A revolutionary invention: they design a stethoscope that detects heart damage using AI.

A group of Mexican scientists designed and built a digital stethoscope that, using artificial intelligence (algorithms), identifies irregular heart sounds and displays them as signs of heart valve damage, which could be a valuable tool for pre-diagnosis.

In a statement, the National Polytechnic Institute (IPN) indicated that the development, carried out by doctors Diana Bueno Hernández and José Alberto Zamora Justo, attached to the Interdisciplinary Professional Unit of Biotechnology (UPIBI), in collaboration with Víctor Manuel Arena Cantoran, currently achieves an accuracy of 96%.

According to the information, the system was trained using neural networks to classify heart sounds and identify possible damage . Specialists explained that conventional stethoscopes work through vibrations and mechanical movements transmitted from the membrane of the bell to the ears.

In contrast, this technology, in addition to the bell, incorporates a microphone, a microprocessor that captures, processes, and classifies heart sounds, and displays them on a 240 x 320 pixel Thin Film Transistor (TFT) display.

The device is portable and fully embedded (it does not depend on a PC or mobile device for operation). It has a 5-volt rechargeable battery and a USB charging port. To ensure proper assembly of the internal components, the stethoscope housing was 3D printed with PLA (polylactic acid) polymer.

Currently, they noted, this medical tool identifies heart sounds such as S3 and S4, both abnormal components of the cardiac cycle and indicators of heart failure, commonly known as murmurs.

"S3 is generated by the rapid influx of blood into the ventricle and can be detected in the region of the mitral (left ventricle) or tricuspid (right ventricle) valve, while S4 occurs during atrial contraction ," they explained.

Scientists do not rule out improving the prototype so that it can also contribute to the pre-diagnosis of other heart pathologies , although they emphasized that the goal is not to replace a specialist's diagnosis, but rather to provide it with a tool that makes detection more precise.

They also noted that the size of the bell can be modified to adapt the device to examine children, and that the specific condition classification can be incorporated on the screen, in addition to the signals. "This is highly relevant, since heart disease is the leading cause of death in Mexico," they explained.

Globally, they recalled, there are machine learning and deep learning- based tools for heartbeat classification; however, there are no fully autonomous (embedded) technologies . Therefore, they will soon begin the patent registration process.