Related Experiment Video
Updated: Nov 27, 2025
10:25
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
48.9K
Multiple Sclerosis lesions detection by a hybrid Watershed-Clustering algorithm.
Lilla Bonanno1, Nadia Mammone1, Simona De Salvo1
1IRCCS Centro Neurolesi "Bonino-Pulejo", Messina, Italy.
Clinical Imaging
|December 5, 2020
Summary
This study introduces a novel Computer Aided Diagnosis (CAD) system using a hybrid Watershed-Clustering algorithm to detect Multiple Sclerosis (MS) lesions. The system achieved 87% diagnostic accuracy, aiding clinical evaluation.
Area of Science:
- Medical Imaging Analysis
- Artificial Intelligence in Medicine
Background:
- Computer Aided Diagnosis (CAD) systems are advancing for disease diagnosis and monitoring.
- A novel CAD system is presented for detecting Multiple Sclerosis (MS) lesions.
Purpose of the Study:
- To develop and evaluate a hybrid Watershed-Clustering algorithm-based CAD system for MS lesion detection.
- To assess the system's accuracy in discriminating MS lesions from non-lesions.
Main Methods:
- Utilized Magnetic Resonance Imaging (MRI) FLAIR sequences from 20 MS patients.
- Employed automated segmentation (Watershed algorithm), feature extraction, and Cluster Analysis for lesion classification.
Main Results:
- Analyzed 316 regions (255 lesions, 61 non-lesions).
- Achieved 87% diagnostic accuracy (sensitivity 77%, specificity 87%) via ROC analysis.
- Identified a significant difference between lesions and non-lesions.
Conclusions:
- A CAD system with a modified automated image segmentation algorithm was developed.
- The system effectively discriminates MS lesions from non-lesions.
- The proposed method provides a detection output to support clinical evaluation.
