Construction and implementation of an AI prediction model for the onset of cardiovascular disease based on 700,000 people and 43 years of large-scale health checkup data

• Project/Area Number: 21K08034
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 53020:Cardiology-related
• Research Institution: Kagoshima University
• Principal Investigator: KAWASOE SHIN 鹿児島大学, 医歯学総合研究科, 特任講師 (00810201)
• Co-Investigator(Kenkyū-buntansha): 窪薗 琢郎 鹿児島大学, 医歯学域医学系, 講師 (00598013) ; 大石 充 鹿児島大学, 医歯学域医学系, 教授 (50335345)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Project Status: Completed (Fiscal Year 2023)
• Budget Amount: ¥4,160,000 (Direct Cost: ¥3,200,000、Indirect Cost: ¥960,000); Fiscal Year 2023: ¥1,170,000 (Direct Cost: ¥900,000、Indirect Cost: ¥270,000); Fiscal Year 2022: ¥1,040,000 (Direct Cost: ¥800,000、Indirect Cost: ¥240,000); Fiscal Year 2021: ¥1,950,000 (Direct Cost: ¥1,500,000、Indirect Cost: ¥450,000)
• Keywords: AI / データベース研究 / 健康診断 / 心血管疾患 / 予測モデル / 健診データ

Outline of Research at the Start

超少子高齢化社会への突入と医療の高度化・高額化に伴い、わが国の医療財政は逼迫しており、疾患の予防および早期発見による医療費の抑制が急務となっている。我々は、統計学的手法を用いて日本人における心血管リスクの知見を報告してきた。しかし、統計手法は網羅的、複合的分析に弱いこと、結果を個人のリスク評価に適用することが難しいことなどの限界があった。本研究では43年間にわたる70万人の健診データをもとにして、個人単位での疾患の早期発見・早期治療に役立つ心血管疾患発症予測の人工知能（AI）モデルを構築・実装する。健康寿命の延伸と医療費の抑制を実証し、世界に先駆けた新たな医療モデルを実現する。

Outline of Final Research Achievements

Visualization of health checkup data and preprocessing to AI application were performed, and LAMP and machine learning algorithms were applied to the data to create models. Hyperparameters were tuned and optimized for several machine learning models (random forest, XGBoosting, logistic regression, neural network, support vector machine, and others). Models were created for hypertension, chronic kidney disease, metabolic syndrome, and atherosclerosis (high baPWV) as outcomes. The results of the research were presented at several domestic and international conferences, and the results are being published in a series of papers.

Academic Significance and Societal Importance of the Research Achievements

超少子高齢化社会への突入と医療の高度化・高額化に伴い、わが国の医療財政は逼迫しており、疾患の予防および早期発見による医療費の抑制が急務である。我々は43年間にわたる70万人の健診データをもとにして、個人単位での疾患の早期発見・早期治療に役立つ心血管疾患発症予測の人工知能（AI）モデルを構築・実装した。さらにそのアルゴリズムを保健指導の場で実際に指導に役立てるべく、アプリ化を現在進めている。健康寿命の延伸と医療費の抑制に寄与する、世界に先駆けた新たな医療モデルであるものと考えている。

Research Products

• [Presentation] J-shaped association between serum uric acid levels and chronic kidney disease: a cross-sectional study using large health examination data. ESC congress, Amsterdam, 2023. (2023)
  • Author(s): Kawasoe S, Kubozono T, Salim AA, Ojima S, Kawabata T, Ikeda Y, Miyahara H, Tokushige K, Ohishi M.
  • Organizer: ESC congress, Amsterdam, 2023.
  • Int'l Joint Research
• [Presentation] Ability of machine learning algorithms to predict stage 2 hypertension using health examination data from the Japanese general population. (2023)
  • Author(s): Kawasoe S, Kubozono T, Ojima S, Kawabata T, Miyahara H, Tokushige K, Ohishi M.
  • Organizer: AHA Scientific Sessions, Boston, USA, 2023
  • Int'l Joint Research
• [Presentation] Development of a risk prediction equation for chronic kidney disease using Japanese health checkup data (2022)
  • Author(s): 川添晋
  • Organizer: 日本循環器学会総会
• [Presentation] 健診データを用いた機械学習アルゴリズムによる高血圧の発症予測 (2022)
  • Author(s): 川添晋
  • Organizer: 日本心臓病学会
• [Presentation] Ability of machine learning algorithms to predict chronic kidney disease using health examination data from the Japanese general population (2022)
  • Author(s): Kawasoe Shin
  • Organizer: American Heart Association
  • Int'l Joint Research
• [Presentation] Ability of machine learning algorithms to predict Hypertension using health examination data from the Japanese general population (2022)
  • Author(s): Kawasoe Shin
  • Organizer: European Society of Cardiology
  • Int'l Joint Research