Regularization via gradient attribution for multiple sclerosis lesion segmentation.
- 1Institute for High-Performance Computing and Networking (ICAR), National Research Council of Italy (CNR), Naples, Italy.
Abstract
Introduction
Accurate segmentation of Multiple Sclerosis (MS) lesions from Magnetic Resonance Imaging (MRI) is fundamental for robust disease monitoring and for the evaluation of therapeutic efficacy. Within Internet of Medical Things (IoMT)-based diagnostic infrastructures, where deep learning models serve as autonomous computational nodes embedded in interconnected clinical workflows, ensuring reliable lesion detection without continuous human supervision is a critical operational and clinical requirement. A persistent methodological challenge is the pronounced class imbalance characteristic of MS neuroimaging datasets. This imbalance causes conventional loss functions to systematically miss small lesions, thereby directly compromising the utility of automated image analysis as an instrument for precision medicine.Methods
This work proposes the Regularization via Gradient Attribution (RGA) framework, a module that leverages gradient-based eXplainable Artificial Intelligence (XAI) attribution maps as active supervisory signals during training. Instead of constraining the learning process exclusively at the prediction-output level, RGA imposes additional penalties on the network when false-negative lesion regions have low internal attribution and when false-positive healthy-tissue regions exhibit elevated attribution. The framework is architecturally agnostic, compatible with both convolutional and transformer-based backbone networks, and accommodating arbitrary spatial attribution techniques. RGA was evaluated on two public MS lesion segmentation datasets, MSLesSeg and 3D-MR-MS, using three backbones (nnU-Net, UNETR, TransBTS) and two XAI methods (LayerCAM, Integrated Gradients) in 5-fold cross-validation.Results
On both datasets, RGA consistently improved DSC, TPR, and LTPR over the baselines. The best configuration, nnU-Net + RGA achieved a DSC of 0.7056 on MSLesSeg and 0.7466 on 3D-MR-MS, with RGA yielding the highest DSC for all backbones. Qualitative analyses further substantiate these findings by demonstrating a marked reduction in missed-lesion regions across all model architectures, without a corresponding substantial increase in false-positive predictions.Discussion
By jointly providing high lesion-level detection completeness, model interpretability, and zero additional inference-time computational cost, RGA represents a methodologically grounded approach compatible with IoMT-oriented precision-medicine pipelines, while requiring dedicated deployment validation before clinical or edge-device implementation.
Related
- Multi-scale Gaussian product-of-experts for missing sequence multiple sclerosis lesion segmentation
- Bridging field strengths: fine-tuned deep learning models for 7T MRI white matter lesion segmentation in multiple sclerosis
- Automated optic chiasm segmentation in magnetic resonance imaging of multiple sclerosis patients using a multi-view, multi-class U-Net with limited annotated data
- Baclofen for Spasticity in Multiple Sclerosis: A Comprehensive Narrative Review
- A 35-Year-Old Woman With Multiple Sclerosis and a New Atypical Thalamic Lesion: A Case Report From the National Multiple Sclerosis Society Case Conference Proceedings.
BibTeX
@article{giugliano2026regularization,
title = {{Regularization via gradient attribution for multiple sclerosis lesion segmentation.}},
author = {Salvatore Giugliano and Giovanna Sannino},
journal = {Frontiers in medicine},
year = {2026},
month = sep,
doi = {10.3389/fmed.2026.1907567},
url = {https://europepmc.org/article/MED/42846083},
note = {CC BY},
}
RIS
TY - JOUR
TI - Regularization via gradient attribution for multiple sclerosis lesion segmentation.
AU - Salvatore Giugliano
AU - Giovanna Sannino
PY - 2026
DA - 2026/09/23
JO - Frontiers in medicine
DO - 10.3389/fmed.2026.1907567
UR - https://europepmc.org/article/MED/42846083
AB - Introduction Accurate segmentation of Multiple Sclerosis (MS) lesions from Magnetic Resonance Imaging (MRI) is fundamental for robust disease monitoring and for the evaluation of therapeutic efficacy. Within Internet of Medical Things (IoMT)-based diagnostic infrastructures, where deep learning models serve as autonomous computational nodes embedded in interconnected clinical workflows, ensuring reliable lesion detection without continuous human supervision is a critical operational and clinical requirement. A persistent methodological challenge is the pronounced class imbalance characteristic of MS neuroimaging datasets. This imbalance causes conventional loss functions to systematically miss small lesions, thereby directly compromising the utility of automated image analysis as an instrument for precision medicine.Methods This work proposes the Regularization via Gradient Attribution (RGA) framework, a module that leverages gradient-based eXplainable Artificial Intelligence (XAI) attribution maps as active supervisory signals during training. Instead of constraining the learning process exclusively at the prediction-output level, RGA imposes additional penalties on the network when false-negative lesion regions have low internal attribution and when false-positive healthy-tissue regions exhibit elevated attribution. The framework is architecturally agnostic, compatible with both convolutional and transformer-based backbone networks, and accommodating arbitrary spatial attribution techniques. RGA was evaluated on two public MS lesion segmentation datasets, MSLesSeg and 3D-MR-MS, using three backbones (nnU-Net, UNETR, TransBTS) and two XAI methods (LayerCAM, Integrated Gradients) in 5-fold cross-validation.Results On both datasets, RGA consistently improved DSC, TPR, and LTPR over the baselines. The best configuration, nnU-Net + RGA achieved a DSC of 0.7056 on MSLesSeg and 0.7466 on 3D-MR-MS, with RGA yielding the highest DSC for all backbones. Qualitative analyses further substantiate these findings by demonstrating a marked reduction in missed-lesion regions across all model architectures, without a corresponding substantial increase in false-positive predictions.Discussion By jointly providing high lesion-level detection completeness, model interpretability, and zero additional inference-time computational cost, RGA represents a methodologically grounded approach compatible with IoMT-oriented precision-medicine pipelines, while requiring dedicated deployment validation before clinical or edge-device implementation.
SN - 2296-858X
ER -