Machine learning identified an Alzheimer’s disease-related FDG-PET pattern which is also expressed in Lewy body dementia and Parkinson’s disease dementia

Abstract

Utilizing the publicly available neuroimaging database enabled by Alzheimer’s disease Neuroimaging Initiative (ADNI; http://adni.loni.usc.edu/), we have compared the performance of automated classification algorithms that differentiate AD vs. normal subjects using Positron Emission Tomography (PET) with fluorodeoxyglucose (FDG). General linear model, scaled subprofile modeling and support vector machines were examined. Among the tested classification methods, support vector machine with Iterative Single Data Algorithm produced the best performance, i.e., sensitivity (0.84) × specificity (0.95), by 10-fold cross-validation. We have applied the same classification algorithm to four different datasets from ADNI, Health Science Centre (Winnipeg, Canada), Dong-A University Hospital (Busan, S. Korea) and Asan Medical Centre (Seoul, S. Korea). Our data analyses confirmed that the support vector machine with Iterative Single Data Algorithm showed the best performance in prediction of future development of AD from the prodromal stage (mild cognitive impairment), and that it was also sensitive to other types of dementia such as Parkinson’s Disease Dementia and Dementia with Lewy Bodies, and that perfusion imaging using single photon emission computed tomography may achieve a similar accuracy to that of FDG-PET.

Quantifying brain metabolism from FDG‐PET images into a probability of Alzheimer's dementia score

“…This is consistent with the classic pattern of impaired metabolism observed in DAT (Del Sole et al, ; Jagust, Reed, Mungas, Ellis, & Decarli, ; Mosconi et al, ; Mosconi et al, ; Sanabria‐Diaz, Martínez‐Montes, & Melie‐Garcia, ). The patterns of hypermetabolism observed in Figure echo those reported in a recent study (Katako et al, ). Although hypermetabolism is seldom reported in association with DAT, the regions showing hypermetabolism have been found to exhibit structural atrophy in DAT.…”

“…Thus, batch normalization requires a larger batch size in order to accurately estimate the mini-batch mean and variance. Training with a larger batch size is memory intensive and may lead to lower generalizability (Keskar, Mudigere, Nocedal, Smelyanskiy, & Tang, 2016). Within the residual blocks, we added instance normalization and leaky ReLU before each convolutional layer as preactivation (He, Zhang, Ren, & Sun, 2016b).…”

Artificial Intelligence in Nuclear Medicine

“…Many studies have already addressed the early detection of Alzheimer disease and mild cognitive impairment using deep learning (32)(33)(34)(35)(36)(37). Ding et al were able to show that a CNN with InceptionV3 architecture (38) could make an Alzheimer disease diagnosis with 82% specificity at 100% sensitivity (AUC, 0.98) on average 75.8 mo before the final diagnosis based on 18 F-FDG PET/CT scans and outperformed human interpreters (majority diagnosis of 5 interpreters) (39).…”

Bringing radiomics into a multi-omics framework for a comprehensive genotype–phenotype characterization of oncological diseases

“…Many subjects possess corresponding genomics data on the GDC (ex TCGA)Medical images in DICOM format, clinical dataOpen/controlled user account (open use studies) or request access by data use application (controlled use studies)Web based, web client and Programmatic2018Neurological and neurodegenerative disorders 1000 Functional Connectomes Project/INDI International NeuroImaging Data-sharing Initiative [52] and curse of dimensionality [4]. https://www.nitrc.org/projects/fcon_1000/It provides the broader imaging community complete access to a large-scale functional imaging dataset such as prospective, retrospective datasetImaging and clinical dataNITRC account for some public datasets and some controlled datasetAmazon Web Services S3 and CyberDuke web client and command line2018 LONI Database (The Laboratory of Neuroimaging at University of Southern California) [53] https://loni.usc.edu/about_loniRepository for sharing and long-term preservation of neuroimaging and biomedical research data especially on neurological, neurodegenerative and psychiatric diseases. Some studies ongoing are: ADNI, ENIGMA, GAAIN, PPMIClinical, imaging (MRI, PET, MRA, DTI and other imaging modalities), genetic and behavioral data from multisite longitudinal studyOpen use data required account controlled access by Image and Data Archive (IDA) request otherwise data use application requestWeb-based Image and Data Archive (IDA)*2018 LRRK2 Cohort consortium (The Michael J.…”