Modeling Disease Progression In Retinal OCTs With Longitudinal Self-Supervised Learning

TL;DR

This paper introduces a self-supervised deep learning method that models disease progression in retinal OCT scans by estimating time intervals between scans, improving prediction of age-related macular degeneration onset.

Contribution

The novel approach leverages a time-interval estimation task for self-supervised learning, enabling effective modeling of disease progression without annotations.

Findings

Improved prediction accuracy for AMD onset using the learned representation.

Method is robust to irregular sampling and poor registration.

Representation focuses on progression-specific information.

Abstract

Longitudinal imaging is capable of capturing the static ana\-to\-mi\-cal structures and the dynamic changes of the morphology resulting from aging or disease progression. Self-supervised learning allows to learn new representation from available large unlabelled data without any expert knowledge. We propose a deep learning self-supervised approach to model disease progression from longitudinal retinal optical coherence tomography (OCT). Our self-supervised model takes benefit from a generic time-related task, by learning to estimate the time interval between pairs of scans acquired from the same patient. This task is (i) easy to implement, (ii) allows to use irregularly sampled data, (iii) is tolerant to poor registration, and (iv) does not rely on additional annotations. This novel method learns a representation that focuses on progression specific information only, which can be transferred to other types of longitudinal problems. We transfer the learnt representation to a clinically highly relevant task of predicting the onset of an advanced stage of age-related macular degeneration within a given time interval based on a single OCT scan. The boost in prediction accuracy, in comparison to a network learned from scratch or transferred from traditional tasks, demonstrates that our pretrained self-supervised representation learns a clinically meaningful information.