Deep learning detection of informative features in tau PET for Alzheimer’s disease classification

Abstract

BackgroundAlzheimer’s disease (AD) is the most common type of dementia, typically characterized by memory loss followed by progressive cognitive decline and functional impairment. Many clinical trials of potential therapies for AD have failed, and there is currently no approved disease-modifying treatment. Biomarkers for early detection and mechanistic understanding of disease course are critical for drug development and clinical trials. Amyloid has been the focus of most biomarker research. Here, we developed a deep learning-based framework to identify informative features for AD classification using tau positron emission tomography (PET) scans.MethodsWe analysed [18F]flortaucipir PET image data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) cohort. We first developed an image classifier to distinguish AD from cognitively normal (CN) older adults by training a 3D convolutional neural network (CNN)-based deep learning model on tau PET images (N=132; 66 CN and 66 AD), then applied the classifier to images from individuals with mild cognitive impairment (MCI; N=168). In addition, we applied a layer-wise relevance propagation (LRP)-based model to identify informative features and to visualize classification results. We compared these results with those from whole brain voxel-wise between-group analysis using conventional Statistical Parametric Mapping (SPM12).ResultsThe 3D CNN-based classification model of AD from CN yielded an average accuracy of 90.8% based on five-fold cross-validation. The LRP model identified the brain regions in tau PET images that contributed most to the AD classification from CN. The top identified regions included the hippocampus, parahippocampus, thalamus, and fusiform. The LRP results were consistent with those from the voxel-wise analysis in SPM12, showing significant focal AD associated regional tau deposition in the bilateral temporal lobes including the entorhinal cortex. The AD probability scores calculated by the classifier were correlated with brain tau deposition in the medial temporal lobe in MCI participants (r=0.43 for early MCI and r=0.49 for late MCI).ConclusionA deep learning framework combining 3D CNN and LRP algorithms can be used with tau PET images to identify informative features for AD classification and may have application for early detection during prodromal stages of AD.

Deep learning improves utility of tau PET in the study of Alzheimer's disease

“…We additionally show that multiple generations of tau radioligands can be integrated into a single framework 7,11 . Earlier deep learning efforts with tau PET have sought to simplify 20 and augment 17 preprocessing steps; and have provided initial proof of feasibility and model interpretation 22 . The main metrics of the models compare similarly with prior deep learning classification tasks that used different sets of PET radioligands, 16,22,46 though comparisons are of course difficult given differences in available clinical outcomes.…”

Explainable, Trustworthy, and Ethical Machine Learning for Healthcare: A Survey

“…They concluded that LRP generates more relevant explanations by describing why any individual patient has the disease. Taeho Jo et al [110] implemented LRP to highlight the areas of tau positron emission tomography (PET) that highly contribute to the classification of Alzheimer's disease using 3D-CNN.…”

Accelerating bioinformatics research with International Conference on Intelligent Biology and Medicine 2020

“…Extensive evaluations showed that the Auto3DCryoMap can accurately align structural particle shapes and can construct a decent 3D density map from only a few thousand aligned particle images while the existing tools require hundreds of thousands of particle images and reconstruct a better 3D density map. Jo et al published “Deep learning detection of informative features in tau PET for Alzheimer’s disease classification” [ 9 ], in which the authors developed a deep learning-based framework to identify informative features for Alzheimer’s disease classification using tau position emission tomography scans. By applying five-fold cross-validation, the authors demonstrated their method yielded an accuracy of 90.8%.…”