Classification of Active and Non-active MS Lesions Using Various Machine Learning Models

Introduction:
Gadolinium-based T1-weighted MRI sequence is the gold standard for the detection of active multiple sclerosis (MS) lesions. The performance of machine learning (ML) models in the classification of active and non- active MS lesions from the non-contrasted T2- weighted MRI images has been investigated in this study.
Methods:
108 Features of 75 active and 100 non- active MS lesions was extracted by using SegmentEditor and Radiomics modules of 3D slicer software. 18 ML models have been made using the 5-fold cross-validation method and each model with its special optimized parameters has been trained by using the training-validation data sets. Performance models in test data set has been evaluated by metric parameters of accuracy, precision, sensitivity, specificity, AUC, and F1 score.
Results:
The highest values of accuracy (91.9%), precision (95.5%), sensitivity (85%), specificity
(95.8%), AUC (94.2%), and F1 score (89.5%)
have been seen in LogisticRegression model.
Conclusion:
The performance of ML models in the classification of active and non-active MS lesions was evaluated. The results of this study show that the LogesticRegression model is the best and reliable ML model for this purpose.
Keywords: Multiple Sclerosis, Machine Learning,