Cardiovascular disease has been the number one killer of adults worldwide for many years, accounting for the loss of more than a million lives per year. This puts it significantly ahead of cancer, suicide, and road accidents, some of the most common causes of death in the world. It is no wonder that the struggle against cardiovascular disease has sought help from nearly every field of science.
One recent development by a company named Eko brings the growing power of artificial intelligence to bear on the problem of heart disease. Diagnosing heart disease can be difficult without an actual cardiovascular event, as many of the symptoms are either hard to notice or rely on the subjective feelings of the patient to understand. With no concrete way to tell if the heart is malfunctioning, doctors must resort to more imprecise methods to diagnose a patient.
The tool that cuts through at least some of these problems is the stethoscope, which has become so prevalent in the medical world that it has become a symbol of the profession in many places. Stethoscopes let a doctor listen to a patient’s heart and lungs to check if they are functioning smoothly and in sync.
In a digital stethoscope, the doctor’s ears are replaced with a digital sensor and speakers. These upgrades prevent the unintentional distortion of the sound through the doctor’s own contact with the stethoscope and allow everyone in an examination to be a bit more relaxed as well; the digital cable connecting the sensor to the earpieces may be any length, letting the doctor use it without getting uncomfortably close to the patient.
Eko takes advantage of the fact that the sounds from the patient’s body are already emanating in digital form and uses a machine learning system to analyze the data for patterns that might indicate cause for concern. The two main symptoms that this machine learning system can detect are atrial fibrillation and heart murmurs, both irregularities in the blood pumping and early signs of potential heart failure. Because a healthy heartbeat has a distinctive, rhythmic tone, often referred to as a ‘lub-dub’ pattern, it is comparatively easy to program a computer to check for heartbeats that do not conform to the pattern and flag them for more thorough investigation. Several other procedures, such as mammary cancer screening, have already taken this approach with considerable success.
Because this tool will be used to make diagnoses on such a vital system in the body, there are rigorous standards for using it to replace the existing, trusted tools that so many doctors already use. The FDA has only recently approved Eko’s invention for widespread use in diagnosing heart disease, clearing it for widespread clinical use. The tests included not only identifying the sounds of a diseased heart, but also pinpointing the chamber where the irregularities occurred.
Connor Landgraf, founder of Eko, said that this machine learning system is only one of the many medical technologies he hopes to make available to the doctors of the world in the near future. He and his company intend to make AI that will be a reliable companion to any doctor for any procedure.