IDx-DR (diabetic retinopathy detection device)

2018

Market Entry and Start of Application

In April 2018, IDx released a device that allows you to diagnose a serious eye disease from a photo of the retina. A solution called IDx-DR is able to detect diabetic retinopathy from a photo of the retina taken by a special Topcon NW400 camera - the AI algorithm determines the image quality, and then, if it considers it acceptable, analyzes the photo.

The software tells the doctor the result: either the patient has moderate to more severe retinopathy and must see an ophthalmologist to select the therapy, or does not have retinopathy and must undergo a second examination after 12 months.

A device has been released that allows you to diagnose serious eye disease from a photo of the retina

The standard procedure still assumes human participation: a trained operator must take pictures and upload them to the cloud. But it could be any clinic employee, technician or trainee, not an ophthalmologist.

Despite the fact that diabetic retinopathy is one of the most disabling complications of diabetes mellitus, which leads to vision loss, it is quite rare - there are only 20 thousand cases a year. As a result, it is more difficult for a doctor to identify this pathology, however, an algorithm that has been deeply trained is able to correctly diagnose diabetic retinopathy in 87% of cases and correctly exclude a photograph of a healthy retina in more than 90%.

IDx-DR is completely autonomous and does not require control by an ophthalmologist, that is, even nurses and doctors of other specialties can use it, explained IDx-DR creator Michael Abràmoff in Science News. Of course, the attending physician is responsible for the diagnosis, but the use of IDx-DR will be more convenient for patients who do not have to wait for an ophthalmologist in a long queue.

IDx-DR is not the first algorithm to use deep learning technology to diagnose diseases. In early 2018, a program was developed to detect senile vision loss due to cataracts, and Google trains AI to diagnose other eye diseases.^[1]

The University of Iowa became the first organization health care USA to introduce IDx-DR technology into clinical practice. According to the press service of the university, this is the world's first autonomous diagnostic system based on artificial intelligence the detection of diabetic retinopathy - a severe complication diabetes mellitus that, without proper observation and treatment, can lead to complete loss of vision. IDx has developed its own algorithm for diagnosing retinopathy in adults over 22 with diabetes from fundus images.

The University of Iowa noted that the system does not require the interpretation of the image or the results obtained by the doctor. This allows health care providers who do not normally engage in eye tests and fundus assessments to conduct a screening study among patients with diabetes during routine medical visits.

The world's first autonomous AI-based diagnostic system has been launched. She identifies a severe form of diabetes

According to experts, early detection of diabetic retinopathy is an important component of the complex treatment of diabetes mellitus. And considering that up to 50% of patients with diabetes do not undergo a mandatory annual examination by an ophthalmologist, that is, they are at risk of developing undiagnosed diabetic retinopathy, the use of the new IDx-DR system in the future will significantly reduce the prevalence of far-reaching untreated cases of this formidable complication.

University of Iowa officials noted that workers were able to use the new diagnostic system after just a few hours of training. The University Hospital serves approximately 7,200 patients with diabetes mellitus and other endocrinological diseases daily. IDx, meanwhile, is working to implement the new system at several other major U.S. clinics.^[2]

Notes