2022 Fiscal Year Final Research Report
Development of markerless lung tumor detection based on the deep learing for real-time tumor-tracking radiotherapy.
Project/Area Number |
18K07753
|
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 | Yamaguchi University |
Principal Investigator |
|
Co-Investigator(Kenkyū-buntansha) |
藤井 文武 山口大学, 大学院創成科学研究科, 准教授 (30274179)
|
Project Period (FY) |
2018-04-01 – 2023-03-31
|
Keywords | 深層学習 / 機械学習 / 肺腫瘍認識 / 医用画像処理 / X線透視画像 / 動体追跡放射線治療 |
Outline of Final Research Achievements |
In radiotherapy, respiratory movement affects the radiation dose delivered. Therefore, we have performed the real-time tumor-tracking radiotherapy for these region, but it is needed to implant the fiducial markers near the tumor, which is highly invasive. To address this, we developed an algorithm using deep learning to directly detect lung tumors from X-ray fluoroscopy images of real-time tumor monitoring system. The proposed method can create the training data by applying four-dimensional computed tomography images, radiation treatment planning, and tumor contouring data in DICOM-RT acquired during the radiotherapy process, making it possible to create individual learning model for each patient.
|
Free Research Field |
医学物理
|
Academic Significance and Societal Importance of the Research Achievements |
本研究で開発したアルゴリズムは,医用画像処理や深層学習を用いることで,従来の手法では困難であった肺腫瘍認識を高精度に実施することを可能とした.放射線治療を行う過程で得られる情報(医用画像情報,治療計画情報,腫瘍輪郭情報)を用いることで,腫瘍認識モデルを構築することができるため,患者個別に対応した予測モデルを構築可能である.また,現在既存の放射線治療機器においても導入可能である.そのため,呼吸により動く腫瘍に対する放射線治療をより高精度に行うことが可能となり,局所制御の向上につながると考えられる.
|