胃低级别上皮内瘤变根除幽门螺杆菌后病理升级的危险因素列线图建立与验证

Establishment and validation of a nomogram model for the progression risk of gastric low-grade intraepithelial neoplasia after the eradication of Helicobacter pylori

摘要:

目的 建立并验证胃低级别上皮内瘤变（LGIN）根除幽门螺杆菌（HP）感染后病理升级的危险因素列线图。方法 选取该院消化内科初诊胃镜活检病理提示为胃LGIN的患者，予以根除HP，并对其进行2年的随访，统计分析病理升级组和未升级组之间临床和内镜特征的差异，通过单因素和多因素分析筛选出有明显差异的独立危险因素，建立胃LGIN病理升级的危险因素列线图模型，并利用C指数、受试者操作特征曲线（ROC）和校正图，予以验证列线图的稳定性及可靠性。结果 共纳入309例患者。其中，有153例发生了病理升级。单因素分析表明，性别、病灶直径、有无边界、病灶数量、形状及是否伴有糜烂溃疡差异有统计学意义；而多因素分析表明，患者为男性、病灶直径 ≥ 1 cm、有边界和凹陷型且无糜烂溃疡是发生病理升级的独立危险因素。建立列线图后，验证该模型C指数为0.775（0.724～0.826），ROC曲线的曲线下面积（AUC）为0.740，且校正图显示该模型有较好的稳定性。结论 该模型具有较好的稳定性和一致性，可用于评估胃LGIN患者中根除HP后发生病理升级的风险大小，指导临床选择合适的治疗方式。

Abstract:

Objective To establish and validate a nomogram model for predicting the progression risk of patients with gastric low-grade intraepithelial neoplasia (LGIN) after the eradication of Helicobacter pylori (HP).Methods Patients who were diagnosed as gastric LGIN by gastroscopic biopsy were collected, followed up for 2 years, with HP eradicated. Independent risk factors screened for significant differences by univariate and multivariate analysis, and establish a gastric LGIN pathological upgrade nomogram model, then use C index, ROC curve, correction plan to verify the stability and reliability of the nomogram.Results 153 of whom were pathological upgrade. Univariate analysis showed that there were statistically significant differences between the two groups in gender, lesion diameter, boundary, number, shape and erosion. Multivariate analysis showed that male patients, lesion diameter ≥ 1 cm, bordered, depressed and non-erosion were independent risk factors for pathological progression. Then the predictive nomogram was established, and the stability of the model was verified by the C index of 0.775 (0.724 ~ 0.826), the AUC of 0.740 and the smart calibration curve.Conclusion This model has good reliability and can be used to evaluate the risk of pathological progression after the eradication of HP in gastric LGIN patients, and guide the selection of treatment modalities.