双能量CT列线图预测甲状腺乳头状癌大容量中央区淋巴结转移的临床价值

Clinical value of dual-energy CT parameters-based nomogram in predicting large-volume central lymph node metastases of papillary thyroid carcinoma

目的 探讨基于双能量CT构建的列线图在术前预测甲状腺乳头状癌(PTC)大容量中央区淋巴结转移(CLNM)中的应用价值。方法 收集2017年9月至2022年12月温州医科大学附属第五医院收治的经病理证实为PTC并接受双能量CT检查的患者274例。根据中央区淋巴结清扫结果,将患者分为小容量CLNM组(198例)和大容量CLNM组(76例)。274例患者按照7∶3比例随机分为训练集(192例)和验证集(82例)。分析两组临床及常规CT特征(年龄、性别、肿瘤最长径、肿瘤位置、包膜外侵犯和钙化)、双能量CT参数[动脉期和静脉期的标准化碘浓度(NIC)、能谱曲线斜率(λ_(HU))和标准化有效原子序数(n_(Zeff))]。采用多因素logistic回归分析与大容量CLNM相关的危险因素,并构建列线图。绘制ROC曲线,根据AUC、灵敏度和特异度评估列线图的诊断效能。结果 两组患者年龄、肿瘤最长径及包膜外侵犯比较,差异均有统计学意义(均P<0.05)。大容量CLNM组PTC动脉期NIC、λ_(HU)、n_(Zeff)及静脉期NIC均明显高于小容量CLNM组(均P<0.05)。多因素logistic回归结果显示,肿瘤最长径、包膜外侵犯、动脉期NIC、动脉期λ_(HU)和动脉期n_(Zeff)是预测PTC患者大容量CLNM的独立危险因素(均P<0.05)。基于上述指标建立的列线图表现出良好的诊断效能,训练集AUC为0.936,灵敏度为0.887,特异度为0.871;验证集AUC为0.912,灵敏度为0.871,特异度为0.808。结论 基于双能量CT的列线图在术前预测PTC患者大容量CLNM具有较好的临床价值。

Objective To explore the application value of a dual-energy CT parameters-based nomogram in preoperative prediction of large-volume central lymph node metastasis(CLNM) of papillary thyroid carcinoma(PTC).Methods The clinical and imaging data of 274 patients with PTC who underwent surgical resection and received dualenergy CT examination before surgery in the Fifth Affiliated Hospital of Wenzhou Medical University from September 2017 to December 2022 were retrospectively analyzed.According to the results of central neck lymph node dissection,the patients were divided into the small-volume CLNM group(n=198) and large-volume CLNM group(n=76).Patients were randomly divided into the training set(n=192) and validation set(n=82) with a ratio of 7:3.The clinical and conventional CT features(age,gender,maximum tumor diameter,tumor location,extracapsular invasion and calcification),dual energy CT parameters[normalized iodine concentration(NIC),slope of the spectral Hounsfield unit curve(λ_(HU)),and normalized effective atomic number(n_(Zeff)) in the arterial phase and venous phase] in the two groups were analyzed.Multivariate logistic analysis was used to analyze the risk factors correlated with the large-volume CLNM,and a nomogram was constructed.The diagnostic performance of the nomogram was evaluated with ROC curve.Results There were significant differences in age,maximum tumor diameter and extracapsular invasion between the two groups(all P<0.05).The NIC,λ_(HU),n_(Zeff)in the arterial phase and NIC in the venous phase in the large-volume CLNM group were significantly higher than those in the small-volume CLNM group(all P<0.05).Multivariate logistic regression results showed that maximum tumor diameter,extracapsular invasion,NIC in arterial phase,λ_(HU) in arterial phase,and n_(Zeff) in arterial phase were independent risk factors for predicting large-volume CLNM in PTC patients(all P<0.05).The nomogram established based on the above indicators showed good prediction performance,the area under ROC curve(AUC) of the nomogram in training set was 0.936 with a sensitivity of 0.887 and a specificity of 0.871;those of the validation set were 0.912,0.871 and 0.808,respectively.Conclusion The nomogram based on dual-energy CT parameters has good clinical value for preoperative evaluation of the large-volume CLNM in PTC patients.