Suppr超能文献

基于图谱的分割提高了勾画术后子宫内膜癌淋巴结体积的一致性并减少了所需时间。

Atlas-based segmentation improves consistency and decreases time required for contouring postoperative endometrial cancer nodal volumes.

机构信息

Department of Radiation Oncology, Beth Israel Medical Center, New York, NY 10019, USA.

出版信息

Int J Radiat Oncol Biol Phys. 2011 Mar 1;79(3):943-7. doi: 10.1016/j.ijrobp.2010.04.063.

Abstract

PURPOSE

Accurate target delineation of the nodal volumes is essential for three-dimensional conformal and intensity-modulated radiotherapy planning for endometrial cancer adjuvant therapy. We hypothesized that atlas-based segmentation ("autocontouring") would lead to time savings and more consistent contours among physicians.

METHODS AND MATERIALS

A reference anatomy atlas was constructed using the data from 15 postoperative endometrial cancer patients by contouring the pelvic nodal clinical target volume on the simulation computed tomography scan according to the Radiation Therapy Oncology Group 0418 trial using commercially available software. On the simulation computed tomography scans from 10 additional endometrial cancer patients, the nodal clinical target volume autocontours were generated. Three radiation oncologists corrected the autocontours and delineated the manual nodal contours under timed conditions while unaware of the other contours. The time difference was determined, and the overlap of the contours was calculated using Dice's coefficient.

RESULTS

For all physicians, manual contouring of the pelvic nodal target volumes and editing the autocontours required a mean±standard deviation of 32±9 vs. 23±7 minutes, respectively (p=.000001), a 26% time savings. For each physician, the time required to delineate the manual contours vs. correcting the autocontours was 30±3 vs. 21±5 min (p=.003), 39±12 vs. 30±5 min (p=.055), and 29±5 vs. 20±5 min (p=.0002). The mean overlap increased from manual contouring (0.77) to correcting the autocontours (0.79; p=.038).

CONCLUSION

The results of our study have shown that autocontouring leads to increased consistency and time savings when contouring the nodal target volumes for adjuvant treatment of endometrial cancer, although the autocontours still required careful editing to ensure that the lymph nodes at risk of recurrence are properly included in the target volume.

摘要

目的

准确勾画淋巴结靶区对于子宫内膜癌辅助治疗的三维适形调强放疗计划至关重要。我们假设基于图谱的分割(“自动勾画”)将节省时间,并使医生之间的勾画更为一致。

方法与材料

使用 15 例术后子宫内膜癌患者的数据构建参考解剖图谱,根据放射治疗肿瘤学组 0418 试验,在模拟 CT 扫描上使用商业软件勾画盆腔淋巴结临床靶区。在另外 10 例子宫内膜癌患者的模拟 CT 扫描上生成了淋巴结临床靶区自动勾画。3 位放射肿瘤学家在计时条件下纠正自动勾画并手动勾画淋巴结轮廓,同时不了解其他轮廓。记录时间差,并使用 Dice 系数计算轮廓重叠。

结果

对于所有医生,手动勾画盆腔淋巴结靶区和编辑自动勾画分别需要 32±9 分钟和 23±7 分钟(p<.000001),可节省 26%的时间。对于每位医生,手动勾画轮廓与纠正自动勾画的时间分别为 30±3 分钟和 21±5 分钟(p=.003)、39±12 分钟和 30±5 分钟(p=.055)以及 29±5 分钟和 20±5 分钟(p=.0002)。从手动勾画(0.77)到纠正自动勾画(0.79;p=.038),平均重叠度增加。

结论

本研究结果表明,自动勾画可提高勾画淋巴结靶区的一致性和节省时间,尽管自动勾画仍需要仔细编辑,以确保将复发风险的淋巴结正确包含在靶区内。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验