TINC: Temporally Informed Non-Contrastive Learning for Disease Progression Modeling in Retinal OCT Volumes

TL;DR

This paper introduces TINC, a non-contrastive learning method that leverages temporal information in longitudinal OCT data to improve disease progression prediction without heavy augmentations or large batch sizes.

Contribution

The work presents a novel temporally informed non-contrastive loss and pair-forming scheme that effectively utilizes longitudinal data for disease progression modeling in retinal OCT images.

Findings

Outperforms existing models in predicting AMD progression.

Effectively uses temporal information without heavy data augmentation.

Improves disease risk prediction accuracy.

Abstract

Recent contrastive learning methods achieved state-of-the-art in low label regimes. However, the training requires large batch sizes and heavy augmentations to create multiple views of an image. With non-contrastive methods, the negatives are implicitly incorporated in the loss, allowing different images and modalities as pairs. Although the meta-information (i.e., age, sex) in medical imaging is abundant, the annotations are noisy and prone to class imbalance. In this work, we exploited already existing temporal information (different visits from a patient) in a longitudinal optical coherence tomography (OCT) dataset using temporally informed non-contrastive loss (TINC) without increasing complexity and need for negative pairs. Moreover, our novel pair-forming scheme can avoid heavy augmentations and implicitly incorporates the temporal information in the pairs. Finally, these representations learned from the pretraining are more successful in predicting disease progression where the temporal information is crucial for the downstream task. More specifically, our model outperforms existing models in predicting the risk of conversion within a time frame from intermediate age-related macular degeneration (AMD) to the late wet-AMD stage.