The algorithm was tested 58 VHD patients and an additional 52 healthy patients, achieving a sensitivity of 93%, specificity of 98%, accuracy of 97% and positive predictive value of 93%. In fact, the group added, their AI model can identify up to five different valvular diseases from a single 10-second burst of sound.
“Our ability to detect multiple diseases simultaneously was a key innovation in this research,” Shokouhmand said. “We aren’t just showing that there’s a valvular problem — we’re able to identify the constellation of problems a patient is suffering from.”
The researchers are far from done exploring the possibilities of this advanced algorithm. They are currently gathering more data with a long-term goal of being able to classify diseases by their severity.
“Instead of showing that you have a particular valvular disorder, we could give a grade out of 10 describing how far the disease has progressed,” Ebadi explained.
Click here to read the full study in IEEE Transactions on Biomedical Engineering.
