文献速递腹部CT

基于双期增强CT的列线图预测局部进展期胃癌新辅助化疗反应

A Dual-phase Enhanced CT-based Nomogram for Predicting the Response to Neoadjuvant Chemotherapy in Locally Advanced Gastric Cancer.
PubMed 文献 · Acad Radiol · 2025-11-19

本研究回顾性纳入143例局部进展期胃癌患者,从动脉期和静脉期增强CT提取影像组学特征构建AP、VP及AP_VP联合模型,并结合临床病理特征建立列线图。列线图预测新辅助化疗反应的训练集和验证集AUC分别为0.845和0.829,优于临床模型。

要点

  • 双期增强CT影像组学联合临床病理特征的列线图可预测局部进展期胃癌新辅助化疗反应。
  • 列线图训练集AUC 0.845(95%CI 0.767-0.924),验证集AUC 0.829(95%CI 0.697-0.96)。
  • 列线图预测效能显著优于基于cT分期和分化的临床模型(DeLong检验P<0.05)。

核心结论

  • 整合双期增强CT影像组学特征与临床病理特征的列线图在预测局部进展期胃癌新辅助化疗反应方面表现出满意性能,有助于个体化治疗。

研究设计与样本

  • 回顾性研究,纳入2018年1月至2024年3月的143例局部进展期胃癌患者,从动脉期和静脉期CT图像提取影像组学特征,构建AP、VP和AP_VP三个影像组学模型,并经单因素和多因素logistic回归筛选临床病理特征后建立列线图。

诊断标准与关键数值

  • AP_VP影像组学模型在训练集AUC为0.810(95%CI 0.721-0.899),验证集AUC为0.745(95%CI 0.588-0.903)。
  • 临床模型由cT分期和分化构建,训练集AUC为0.723(95%CI 0.619-0.827),验证集AUC为0.754(95%CI 0.591-0.916)。
  • 列线图训练集AUC为0.845(95%CI 0.767-0.924),验证集AUC为0.829(95%CI 0.697-0.96),显著优于临床模型(DeLong检验P<0.05)。

临床应用与阅片要点

  • 可基于动脉期和静脉期增强CT提取的影像组学特征联合cT分期与分化构建列线图,用于预测局部进展期胃癌新辅助化疗反应。
  • 列线图预测效能显著优于仅基于cT分期和分化的临床模型(DeLong检验P<0.05)。

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)。

英文原文(摘要)· 点击展开

RATIONALE AND OBJECTIVES: Early prediction of response to neoadjuvant chemotherapy (NAC) in patients with locally advanced gastric cancer (LAGC) helps guide treatment decisions and optimize treatment. This study aimed to establish a radiomics nomogram for predicting NAC response in LAGC patients using dual-phase contrast-enhanced CT (CECT) images.

MATERIALS AND METHODS: This retrospective study recruited 143 patients with LAGC from January 2018 to March 2024. Radiomics features were extracted from arterial phase (AP) and venous phase (VP) CT images, and were used to develop three radiomics models: AP, VP, and a combined AP_VP model. Clinicopathological characteristics were selected via univariate and multivariate logistic regression. A nomogram was then constructed by integrating the AP_VP radiomics signature with clinicopathological characteristics. The predictive performance of the model was evaluated using receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and calibration curves. The clinical utility and benefits were further quantified by comparing the nomogram to the clinical model using the net reclassification index (NRI) and integrated discrimination improvement (IDI). Stratified analyses were conducted to explore the model's performance across different patient subgroups.

RESULTS: The AP_VP model performed well in the radiomics models, with AUCs of 0.810 (95% CI, 0.721-0.899) and 0.745 (95% CI, 0.588-0.903) in the training and validation cohorts, respectively. The clinical model was constructed by cT stage and differentiation, with AUCs of 0.723 (95% CI, 0.619-0.827) and 0.754 (95% CI, 0.591-0.916). The nomogram combining radiomics and clinicopathological characteristics achieved AUCs of 0.845 (95% CI, 0.767-0.924) and 0.829 (95% CI, 0.697-0.96), significantly outperforming the clinical model (DeLong test p < 0.05).

CONCLUSION: The nomogram, incorporating dual-phase enhanced CT and clinicopathological characteristics, demonstrated satisfactory performance in predicting NAC response in LAGC patients, assisting with individualized treatment.

来源:Acad Radiol,查看原文。版权归原作者所有。

引用

  1. Xiao Z, Chen S, Wang H, et al. A Dual-phase Enhanced CT-based Nomogram for Predicting the Response to Neoadjuvant Chemotherapy in Locally Advanced Gastric Cancer. Acad Radiol. 2025. PMID: 41260957. doi:10.1016/j.acra.2025.10.060

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。