Development of an Integrated Alzheimer's Disease Diagnostic Approach Based on Neuropathological Insights from PET and MRI Imaging

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K18097
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 90140:Medical technology assessment-related
• Research Institution: Fukushima Medical University
• Principal Investigator: Miwa Kenta 福島県立医科大学, 保健科学部, 教授 (40716594)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: アミロイドPET / タウPET / ATN / deep learning

Outline of Final Research Achievements

To assess the standardized and quantitative AT(N) classification, we calculated the Centiloid Scale (CL) for amyloid PET images, the CenTauR for tau PET images, and the Hippocampal Ratio for MRI images using ADNI3 study data.
For amyloid PET, we developed a deep learning method to automatically calculate CL, improving analysis efficiency. We divided the whole-brain region into 12 regions to create region-specific CL formulas. For tau PET, using the cerebellar cortex as the reference, we calculated SUVRs for a universal region and four subregions, then obtained CenTauR using specific formulas. For MRI, we segmented gray matter and calculated volumes for 79 regions. Using these volumes, we calculated the normalized hippocampal volume (hippocampal ratio) to assess brain atrophy.
We established cut-offs for each index and conducted AT(N) classification evaluations, showing a high correlation between our standardized classification and clinical diagnosis.

Free Research Field

核医学技術

Academic Significance and Societal Importance of the Research Achievements

本研究でアルツハイマー病（AD）の定量評価法を提案し、ADの病態把握における精度向上が期待される。学術的意義として、標準化されたバイオマーカーの導入により、AD研究の再現性と信頼性が向上させる可能性がある。社会的意義として、提案した標準化された定量的AT（N）分類により、AD診断における異なる施設や研究間での画像データの直接的な比較・解釈、縦断的な評価、正常例と異常例の正確な閾値の定義が可能となり、早期診断や治療の改善に寄与する。