磁共振参数结合临床数据构建胃癌伴淋巴结转移的预测模型

A predictive model for gastric cancer with lymph node metastasis based on magnetic resonance parameters and clinical data

目的基于磁共振参数和临床数据构建胃癌患者淋巴结转移的风险预测模型并进行验证。方法回顾性总结2018年6月至2022年6月病理确诊胃癌患者160例为研究对象,术前检测外周血癌胚抗原(CEA)和CA199,采用3.0T MRI获得T1-横轴位、T2-冠状位和横轴位以及扩散加权成像(DWI),判断肿瘤T和N分期,测量表观扩散系数(ADC)_(mean)。根据病理结果证实有无淋巴结转移。结果160例患者中52例病理证实淋巴结转移,发生率为32.5%(52/160)。单因素比较发现,淋巴结转移阳性组患者血清CEA和CA199明显高于阴性组,ADC_(mean)降低,T分期(T3~T4)、N(阳性)、切缘阳性、壁外血管侵犯和低分化增多(P<0.05)。多因素Logistic回归分析显示,CA199和ADC_(mean)、T分期、N(阳性)和分化级别是淋巴结转移的独立预测因素(P<0.05)。构建预测模型Y=0.562+0.123×(CA199)-0.285×(ADC_(mean))+0.896×(T分期)+1.256×N+0.758×(分化级别)。受试者工作曲线(ROC)显示,预测模型诊断淋巴结转移的曲线下面积(AUC)为0.895(P<0.05)。结论术前胃癌磁共振检查有助于判断淋巴结转移,血清CA199升高、ADC_(mean)降低、T分期升高、N阳性和低分化是淋巴结转移的独立预测因素,建立预测模型有较高的诊断潜力,便于临床推广应用。

Objective To construct a risk predictive model for gastric cancer with lymph node metastasis based on magnetic resonance parameters and clinical data,and to verify its potential.Methods A total of 160 patients with pathologically diagnosed gastric cancer in our hospital from June 2018 to June 2022 were retrospectively summarized.Peripheral blood levels of preoperative carcinoembryonic antigen(CEA)and carbohydrate antigen 19-9(CA19-9)were detected.The axial T1,coronal and axial T2 images,and diffusion weighted imaging(DWI)were obtained by 3.0T magnetic resonance imaging(MRI)to determine the T and N stages of tumor.The mean apparent diffusion coefficient(ADC_(mean))was measured.The presence or absence of lymph node metastasis was pathologically determined.Results There were 52/160(32.5%)gastric cancer patients with pathologically confirmed positive lymph node metastasis.Univariate logistic regression analysis showed that gastric cancer patients with positive lymph node metastasis presented significantly higher serum levels of CEA and CA199,lower ADC_(mean)and high proportions of T3-4 staging,positive N staging,positive tumor margin,extramural vascular invasion and low differentiation than those with negative lymph node metastasis(P<0.05).Multivariate logistic regression analysis showed that CA199,ADC_(mean),T staging,positive N staging and differentiation level were the independent predictors to lymph node metastasis in gastric cancer patients(P<0.05).A predictive model was constructed as follows:Y=0.562+0.123×CA199 level-0.285×ADC_(mean)+0.896×T staging+1.256×N staging+0.758×differentiation degree level.Receiver operating characteristic(ROC)curves showed that the area under curve(AUC)of the model for predicting lymph node metastasis in gastric cancer patients was 0.895(P<0.05).Conclusion Preoperative MRI examination of gastric cancer is helpful to determine lymph node metastasis.The increased serum CA199,decreased ADC_(mean),increased T staging,positive N staging and low differentiation level are the independent predictors of lymph node metastasis in gastric cancer.The constructed model has a high predictive potential and is convenient for clinical application.