2022 Fiscal Year Final Research Report
Establishment of a method for estimating stopping power ratio in the patient's body for adaptive particle therapy
Project/Area Number |
19K17192
|
Research Category |
Grant-in-Aid for Early-Career Scientists
|
Allocation Type | Multi-year Fund |
Review Section |
Basic Section 52040:Radiological sciences-related
|
Research Institution | Tokyo Women's Medical University (2022) Yamagata University (2019-2021) |
Principal Investigator |
Kanai Takayuki 東京女子医科大学, 医学部, 講師 (40764139)
|
Project Period (FY) |
2019-04-01 – 2023-03-31
|
Keywords | 放射線治療 / 粒子線治療 / 適応放射線治療 / MR画像誘導放射線治療 / 畳み込みニューラルネットワーク |
Outline of Final Research Achievements |
In order to establish a method for estimating the stopping power ratio for "adaptive particle radiotherapy", we developed a method using convolutional neural networks. We compared U-net, CycleGAN, and pix2pix and found that pix2pix was the best in terms of accuracy and applicability. The proposed method was accurate enough to be applied to tumors in a shallow region, and conversion time was about 1.3 seconds per case, which is fast enough to be applied to the on-line dose calculation. For reducing the range uncertainty, we also developed a method to utilize a body surface image guidance system. The developed method was able to quantitatively calculate the magnitude of deformation of the patient's body surface with a high accuracy of less than 0.1 mm.
|
Free Research Field |
医学物理学
|
Academic Significance and Societal Importance of the Research Achievements |
本研究で開発した手法は、患者の体型変化や臓器の変化などに応じて粒子線治療計画を変更する「適応粒子線治療」の実現に寄与すると期待される。また、粒子線の飛程と飛程誤差が比例関係でなく2乗根に比例する関係であったことはこれまでの研究報告とは異なる結果であり、今後の粒子線治療における飛程誤差の考え方に影響を及ぼすものである。更に、本研究の研究成果は学術大会のインターナショナルセッションで学術的な賞を受賞するなど、高い評価を受けている。
|