超声检查评分系统在预测甲状腺乳头状癌隐匿性中央区淋巴结转移中的应用

The application of scoring system of ultrasonography in prediction for occult central lymph node metastasis of papillary thyroid carcinoma

目的:探索甲状腺乳头状癌(papillary thyroid carcinoma,PTC)的超声特征和内分泌指标等信息与隐匿性中央区淋巴结转移(occult central lymph node metastasis,OCLNM)之间的联系,并建立预测评分系统。方法:对2016年01月至2018年01月期间的随机选取的283例接受甲状腺全切除术与预防性中央区淋巴结清扫的PTC患者信息进行归纳,通过单因素和多因素分析确定预测因子。使用二元Logistic回归分析筛选独立影响因素,并建立评分预测系统。结果:多因素分析显示年龄、男性、中下部存在肿瘤、峡部存在肿瘤、超声所测肿瘤最大径、存在微钙化、CDFI可见周边型血流信号以及血清甲状腺球蛋白(thyroglobulin,TG)水平是OCLNM的独立影响因素。利用以上因素建立评分预测模型:预测评分Y(P)=1/[1+exp(1.001+0.079×年龄-0.764×男性-2.739×峡部存在肿瘤-1.244×腺体的中下部存在肿瘤-0.117×肿瘤最大径-1.041×存在微钙化-0.728×CDFI可见周边型血流信号-0.016×血清TG)]。接受者操作特征曲线下面积为0.815。通过评价发现0.257作为截止值时预测OCLNM效果最佳,特异度和敏感度分别为87.5%和61.5%,阳性预测值和阴性预测值分别为56.9%和89.4%,预测准确率为71.0%。结论:当Y(P)≥0.257时,预测PTC患者发生OCLNM风险较高。通过评分预测系统的建立,利用临床中术前易得到的相关信息预测PTC患者的OCLNM。为术前无法确定OCLNM的患者确定合理的手术范围,减少不必要的中央区淋巴结清扫(central lymph node dissection,CLND),最大化患者在治疗中的获益。

Objective:To explore the factors that can predict occult central lymph node metastasis(OCLNM)of papillary thyroid carcinoma(PTC)independently from ultrasound characteristics and endocrine indicators.A predictive scoring system for OCLNM was established.Methods:283 PTC patients undergoing thyroidectomy and prophylactic central lymph node dissection(pCLND)from January 2016 to January 2018 were collected randomly to determine the predictors of OCLNM by single-factor and multivariate Logistic regression analysis.Logistic regression analysis was used to screen independent influence factors and a scoring prediction system was established.Results:Multivariate Logistic regression analysis showed independent predictors of OCLNM,age,male,tumors in the middle or lower poles,tumors in the isthmus,maximum tumor diameter measured by ultrasound,microcalcification,visible surrounding blood flow signal,and thyroglobulin(TG).We used the factors above to establish a scoring prediction model:predictive score Y(P)=1/[1+exp(1.001+0.079×age-0.764×male-2.739×tumors in the isthmus-1.244×tumors in the middle or lower poles-0.117×tumor maximum diameter-1.041×microcalcification-0.728×visible surrounding blood flow signal-0.016×TG)].The area under the curve(AUC)of the receiver operating characteristic(ROC)was 0.815.It was found that 0.257 was the best cut-off value for predicting OCLNM,with specificity and sensitivity of 87.5%and 61.5%,respectively.Positive and negative predictive values were 56.9%and 89.4%,respectively.Predictive accuracy was 71.0%.Conclusion:When Y(P)≥0.257,the risk of OCLNM in patients with PTC is predicted to be high.With the scoring prediction system,the clinical information obtained before surgery can be used to predict the risk of OCLNM in PTC patients.The significance lies in determining a reasonable extent of surgery for patients who cannot determine OCLNM before surgery,reducing unnecessary central lymph node dissection(CLND)and maximizing the patient’s benefit in treatment.