UNISELF: A Unified Network with Instance Normalization and Self-Ensembled Lesion Fusion for Multiple Sclerosis Lesion Segmentation

2025-08-05

AI Summary

Existing MRI lesion segmentation methods for multiple sclerosis (MS) struggle to simultaneously achieve high in-domain accuracy and strong cross-domain generalization due to domain shifts arising from variations in scanning protocols, imaging devices, and missing contrast modalities. Method: We propose a robust single-source training framework that integrates test-time instance normalization (TTIN) within a unified network to mitigate domain shift, coupled with a test-time self-ensembling lesion fusion mechanism to enhance prediction stability. Contribution/Results: To our knowledge, this is the first approach achieving both high in-domain accuracy and superior cross-domain generalization without requiring multi-center data collaboration during training. Experiments demonstrate state-of-the-art performance on the ISBI 2015 test set and consistent superiority over baseline models on MICCAI 2016, UMCL, and multiple private multi-center datasets. The method significantly improves robustness and practicality for clinical deployment.

Technology Category

• Machine Learning: Multi-instance/Multi-view Learning
• Computer Vision: Segmentation
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Automated segmentation of multiple sclerosis (MS) lesions using multicontrast magnetic resonance (MR) images improves efficiency and reproducibility compared to manual delineation, with deep learning (DL) methods achieving state-of-the-art performance. However, these DL-based methods have yet to simultaneously optimize in-domain accuracy and out-of-domain generalization when trained on a single source with limited data, or their performance has been unsatisfactory. To fill this gap, we propose a method called UNISELF, which achieves high accuracy within a single training domain while demonstrating strong generalizability across multiple out-of-domain test datasets. UNISELF employs a novel test-time self-ensembled lesion fusion to improve segmentation accuracy, and leverages test-time instance normalization (TTIN) of latent features to address domain shifts and missing input contrasts. Trained on the ISBI 2015 longitudinal MS segmentation challenge training dataset, UNISELF ranks among the best-performing methods on the challenge test dataset. Additionally, UNISELF outperforms all benchmark methods trained on the same ISBI training data across diverse out-of-domain test datasets with domain shifts and missing contrasts, including the public MICCAI 2016 and UMCL datasets, as well as a private multisite dataset. These test datasets exhibit domain shifts and/or missing contrasts caused by variations in acquisition protocols, scanner types, and imaging artifacts arising from imperfect acquisition. Our code is available at https://github.com/uponacceptance.

Problem

• Improving MS lesion segmentation accuracy and generalization
• Addressing domain shifts and missing input contrasts
• Optimizing performance across diverse out-of-domain datasets

Innovation

• Test-time self-ensembled lesion fusion
• Test-time instance normalization (TTIN)
• Handles domain shifts and missing contrasts