文献速递腹部技术与安全CT

胃癌侵袭性可解释CT评分模型的开发与验证

Development and validation of an interpretable CT-based scoring model for gastric cancer aggressiveness.
PubMed 文献 · Abdom Radiol (NY) · 2026-03-10

这项双中心回顾性研究纳入2014至2023年间接受根治性胃切除术的1164例胃癌患者,基于四个语义CT特征(临床淋巴结状态、最长径、肿瘤厚度、浆膜状态)构建机器学习模型预测不良组织病理学状态,XGBoost模型在训练、测试和验证队列的AUC分别为0.803、0.848和0.764,衍生的I-score将患者分为低风险(43%)和高风险(57%)组,高风险组1000天总生存率显著更低(55.7% vs 73.8%,p<0.001)。

要点

  • 基于cN、最长径、肿瘤厚度、浆膜状态四个语义CT特征的XGBoost模型预测胃癌不良组织病理学状态,验证队列AUC为0.764。
  • I-score将患者分为低风险(43%)和高风险(57%)组,高风险组1000天总生存率为55.7%,低于低风险组的73.8%(p<0.001)。
  • 多因素Cox分析显示I-score为独立预后因素(HR=1.02,p=0.004)。

核心结论

  • 基于语义影像特征的简化CT机器学习模型对胃癌不良组织病理学状态具有较高预测性能,可解释的I-score可实现有效的术前风险分层和个体化治疗规划。

研究设计与样本

  • 双中心回顾性研究,纳入2014至2023年间接受根治性胃切除术的1164例胃癌患者(中位年龄62岁;男847例,女317例),分为训练队列(n=396)、测试队列(n=618)和验证队列(n=150);由放射科医师独立评估四个语义CT特征(cN、最长径、肿瘤厚度、浆膜状态),并从动脉期和门静脉期感兴趣体积提取影像组学特征;训练逻辑回归、支持向量机、随机森林和极端梯度提升(XGBoost)模型预测不良组织病理学状态,以最佳模型构建I-score并在独立队列中验证预后评估。

关键影像表现/征象

  • 评估的四个语义CT特征为临床淋巴结状态(cN)、最长径(LD)、肿瘤厚度(TT)和浆膜状态。
  • 采用语义特征的XGBoost模型预测不良组织病理学状态性能最高,训练、测试和验证队列AUC分别为0.803、0.848和0.764。

诊断标准与关键数值

  • XGBoost模型(语义特征)预测不良组织病理学状态的AUC在训练队列为0.803、测试队列为0.848、验证队列为0.764。
  • I-score将患者分为低风险组(43%)和高风险组(57%)。
  • 高风险组1000天总生存率为55.7%,低风险组为73.8%(p<0.001)。
  • 多因素Cox分析中I-score为独立预后因素(HR=1.02,p=0.004)。

临床应用与阅片要点

  • 作者认为,可解释的I-score可实现有效的术前风险分层和个体化治疗规划。

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

OBJECTIVES: Accurate prediction of adverse histopathological status (AHS) in gastric cancer (GC) is clinically crucial due to its strong association with poor prognosis. This study aims to develop and validate a CT-based machine learning model for AHS prediction, and establish an interpretable imaging score (I-score) for prognostic stratification.

METHODS: In this dual-center retrospective study, 1164 GC patients undergoing radical gastrectomy between 2014 and 2023 were included. Four semantic CT features, including clinical lymph node status (cN), longest diameter (LD), tumor thickness (TT), and serosal status, were independently evaluated by radiologists. Radiomic features were extracted from arterial and portal venous phase volumes of interest. Logistic regression, support vector machine, random forest, and extreme gradient boosting (XGBoost) models were trained to predict AHS. The best-performing model was used to construct the I-score, which was validated in an independent cohort for prognostic assessment.

RESULTS: Among the 1164 patients (median age: 62 years old; 847 men, 317 women), data were divided into training (n = 396), test (n = 618), and validation (n = 150) cohorts. The XGBoost model using semantic features achieved the highest predictive performance (AUC: 0.803, 0.848 and 0.764, respectively). The derived I-score stratified patients into low-risk (43%) and high-risk (57%) groups, with significantly poorer 1000-day overall survival in the high-risk group (55.7% vs. 73.8%, p < 0.001). Multivariate Cox analysis confirmed the I-score as an independent prognostic factor (HR = 1.02, p = 0.004).

CONCLUSION: The simplified CT-based machine learning model using semantic imaging features achieved high predictive performance for AHS in GC. The interpretable I-score enables effective preoperative risk stratification and individualized treatment planning.

来源:Abdom Radiol (NY),查看原文。版权归原作者所有。

引用

  1. Zhang YQ, Zhang J, Jin YY, et al. Development and validation of an interpretable CT-based scoring model for gastric cancer aggressiveness. Abdom Radiol (NY). 2026. PMID: 41803545. doi:10.1007/s00261-026-05458-5

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。