甲状腺结节
医学
人口
甲状腺癌
内科学
病理
甲状腺
环境卫生
作者
Min Ding,Gaosong Wu,S X Zhang,Rongli Xie,Jianming Yuan,Guohui Xiao,Xiaoyue Zhang,Jiaqi Dai,Jian Chen,Sanming Wang,Rui Zhou,Xingchen Li,Jiuzhou Zhao,Feifei Ning,Zeyu Wu,Zhiqiang Yin,Jian Fei
标识
DOI:10.1158/1078-0432.ccr-25-0270
摘要
Ultrasound-guided fine-needle aspiration (FNA) biopsy is the gold standard for diagnosing thyroid cancer (TC), but 20%-35% of nodules remain cytologically indeterminate. This study evaluated the value of molecular testing in diagnosing TC and its preoperative risk stratification in the Chinese population. This multicenter observational study included patients with thyroid nodules admitted to five research centers from January 2021 to June 2023. All samples underwent molecular testing using quantitative polymerase chain reaction (qPCR) or next-generation sequencing (NGS). The study assessed genetic variations in nodules and the diagnostic performance of each test, using stepwise multivariable logistic regression to explore factors affecting lymph node metastases (LNM) and tumor stage. A total of 1984 patients with 2027 thyroid nodules were analyzed. The most common genetic alteration detected was BRAF V600E, followed by TERT promoter and CCDC6-RET fusion. For Bethesda categories II and VI, molecular tests combined with cytology significantly enhanced diagnostic performance, yielding Youden Index values of 0.97 for PCR-8 genes and 0.94 for NGS-28 genes. In cytologically indeterminate nodules, PCR-3 genes and NGS-8 genes exhibited 85% sensitivity and 100% specificity. Independent risk factors for LNM included age (OR=0.97), male sex (OR=1.45), higher TI-RADS grading (OR=1.50), larger tumor size (OR=1.13), and multifocal nodules (OR=1.60). Combined mutations in the 8-gene (OR=6.43) were significant for the advanced tumor stage. Molecular testing substantially improves diagnostic accuracy when integrated with cytology, facilitating the diagnosis of cytologically indeterminate nodules and offering prognostic insights for TC patients.
科研通智能强力驱动
Strongly Powered by AbleSci AI