Newly Developed Artificial Intelligence Method Evaluates Vision Loss 

National Eye Institute researchers created and validated an artificial intelligence-based method to evaluate patients with Stargardt, an eye disease that can lead to childhood vision loss.

The AI method quantifies the disease-related loss of light-sensing retina cells, generating data and information for monitoring patients, understanding genetic causes, and developing potential treatments.

“These results provide a framework to evaluate Stargardt disease progression, which will help control for the significant variability from patient to patient and facilitate therapeutic trials,” Michael F. Chiang, MD, director of the NEI, which is part of the National Institutes of Health, said in a press release.

According to researchers, about 1 in 9,000 people develop the most common form of Stargardt, or ABCA4-associated retinopathy. Individuals develop Stargardt when they inherit two mutated copies of the ABCA4 gene, one from each parent. Those who inherit just one mutated copy of ABCA4 are genetic carriers but do not develop the disease.

Among patients who have all ABCA4 gene variants, researchers found that there can be a wide spectrum regarding the age of onset and disease progression.

“Different variants of the ABCA4 gene are likely driving the different disease characteristics, or phenotypes. However, conventional approaches to analyzing structural changes in the retina have not allowed us to correlate genetic variants with phenotype,” said the study’s co-leader, Brian P. Brooks, MD, PhD, chief of the NEI Ophthalmic Genetics & Visual Function Branch, in a news release.

The study followed 66 Stargardt patients for five years using a retinal imaging technology called spectral-domain optical coherence tomography (SD-OCT). The 3D SD-OCT retinal images were segmented and examined using deep learning.

With the deep-learning method, the team quantified and compared the loss of photoreceptors and various layers of the retina according to the patient’s phenotype and ABCA4 variant.

The researchers focused on the health of photoreceptors in an area known as the ellipsoid zone and the outer nuclear layer in the immediate region surrounding the area of ellipsoid zone loss.

The team found that the loss of the ellipsoid zone and thinning of the outer nuclear layer followed a predictable temporal and spatial pattern. With predictive analytics, the research could generate a way to classify the severity of 31 different ABCA4 variants.

Additionally, the researchers found that photoreceptor degeneration was not limited to the area of the ellipsoid zone loss. Rather, progressive photoreceptor layer thinning was apparent in areas distant from the ellipsoid zone loss boundary, suggesting that it would be an important area to monitor to determine the effectiveness of a new therapy.

“We now have sensitive structural outcome measures for Stargardt disease, applicable to a wide range of patients which is essential for forging ahead with therapeutic trials,” co-leader of the study and Head of the Human Visual Function Core of the NEI’s Ophthalmic Genetics and Visual Function Branch Brett G. Jeffrey, PhD, said.