B超特征预测甲状腺乳头状癌统计学模型应用价值分析

Application Value of Statistical Model Based on B-Ultrasound Characteristics in Predicting Papillary Thyroid Carcinoma

目的探讨和分析基于B超特征预测甲状腺乳头状癌的统计学模型在不同大小结节中的诊断价值。方法将2015年1月—2022年8月在福建省福清市第二医院和福建省福清市医院进行临床诊治的1385例甲状腺乳头状癌(papillary thyroid carcinoma,PTC)纳入研究范围,以PTC患者500例、良性病变患者192例进行统计学模型的随机选择验证。分析临床特征与5种超声特征相匹配度,同时计算和分析单一指标诊断恶性情况及95%置信区间。超声特征包括回声(ECHO)、高宽比(aspect ratio,AR)>1 cm、不规则边缘(irregular edges,IRE)、微钙化(microcalcification,MCAL)、甲状腺囊侵犯(thyroid cyst invasion,TCI)。结果结节超声特征与PTC的单因素和多因素Logistic回归分析结果表明,AR和TCI的OR值明显高于其他超声特征。PTC结节大小与超声特征关联分析结果表明,≥0.3 cm组IRE模式随着结节大小的增加而减少。四组AR的OR值相似,回声焦点(echo focus,EF)和TCI的OR值随着结节大小的增加而增加。不同结节大小患者模型的AUC分析结果表明<0.5 cm组及≥0.5 cm组中AUC分别为0.838和0.882。超声特征数量模型的概率分析结果表明,基于B超特征预测甲状腺乳头状癌的统计学模型PTC指示风险优于2015年美国甲状腺学会(American Thyroid Association,ATA)指南。结论本研究模型准确计算出针对各个超声特征的风险概率,具有更多超声特征的结节恶性概率明显更高。应用ECHO、AR、MCAL、TCI和结节大小的建立的模型可以为预测PTC提供较高的准确度。

Objective To explore and analyze the diagnostic value of statistical model based on B-ultrasound characteristics in predicting papillary thyroid carcinoma(PTC)with different sizes of nodules.Methods A total of 1385 patients with PTC treated in Fuqing Second Hospital,Fujian Province and Fuqing Hospital,Fujian Province were enrolled from January 2015 to August 2022.There were 500 cases with PTC and 192 cases with benign lesions randomly selected for validation.The matching degree between clinical characteristics and the 5 kinds of ultrasound characteristics was analyzed,the conditions and 95%confidence interval(CI)of single index in the diagnosis of malignant lesions were calculated and analyzed.The ultrasound characteristics included echo,aspect ratio(AR)>1 cm,irregular edges(IRE),microcalcification(MCAL)and thyroid cyst invasion(TCI).Results The results of univariate and multivariate Logistic regression analysis showed that OR values of AR and TCI were significantly greater than those of the other ultrasound characteristics.The results of correlation analysis between PTC nodule size and ultrasound characteristics showed that IRE pattern was decreased with the increase of nodule size in≥0.3 cm group.OR values of AR were similar in the four groups.OR values of echo focus(EF)and TCI were increased with the increase of nodule size.AUC analysis in patient with different nodule sizes showed that AUC values in<0.5 cm group and≥0.5 cm group were 0.838 and 0.882,respectively.The results of probabilistic analysis on the quantitative model of ultrasound characteristics showed that the indicated risk of statistical model based on B-ultrasound characteristics for predicting PTC was superior to 2015 American Thyroid Association(ATA)guidelines.Conclusion This model can accurately calculate the risk probability based on different ultrasound characteristics.The incidence of malignant nodules with more ultrasound characteristics is significantly higher.The model constructed by ECHO,AR,MCAL,TCI and nodule size can provide higher accuracy to predict PTC.