DiffSight-Former: Modeling Structural Differences and Temporal Dynamics for Glaucoma Progression Prediction

2026-06-08

AI Summary

Current approaches to glaucoma progression prediction predominantly rely on single-timepoint fundus images, limiting their ability to capture longitudinal structural and vascular changes and reducing sensitivity to subtle early-stage progression. To address these limitations, this work proposes DiffSight-Former, a novel framework that integrates a fundus-specific foundation model, multi-structure difference quantification, and temporal interval embedding within a time-aware Transformer architecture. This enables joint modeling of structural evolution and temporal dynamics in both the optic disc/cup regions and retinal vasculature, supporting variable-length input sequences without requiring pre-diagnosed samples. Evaluated on the SIGF and GRAPE datasets, the method achieves an AUC of 91.54% and a cross-criterion average accuracy of 87.48%, significantly outperforming existing state-of-the-art approaches.

Abstract

Glaucoma is a leading cause of irreversible blindness worldwide, and early detection from fundus images is critical for effective disease management. While deep learning has achieved promising performance in fundus image analysis, most existing methods rely on single time-point images and fail to capture longitudinal structural and vascular changes associated with disease progression. Sequential fundus images acquired during clinical follow-up provide valuable temporal information; however, current sequential models often struggle to detect subtle early progression signals and commonly depend on fixed-length inputs or diagnostic cues from already glaucomatous images, limiting their clinical utility for early prediction. To address these limitations, we propose DiffSight-Former, a framework for glaucoma progression prediction from sequential fundus images. It incorporates a time-variant feature extraction module based on a fundus-specific foundation model to obtain robust anatomical representations. A multi-structure difference modeling module is introduced to quantify progression-related changes in the optic disc/cup region and retinal vasculature. These representations are integrated with temporal interval embeddings and processed by a time-aware Transformer to model disease progression and estimate the probability of future glaucoma onset. Experiments were conducted on two longitudinal datasets, SIGF (405 sequences) and GRAPE (263 sequences). On SIGF, DiffSight-Former achieved an AUC of 91.54% and a sensitivity of 92.16% for progression prediction. On GRAPE, it achieved an average accuracy of 87.48% across three clinical visual-field progression criteria. Compared with existing approaches, DiffSight-Former demonstrates strong performance and robustness across different temporal settings, highlighting its potential for longitudinal glaucoma monitoring and early risk prediction.

Problem

• glaucoma progression prediction
• longitudinal fundus images
• structural differences
• temporal dynamics
• early detection

Innovation

• glaucoma progression prediction
• longitudinal fundus imaging
• structural difference modeling
• time-aware Transformer
• foundation model