2023 Fiscal Year Final Research Report
Prediction of uterus volume shrinkage for cervical cancer patients during radiotherapy using machine-learning approach with treatment planning-CT radiomic features
| Project/Area Number |
18K15569
|
|---|
| Research Category |
Grant-in-Aid for Early-Career Scientists
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
Basic Section 52040:Radiological sciences-related
|
|---|
| Research Institution | Kitasato University (2020-2023)
Japanese Foundation for Cancer Research (2018-2019) |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2018-04-01 – 2024-03-31
|
| Keywords | 子宮頸がん / Radiomics / 適応放射線治療 / 画像特徴量 / 機械学習 / 臨床標的体積 / 医学物理学 |
|---|
| Outline of Final Research Achievements |
In some patients of cervical cancer, uterus commonly shows large shrinkage during concurrent chemo-radiotherapy (CCRT), and shrinkage prediction might be a clue of adaptive planning of radiotherapy (RT) and adoption of intensity-modulated radiotherapy (IMRT). This study aimed to predict uterus volume shrinkage for one month using machine-learning (ML) approach with radiomic features extracted from treatment planning-CT (pCT) for external beam radiotherapy (EBRT).
Logistic regression with L1-regularization and L2-regularization terms and a support vector machine were used to predict uterine volume shrinkage, and the model using the support vector machine showed the best results (accuracy 87.2%, AUC value 0.914). The results suggested that a machine learning approach using CT image features can predict uterine volume reduction and may be useful for future applications of intensity-modulated radiation therapy and modification of adaptive treatment plans.
|
| Free Research Field |
医学物理学
|
| Academic Significance and Societal Importance of the Research Achievements |
局所進行子宮頸がんの同時化学放射線療法における放射線治療の期間中に、primary CTVである子宮頸部および体部の体積が大幅に縮小する場合が多いが、一方で体積がそれほど変化しない症例もある程度存在する。縮小／非縮小を事前に予測できれば、放射線治療のアダプティブな治療計画修正や強度変調放射線治療の適用、また線量増加もしくは減少の手がかりとなる可能性がある。本研究で提案する手法は、外部照射放射線治療で必ず取得する治療計画CT画像から画像特徴量を抽出し、患者に追加で放射線被曝をさせることなく特徴量を取得し、機械学習アプローチにより1ヶ月後の子宮体積収縮を予測することが可能であることを示した。
|
