双能量CT多参数成像预测胃腺癌淋巴结转移的价值

The Value of Dual-Energy CT Multi-Parameter Imaging in Predicting Lymph Node Metastasis of Gastric Adenocarcinoma

目的:探讨双能量CT双期增强扫描的碘浓度、电子云密度/等效原子序数、能谱曲线率等参数预测胃腺癌淋巴结转移的价值。方法:回顾性分析2019年5月至2021年1月在莆田市第一医院接受双能量CT增强扫描的63例胃腺癌患者的临床影像资料,共纳入194枚淋巴结。根据淋巴结转移情况分为淋巴结转移组(124枚)与非淋巴结转移组(70枚),淋巴结转移组又分为分化良好型胃腺癌组(57枚)与分化不良型胃腺癌组(67枚)。测量并计算淋巴结双期碘浓度(IC)、标准化碘浓度(NIC)、电子云密度/等效原子序数(Rho/Z)及不同区间能谱曲线率(K)。用t检验或Mann-Whitney U检验比较各参数的组间差异,用二元Logistic回归建立联合模型,用受试者工作特征(ROC)曲线确定各参数及联合模型的诊断效能,用Delong检验进行ROC曲线比较。结果:淋巴结转移组的动静脉期IC、NIC、Rho/Z及K均显著高于非淋巴结转移组(P均<0.05)。其中动脉期IC、NIC、Z、K_(40-150KeV)及静脉期Rho、K_(40-140KeV)的ROC曲线下面积(AUC)依次为0.867、0.923、0.905、0.852、0.775、0.843。由动脉期NIC和静脉期K_(40-140KeV)构建的联合模型显著优于单一参数(P均<0.05),AUC为0.949,阈值为0.588时,灵敏度、特异度、准确度分别为88.7%、90.0%、89.2%。分化不良型胃腺癌组的动静脉期IC、Z及K均显著高于分化良好型胃腺癌组(P均<0.05),其中动脉期IC、Z的AUC分别为0.629、0.671,静脉期IC、Z的AUC分别为0.736、0.735,动静脉期K的AUC为0.665-0.746。结论:双能量CT动静脉期IC、NIC、Rho/Z及K有助于术前预测胃腺癌淋巴结转移,联合模型具有更高的诊断效能。动静脉期IC、Z及K对胃腺癌分化程度的预测具有一定价值。

Objective:To evaluate the value of iodine concentration,electron cloud density/equivalent atomic number and energy spectrum curve rate in predicting lymph node metastasis of gastric adenocarcinoma by dual-energy CT dual-phase enhanced scan.Methods:The clinical imaging data of 63 patients with gastric adenocarcinoma who underwent dual-energy CT enhanced scan in the First Hospital of Putian from May 2019 to January 2021 were analyzed retrospectively.According to the situation of lymph node metastasis,the patients were divided into lymph node metastasis group(124 cases) and non-lymph node metastasis group(70 cases).Lymph node metastasis group was subdivided into well-differentiated gastric adenocarcinoma group(57 cases) and poorly differentiated gastric adenocarcinoma group(67 cases).The biphasic iodine concentration(IC),standardized iodine concentration(NIC),electron cloud density/equivalent atomic number(Rho/Z) and energy spectrum curve rate(K) in different regions were measured and calculated.T-test or Mann-Whitney U test was used to compare the differences between groups of parameters,binary Logistic regression was used to establish a joint model,receiver operating characteristic(ROC) curve was used to determine the diagnostic efficiency of parameters and the joint model,and Delong test was used to compare ROC curves.Results:IC,NIC,Rho/Z and K in arterial phase and venous phase in lymph node metastasis group were significantly higher than those in non-lymph node metastasis group.The area under ROC curve(AUC) of arterial phase IC,NIC,Z,K_(40-150KeV) and venous phase Rho,K_(40-140KeV) were 0.867,0.923,0.905,0.852,0.775 and 0.843,respectively.The combined model constructed by arterial phase NIC and venous phase K_(40-140KeV) was significantly better than a single parameter(all P<0.05).The AUC was 0.949.When the threshold was 0.588,the sensitivity,specificity and accuracy were 88.7%,90.0% and 89.2%,respectively.The IC,Z and K in arterial phase and venous phase in poorly differentiated gastric adenocarcinoma group were significantly higher than those in well differentiated gastric adenocarcinoma group.The AUC of IC and Z in arterial phase were 0.629 and 0.671,the AUC of IC and Z in venous phase were 0.736 and 0.735,and the AUC of K in arterial phase and venous phase were 0.665-0.746.Conclusion:Dual-energy CT IC,NIC,Rho/Z and K in arterial phase and venous phase are helpful to predict lymph node metastasis of gastric adenocarcinoma before operation,and the combined model has higher diagnostic efficiency.IC,Z and K in arterial phase and venous phase have certain value in predicting the differentiation degree of gastric adenocarcinoma.