Recurrent prediction within 1, 3, and 5 years after acute ischemic stroke based on machine learning using 10 years J-ASPECT study
Authors:
Open Access
- 1 January 2025
- journal article
- Published by Japan Stroke Society in Japanese Journal of Stroke
- Vol. 47 (1) , 17-24
Abstract
【背景および目的】本研究の目的は急性期脳梗塞発症患者の退院後1,3,5年以内の再発を予測することである.【方法】対象は2010–2019年度にJ-ASPECT Study参加施設に急性期脳梗塞で入院した患者とした.117個の臨床情報を変数として機械学習モデルを作成し,Stroke Prognosis Instrument II (SPI-II),Essen Stroke Risk Score(ESRS)と比較した.【結果】機械学習による再発予測の平均受信者操作特性曲線のArea Under the Curve(ROC AUC)は1,3,5年以内でそれぞれ0.62,0.63,0.63となり,SPI-IIは0.54,0.54,0.54,ESRSは0.55,0.54,0.53となった.【結論】機械学習を用いた脳梗塞再発予測は既存のリスクスコアよりも高い予測精度が得られた.Keywords
Funding Information
- Japan Agency for Medical Research and Development (JP22ek0210147)
- Idorsia Pharmaceuticals Japan (#D602)
- Idorsia Pharmaceuticals Japan (#D712)
This publication has 12 references indexed in Scilit:
- One‐Year Disability Trajectories and Long‐Term Cardiovascular Events, Recurrent Stroke, and Mortality After Ischemic StrokeJournal of the American Heart Association, 2024
- Stroke health management: Novel strategies for the prevention of recurrent ischemic strokeFrontiers in Neurology, 2022
- The number of risk factors increases the recurrence events in ischemic strokeEuropean Journal of Medical Research, 2022
- Prediction of Recurrent Ischemic Stroke Using Registry Data and Machine Learning Methods: The Erlangen Stroke RegistryStroke, 2022
- Optuna2019
- Three-year survival and recurrence after first-ever stroke: the Joinville stroke registryBMC Neurology, 2015
- Effects of Comprehensive Stroke Care Capabilities on In-Hospital Mortality of Patients with Ischemic and Hemorrhagic Stroke: J-ASPECT StudyPLOS ONE, 2014
- Class Imbalance, Redux2011
- The Essen Stroke Risk Score Predicts Recurrent Cardiovascular EventsStroke, 2009
- The Stroke Prognosis Instrument II (SPI-II)Stroke, 2000