Development of a Prognostic Method Integrating AI and Radiomics Analysis in Cerebral Ischemic Lesions

2023 Fiscal Year Final Research Report

• Project/Area Number: 21K07586
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 52040:Radiological sciences-related
• Research Institution: Hokkaido University
• Principal Investigator: Sugimori Hiroyuki 北海道大学, 保健科学研究院, 准教授 (20711899)
• Project Period (FY): 2021-04-01 – 2024-03-31
• Keywords: 医用画像解析 / 深層学習 / 予後予測

Outline of Final Research Achievements

This study integrated deep learning-based AI technology for detecting cerebral ischemic lesions with machine learning (ML) methods aimed at predicting treatment outcomes for patients with acute ischemic stroke (AIS) as a prognostic prediction method. Specifically, it developed a model that identifies brain infarctions from "fusion images" integrating DWI, FLAIR, and T2WI, and predicts the results and time limits of mechanical thrombectomy (MT) based on the apparent diffusion coefficient (ADC). This technology demonstrated the potential to significantly assist clinical decision-making by enabling accurate diagnosis and prognosis prediction through the interaction of AI and ML.

Free Research Field

医用画像解析

Academic Significance and Societal Importance of the Research Achievements

本研究では、脳梗塞を高精度で検出し治療効果を予測するために、AIと機械学習技術を統合した新しい手法を開発しました。特に、複数の画像データを組み合わせた「fusion画像」を用いて脳梗塞の位置を特定し、治療後の結果と必要な時間を予測するモデルを構築した。この技術は、脳卒中治療の計画と評価に役立ち、迅速かつ正確な診断支援が可能となり、医療現場における意思決定を支援することができる可能性がある。