White Matter Lesion Segmentation for Multiple Sclerosis Patients implementing deep learning.

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Neurology

Background:

• Multiple Sclerosis (MS) is a demyelinating disease affecting the central nervous system.
• Accurate segmentation of White Matter Lesions (WML) in Magnetic Resonance Imaging (MRI) is crucial for monitoring MS progression.
• Existing segmentation methods may face challenges with lesion variability and image quality.

Purpose of the Study:

• To develop and evaluate a deep learning model for automated WML segmentation in MS patients.
• To apply a U-net based architecture for precise pixel-wise classification of lesions.
• To assess the model's performance using quantitative metrics on FLAIR MRI data.

Main Methods:

• Implementation of a U-net architecture with contrastive and expanding paths.
• Utilized Fluid-Attenuated Inversion Recovery (FLAIR) MRI images from 30 MS patients.
• Data included baseline and follow-up scans from Ippokratio Radiology Center.

Main Results:

• The deep learning model achieved a Dice coefficient of 0.7292.
• Precision was recorded at 75.92% and Recall at 70.16%.
• Significant pixel-wise classification performance was observed on 2D FLAIR MRI slices.

Conclusions:

• The U-net based deep learning model shows promise for automated WML segmentation in MS.
• The model demonstrates effective performance in identifying and classifying lesions in MRI scans.
• Further validation on larger datasets could enhance clinical applicability for MS patient management.