GONet: A Generalizable Deep Learning Model for Glaucoma Detection

2025-02-26

AI Summary

Glaucoma deep learning models exhibit limited generalizability across racial groups, imaging devices, and clinical settings, hindering real-world deployment. To address this, we propose GONet—a highly generalizable model that uniquely integrates DINOv2 self-supervised pretraining with a multi-source domain collaborative fine-tuning strategy. We establish a multicenter training paradigm encompassing seven independent cohorts and 119,000 expert-annotated fundus images with gold-standard glaucomatous optic neuropathy (GON) labels. Furthermore, we release the first high-quality, pixel-level GON lesion annotation dataset (768 images). Built upon a Vision Transformer architecture, GONet jointly optimizes glaucoma classification and lesion discrimination tasks. Under diverse out-of-distribution scenarios—including cross-race, cross-device, and cross-site evaluations—GONet achieves AUCs of 0.85–0.99, substantially outperforming state-of-the-art methods and improving upon conventional cup-to-disc ratio metrics by up to 21.6%.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Deep Neural Architectures and Foundation Models
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

Glaucomatous optic neuropathy (GON) is a prevalent ocular disease that can lead to irreversible vision loss if not detected early and treated. The traditional diagnostic approach for GON involves a set of ophthalmic examinations, which are time-consuming and require a visit to an ophthalmologist. Recent deep learning models for automating GON detection from digital fundus images (DFI) have shown promise but often suffer from limited generalizability across different ethnicities, disease groups and examination settings. To address these limitations, we introduce GONet, a robust deep learning model developed using seven independent datasets, including over 119,000 DFIs with gold-standard annotations and from patients of diverse geographic backgrounds. GONet consists of a DINOv2 pre-trained self-supervised vision transformers fine-tuned using a multisource domain strategy. GONet demonstrated high out-of-distribution generalizability, with an AUC of 0.85-0.99 in target domains. GONet performance was similar or superior to state-of-the-art works and was significantly superior to the cup-to-disc ratio, by up to 21.6%. GONet is available at [URL provided on publication]. We also contribute a new dataset consisting of 768 DFI with GON labels as open access.

Problem

• Detects glaucoma using deep learning.
• Improves cross-ethnicity diagnostic generalizability.
• Leverages diverse datasets for robust performance.

Innovation

• Self-supervised vision transformers
• Multisource domain fine-tuning
• High out-of-distribution generalizability