Automated Detection of Central Retinal Artery Occlusion Using OCT Imaging via Explainable Deep Learning

• Published in: Ophthalmology science (Online) Vol. 5; no. 2; p. 100630
• Main Authors: Beuse, Ansgar, Wenzel, Daniel Alexander, Spitzer, Martin Stephan, Bartz-Schmidt, Karl Ulrich, Schultheiss, Maximilian, Poli, Sven, Grohmann, Carsten
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier Inc 01.03.2025; Elsevier
• Subjects: AI OCT; CRAO; Deep learning retina; OCT imaging; Ophthalmology; Ophthalmology deep learning; Original; SD; CRAO; OCT imaging; MLA; Deep learning retina; CI; AI OCT; px; Ophthalmology deep learning; AUC; area under the curve; confidence interval; central retinal artery occlusion; machine learning algorithm; pixel; standard deviation
• ISSN: 2666-9145; 2666-9145
• DOI: 10.1016/j.xops.2024.100630

To demonstrate the capability of a deep learning model to detect central retinal artery occlusion (CRAO), a retinal pathology with significant clinical urgency, using OCT data. Retrospective, external validation study analyzing OCT and clinical baseline data of 2 institutions via deep learning classification analysis. Patients presenting to the University Medical Center Tübingen and the University Medical Center Hamburg-Eppendorf in Germany. OCT data of patients suffering from CRAO, differential diagnosis with (sub) acute visual loss (central retinal vein occlusion, diabetic macular edema, nonarteritic ischemic optic neuropathy), and from controls were expertly graded and distinguished into 3 groups. Our methodological approach involved a nested multiclass five fold cross-validation classification scheme. Area under the curve (AUC). The optimal performance of our algorithm was observed using 30 epochs, complemented by an early stopping mechanism to prevent overfitting. Our model followed a multiclass approach, distinguishing among the 3 different classes: control, CRAO, and differential diagnoses. The evaluation was conducted by the “one vs. all” area under the receiver operating characteristics curve (AUC) method. The results demonstrated AUC of 0.96 (95% confidence interval [CI], ± 0.01); 0.99 (95% CI, ± 0.00); and 0.90 (95% CI, ± 0.03) for each class, respectively. Our machine learning algorithm (MLA) exhibited a high AUC, as well as sensitivity and specificity in detecting CRAO and the differential classes, respectively. These findings underscore the potential for deploying MLAs in the identification of less common etiologies within an acute emergency clinical setting. Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.