Inferred retinal sensitivity in recessive Stargardt disease using machine learning

Applying artificial intelligence to rare diseases: a literature review highlighting lessons from Fabry disease

BACKGROUND

Use of artificial intelligence (AI) in rare diseases has grown rapidly in recent years. In this review we have outlined the most common machine-learning and deep-learning methods currently being used to classify and analyse large amounts of data, such as standardized images or specific text in electronic health records. To illustrate how these methods have been adapted or developed for use with rare diseases, we have focused on Fabry disease, an X-linked genetic disorder caused by lysosomal α-galactosidase. A deficiency that can result in multiple organ damage.

METHODS

We searched PubMed for articles focusing on AI, rare diseases, and Fabry disease published anytime up to 08 January 2025. Further searches, limited to articles published between 01 January 2021 and 31 December 2023, were also performed using double combinations of keywords related to AI and each organ affected in Fabry disease, and AI and rare diseases.

RESULTS

In total, 20 articles on AI and Fabry disease were included. In the rare disease field, AI methods may be applied prospectively to large populations to identify specific patients, or retrospectively to large data sets to diagnose a previously overlooked rare disease. Different AI methods may facilitate Fabry disease diagnosis, help monitor progression in affected organs, and potentially contribute to personalized therapy development. The implementation of AI methods in general healthcare and medical imaging centres may help raise awareness of rare diseases and prompt general practitioners to consider these conditions earlier in the diagnostic pathway, while chatbots and telemedicine may accelerate patient referral to rare disease experts. The use of AI technologies in healthcare may generate specific ethical risks, prompting new AI regulatory frameworks aimed at addressing these issues to be established in Europe and the United States.

CONCLUSION

AI-based methods will lead to substantial improvements in the diagnosis and management of rare diseases. The need for a human guarantee of AI is a key issue in pursuing innovation while ensuring that human involvement remains at the centre of patient care during this technological revolution.

Inner retinal thickness in Stargardt disease

PURPOSE

To analyze the alterations at the level of the inner retina in patients affected by Stargardt disease (STGD1).

METHODS

Cross-sectional investigation involving STGD1 patients with genetically confirmed diagnosis, who underwent optical coherence tomography (OCT), optical coherence tomography angiography (OCTA), and microperimetry.

RESULTS

Overall, 31 patients (62 eyes) with genetically confirmed STGD1 were included in the study. Mean inner retinal thickness, vessel density of plexa, and retinal sensitivity resulted significantly reduced in STGD patients, compared with healthy controls (p < 0.05), both in the outer and in the inner ETDRS rings. Overall, 43% of eyes revealed an inner retinal thinning, whereas 21% and 35% showed a thicker or within normal range inner retina.

CONCLUSIONS

Inner retina is irregularly altered in STGD1, showing variable quantitative alterations as detected on OCT. Inner retinal status might represent a useful biomarker to better characterize STGD1 and to ascertain the effects of new treatment approaches.

Photoreceptor degeneration in ABCA4-associated retinopathy and its genetic correlates

BACKGROUND

Outcome measures sensitive to disease progression are needed for ATP-binding cassette, sub-family A, member 4–associated (ABCA4-associated) retinopathy. We aimed to quantify ellipsoid zone (EZ) loss and photoreceptor degeneration beyond EZ-loss in ABCA4-associated retinopathy and investigate associations between photoreceptor degeneration, genotype, and age.

METHODS

We analyzed 132 eyes from 66 patients (of 67 enrolled) with molecularly confirmed ABCA4-associated retinopathy from a prospective natural history study with a median [IQR] follow-up of 4.2 years [3.1, 5.1]. Longitudinal spectral-domain optical coherence tomography volume scans (37 B-scans, 30° × 15°) were segmented using a deep learning (DL) approach. For genotype-phenotype analysis, a model of ABCA4 variants was applied with the age of criterion EZ-loss (6.25 mm²) as the dependent variable.

RESULTS

Patients exhibited an average (square-root-transformed) EZ-loss progression rate of [95% CI] 0.09 mm/y [0.06, 0.11]. Outer nuclear layer (ONL) thinning extended beyond the area of EZ-loss. The average distance from the EZ-loss boundary to normalization of ONL thickness (to ±2 z score units) was 3.20° [2.53, 3.87]. Inner segment (IS) and outer segment (OS) thinning was less pronounced, with an average distance from the EZ-loss boundary to layer thickness normalization of 1.20° [0.91, 1.48] for the IS and 0.60° [0.49, 0.72] for the OS. An additive model of allele severity explained 52.7% of variability in the age of criterion EZ-loss.

CONCLUSION

Patients with ABCA4-associated retinopathy exhibited significant alterations of photoreceptors outside of EZ-loss. DL-based analysis of photoreceptor laminae may help monitor disease progression and estimate the severity of ABCA4 variants.

TRIAL REGISTRATION

ClinicalTrials.gov identifier: NCT01736293.

FUNDING

National Eye Institute Intramural Research Program and German Research Foundation grant PF950/1-1.

Estimation of current and post-treatment retinal function in chronic central serous chorioretinopathy using artificial intelligence

Refined understanding of the association of retinal microstructure with current and future (post-treatment) function in chronic central serous chorioretinopathy (cCSC) may help to identify patients that would benefit most from treatment. In this post-hoc analysis of data from the prospective, randomized PLACE trial (NCT01797861), we aimed to determine the accuracy of AI-based inference of retinal function from retinal morphology in cCSC. Longitudinal spectral-domain optical coherence tomography (SD-OCT) data from 57 eyes of 57 patients from baseline, week 6-8 and month 7-8 post-treatment were segmented using deep-learning software. Fundus-controlled perimetry data were aligned to the SD-OCT data to extract layer thickness and reflectivity values for each test point. Point-wise retinal sensitivity could be inferred with a (leave-one-out) cross-validated mean absolute error (MAE) [95% CI] of 2.93 dB [2.40-3.46] (scenario 1) using random forest regression. With addition of patient-specific baseline data (scenario 2), retinal sensitivity at remaining follow-up visits was estimated even more accurately with a MAE of 1.07 dB [1.06-1.08]. In scenario 3, month 7-8 post-treatment retinal sensitivity was predicted from baseline SD-OCT data with a MAE of 3.38 dB [2.82-3.94]. Our study shows that localized retinal sensitivity can be inferred from retinal structure in cCSC using machine-learning. Especially, prediction of month 7-8 post-treatment sensitivity with consideration of the treatment as explanatory variable constitutes an important step toward personalized treatment decisions in cCSC.