Multimodal and Multiscale Deep Neural Networks for the Early Diagnosis of Alzheimer’s Disease using structural MR and FDG-PET images

TLDR

Alzheimer’s disease is a progressive neurodegenerative disorder for which biomarkers derived from MRI and FDG‑PET neuroimaging can provide objective measures of structural and metabolic changes useful for diagnosis and staging. The study aims to test whether integrating multimodal imaging data can enhance early diagnosis of Alzheimer’s disease. The authors develop a multimodal, multiscale deep neural network that fuses MRI and FDG‑PET data to classify individuals with Alzheimer’s disease. The model achieves 82.4 % accuracy in identifying mild cognitive impairment patients who will convert to AD within three years, 86.4 % combined accuracy for conversion within 1–3 years, 94.23 % sensitivity for probable AD, and 86.3 % specificity for non‑demented controls, outperforming published results.

Abstract

Abstract Alzheimer’s Disease (AD) is a progressive neurodegenerative disease where biomarkers for disease based on pathophysiology may be able to provide objective measures for disease diagnosis and staging. Neuroimaging scans acquired from MRI and metabolism images obtained by FDG-PET provide in-vivo measurements of structure and function (glucose metabolism) in a living brain. It is hypothesized that combining multiple different image modalities providing complementary information could help improve early diagnosis of AD. In this paper, we propose a novel deep-learning-based framework to discriminate individuals with AD utilizing a multimodal and multiscale deep neural network. Our method delivers 82.4% accuracy in identifying the individuals with mild cognitive impairment (MCI) who will convert to AD at 3 years prior to conversion (86.4% combined accuracy for conversion within 1–3 years), a 94.23% sensitivity in classifying individuals with clinical diagnosis of probable AD, and a 86.3% specificity in classifying non-demented controls improving upon results in published literature.