How can a person prevent Alzheimer’s Disease, and is there any treatment for it? Early diagnosis may not be an answer but helps early interventions that are delaying the development of the disease. For this reason, early diagnosis of Alzheimer’s became a priority in research. There is substantial evidence that white matter alterations may help in early Alzheimer’s diagnosis.
The interinstitutional team of researchers, led by Yong Liu, professor in the School of Artificial Intelligence, Beijing University of Posts and Telecommunications, Beijing, China, organized competition, and through the competition, they created a multisite Diffusion Tensor Imaging (DTI) image platform and validated the feasibility of white matter measures for Alzheimer’s diagnosis. The article is online now in Brain Disorders.
“We conducted a competition with diffusion measurements along 18 fiber tracts as features extracted via the automated fiber quantification (AFQ) method based on one of the largest worldwide DTI multisite biobanks,” Professor Liu said. “The present dataset combined data from 7 Magnetic Resonance Imaging scanners in 4 hospitals in China, which contained a total of 862 individuals with DTI images, T1 images, and demographic and psychological information.”
For the extraction of white matter features, the team performed an AFQ pipeline which consisted of three steps. Firstly, they tracked whole-brain fibers with a deterministic streamline tracking algorithm. Then they segmented fiber tracts with waypoint regions of interest procedure and refined them employing a fiber tract probability map. A second procedure was to repeatedly remove abnormal fibers that evolved far from the fiber tract’s core. At each node of each fiber, a nodal diffusion property was determined by re-sampling each fiber to 100 nodes equally spaced between the two regions of interest.
In the competition, Professor Liu and colleagues got solutions from scientists from universities/institutes in China, the United States, and the United Kingdom. The objective of the competition was to assess and develop an analytical framework to optimize the performance of Alzheimer’s binary classification using diffusion measurements along major white matter tracts extracted via AFQ. Based on the results, it was clear that DTI measures of white matter fibers demonstrated their utility in detecting Alzheimer’s in the early stages. The authors have suggested that a post hoc analysis for a model explanation, a larger multimodality multisite dataset, and better algorithms will be far more efficient for the early detection of Alzheimer’s.
The research team successfully determined that white matter could be a biomarker for the early diagnosis of Alzheimer’s. Moreover, Professor Liu said to Science Feature that: “The dataset and portions of the codes are available as open sources. The project we present in this study would benefit from having more researchers involved for sharing data or machine learning models, as well as for framing biomarker extraction as an open, international challenge to predict Alzheimer’s with the largest available DTI dataset biobank.” He also added that they concentrated on Alzheimer’s diagnosis for the current challenge. Still, a classification of numerous mental illnesses might be more beneficial for future clinical applications, hence delivering benefits for both revealing pathological differences between disorders and boosting the accuracy of a precise diagnosis.
Journal Reference and image credits:
Yida Qu, Pan Wang, Bing Liu, Chengyuan Song, Dawei Wang, Hongwei Yang, Zengqiang Zhang…, Yong Liu et al. “AI4AD: artificial intelligence analysis for Alzheimer’s disease classification based on a multisite DTI database.” Brain Disorders 1 (2021): 100005. https://doi.org/10.1016/j.dscb.2021.100005