Predicting Stroke through Retinal Graphs and Multimodal Self-supervised Learning

TL;DR

This paper introduces a novel self-supervised multimodal framework using retinal vessel graphs and clinical data to improve stroke prediction accuracy, reducing reliance on labeled datasets and enhancing efficiency.

Contribution

It presents one of the first contrastive frameworks integrating graph and tabular data for stroke prediction, leveraging self-supervised learning with retinal vessel graphs.

Findings

AUROC improved by 3.78% with self-supervised learning

Graph-level representations outperform image encoders

Reduced pre-training and fine-tuning runtimes

Abstract

Early identification of stroke is crucial for intervention, requiring reliable models. We proposed an efficient retinal image representation together with clinical information to capture a comprehensive overview of cardiovascular health, leveraging large multimodal datasets for new medical insights. Our approach is one of the first contrastive frameworks that integrates graph and tabular data, using vessel graphs derived from retinal images for efficient representation. This method, combined with multimodal contrastive learning, significantly enhances stroke prediction accuracy by integrating data from multiple sources and using contrastive learning for transfer learning. The self-supervised learning techniques employed allow the model to learn effectively from unlabeled data, reducing the dependency on large annotated datasets. Our framework showed an AUROC improvement of 3.78% from supervised to self-supervised approaches. Additionally, the graph-level representation approach achieved superior performance to image encoders while significantly reducing pre-training and fine-tuning runtimes. These findings indicate that retinal images are a cost-effective method for improving cardiovascular disease predictions and pave the way for future research into retinal and cerebral vessel connections and the use of graph-based retinal vessel representations.