A Robust Dual-Stage Learning-Based Pipeline for Multiclass Segmentation of Multiple Sclerosis Lesions in MRI.
Authors
Affiliations (7)
Affiliations (7)
- Medical Physics and Biomedical Engineering Department, Faculty of Medicine, Tehran University of Medical Sciences, Tehran 1417653761, Iran.
- Research Center for Intelligent Technologies in Medicine (RCITM), Advanced Medical Technologies and Equipment Institute (AMTEI), Tehran University of Medical Sciences, Tehran 1417653761, Iran.
- Research Center of Biomedical Technology and Robotics (RCBTR), Advanced Medical Technologies and Equipment Institute (AMTEI), Tehran University of Medical Sciences, Tehran 1417653761, Iran.
- Research and Development Department, Parseh Intelligent Surgical Systems Company (Parsiss), Tehran 1418693913, Iran.
- Department of Molecular Imaging, Faculty of Advanced Technologies in Medicine, Iran University of Medical Sciences, Tehran 1449614535, Iran.
- Finetech in Medicine Research Center, School of Medicine, Iran University of Medical Sciences, Tehran 1449614535, Iran.
- Medical Physics Department, School of Medicine, Iran University of Medical Sciences, Tehran 1449614535, Iran.
Abstract
<b>Background</b>: Accurate segmentation and classification of multiple sclerosis (MS) lesions are vital for a reliable diagnosis and disease monitoring. However, lesion heterogeneity in size, location, and intensity poses significant challenges to automated analysis. <b>Methods</b>: To address this, we developed a dual-stage pipeline integrating deep learning (DL) for precise spatial delineation and machine learning (ML) for robust classification of MS lesions. Two advanced DL models, nnU-Net and UNETR++, were optimized for lesion segmentation. Moreover, UNETR++ and several conventional ML methods were considered for the classification task, and Random Forest was found to be the best choice. <b>Results</b>: Experimental results indicate that nnU-Net outperformed UNETR++ for lesion segmentation across all cases, achieving a maximum improvement of 12.8%. During classification, Random Forest consistently outperformed advanced DL models, achieving at least 12% higher performance. Under practical conditions, an optimized hybrid pipeline that integrates nnU-Net for precise segmentation with Random Forests for robust classification delivers the best overall performance. Furthermore, qualitative analysis indicates that some apparent false positives may correspond to lesions missed during annotation, highlighting potential limitations in ground truth labeling. <b>Conclusions</b>: Overall, the proposed pipeline effectively leverages the complementary strengths of DL and ML, offering a promising, accurate framework for automated MS lesion analysis with potential clinical utility.