Feature Integration of [¹⁸F]FDG PET Brain Imaging Using Deep Learning for Sensitive Cognitive Decline Detection

Area of Science:

• Neuroimaging
• Machine Learning
• Medical Diagnostics

Background:

• Accurate diagnosis of cognitive decline (CD), including early Alzheimer's disease, is crucial for timely intervention.
• Positron emission tomography (PET) reveals functional brain changes in CD but faces limitations due to cost and radiation.
• A data-efficient multi-representational learning framework is proposed to enhance PET's clinical utility.

Purpose of the Study:

• To develop and validate a novel deep learning framework for distinguishing cognitively normal (CN) individuals from those with cognitive decline (CD).
• To leverage both voxel-level imaging data and region-level quantification for improved diagnostic accuracy.
• To enhance the data efficiency and clinical applicability of PET in diagnosing CD.

Main Methods:

• Extracted voxel-level features from [¹⁸F]FDG PET using CNNs and PCANets.
• Derived region-level features using DNNs from standardized uptake value ratio measurements.
• Integrated voxel- and region-level features via direct concatenation and applied various machine learning models for prediction, validated on 252 ADNI participants.

Main Results:

• The integrated DNN-CNN model achieved the highest classification accuracy (0.87 ± 0.05), outperforming individual models.
• This fusion approach demonstrated a 6.10% accuracy improvement, increased Recall by 14.29%, and F1-Score by 7.32% compared to the DNN-only model.
• Model performance showed significant correlation with MMSE scores and surpassed MMSE-based classification in accuracy, recall, and F1-score.

Conclusions:

• Combining PET imaging, region-level quantification, and deep learning significantly improves diagnostic performance for cognitive decline.
• Fusion-based deep learning strategies enhance sensitivity in detecting cognitive decline, offering a more accurate and data-efficient approach.
• This multimodal strategy supports broader clinical application of PET for diagnosing cognitive decline and related conditions like Alzheimer's disease.