Machine learning for diagnosis of early Alzheimer's disease

2022 Fiscal Year Final Research Report

• Project/Area Number: 17K10414
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Radiation science
• Research Institution: Kyoto Prefectural University of Medicine
• Principal Investigator: Yamada Kei 京都府立医科大学, 医学(系)研究科(研究院), 教授 (80315960)
• Project Period (FY): 2017-04-01 – 2023-03-31
• Keywords: MRI / 拡散強調画像 / アルツハイマー病 / 教師付き機械学習 / deep learning / 拡散テンソル画像

Outline of Final Research Achievements

Our research focused on DWI and q-space imaging (QSI). We shortened the scan time of QSI[1] and also the regular DWI using Simultaneous Multi-Slice technique[2]. We have also worked on post-processing techniques, namely TRACULA for Alzheimer’s disease (AD) patients[3] and automated segmentation tool [4]. We also did temperature measurements[5-7]. We also used high-resolution T1-weighted images and showed that AD patients tend to have smaller pineal gland [8] and volume loss in precuneous regions[9]. In another research, we have shown that specific regions of the brain are correlated to specific test results [10]. We also conducted a study on serum biomarker [11]. We also participated in a trial of amyloid PET studies [12].
Our research also focuses on machine learning (ML). We worked on developing ML that aids in diagnosis of AD. We first investigated the publications in the past 5 years[13]. We have also made predictions as to how this technique is going to affect us[14].

Free Research Field

放射線科学（神経放射線）

Academic Significance and Societal Importance of the Research Achievements

日本国民の急速な高齢化に伴い、認知症は万人の関心事項となった。中でもアルツハイマー病は最も頻度の高い認知症の一つである。その原因として有力視されているアミロイド沈着に対する薬剤の開発が進んでおり、早期発見の意義が高まっている。すなわち、アルツハイマーの可能性を早く確実に見つける事ができると先制治療が可能かもしれない。本研究は画像を用いた早期発見への礎であり、社会的に重要な意義が存在する。