To compare the diagnostic accuracy and explainability of a Vision Transformer deep learning technique, Data-efficient image Transformer (DeiT), and ResNet-50, trained on fundus photographs from the Ocular Hypertension Treatment Study (OHTS) to detect primary open-angle glaucoma (POAG) and identify the salient areas of the photographs most important for each model’s decision-making process. Evaluation of a diagnostic technology. Overall 66 715 photographs from 1636 OHTS participants and an additional 5 external datasets of 16 137 photographs of healthy and glaucoma eyes. Data-efficient image Transformer models were trained to detect 5 ground-truth OHTS POAG classifications: OHTS end point committee POAG determinations because of disc changes (model 1), visual field (VF) changes (model 2), or either disc or VF changes (model 3) and Reading Center determinations based on disc (model 4) and VFs (model 5). The best-performing DeiT models were compared with ResNet-50 models on OHTS and 5 external datasets. Diagnostic performance was compared using areas under the receiver operating characteristic curve (AUROC) and sensitivities at fixed specificities. The explainability of the DeiT and ResNet-50 models was compared by evaluating the attention maps derived directly from DeiT to 3 gradient-weighted class activation map strategies. Compared with our best-performing ResNet-50 models, the DeiT models demonstrated similar performance on the OHTS test sets for all 5 ground-truth POAG labels; AUROC ranged from 0.82 (model 5) to 0.91 (model 1). Data-efficient image Transformer AUROC was consistently higher than ResNet-50 on the 5 external datasets. For example, AUROC for the main OHTS end point (model 3) was between 0.08 and 0.20 higher in the DeiT than ResNet-50 models. The saliency maps from the DeiT highlight localized areas of the neuroretinal rim, suggesting important rim features for classification. The same maps in the ResNet-50 models show a more diffuse, generalized distribution around the optic disc. Vision Transformers have the potential to improve generalizability and explainability in deep learning models, detecting eye disease and possibly other medical conditions that rely on imaging for clinical diagnosis and management.

The aim is to assess GPT-4V's (OpenAI) diagnostic accuracy and its capability to identify glaucoma-related features compared to expert evaluations. Evaluation of multimodal large language models for reviewing fundus images in glaucoma. A total of 300 fundus images from 3 public datasets (ACRIMA, ORIGA, and RIM-One v3) that included 139 glaucomatous and 161 nonglaucomatous cases were analyzed. Preprocessing ensured each image was centered on the optic disc. GPT-4's vision-preview model (GPT-4V) assessed each image for various glaucoma-related criteria: image quality, image gradability, cup-to-disc ratio, peripapillary atrophy, disc hemorrhages, rim thinning (by quadrant and clock hour), glaucoma status, and estimated probability of glaucoma. Each image was analyzed twice by GPT-4V to evaluate consistency in its predictions. Two expert graders independently evaluated the same images using identical criteria. Comparisons between GPT-4V's assessments, expert evaluations, and dataset labels were made to determine accuracy, sensitivity, specificity, and Cohen kappa. The main parameters measured were the accuracy, sensitivity, specificity, and Cohen kappa of GPT-4V in detecting glaucoma compared with expert evaluations. GPT-4V successfully provided glaucoma assessments for all 300 fundus images across the datasets, although approximately 35% required multiple prompt submissions. GPT-4V's overall accuracy in glaucoma detection was slightly lower (0.68, 0.70, and 0.81, respectively) than that of expert graders (0.78, 0.80, and 0.88, for expert grader 1 and 0.72, 0.78, and 0.87, for expert grader 2, respectively), across the ACRIMA, ORIGA, and RIM-ONE datasets. In Glaucoma detection, GPT-4V showed variable agreement by dataset and expert graders, with Cohen kappa values ranging from 0.08 to 0.72. In terms of feature detection, GPT-4V demonstrated high consistency (repeatability) in image gradability, with an agreement accuracy of ≥89% and substantial agreement in rim thinning and cup-to-disc ratio assessments, although kappas were generally lower than expert-to-expert agreement. GPT-4V shows promise as a tool in glaucoma screening and detection through fundus image analysis, demonstrating generally high agreement with expert evaluations of key diagnostic features, although agreement did vary substantially across datasets. Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Purpose: To develop and evaluate a deep learning (DL) model to assess fundus photograph quality, and quantitatively measure its impact on automated POAG detection in independent study populations. Methods: Image quality ground truth was determined by manual review of 2815 fundus photographs of healthy and POAG eyes from the Diagnostic Innovations in Glaucoma Study and African Descent and Glaucoma Evaluation Study (DIGS/ADAGES), as well as 11,350 from the Ocular Hypertension Treatment Study (OHTS). Human experts assessed a photograph as high quality if of sufficient quality to determine POAG status and poor quality if not. A DL quality model was trained on photographs from DIGS/ADAGES and tested on OHTS. The effect of DL quality assessment on DL POAG detection was measured using area under the receiver operating characteristic (AUROC). Results: The DL quality model yielded an AUROC of 0.97 for differentiating between high- and low-quality photographs; qualitative human review affirmed high model performance. Diagnostic accuracy of the DL POAG model was significantly greater (P < 0.001) in good (AUROC, 0.87; 95% CI, 0.80–0.92) compared with poor quality photographs (AUROC, 0.77; 95% CI, 0.67–0.88). Conclusions: The DL quality model was able to accurately assess fundus photograph quality. Using automated quality assessment to filter out low-quality photographs increased the accuracy of a DL POAG detection model. Translational Relevance: Incorporating DL quality assessment into automated review of fundus photographs can help to decrease the burden of manual review and improve accuracy for automated DL POAG detection.