Interaction of a priori Anatomic Knowledge with Self-Supervised Contrastive Learning in Cardiac Magnetic Resonance Imaging

TL;DR

This study investigates how integrating anatomical prior knowledge with self-supervised contrastive learning affects cardiac MRI analysis, finding that anatomical guidance improves diagnostic accuracy and saliency, especially with in-domain data.

Contribution

It evaluates the impact of incorporating a priori anatomical knowledge into SSCL for cardiac MRI, highlighting benefits for downstream diagnostic performance.

Findings

Anatomical knowledge improves diagnostic accuracy.

In-domain SSCL pretraining enhances performance.

Anatomical guidance has limited impact when combined with SSCL.

Abstract

Training deep learning models on cardiac magnetic resonance imaging (CMR) can be a challenge due to the small amount of expert generated labels and inherent complexity of data source. Self-supervised contrastive learning (SSCL) has recently been shown to boost performance in several medical imaging tasks. However, it is unclear how much the pre-trained representation reflects the primary organ of interest compared to spurious surrounding tissue. In this work, we evaluate the optimal method of incorporating prior knowledge of anatomy into a SSCL training paradigm. Specifically, we evaluate using a segmentation network to explicitly local the heart in CMR images, followed by SSCL pretraining in multiple diagnostic tasks. We find that using a priori knowledge of anatomy can greatly improve the downstream diagnostic performance. Furthermore, SSCL pre-training with in-domain data generally improved downstream performance and more human-like saliency compared to end-to-end training and ImageNet pre-trained networks. However, introducing anatomic knowledge to pre-training generally does not have significant impact.