SegHeD: Segmentation of Heterogeneous Data for Multiple Sclerosis Lesions with Anatomical Constraints

TL;DR

SegHeD is a novel multi-dataset segmentation model for multiple sclerosis lesions that integrates heterogeneous data and domain knowledge, achieving high accuracy across diverse datasets and lesion types.

Contribution

The paper introduces SegHeD, a multi-task model that handles heterogeneous MS datasets and incorporates anatomical constraints for improved lesion segmentation.

Findings

Outperforms state-of-the-art methods on five MS datasets.

Effectively segments all, new, and vanishing lesions.

Successfully integrates diverse datasets with different formats and annotations.

Abstract

Assessment of lesions and their longitudinal progression from brain magnetic resonance (MR) images plays a crucial role in diagnosing and monitoring multiple sclerosis (MS). Machine learning models have demonstrated a great potential for automated MS lesion segmentation. Training such models typically requires large-scale high-quality datasets that are consistently annotated. However, MS imaging datasets are often small, segregated across multiple sites, with different formats (cross-sectional or longitudinal), and diverse annotation styles. This poses a significant challenge to train a unified MS lesion segmentation model. To tackle this challenge, we present SegHeD, a novel multi-dataset multi-task segmentation model that can incorporate heterogeneous data as input and perform all-lesion, new-lesion, as well as vanishing-lesion segmentation. Furthermore, we account for domain knowledge about MS lesions, incorporating longitudinal, spatial, and volumetric constraints into the segmentation model. SegHeD is assessed on five MS datasets and achieves a high performance in all, new, and vanishing-lesion segmentation, outperforming several state-of-the-art methods in this field.