文献速递消化道肿瘤放疗

基于新辅助放化疗后淋巴结退缩分级的局部进展期直肠癌影像组学模型的开发与验证

Development and Validation of a Radiomics Model Based on Lymph-Node Regression Grading After Neoadjuvant Chemoradiotherapy in Locally Advanced Rectal Cancer.
PubMed 文献 · Int J Radiat Oncol Biol Phys · 2023-05-24

本研究纳入78例临床分期T3-T4、N1-2、M0且接受长程新辅助放疗后手术的直肠腺癌患者,病理评估243枚淋巴结。从新辅助放化疗前高分辨率T2WI提取3641个影像组学特征构建模型,影像组学标签在训练队列AUC为0.908、验证队列AUC为0.865,联合淋巴结形态特征的列线图训练队列AUC为0.925、验证队列AUC为0.918。结论认为基于淋巴结的影像组学模型可有效预测局部进展期直肠癌新辅助放化疗后淋巴结治疗反应。

要点

  • 基于新辅助放化疗前MRI淋巴结的影像组学标签可预测局部进展期直肠癌淋巴结治疗反应。
  • 联合影像组学标签与淋巴结短轴直径、边缘轮廓的列线图在训练队列AUC为0.925、验证队列AUC为0.918。
  • 决策曲线分析证实列线图具有最高的临床效用。

核心结论

  • 基于淋巴结的影像组学模型可有效预测局部进展期直肠癌患者新辅助放化疗后淋巴结的治疗反应,可能有助于个体化治疗计划并指导观察等待策略的实施。

研究设计与样本

  • 研究纳入78例临床分期T3-T4、N1-2、M0且接受长程新辅助放疗后手术的直肠腺癌患者。
  • 病理科医师评估243枚淋巴结,其中173枚和70枚分别归入训练队列和验证队列。
  • 每枚淋巴结从新辅助放化疗前高分辨率T2WI磁共振成像的兴趣区提取3641个影像组学特征。
  • 采用最小绝对收缩和选择算子回归模型进行特征选择和影像组学标签构建。
  • 基于多因素logistic分析构建联合影像组学标签与所选淋巴结形态特征的预测模型,并绘制列线图进行可视化。

主要发现

  • 影像组学标签由5个所选特征组成,在训练队列中有效判别(AUC为0.908,95%CI为0.857%-0.958%),在验证队列中有效判别(AUC为0.865,95%CI为0.757%-0.973%)。
  • 由影像组学标签和淋巴结形态特征(短轴直径和边缘轮廓)组成的列线图在训练队列(AUC为0.925,95%CI为0.880%-0.969%)和验证队列(AUC为0.918,95%CI为0.854%-0.983%)中显示出更好的校准和判别能力。
  • 决策曲线分析证实列线图具有最高的临床效用。

临床怎么用

  • 作者认为该基于淋巴结的影像组学模型可有效预测局部进展期直肠癌患者新辅助放化疗后淋巴结的治疗反应,可能有助于个体化治疗计划并指导观察等待策略的实施。

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: The response to neoadjuvant chemoradiotherapy (nCRT) varies among patients with locally advanced rectal cancer (LARC), and the treatment response of lymph nodes (LNs) to nCRT is critical in implementing a watch-and-wait strategy. A robust predictive model may help personalize treatment plans to increase the chance that patients achieve a complete response. This study investigated whether radiomics features based on prenCRT magnetic resonance imaging nodes could predict treatment response in preoperative LARC LNs. METHODS AND MATERIALS: The study included 78 patients with clinical stage T3-T4, N1-2, and M0 rectal adenocarcinoma who received long-course neoadjuvant radiotherapy before surgery. Pathologists evaluated 243 LNs, of which 173 and 70 were assigned to training and validation cohorts, respectively. For each LN, 3641 radiomics features were extracted from the region of interest in high-resolution T2WI magnetic resonance imaging before nCRT. The least absolute shrinkage and selection operator regression model was used for feature selection and radiomics signature building. A prediction model based on multivariate logistic analysis, combining radiomics signature and selected LN morphologic characteristics, was developed and visualized by drawing a nomogram. The model's performance was assessed by receiver operating characteristic curve analysis and calibration curves. RESULTS: The radiomics signature consists of 5 selected features that were effectively discriminated within the training cohort (area under the curve [AUC], 0.908; 95% CI, 0.857%-0.958%) and the validation cohort (AUC, 0.865; 95% CI, 0.757%-0.973%). The nomogram, which consisted of radiomics signature and LN morphologic characteristics (short-axis diameter and border contours), showed better calibration and discrimination in the training and validation cohorts (AUC, 0.925; 95% CI, 0.880%-0.969% and AUC, 0.918; 95% CI, 0.854%-0.983%, respectively). The decision curve analysis confirmed that the nomogram had the highest clinical utility. CONCLUSIONS: The nodal-based radiomics model effectively predicts LNs treatment response in patients with LARC after nCRT, which could help personalize treatment plans and guide the implementation of the watch-and-wait approach in these patients.

来源:Int J Radiat Oncol Biol Phys,查看原文。版权归原作者所有。

引用

  1. Zhang S, Tang B, Yu M, et al. Development and Validation of a Radiomics Model Based on Lymph-Node Regression Grading After Neoadjuvant Chemoradiotherapy in Locally Advanced Rectal Cancer. Int J Radiat Oncol Biol Phys. 2023. PMID: 37230433. doi:10.1016/j.ijrobp.2023.05.027

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。