AI could save your dog from a heart attack

University of Cambridge researchers have developed a machine learning algorithm that detects heart murmurs in dogs – one of the key heart disease signs – with the precision of expert cardiologists.

Tests showed that the algorithm detected heart murmurs with a sensitivity of 90%, a similar accuracy to human specialists.

The algorithm, originally designed for humans, was adapted to automatically detect and grade heart murmurs in dogs based on audio recordings from digital stethoscopes.

“Heart disease in humans is a huge health issue, but in dogs it’s an even bigger problem,” said first author Dr Andrew McDonald from Cambridge’s Department of Engineering.

Heart murmurs are the main indicator of mitral valve disease, the most common heart condition in adult dogs. About one in 30 dogs seen by a veterinarian has a heart murmur, which is more frequent in small breeds and older dogs.

“Most smaller dog breeds will have heart disease when they get older, but obviously dogs can’t communicate in the same way that humans can, so it’s up to primary care vets to detect heart disease early enough so it can be treated,” McDonald said.

The only treatment for humans with valve disease is surgery, but effective medication is available for dogs. However, only timely medication can prolong their lives and early detection is crucial.

The researchers believe that their technology could offer an affordable and effective screening tool for primary care veterinarians.

Grading a heart murmur and deciding on treatment now require significant experience, referral to a veterinary cardiologist, and costly heart scans, according to study co-author Professor Jose Novo Matos from Cambridge’s Department of Veterinary Medicine.

“We want to empower general practitioners to detect heart disease and assess its severity to help owners make the best decisions for their dogs,” Novo Matos said.

Building ‘largest’ database of dog heart sounds

The team collected data from almost 800 dogs undergoing routine heart examination at four veterinary specialist centers in the UK, building what it said is the largest dataset of dog heart sounds “ever created.”

Heart sounds were recorded using an electronic stethoscope and every dog received a full physical examination and heart scan to grade heart murmurs and identify heart disease.

While mitral valve disease mainly affects smaller dogs, the researchers gathered data from dogs of all shapes, sizes, and ages, according to Novo Matos.

“The more data we have to train it, the more useful our algorithm will be, both for vets and for dog owners,” he said.

The team first developed an algorithm based on a database of heart sounds from about 1000 human patients, which they then adapted to the dogs.

“As far as we’re aware, there are no existing databases of heart sounds in dogs, which is why we started out with a database of heart sounds in humans,” lead researcher, Professor Anurag Agarwal, said.

“Mammalian hearts are fairly similar, and when things go wrong, they tend to go wrong in similar ways,” he added.

The algorithm was fine-tuned to detect and grade heart murmurs based on the audio recordings of dog heart sounds, as well as differentiate between murmurs associated with mild disease and those signaling advanced conditions.

“So many people talk about AI as a threat to jobs, but for me, I see it as a tool that will make me a better cardiologist,” Novo Matos said, adding that “tools like these could help vets and owners, so we can quickly identify those dogs who are most in need of treatment.”

The results of the study were published in the Journal of Veterinary Internal Medicine.