基于小探头超声内镜下胃小间质瘤特征表现的诊断价值及预测模型研究  

Study on Diagnostic Value and Prediction Model of Characteristics of Gastric Small Stromal Tumor under Endoscopic Ultrasonography with Small Probe

作　　者：陈正彦 刘君颖[1] 杨坤 张然[1] 余海洋[1] 张红娟 刘玮[1] CHEN Zheng-yan;LIU Jun-ying;YANG Kun;ZHANG Ran;YU Hai-yang;ZHANG Hong-juan;LIU Wei(Digestive Endoscopy Center,the First Affiliated Hospital of Henan University of Chinese Medicine,Zhengzhou 450000,Henan Province,China)

机构地区：[1]河南中医药大学第一附属医院消化内镜中心,河南郑州450000

出　　处：《罕少疾病杂志》2025年第3期130-132,共3页Journal of Rare and Uncommon Diseases

摘　　要：目的 探讨基于小探头超声内镜下胃小间质瘤特征表现的诊断价值及预测模型的价值分析。方法 选择2016年1月到2022年12月之间在我院接受小探头超声内镜检查的胃粘膜下占位性病变患者193例的数据回顾性分析。依据患者的病理结果实施分组,对照组(非胃小间质瘤)以及病例组(胃小间质瘤)。描述患者的分组情况。超声内镜下诊断胃小间质瘤的单因素分析、多因素分析。超声内镜下诊断胃小间质瘤的预测模型效能分析。结果本次被调查患者人群中,病理诊断为胃小间质瘤的68人,诊断为非胃小间质瘤的125人。在诊断为非胃小间质瘤的患者人群分布中,平滑肌瘤47人,神经鞘膜瘤38人,慢性炎性病变22人,息肉8人,纤维瘤4人,血管球瘤4人,脂肪瘤2人。单因素分析结果显示,不同病理学诊断结果的患者年龄、病灶部位、起源肌层、生长方式的数据差异有统计学意义(均P<0.05)。多因素分析结果显示,年龄、病灶部位、起源肌层、生长方式为超声内镜下诊断胃小间质瘤的独立影响因素(均P<0.05)。经过验证,logistic多因素回归模型具有良好的拟合优度(Hosmer-Lemeshowχ^(2)=5.769,P>0.05)。利用年龄、病灶部位、起源肌层、生长方式、Logistic回归模型P值对于胃小间质瘤诊断情况进行预测,约登指数为35.88%、32.16%、42.22%、17.39%、59.54%。结论依据小探头超声内镜下胃小间质瘤的特征表现,能够建立起诊断该类疾病的数学模型。利用该模型实施疾病的诊断预测,也具有良好的效能,这为临床工作提供了参考。Objective To explore the diagnostic value and predictive model of characteristics of gastric small stromal tumors under endoscopic ultrasonography with small probes.Methods A retrospective analysis was performed on 193 patients with gastric submucosal space occupying lesions who underwent endoscopic ultrasonography with small probes in our hospital from January 2016 to December 2022.Patients were divided into control group(non-small gastric stromal tumor)and case group(small gastric stromal tumor)according to pathological findings.Describe the group of patients.Single factor analysis and multi-factor analysis of endoscopic diagnosis of gastric small stromal tumor.Efficacy analysis of prediction model for diagnosing gastric small stromal tumor under endoscopic ultrasonography.Results Among the surveyed patients,68 were pathologically diagnosed with small gastric stromal tumors and 125 were diagnosed with non-small gastric stromal tumors.Among the patients diagnosed with non-gastric small stromal tumors,47 were leiomyomas,38 were schwannomas,22 were chronic inflammatory lesions,8 were polyps,4 were fibromas,4 were glomus tumors,and 2 were lipomas.The results of single factor analysis showed that the data of age,lesion location,origin muscle layer and growth mode of patients with different pathological diagnosis results had statistical significance(all P<0.05).The results of multi-factor analysis showed that age,lesion location,myometria of origin and growth mode were independent factors for the diagnosis of small gastric stromal tumor under endoscopic ultrasonography(all P<0.05).It was verified that the multivariate logistic regression model had good goodness of fit(Hosmer-Lemeshowχ^(2)=5.769,P>0.05).Age,lesion location,myometria of origin,growth mode and Logistic regression model were used to predict the diagnosis of small gastric stromal tumor.The Yoden index was 35.88%,32.16%,42.22%,17.39%and 59.54%.Conclusion According to the characteristics of gastric small stromal tumor under endoscopic ultrasonography with

关 键 词：超声内镜 胃小间质瘤 预测模型 消化系统 回声 

分 类 号：R735.2[医药卫生—肿瘤]