临床-CT模型评估食管鳞状细胞癌侵犯脉管和/或神经

Clinical-CT model for evaluating lymphovascular and/or perineural invasion of esophageal squamous cell carcinoma

目的观察临床-CT模型评估食管鳞状细胞癌(ESCC)侵犯脉管(LVI)和/或神经(PNI)的价值。方法回顾性分析156例ESCC患者资料,根据术后病理结果将LVI和/或PNI阳性者归为阳性组(n=58)、LVI及PNI均阴性者归为阴性组(n=98);比较组间临床及CT资料,行logistic回归分析并建立模型,观察其评估LVI和/或PNI的效能。结果组间癌胚抗原(CEA)、糖类抗原199(CA199)、肿瘤厚度、肿瘤体积、静脉期CT值(CTV)、CTV与平扫CT值(CTP)差值(ΔCT_(V-P))及静脉期增强率(V%)差异均有统计学意义(P均<0.05),其评估ESCC LVI和/或PNI的曲线下面积(AUC)分别为0.702、0.690、0.731、0.744、0.621、0.631及0.599。CEA、CA199、肿瘤厚度、肿瘤体积及CTV为ESCC LVI和/或PNI的独立预测因素,以之建立的联合模型评估ESCC LVI和/或PNI的准确率、敏感度及特异度分别为82.05%、65.52%及91.84%,且其AUC为0.838,高于各单一参数(P均<0.05)。结论所获临床-CT模型可有效评估ESCC LVI和/或PNI。

Objective To establish a clinical-CT model,and to observe its value for evaluating lymphovascular invasion(LVI)and/or perineural invasion(PNI)in esophageal squamous cell carcinoma(ESCC).Methods Data of 156 ESCC patients were retrospectively analyzed.The patients were divided into positive group(n=58,LVI[+]and/or PNI[+])and negative group(n=98,LVI[-]and PNI[-])according to postoperative pathological results.Clinical and CT data were compared between groups.Logistic regression analysis was performed to establish a model,and its efficacy of evaluating ESCC LVI and/or PNI was analyzed.Results Significant differences of carcinoembryonic antigen(CEA),carbohydrate antigen 199(CA199),tumor thickness,tumor volume and CT venous phase value(CTV),the difference between CTV and CT plain phase value(CTP)(ΔCT_(V-P))and venous phase enhancement rate(V%)were found between groups(all P<0.05),and the area under the curve(AUC)of the above parameters for evaluating ESCC LVI and/or PNI was 0.702,0.690,0.731,0.744,0.621,0.631 and 0.599,respectively.CEA,CA199,tumor thickness,tumor volume and CTV were all independent predictive factors for ESCC LVI and/or PNI.A combined model was established based on the above features,and its accuracy,sensitivity and specificity for evaluating ESCC LVI and/or PNI was 82.05%,65.52%and 91.84%,respectively,with AUC of 0.838,higher than that of each single parameter(all P<0.05).Conclusion The established clinical-CT model could effectively evaluate ESCC LVI and/or PNI.