Artificial Intelligence to help avert blindness

BY W. GIFFORD- JONES MD & DIANA GIFFORD-JONES

How can doctors diagnose and treat 425 million worldwide diabetes patients? That number keeps going up and up, projected to reach 700 million by 2045. There are millions more with undiagnosed prediabetes. Add millions with undiagnosed hypertension. All these people are destined to lives defined by cardiovascular problems and complications that include debilitating conditions like blindness.

Diabetes is swamping healthcare systems worldwide. Let us be clear: whatever we have been doing to fight the problem, it is not working.

Artificial Intelligence (AI) is offering new possibilities. Using new technologies, data science, vast quantities of medical images, and computer algorithms, it is possible to fight diseases differently. The medical model of a patient and a doctor is outdated. We need to put AI on our healthcare team and use analytical methods to predict problems before they occur and to help doctors and patients make better decisions.

Computer-assisted retinal analysis (CARA) is one such technology.  Developed by DIAGNOS, a Montreal-based company, CARA uses retina scans to detect early warning signs of big health problems. And CARA can do it on a scale that will make a big difference in fighting the diabetes epidemics.

The retina, the back part of the eye, is the only area of the body where doctors can easily see the condition of arteries and veins without invasive procedures.  Early detection of atherosclerosis (hardening of arteries) in the retinas of diabetes patients signals a warning that the same problem is occurring in coronary arteries. This is why the retina is called, “the window to the heart”.

Prevention is always better than cure. But this is easier said than done in many parts of the world where highly trained retinal specialists are in short supply. We are more fortunate in North America, but retinal checkups are mainly the purview of ophthalmologists focused on your eyes, not your cardiovascular system.

Type 2 diabetes has become a worldwide epidemic and an expensive problem for every health care system. Type 2 diabetes is not just a singular disease. Rather, by triggering atherosclerosis, it decreases blood supply to many parts of the body with catastrophic results. For example, long standing diabetes increases the risk of blindness, heart attack, and kidney failure, which may require renal dialysis or a kidney transplant.

Doctors can only treat so many patients. So this problem is an example of where we can leverage technology to screen millions of people. CARA can scan an eye in two seconds. Furthermore, it can scan hundreds of patients for hours without getting tired or making errors. We need to use AI to detect retina changes and prevent diabetes – averting countless cases of blindness and other problems, improving lives, and saving dollars.

Andre Larente, president of DIAGNOS, recently remarked, “CARA can now look at a patient’s retina, discover the presence of hypertension and predict a chance of stroke in 12 to 24 months.” Given that CARA can do this across very large populations of patients, at low cost, it’s easy to see the appeal of this technology from a health care and economic perspective, not to mention the incentive to individual patients to reduce their risk profile.

There’s no doubt that the capacities of artificial intelligence are changing the way we can fight illness, and companies like DIAGNOS are important partners in medical practice. The key is in scaling up. CARA has accumulated a vast database of retinal photos of patients worldwide.  This data can be used for predictive modeling.  So the next step will be in getting this data into the hands of those who can take steps to stop the progression of illness, change conditions leading to disease, and prevent these avoidable health problems in the first place.