Recent Posts
Connect with:
Tuesday / August 9.
HomeminewsAI to Evaluate Stargardt Disease

AI to Evaluate Stargardt Disease

An artificial-intelligence-based method to evaluate patients with Stargardt disease has been developed and validated by National Eye Institute (NEI) researchers. The method quantifies disease-related loss of light-sensing retina cells, yielding information for monitoring patients, understanding genetic causes of the disease, and developing therapies to treat it.

Published in JCI Insight, researchers followed 66 Stargardt patients (132 eyes) for five years using spectral-domain optical coherence tomography (SD-OCT). The cross-sectional, 3D SD-OCT retinal images were segmented and analysed using deep learning, enabling researchers to quantify and compare loss of photoreceptors and various layers of the retina according to the patient’s phenotype and ABCA4 variant.

The researchers zeroed in on the health of photoreceptors in the ellipsoid zone – a feature of the photoreceptor inner/outer segment border that is diminished or lost due to disease. They also examined the outer nuclear layer in the immediate region surrounding the area of ellipsoid zone loss.

They found loss of the ellipsoid zone and thinning of the outer nuclear layer beyond those areas followed a predictable temporal and spatial pattern. On the basis of that, they could generate a way to classify the severity of 31 different ABCA4 variants.

Importantly, they found progressive photoreceptor layer thinning – subtle to the eye, but quantitatively measurable – was evident in areas distant to the ellipsoid zone loss boundary. This represented the leading front of the disease, suggesting it would be an area to monitor closely to determine if a new therapy was having an effect.


Pfau M, Cukras CA, Huryn LA, Zein WM, Ullah E, Boyle MP, Turriff A, Chen MA, Hinduja AS, Siebel HEA, Hufnagel RB, Jeffrey BG, Brooks BP. “Photoreceptor degeneration in ABCA4-associated retinopathy and its genetic correlates,” published in press review February 25, 2022 in JCI Insight. 2022;7(2):e155373.https://doi.org/10.1172/jci.insight.155373.


By agreeing & continuing, you are declaring that you are a registered Healthcare professional with an appropriate registration. In order to view some areas of this website you will need to register and login.
If you are not a Healthcare professional do not continue.