Segmenting Small Stroke Lesions with Novel Labeling Strategies

TL;DR

This paper introduces two new labeling strategies, MSL and DBL, that improve the segmentation accuracy of small stroke lesions in neural networks, outperforming previous methods on the ATLAS v2.0 dataset.

Contribution

The paper presents two novel labeling approaches, MSL and DBL, that enhance small lesion segmentation accuracy and can be integrated into various neural network architectures.

Findings

Ensemble of MSL and DBL improves recall, F1, and Dice scores.

Single MSL model surpasses previous best on small lesions.

Code is publicly available for replication.

Abstract

Deep neural networks have demonstrated exceptional efficacy in stroke lesion segmentation. However, the delineation of small lesions, critical for stroke diagnosis, remains a challenge. In this study, we propose two straightforward yet powerful approaches that can be seamlessly integrated into a variety of networks: Multi-Size Labeling (MSL) and Distance-Based Labeling (DBL), with the aim of enhancing the segmentation accuracy of small lesions. MSL divides lesion masks into various categories based on lesion volume while DBL emphasizes the lesion boundaries. Experimental evaluations on the Anatomical Tracings of Lesions After Stroke (ATLAS) v2.0 dataset showcase that an ensemble of MSL and DBL achieves consistently better or equal performance on recall (3.6% and 3.7%), F1 (2.4% and 1.5%), and Dice scores (1.3% and 0.0%) compared to the top-1 winner of the 2022 MICCAI ATLAS Challenge on both the subset only containing small lesions and the entire dataset, respectively. Notably, on the mini-lesion subset, a single MSL model surpasses the previous best ensemble strategy, with enhancements of 1.0% and 0.3% on F1 and Dice scores, respectively. Our code is available at: https://github.com/nadluru/StrokeLesSeg.