• Search Research Projects
  • Search Researchers
  • How to Use
  1. Back to project page

2022 Fiscal Year Final Research Report

Automatic treatment planning based on patient-specific dose distribution using deep learning

Research Project

  • PDF
Project/Area Number 20K16742
Research Category

Grant-in-Aid for Early-Career Scientists

Allocation TypeMulti-year Fund
Review Section Basic Section 52040:Radiological sciences-related
Research InstitutionKansai Medical University

Principal Investigator

KOIKE Yuhei  関西医科大学, 医学部, 助教 (90866154)

Project Period (FY) 2020-04-01 – 2023-03-31
Keywords医学物理 / 高精度放射線治療 / 個別化治療 / 人工知能 / 深層学習 / 機械学習 / 適応放射線治療
Outline of Final Research Achievements

By applying artificial intelligence, we have developed a method for estimating patient-specific dose distribution in radiotherapy based on CT images and individual patient anatomical information. In this study, a new loss function based on each input contour was introduced for prostate cancer cases, and a prediction model with higher accuracy than the conventional method was successfully constructed. We believe that the automatic treatment planning method based on deep learning developed in this work will contribute to a new technology for adaptive radiotherapy that can cope with daily changes in body shape and organ location during radiotherapy.

Free Research Field

医学物理学

Academic Significance and Societal Importance of the Research Achievements

人工知能の回帰問題では損失関数として一般的に平均二乗誤差が用いられているが、本研究では、入力する臓器の輪郭内の誤差が最小となるような損失関数を新たに導入することで精度の高い線量分布を作成した。開発した方法は、従来の線量体積ヒストグラムに基づく最適化で考慮されていなかった空間情報を加味した最適化法である。本研究成果の応用が期待される適応放射線治療において、その時の体型や臓器の位置に応じた線量分布が短時間で推定可能となる。治療効果の向上、有害事象の低減が期待でき、社会的意義も大きい。

URL: 

Published: 2024-01-30  

Information User Guide FAQ News Terms of Use Attribution of KAKENHI

Powered by NII kakenhi