The AI ​​could soon tell you how often you should visit the eye doctor

Three of the most common chronic eye conditions require regular medical checkups and eye injections by ophthalmologists to keep the blindness that appears at bay. A study by the University of Bern and the Inselspital, in collaboration with an AI in the implementation of visual care, now shows that the ideal individual frequency of patients for these visits can be accurately predicted by the machine learning, which has a triple benefit.

Age-related macular degeneration (AMD) is the most common cause of vision loss in people over 50 years of age. Up to 12% of people over the age of 80 have chronic illness. An estimated 16.4 million adults are affected by retinal vein occlusion (RVO) worldwide, a condition caused by retinal vein thrombosis. It is the second most common cause of blindness due to vascular disease of the retina after diabetic retinopathy (DR). DR, in turn, is the leading cause of blindness in developed countries and affects up to 80 percent of people with more than 20 years of diabetes. It can cause inflammation of the macula (diabetic macular edema or DME), which can cause partial or complete loss of vision.

All three conditions are treated by injections of the so-called anti-vascular endothelial growth factor (anti-VEGF) into the eye at intervals to slow the progression of the disease and prevent blindness. Because with sight, a central human sense is in jeopardy, patients want to know that they are being treated often enough to prevent rapid worsening. And doctors want to make sure they see each patient often enough so they don’t miss important events.

The challenge

With the aging population, cases of AMD, RVO or DME are increasing worldwide, making it difficult for specialist eye clinics to keep pace with the growing demand for regular treatments. “As physicians, we want to give each patient the care and frequency of treatment they need,” says Sebastian Wolf, head of Inselspital’s ophthalmology department, which currently sees 6,000 patients with AMD, RVO and DR. “But it is also an organizational challenge to meet the needs of all patients and to be able to study all relevant ocular imaging data to assess individual disease progression and make treatment decisions in the short term.”

To monitor the progression of chronic eye conditions, icalptical Coherence Tomography (OCT), an imaging tool which usually generates 3D images of the eye at an extremely high resolution. In collaboration with the ARTORG Biomedical Engineering Research Center, Inselspital has developed automated OCT analysis tools based on artificial intelligence, which can help eye doctors evaluate an entire established patient with OCT in seconds. Together with RetinAI, a startup specializing in AI-based eye care technologies, they have now conducted a retrospective study of patients to assess the extent to which AI can predict the demand for anti-VEGF treatment from the outset.

The configuration

The study examined OCT data from 340 patients with AMD and 285 patients with RVO or DME, treated with anti-VEGF at Inselspital between 2014 and 2018. Based on morphological features automatically extracted from OCT volumes at the baseline and after two consecutive visits, as well as patient demographic information, two machine learning models were trained to predict the likelihood of long-term treatment frequency demand of a new patient (one for AMD and one for RVO and DME).

Based on the first three visits, it was possible to predict whether a patient had a low or high treatment demand for both the AMD and RVO and DME groups with similar accuracy. More importantly, the study revealed that it is possible to predict reasonably well at the initial visit and even before the first injection if a patient will need injections less frequently.

Three advantages

“We have shown that machine learning classifiers can predict the demand for treatment when a patient is first diagnosed with chronic eye disease,” says Mathias Gallardo, a postdoctoral researcher at ARTORG AI in Medical Imaging (AIMI) and a member of the new Center for Artificial Intelligence in Medicine (CAIM). “Therefore, artificial intelligence can help establish patient-specific treatment plans for the most common chronic eye conditions in the near future.”

Planning the ideal treatment frequency for each patient has multiple advantages. First, patients can make sure that their disease is treated in the best possible way without subjecting them to excessive visits and unpleasant injections into the eye. Second, individualized planning can help clinics cope with an increasing number of patients, enabling the use of the highest possible capacity of specialized medical skills and infrastructure. Third, objectified demand planning helps prevent oversupply and can lead to better cost efficiency and reduced health care costs.

High-performance confluence of clinical research, data science, and industry

This study illustrates once again the visually proven collaboration between clinicians and data scientists from Inselspital and the ARTORG Center, which produces technological solutions suitable for everyday use because they were designed directly in response. to clinical needs. Another important aspect to provide a roadmap for the clinical implementation of this technology was the implementation of RetinAI.

“We are very pleased to apply the EU funding we have received to create patient-centered solutions in ophthalmology, ensuring that technology can be transformed into products that can truly benefit patients and improve treatment on a large scale.” , says Carlos Ciller, CEO of RetinAI. With headquarters in sitem-insel, the launch is also located exactly at the interface between clinic and science. This unique Bernese environment for clinically driven artificial intelligence technologies will be further capitalized by the new Center for Artificial Intelligence in Medicine (CAIM), which combines the best of the three worlds of health, science and industry. for the benefit of patients.

Provided by Inselspital, a university hospital in Bern