文献速递消化道肿瘤放疗放疗物理与技术

基于深度学习的直肠癌新辅助放疗靶区全自动勾画:分而治之策略的多中心盲法随机验证研究

Towards deep-learning (DL) based fully automated target delineation for rectal cancer neoadjuvant radiotherapy using a divide-and-conquer strategy: a study with multicenter blind and randomized validation.
PubMed 文献 · Radiat Oncol · 2023-10-06

回顾性纳入141例II-III期中低位直肠癌患者,随机分为训练集121例和测试集20例,采用分而治之策略分别构建CT域CTV和MRI域GTV深度学习分割模型,并通过三级多中心盲法随机方案验证。测试集CTV的DSC为0.85±0.06、95HD为7.75±6.42 mm,96.4%病例达到临床可行评分(≥2分),图灵测试阳性率52.3%;GTV的DSC为0.87±0.07、95HD为4.07±1.67 mm,100%达到临床可行评分(≥2分),图灵测试阳性率52.0%。

要点

  • 分而治之策略分别用深度学习模型勾画CT域CTV和MRI域GTV,测试集DSC分别为0.85±0.06和0.87±0.07。
  • 测试集CTV有96.4%病例、GTV有100%病例达到临床可行评分(≥2分)。
  • 三级多中心盲法随机验证中,CTV和GTV的图灵测试阳性率分别为52.3%和52.0%。

核心结论

  • 所提出的深度学习工作流程在直肠癌新辅助放疗的CTV和GTV自动勾画中表现出有前景的准确性和出色的临床可行性。

研究设计与样本

  • 回顾性纳入141例II-III期中低位直肠癌患者,随机分为训练集121例和测试集20例。
  • 采用分而治之策略,使用两个独立的深度学习模型(后端均为DpuUnet),一个在CT域分割CTV,另一个在MRI域分割GTV。
  • 采用三级多中心盲法随机评价方案验证工作流程:第一级计算DSC和95HD,第二级进行四级专家评分,第三级进行头对头图灵测试。

主要发现

  • 测试集基于深度学习的CTV勾画:DSC为0.85±0.06,95HD为7.75±6.42 mm(均值±标准差),96.4%病例达到临床可行评分(≥2分),图灵测试阳性率为52.3%。
  • 测试集GTV勾画:DSC为0.87±0.07,95HD为4.07±1.67 mm(均值±标准差),100%的深度学习勾画达到临床可行评分(≥2分),图灵测试阳性率为52.0%。

关键数值与推荐

  • 测试集CTV的DSC为0.85±0.06,95HD为7.75±6.42 mm。
  • 测试集GTV的DSC为0.87±0.07,95HD为4.07±1.67 mm。
  • 临床可行评分定义为≥2分,CTV有96.4%病例达到,GTV有100%病例达到。
  • 图灵测试阳性率:CTV为52.3%,GTV为52.0%。

临床怎么用

  • 原文未据此给出推荐。

局限性

  • 原文未报告局限性。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

原文

PURPOSE: Manual clinical target volume (CTV) and gross tumor volume (GTV) delineation for rectal cancer neoadjuvant radiotherapy is pivotal but labor-intensive. This study aims to propose a deep learning (DL)-based workflow towards fully automated clinical target volume (CTV) and gross tumor volume (GTV) delineation for rectal cancer neoadjuvant radiotherapy. MATERIALS & METHODS: We retrospectively included 141 patients with Stage II-III mid-low rectal cancer and randomly grouped them into training (n = 121) and testing (n = 20) cohorts. We adopted a divide-and-conquer strategy to address CTV and GTV segmentation using two separate DL models with DpuUnet as backend-one model for CTV segmentation in the CT domain, and the other for GTV in the MRI domain. The workflow was validated using a three-level multicenter-involved blind and randomized evaluation scheme. Dice similarity coefficient (DSC) and 95th percentile Hausdorff distance (95HD) metrics were calculated in Level 1, four-grade expert scoring was performed in Level 2, and head-to-head Turing test in Level 3. RESULTS: For the DL-based CTV contours over the testing cohort, the DSC and 95HD (mean ± SD) were 0.85 ± 0.06 and 7.75 ± 6.42 mm respectively, and 96.4% cases achieved clinical viable scores (≥ 2). The positive rate in the Turing test was 52.3%. For GTV, the DSC and 95HD were 0.87 ± 0.07 and 4.07 ± 1.67 mm respectively, and 100% of the DL-based contours achieved clinical viable scores (≥ 2). The positive rate in the Turing test was 52.0%. CONCLUSION: The proposed DL-based workflow exhibited promising accuracy and excellent clinical viability towards automated CTV and GTV delineation for rectal cancer neoadjuvant radiotherapy.

来源:Radiat Oncol,查看原文。版权归原作者所有。

引用

  1. Geng J, Zhu X, Liu Z, et al. Towards deep-learning (DL) based fully automated target delineation for rectal cancer neoadjuvant radiotherapy using a divide-and-conquer strategy: a study with multicenter blind and randomized validation. Radiat Oncol. 2023. PMID: 37803462. doi:10.1186/s13014-023-02350-0

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。