基于常规检验学指标构建胃部疾病患者恶性风险评估模型

Construction of a model for assessing the malignant risk of patients with gastric disease based on conventional laboratory indicators

目的利用肿瘤标志物、血常规、凝血等常规检验学指标构建预测胃部疾病患者良恶性判断的风险预测模型并进行预测价值验证。方法回顾性分析2018年1月至2023年1月就诊于广州医科大学附属第一医院的胃部疾病患者的病历资料,根据病理结果分为胃癌组(134例)和非萎缩性胃炎组(298例),收集两组患者的血清及全血检测等常规检验学指标数据,利用R 4.2.3软件进行统计分析,构建胃癌风险列线图预测模型并进行验证。结果采用Logistic回归分析构建了包含D-二聚体(DD)、癌胚抗原(CEA)、糖类抗原72-4(CA72-4)、血红蛋白(Hb)4个检验学指标的风险预测模型,并绘制可视化列线图作为最终预测模型。该模型在训练集和测试集的ROC曲线下面积(AUC^(ROC))分别为0.809(95%CI:0.754~0.864)和0.808(95%CI:0.724~0.892),敏感性和特异性分别为58.5%和93.3%,提示该模型具有良好的预测能力。结论通过常规检验学指标建立了胃部疾病患者恶性风险预测模型,该模型准确性良好,可有效预估胃部疾病转化为胃癌的风险,有助于临床尽早发现早期胃癌患者并采取针对性的预防和干预措施。

Objective To construct a model for predicting the benign and malignant risk of the patients with gastric diseases using conventional laboratory indicators such as tumor markers,blood routine,and coagulation indicators,and validate its predictive value.Methods The medical records of patients with gastric diseases who visited the First Affiliated Hospital of Guangzhou Medical University from January 2018 to January 2023 were analyzed retrospectively.According to the pathological results,the patients were divided into the gastric cancer(GC)group(n=134)and chronic non-atrophic gastritis(CNAG)group(n=298).Their routine test data such as serum and whole blood tests were collected.The statistical analysis was conducted using the R 4.2.3 software,and a model for predicting the risk of GC was constructed and validated.Results A model for predicting the risk of GC was constructed successfully using the Logistic regression analysis,which included D-dimer,carcinoembryonic antigen(CEA),carbohydrate antigen 72-4(CA72-4)and hemoglobin(Hb).A visual nomogram was plotted as the final prediction model.The areas under the receiver operating characteristic(ROC)curve(AUC^(ROC))of the model in the training and testing sets were 0.809(95%CI:0.754-0.864)and 0.808(95%CI:0.724-0.892),respectively.The sensitivity and specificity of the model were 58.5%and 93.3%,respectively,indicating that it had good predictive ability.Conclusion The model for predicting the malignant risk of patients with gastric diseases constructed using routine testing indicators has good accuracy and can effectively predict the risk of gastric disease transforming into gastric cancer.It helps to find early gastric cancer patients in clinical practice and take targeted prevention and intervention measures.