Related Experiment Videos
Corpus Callosum Radiomics-Based Classification Model in Alzheimer's Disease: A Case-Control Study
Qi Feng1,2, Yuanjun Chen3, Zhengluan Liao4
1Bengbu Medical College, Bengbu, China.
Frontiers in Neurology
|August 11, 2018
Summary
Radiomics analysis of the corpus callosum (CC) in Alzheimer's disease (AD) identified texture features for diagnosis. A machine learning model showed potential for distinguishing AD patients from healthy controls.
Area of Science:
- Neuroimaging
- Radiomics
- Machine Learning
Background:
- Alzheimer's disease (AD) is a progressive neurodegenerative disorder impacting cognitive functions.
- The corpus callosum (CC) is a critical white matter tract potentially affected in AD.
Purpose of the Study:
- To identify radiomic features of the corpus callosum (CC) associated with Alzheimer's disease (AD) diagnosis.
- To develop and validate a machine learning classification model for AD detection using CC features.
Main Methods:
- Radiomics analysis was performed on 3D T1-weighted MPRAGE MRI scans from 78 AD patients and 44 healthy controls (HC).
- The corpus callosum (CC) was manually segmented, and 385 radiomic features were extracted.
- Feature selection using LASSO identified key features, and a logistic regression model was built for classification.
Main Results:
- Eleven radiomic features from the CC were selected using the LASSO model.
- The logistic regression model achieved an AUC of 0.720, with sensitivity 0.792, specificity 0.500, and accuracy 0.684.
- The model demonstrated potential in distinguishing AD subjects from HC.
Conclusions:
- Corpus callosum (CC) texture features show promise as imaging biomarkers for Alzheimer's disease (AD) diagnosis.
- A radiomics-based machine learning model offers a valuable approach for AD detection.