for reassurance could let real disease go undetected.”

The researchers cautioned that a negative result from the device, particularly in cats, should not be interpreted as evidence that a heart murmur or other cardiac problem is absent.

Researchers urge caution with automated diagnoses

The findings do not necessarily mean that digital stethoscopes have no role in veterinary medicine. The device can record heart sounds and produce electrocardiogram (ECG) tracings, features that may provide additional information during an examination. The researchers said those recordings can be useful when interpreted by a trained veterinary professional.

“That said, the ECG tracing it captures is genuinely good quality,” Johnson said. “This tool works best as an adjunct—a quick ECG or a flag worth a second look—not a stand-alone diagnosis.”

The study was prompted in part by reports from veterinarians who were questioning their own clinical assessments after receiving conflicting results from AI-enabled stethoscopes, according to Pierce. The researchers said the experience illustrates a broader issue with applying diagnostic technologies outside the populations for which they were developed and evaluated.

Pierce said veterinarians should continue to use their clinical training when interpreting results from AI-assisted devices.

“We want to encourage veterinarians to rely on the expertise they’ve gained through their training and to understand what this tool can and cannot provide in a veterinary setting,” she said.

The authors noted several limitations to the study, including its relatively small sample size and the use of clinician assessment as the reference standard for auscultation. The study also evaluated a single AI-enabled stethoscope, so its findings should not be generalized to all digital or AI-assisted stethoscopes.

The study's results indicate that the particular device evaluated may provide useful cardiac recordings while its automated interpretations require caution in veterinary patients. Further validation in dogs and cats would be needed to determine how reliably such technology can be used across different species and clinical settings.

Reference

1. Johnson JH, Stern JA, DeFrancesco TC, Pierce KV. An artificial intelligence–enabled digital stethoscope demonstrates moderate murmur detection in dogs but not cats and unreliable arrhythmia classification in both species. J Am Vet Med Assoc. Published online August 5, 2026. doi:10.2460/javma.26.05.0353.
2. North Carolina State University. AI-enabled stethoscope can miss the beat in pets. EurekAlert! Published August 12, 2026. Accessed August 18, 2026.