Development of radiation therapy support system based on radiomics

2021 Fiscal Year Final Research Report

• Project/Area Number: 18K15604
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 52040:Radiological sciences-related
• Research Institution: Komazawa University
• Principal Investigator: Magome Taiki 駒澤大学, 医療健康科学部, 准教授 (60725977)
• Project Period (FY): 2018-04-01 – 2022-03-31
• Keywords: 機械学習 / 深層学習 / 放射線治療 / 個別化治療 / 最適化 / Radiomics

Outline of Final Research Achievements

The purpose of this study was to develop a radiotherapy support system that combines image features calculated from medical images and artificial intelligence technology to provide various types of support in the field of radiotherapy. In the current radiotherapy, the same prescribed dose is often administered to all patients once the stage of disease is determined. However, there is a possibility to determine the optimal prescribed dose for each patient by integrating and analyzing medical images and other various patient information. In this study, we developed an image generation method to calculate robust image features and a prediction method for patient prognosis based on radiomics and artificial intelligence techniques for various diseases.

Free Research Field

医学物理

Academic Significance and Societal Importance of the Research Achievements

現在の放射線治療では、病期等が決まれば全ての患者に同一の処方線量が投与される場合が多い。しかし、医用画像やその他の様々な患者情報を統合して解析することで、患者個別に最適な処方線量等を決定できる可能性がある。本研究の成果は、個別化医療実現のための支援システムとして利用できる可能性がある。また、提案システムを用いることで、実際の臨床試験を行う前に試験の成功率を予測できる可能性がある。