For the first time, the US Food and Drug Administration. UU He has approved an artificial intelligence diagnostic device that does not need a specialized doctor to interpret the results. The software program, called IDx-DR, can detect a form of eye disease by observing photos of the retina.
It works like this: a nurse or a doctor uploads photos of the patient's retina taken with a special retina camera. The IDx-DR software algorithm first indicates whether the loaded image is of high enough quality to obtain a result. Then, analyze the images to determine whether or not the patient has diabetic retinopathy, a form of eye disease where too much sugar in the blood damages the blood vessels in the back of the eye. Diabetic retinopathy is the most common vision complication for people with diabetes, but it is still quite rare: there are approximately 200,000 cases per year.
In a clinical trial that used more than 900 images, IDx-DR correctly detected retinopathy approximately 87 percent of the time and could correctly identify those who did not have the disease approximately 90 percent of the time.
The software is unique because it is autonomous and there is "a specialist who looks over the shoulder of the algorithm [this]," IDx-DR founder Michael Abràmoff told Science News. "Take the clinical decision for yourself." This means that the technology can be used by a nurse or a doctor who is not an eye doctor, which makes the diagnosis more accessible. For example, patients would not need to wait for an eye doctor to be available to obtain a diagnosis.
IDx-DR is part of a growing trend of algorithms that learn to detect and diagnose diseases. Earlier this year, scientists trained a different algorithm to learn how to recognize conditions, including vision loss related to age and diabetic retinopathy. Google is also training its DeepMind AI to detect eye diseases. Now that the FDA has approved IDx-DR, it could open the way to a new series of autonomous diagnostic tests and the compensations they provide. These diagnoses may be more convenient for patients (and possibly even more accurate than doctors). But, of course, not having a specialist "looking over his shoulder," as Abràmoff puts it, raises the question of who will be responsible when the diagnosis is incorrect.