Department of Oncology, Cambridge University Hospitals, UK; University of Nottingham, UK.
Department of Oncology, Cambridge University Hospitals, UK.
Radiother Oncol. 2020 Jun;147:22-29. doi: 10.1016/j.radonc.2020.02.017. Epub 2020 Mar 30.
EMBRACE-II is an international prospective study of IMRT and MRI-guided adaptive brachytherapy (IGABT) in locally advanced cervix cancer. An online radiotherapy quality assurance (RTQA) programme with minimal data transfer and supporting continuing medical education (CME) was implemented for IMRT contouring.
Participant contours for six volumes-of-interest (VOIs) on one benchmark case were scored (2 = excellent, 1 = fair, 0 = revision required) against a consensus reference contour. For contours receiving a 0 or 1 score, additional qualitative comments were provided. The Jaccard conformity index (JCI) was retrospectively calculated. User interaction with CME content (pre-accreditation questionnaire, contouring atlas, practice cases, quizzes, internal target volume (ITV-T) guide) was analysed.
78 clinicians submitted contours for evaluation. 41% passed at the first attempt, 44% after one revision and 6% after two or more revisions. 9% did not re-submit after failing. The lowest mean scores were for the elective nodal CTV (CTV-E) (1.01/2) and ITV-T (1.06/2). 60 different errors across the six VOIs were identified; five potentially had high impact on loco-regional control. A JCI cut-off of 0.7 would have identified 87% contours that failed expert assessment, but also excluded 54% of passing contours. 39 clinicians responded to the pre-accreditation questionnaire - 36% anticipated difficulties with the ITV-T and 13% with the CTV-E. 35% clinicians contoured on the practice cases, 17% answered a quiz, 96% used the atlas and 38% the ITV-T guide.
Expert evaluation with qualitative feedback improved contouring compliance. The JCI is not a reliable alternative to expert assessment. Moderate uptake of optional CME content limited evaluation.
EMBRACE-II 是一项关于局部晚期宫颈癌调强放疗（IMRT）和 MRI 引导自适应近距离放疗（IGABT）的国际性前瞻性研究。为了实现 IMRT 勾画，我们实施了一个在线放射治疗质量保证（RTQA）计划，该计划数据传输量最小，并支持继续医学教育（CME）。
针对一个基准病例的六个感兴趣体积（VOI），对 78 位临床医生提交的勾画轮廓进行了评分（2=优秀，1=良好，0=需要修改），评分标准与共识参考轮廓相对照。对于评分 0 或 1 的轮廓，提供了额外的定性评论。回顾性计算了 Jaccard 一致性指数（JCI）。分析了用户对 CME 内容（预认证问卷、勾画图谱、练习病例、测验、内部靶区（ITV-T）指南）的交互情况。
78 位临床医生提交了勾画轮廓进行评估。41%的人第一次尝试就通过了，44%的人在第一次修改后通过，6%的人在两次或更多次修改后通过。有 9%的人在第一次失败后没有重新提交。CTV-E 和 ITV-T 的平均得分最低（分别为 1.01/2 和 1.06/2）。在六个 VOI 中发现了 60 种不同的错误；其中 5 种可能对局部区域控制有很大影响。JCI 截断值为 0.7 可以识别出 87%的失败专家评估的轮廓，但也排除了 54%的通过轮廓。有 39 位临床医生回复了预认证问卷-36%的人预计 ITV-T 有困难，13%的人预计 CTV-E 有困难。35%的临床医生在练习病例上进行了勾画，17%的人参加了测验，96%的人使用了图谱，38%的人使用了 ITV-T 指南。
有定性反馈的专家评估提高了勾画的一致性。JCI 不能替代专家评估。可选 CME 内容的中等使用率限制了评估的范围。
