摘要
目的基于桥本甲状腺炎背景下甲状腺乳头状癌的超声特征和桥本甲状腺炎特异性抗体状态,预测2015版美国甲状腺学会(ATA)复发风险分层。资料与方法回顾性分析2017年1月—2019年12月于解放军总医院第一医学中心初次行甲状腺手术的479例甲状腺乳头状癌与桥本甲状腺炎共存患者的超声及临床资料,按照时间分为训练组327例及验证组152例。采用多因素Logistic回归分析与ATA高复发风险分层相关的独立影响因素,构建、筛选预测模型,采用曲线下面积、校准曲线、布里尔评分评估模型效能。结果多因素Logistic分析显示,结节恶性多灶(OR=3.812,95%CI 1.275~11.397,P=0.017)、结节接触被膜(OR=8.012,95%CI 1.647~38.972,P=0.010)、微钙化(OR=4.220,95%CI 1.302~13.678,P=0.016)、纵横比>1(OR=4.017,95%CI 1.286~12.548,P=0.017)、结节血流丰富(OR=6.120,95%CI 2.225~16.832,P<0.001)、结节最大径≥1 cm(OR=4.784,95%CI 1.360~16.833,P=0.015)、腺体呈典型桥本回声(OR=0.114,95%CI 0.039~0.330,P<0.001)、抗甲状腺过氧化物酶抗体单阳性(OR=0.088,95%CI 0.006~1.299,P=0.077)是高复发风险分层的独立预测因素。预测模型在训练组及验证组中曲线下面积分别为0.942(95%CI 0.911~0.972)和0.933(95%CI 0.878~0.990),两组校准曲线均拟合良好,布里尔评分分别为0.054和0.058,模型预测效能良好。结论基于超声特征联合抗体状态的术前预测模型评估甲状腺乳头状癌与桥本甲状腺炎共存患者ATA复发风险分层具有良好的效能,可为临床拟定治疗方案提供帮助。
Purpose To predict the 2015 American thyroid association(ATA)recurrence risk stratification based on ultrasonographic features and Hashimoto's thyroiditis(HT)-specific antibody status of papillary thyroid carcinoma(PTC)in the context of HT.Materials and Methods A retrospective analysis was conducted on the ultrasonographic and clinical data of 479 patients with coexisting PTC and HT,who underwent their first thyroid surgery at the First Medical Center of Chinese PLA General Hospital from January 2017 to December 2019.All patients were divided in chronological order into a training group(n=327)and a validation group(n=152).Multivariate Logistic regression analysis was utilized to identify independent factors associated with high recurrence risk stratification according to the ATA guidelines.Predictive models were constructed and screened,and the efficacy of these models was evaluated using the area under the curve,calibration curves and Brier scores.Results Multivariate Logistic regression analysis identified the following as independent predictive factors for high recurrence risk stratification:multifocal malignancy of nodules(OR=3.812,95%CI 1.275-11.397,P=0.017),nodule contact with the capsule(OR=8.012,95%CI 1.647-38.972,P=0.010),microcalcifications(OR=4.220,95%CI 1.302-13.678,P=0.016),an aspect ratio>1(OR=4.017,95%CI 1.286-12.548,P=0.017),abundant nodule vascularity(OR=6.120,95%CI 2.225-16.832,P<0.001),maximum nodule diameter≥1 cm(OR=4.784,95%CI 1.360-16.833,P=0.015),a glandular echo characteristic of typical HT(OR=0.114,95%CI 0.039-0.330,P<0.001),and anti-thyroid peroxidase antibody monopositivity(OR=0.088,95%CI 0.006-1.299,P=0.077).The predictive model demonstrated strong performance,as evidenced by the area under the curve of 0.942(95%CI 0.911-0.972)in the training set and 0.933(95%CI 0.878-0.990)in the validation set.Both groups exhibited well-fitting calibration curves.The Brier scores were 0.054 and 0.058 for the training and validation sets,respectively,indicating excellent predictive efficacy of the m
作者
葛乃侨
王月香
兰雨
姜波
李墨琳
邢光辉
罗渝昆
GE Naiqiao;WANG Yuexiang;LAN Yu;JIANG Bo;LI Molin;XING Guanghui;LUO Yukun(Chinese PLA Medical School,Beijing 100853,China;Department of Ultrasound,the First Medical Center of Chinese PLA General Hospital,Beijing 100853,China)
出处
《中国医学影像学杂志》
CSCD
北大核心
2024年第9期891-896,共6页
Chinese Journal of Medical Imaging
关键词
桥本病
甲状腺癌
乳头状
超声检查
抗体
复发
风险
预测
Hashimoto disease
Thyroid cancer,papillary
Ultrasonography
Antibodies
Recurrence
Risk
Forecasting
