Detection of diabetic retinopathy using longitudinal self-supervised learning

TL;DR

This paper explores the use of longitudinal self-supervised learning to improve early detection of diabetic retinopathy progression from retinal images, achieving higher accuracy than traditional methods.

Contribution

It introduces a novel longitudinal self-supervised learning approach for DR diagnosis, leveraging disease progression information from consecutive retinal images.

Findings

LSSL improves AUC from 0.875 to 0.96 in DR detection.

LSSL encodes dynamic disease progression effectively.

Frozen LSSL weights enhance model performance.

Abstract

Longitudinal imaging is able to capture both static anatomical structures and dynamic changes in disease progression towards earlier and better patient-specific pathology management. However, conventional approaches for detecting diabetic retinopathy (DR) rarely take advantage of longitudinal information to improve DR analysis. In this work, we investigate the benefit of exploiting self-supervised learning with a longitudinal nature for DR diagnosis purposes. We compare different longitudinal self-supervised learning (LSSL) methods to model the disease progression from longitudinal retinal color fundus photographs (CFP) to detect early DR severity changes using a pair of consecutive exams. The experiments were conducted on a longitudinal DR screening dataset with or without those trained encoders (LSSL) acting as a longitudinal pretext task. Results achieve an AUC of 0.875 for the baseline (model trained from scratch) and an AUC of 0.96 (95% CI: 0.9593-0.9655 DeLong test) with a p-value < 2.2e-16 on early fusion using a simple ResNet alike architecture with frozen LSSL weights, suggesting that the LSSL latent space enables to encode the dynamic of DR progression.