WEKO3
アイテム
Automatic organ detection for radiotherapy planning by standard human body map
https://repo.qst.go.jp/records/68289
https://repo.qst.go.jp/records/68289da2f6852-30ee-4643-89cd-8c6a540e24e7
| Item type | 会議発表用資料 / Presentation(1) | |||||
|---|---|---|---|---|---|---|
| 公開日 | 2005-12-19 | |||||
| タイトル | ||||||
| タイトル | Automatic organ detection for radiotherapy planning by standard human body map | |||||
| 言語 | ||||||
| 言語 | eng | |||||
| 資源タイプ | ||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_c94f | |||||
| 資源タイプ | conference object | |||||
| アクセス権 | ||||||
| アクセス権 | metadata only access | |||||
| アクセス権URI | http://purl.org/coar/access_right/c_14cb | |||||
| 著者 |
Tsukamoto, Nobuhiro
× Tsukamoto, Nobuhiro× Ando, Yutaka× Kitamura, Masayuki× Kawaguchi, Osamu× et.al× 塚本 信宏× 安藤 裕× 川口 修 |
|||||
| 抄録 | ||||||
| 内容記述タイプ | Abstract | |||||
| 内容記述 | 1. Introduction Recently Radiotherapy is increasing the precision of tumor localization and dose delivery. In image-guided radiotherapy, images of the target are collected at the time of treatment and compared with treatment plan to calculate how to move the patient to bring both images into alignment. In this way, safety margins can be reduced and fewer complication expected. The purpose of this study is development of automatic organ recognition methods using pixel-based standard human body map for the image-guided radiotherapy. 2. Methods Pelvic axial CT scan (10mm thickness, 10mm interval) of four males with prostatic cancer were used. Each patient performed CT scan again two weeks after the first CT scan. All images were collected for ordinary radiotherapy planning. Pixel-based standard human body map were constructed two planes: usual CT scan image plane which contains typical Hounsfield Units. and code plane which contains ICD-O code corresponding the anatomical feature. First of all, pixel-based standard human body map were made. Radiologists draw manually the outline in the internal organs of the another patient's CT scan image taken beforehand. The pixels in that were filled with the organ's ICD-O code. Organ-detection software was developed in java. Organ detection were proceeding as the following three steps : (1) Anatomical Normalization using pixel value of target CT image (2) Comparison with standard body map (3) Determination organ by reading the corresponding pixel's ICD-O code on the body map. Thus, it can know what kind of which pixel is internal organs. Radiologists verified positional accuracy of automatically detected contours of organs. 3. Results It takes to normalize the images and detect organ for in each study about few seconds in average, and most processes were performed automatically. It is sufficiently brief to perform in daily radiotherapy. The gap of the automatically detected outline of prostate and manual outline was two or three millimeter at most, and 85-100% overlapped with the area. It is enough in positional accuracy, even if safety margin is set to five millimeter. |
|||||
| 会議概要（会議名, 開催地, 会期, 主催者等） | ||||||
| 内容記述タイプ | Other | |||||
| 内容記述 | 19th International Congress and Exhibition of Computer Assisted Radiology and Surgery (CARS2005) | |||||
| 発表年月日 | ||||||
| 日付 | 2005-06-25 | |||||
| 日付タイプ | Issued | |||||
