Deep Learning-Based Diagnosis Algorithm for Alzheimer's Disease.

Area of Science:

• Neuroimaging and Artificial Intelligence
• Medical Diagnostics

Background:

• Alzheimer's disease (AD) prevalence is increasing with the aging population.
• Medical imaging and AI integration enhances brain disease diagnosis efficiency.
• Current diagnostic methods require improvement for accuracy and speed.

Purpose of the Study:

• To develop an innovative two-stage automatic auxiliary diagnosis algorithm for AD.
• To improve the accuracy and efficiency of diagnosing Alzheimer's disease using brain MR images.
• To leverage deep learning models for enhanced AD detection.

Main Methods:

• An improved 3D DenseNet segmentation model and an improved MobileNetV3 classification model were utilized.
• Modifications included backbone simplification, activation/loss function replacement, and integration of 3D GAM and CA attention mechanisms.
• Dilated convolutions and transfer learning were employed to enhance feature extraction capabilities.

Main Results:

• The proposed algorithm achieved high classification accuracies: 97.85% for AD/NC, 95.31% for MCI/NC, 93.96% for AD/MCI, and 92.63% for AD/MCI/NC.
• Accuracy improvements ranged from 2.6 to 3.1 percentage points compared to the baseline models.
• Comparative and ablation studies validated the superior classification performance.

Conclusions:

• The developed deep learning-based algorithm offers an accurate and efficient automated auxiliary diagnosis for AD.
• The enhanced segmentation and classification models demonstrate significant improvements in detecting AD and related cognitive impairments.
• This approach provides a promising solution for early and reliable Alzheimer's disease diagnosis.