Distinguishing amyotrophic lateral sclerosis from radiculopathy using machine learning to analyze nerve conduction...

Area of Science:

• Neurology
• Biomedical Engineering
• Data Science

Background:

• Amyotrophic lateral sclerosis (ALS) and radiculopathy present similar clinical symptoms, complicating diagnosis.
• Distinguishing between ALS and radiculopathy is crucial for appropriate patient management and treatment.
• Current diagnostic methods may not always provide definitive differentiation.

Purpose of the Study:

• To develop a reliable method for differentiating ALS from radiculopathy.
• To identify novel clinical biomarkers using machine learning from nerve conduction study (NCS) data.
• To enhance the accuracy of differential diagnosis for these neurological conditions.

Main Methods:

• Utilized machine learning algorithms, including random forest and XGBoost, for feature selection and classification.
• Applied data preparation techniques and a confusion matrix for model selection.
• Employed grid search cross-validation to optimize algorithm hyperparameters.
• Analyzed 77 ranked features from NCS data to identify the most informative ones.

Main Results:

• The XGB algorithm, utilizing 35 key NCS features, achieved the highest performance metrics.
• Achieved high accuracy (0.871), precision (0.923), recall (0.850), and F1-score (0.857) in differentiating ALS from radiculopathy.
• Demonstrated superior performance, particularly in the recall parameter, for identifying ALS patients.

Conclusions:

• Machine learning, specifically the XGB algorithm with 35 NCS features, offers a highly accurate method for differentiating ALS from radiculopathy.
• This approach can serve as a valuable tool for improving the differential diagnosis of these complex neurological disorders.
• The identified NCS features hold potential as novel biomarkers for distinguishing ALS from radiculopathy.