能谱CT术前预测胃腺癌淋巴结转移的价值

Clinical Value of Spectral CT in Preoperative Prediction of Lymph Node Metastasis of Gastric Adenocarcinoma

目的探讨能谱CT参数术前预测胃腺癌淋巴结转移的价值。方法回顾性分析2015年1月至2017年1月经病理确诊为胃腺癌并行能谱CT扫描的210例患者的临床及影像学资料,根据有无淋巴结转移分为淋巴结转移组(123例)和非转移组(87例)。记录胃癌患者的临床及病理学资料,并在后处理工作站中测量和计算胃癌原发灶的动静脉期CT值、碘浓度(IC)值和标准化碘浓度(nIC)值。采用组内相关系数(ICC)评价两名医师测量各参数值的一致性;采用独立样本t检验比较两组间定量资料;采用卡方检验或Kruskal-Wallis H检验比较两组间定性资料;采用受试者工作特征(ROC)曲线分析能谱IC值鉴别淋巴结转移的诊断效能,并确定其最佳阈值。采用Logistic回归分析筛选有预测意义的指标并计算其优势比(OR)。结果淋巴结转移组和非转移组患者肿瘤原发灶的分化程度、Borrman分型、T分期、最厚径、动静脉期IC值和动静脉期nIC值差异均有统计学意义(P<0.05),而患者的性别、年龄、肿瘤的部位、Lauren分型和动静脉期CT值差异均无统计学意义(P>0.05)。肿瘤原发灶的动静脉期IC和nIC值的ROC曲线下面积(AUC)分别为0.647、0.764、0.695、0.849,并以静脉期nIC=0.47为临界值判断淋巴结的转移性效能最佳。Logistic回归分析显示,肿瘤Borrman分型、最厚径、T分期及静脉期nIC值是影响胃腺癌淋巴结转移状态的独立预测指标(P<0.05)。结论肿瘤原发灶的Borrman分型、最厚径、T分期及能谱CT参数静脉期nIC值可以评价并预测胃腺癌淋巴结转移,提高CT胃癌术前诊断N分期的准确度,为临床胃癌患者个体化治疗提供更多依据。

Objective To investigate the clinical value of spectral CT parameters in predicting thelymph node metastasis of gastric adenocarcinoma.Methods From January 2015 to January 2017,the clinical and imaging data of 210 gastric adenocarcinoma patients confirmed by pathology and underwent preoperative spectral CT scanning were analyzed retrospectively.According to the status of lymph node metastasis,the patients were divided into nodular metastasis group(123 cases)and non-metastasis group(87 cases).The clinicopathological data of these patients was recorded,and the CT values,iodine concentration(IC)values and normalized iodine concentration(nIC)values of primary lesions in dual-phases were measured and calculated at the postprocessing workstation.Intraclass correlation coefficient(ICC)was used to evaluate interobserver agreements for IC and nIC values measurements.The Student’s t-test and the Chi-square test(or Kruskal-Wallis H test)were used to compare the differences of quantitative data and qualitative data,respectively.The receiver operating characteristic(ROC)curves were applied to evaluate the diagnostic efficiency ofIC and nIC values in differentiating lymph node metastasis of gastric cancer,and the optimal threshold values of IC and nIC were determined.Logistic regression analysis was used to screen predictive indicators for nodular metastasis in gastricadenocarcinoma and to calculate their odds ratio(OR).Results The degree of differentiation,Borrman classification,T staging,tumor thickness,IC and nIC values in dual-phases of primary lesions show significant differences between the nodular metastasis group and non-metastasis group(P<0.05).However,there were no significant differences between the two groups in the gender,age,location of tumor,Lauren classification and CT values in dual-phases(P>0.05).The areas under the curve(AUC)of IC and nIC values in dual-phases were 0.647,0.764,0.695 and 0.849,respectively.Moreover,the nIC=0.47 in venous phase was the best cut-off value for diagnosing lymph node metastasis of gastric adenocarcinoma.The logistics regression analysis revealed that Borrman classification,tumor thickness,T staging and nIC values in venous phase were independent indicators for prediction of preoperative nodular status in gastric adenocarcinoma(P<0.05).Conclusion The Borrman classification,tumor thickness,T staging and nIC values in venous phase of primary lesions can be regarded as effective factors for evaluating and predicting lymph node metastasis in gastricadenocarcinoma.Based on this,the accuracy of the preoperative diagnosis of N staging of gastric cancer by spectral CT can be increasingly promoted.Therefore,more information can be used to develop an individualized treatment scheme for patients with clinical gastric cancer.