天津医药 ›› 2026, Vol. 54 ›› Issue (6): 637-642.doi: 10.11958/20253436

• 应用研究 • 上一篇    下一篇

基于超声内镜与增强CT构建胃肠道间质瘤恶性潜能的术前评估模型

康晓娜(), 刘亚萍, 冯云, 罗玉梅, 贾皑()   

  1. 西安交通大学第一附属医院消化内科 (邮编710048)
  • 收稿日期:2025-11-27 修回日期:2026-02-25 出版日期:2026-06-15 发布日期:2026-06-15
  • 通讯作者: E-mail:JiaAi2663@163.com
  • 作者简介:康晓娜(1982),女,主管护师,主要从事内镜操作技术的护理配合方面研究。E-mail:kang123xiaona@163.com
  • 基金资助:
    陕西省卫生健康委卫生健康科研项目(2022D441)

Development of a preoperative assessment model for the malignant potential in gastrointestinal stromal tumor based on endoscopic ultrasound and contrast-enhanced computed tomography

KANG Xiaona(), LIU Yaping, FENG Yun, LUO Yumei, JIA Ai()   

  1. Department of Gastroenterology, First Affiliated Hospital of Xi'an Jiaotong University, Xi'an 710048, China
  • Received:2025-11-27 Revised:2026-02-25 Published:2026-06-15 Online:2026-06-15
  • Contact: E-mail:JiaAi2663@163.com

摘要:

目的 探讨超声内镜(EUS)与增强CT对胃肠道间质瘤(GIST)恶性潜能的术前评估价值,构建GIST恶性潜能的术前评估模型并验证模型效能。方法 收集191例GIST患者的临床与病理资料,分析所有患者术前EUS及增强CT的影像特征,并根据美国国立卫生研究院(NIH)分类系统进行分组,将极低危级/低危级归为低恶性潜能组(145例),将中危级/高危级归为高恶性潜能组(46例)。比较2组临床资料、EUS及增强CT影像特征;多因素Logistic回归分析GIST患者高恶性潜能的影响因素,构建GIST患者高恶性潜能的风险评估模型,采用受试者工作特征(ROC)曲线及校准曲线对该模型进行验证。结果 与低恶性潜能组相比,高恶性潜能组患者EUS测得的肿瘤直径更长,肿瘤边界不清晰占比更高,增强CT影像中形态不规则、不均匀强化的占比更高(P<0.05)。多因素Logistic回归分析显示,肿瘤直径较长、EUS肿瘤边界不清晰,增强CT肿瘤形态不规则、不均匀强化是GIST患者高恶性潜能的危险因素(P<0.05)。基于影响因素构建评估模型,ROC曲线结果显示,该模型评估患者高恶性潜能的曲线下面积(AUC)为0.880(95%CI:0.826~0.923),敏感度82.61%、特异度90.34%。Bootstrap校准曲线显示该评估模型的校准曲线与实际曲线接近重合(Hosmer-Lemeshow χ2=0.396,P=0.508)。结论 基于EUS与增强CT构建的术前评估模型对于GIST患者高恶性潜能具有较高的评估效能。

关键词: 胃肠道间质肿瘤, 体层摄影术, X线计算机, 超声检查, 内窥镜检查, 恶性潜能, 评估模型

Abstract:

Objective To explore the preoperative evaluation value of endoscopic ultrasonography (EUS) and enhanced CT for the malignant potential of gastrointestinal stromal tumor (GIST), and to construct a preoperative evaluation model of the malignant potential of GIST based on EUS and CT and verify the model efficacy. Methods Clinical, imaging and histopathological data were retrospectively collected from 191 consecutive patients with pathologically confirmed GIST. Preoperative EUS and CECT image features-including tumor size, boundary definition and enhancement pattern were systematically assessed. Patients were stratified according to the classfication system of the National Institutes of Health (NIH) of the United States. Those classified as "extremely low-risk" or "low-risk" (n = 145) constituted the low-malignant-potential group, those classified as "intermediate-risk" or "high-risk" (n = 46) formed the high-malignant-potential group. The clinical data, EUS and enhanced CT imaging features of the two groups were compared. Multivariable Logistic regression was employed to identify independent predictors of high malignant potential, and a predictive model was constructed based on statistically significant variables. Model performance was evaluated using receiver operating characteristic (ROC) curve analysis and calibration assessment. Results Compared with the low-malignant-potential group, the high-malignant-potential group exhibited significantly longer tumor diameters measured by EUS, a higher prevalence of ill-defined tumor margins on EUS and a greater frequency of irregular and heterogeneous enhancement patterns on CECT(P<0.05). Multivariable Logistic regression identified EUS tumor diameter, ill-defined EUS tumor boundaries and irregular/heterogeneous CECT enhancement as independent predictors of high malignant potential (P<0.05). The result of predictive model achieved an AUC of 0.880 (95%CI: 0.826-0.923), with sensitivity of 82.6% and specificity of 90.3%. Calibration analysis demonstrated excellent agreement between predicted and observed probabilities (Hosmer-Lemeshow χ2=0.396, P=0.508). Conclusion The preoperative assessment model based on EUS and enhanced CT has a high assessment efficacy for the high malignant potential of GIST patients.

Key words: gastrointestinal stromal tumors, tomography, X-ray computed, ultrasonography, endoscopy, malignant potential, evaluation model

中图分类号: