MIRAGE: Multimodal foundation model and benchmark for comprehensive retinal OCT image analysis

TL;DR

MIRAGE is a new multimodal foundation model and benchmark that significantly improves retinal OCT image analysis by integrating OCT and SLO data, outperforming existing models in classification and segmentation tasks.

Contribution

This work introduces MIRAGE, a novel multimodal foundation model for OCT and SLO images, along with a comprehensive benchmark for evaluation, addressing limitations of existing models.

Findings

MIRAGE outperforms general and specialized FMs in classification tasks.

MIRAGE achieves superior segmentation performance on retinal OCT images.

The benchmark facilitates robust evaluation of AI models in ophthalmology.

Abstract

Artificial intelligence (AI) has become a fundamental tool for assisting clinicians in analyzing ophthalmic images, such as optical coherence tomography (OCT). However, developing AI models often requires extensive annotation, and existing models tend to underperform on independent, unseen data. Foundation models (FMs), large AI models trained on vast unlabeled datasets, have shown promise in overcoming these challenges. Nonetheless, available FMs for ophthalmology lack extensive validation, especially for segmentation tasks, and focus on a single imaging modality. In this context, we propose MIRAGE, a novel multimodal FM for the analysis of OCT and scanning laser ophthalmoscopy (SLO) images. Additionally, we propose a new evaluation benchmark with OCT/SLO classification and segmentation tasks. The comparison with general and specialized FMs and segmentation methods shows the superiority of MIRAGE in both types of tasks, highlighting its suitability as a basis for the development of robust AI systems for retinal OCT image analysis. Both MIRAGE and the evaluation benchmark are publicly available: https://github.com/j-morano/MIRAGE.