|
Hokkaido University Collection of Scholarly and Academic Papers >
Graduate School of Medicine / Faculty of Medicine >
Peer-reviewed Journal Articles, etc >
Prediction of liver D-mean for proton beam therapy using deep learning and contour-based data augmentation
| Title: | Prediction of liver D-mean for proton beam therapy using deep learning and contour-based data augmentation |
| Authors: | Jampa-ngern, Sira Browse this author | | Kobashi, Keiji Browse this author | | Shimizu, Shinichi Browse this author | | Takao, Seishin Browse this author | | Nakazato, Keiji Browse this author | | Shirato, Hiroki Browse this author →KAKEN DB |
| Keywords: | proton beam therapy (PBT) | | liver cancer | | deep learning | | dose prediction |
| Issue Date: | Nov-2021 |
| Publisher: | Oxford University Press |
| Journal Title: | Journal of Radiation Research |
| Volume: | 62 |
| Issue: | 6 |
| Start Page: | 1120 |
| End Page: | 1129 |
| Publisher DOI: | 10.1093/jrr/rrab095 |
| Abstract: | The prediction of liver D-mean with 3-dimensional radiation treatment planning (3DRTP) is time consuming in the selection of proton beam therapy (PBT), and deep learning prediction generally requires large and tumor-specific databases. We developed a simple dose prediction tool (SDP) using deep learning and a novel contour-based data augmentation (CDA) approach and assessed its usability. We trained the SDP to predict the liver D mean immediately. Five and two computed tomography (CT) data sets of actual patients with liver cancer were used for the training and validation. Data augmentation was performed by artificially embedding 199 contours of virtual clinical target volume (CTV) into CT images for each patient. The data sets of the CTVs and OARs are labeled with liver D-mean for six different treatment plans using two-dimensional calculations assuming all tissue densities as 1.0. The test of the validated model was performed using 10 unlabeled CT data sets of actual patients. Contouring only of the liver and CTV was required as input. The mean relative error (MRE), the mean percentage error (MPE) and regression coefficient between the planned and predicted D-mean was 0.1637, 6.6%, and 0.9455, respectively. The mean time required for the inference of liver D-mean of the six different treatment plans for a patient was 4.47 +/- 0.13 seconds. We conclude that the SDP is cost-effective and usable for gross estimation of liver Dmean in the clinic although the accuracy should be improved further if we need the accuracy of liver D-mean tobe compatible with 3DRTP. |
| Type: | article |
| URI: | http://hdl.handle.net/2115/84499 |
| Appears in Collections: | 医学院・医学研究院 (Graduate School of Medicine / Faculty of Medicine) > 雑誌発表論文等 (Peer-reviewed Journal Articles, etc)
|
|
