AI TI-RADS与ACR TI-RADS的诊断价值及非必要穿刺率对比研究  

Comparison of Diagnostic Value and Unnecessary FNA Rate of AI TI-RADS and ACR TI-RADS

作　　者：康亚宁 付超[1] 马笑[1] 郭艳飞[1] 许聪颖[1] 李静[1] 崔可飞[1] Kang Yaning;Fu Chao;Ma Xiao;Guo Yanfei;Xu Congying;Li Jing;Cui Kefei(Department of Ultrasound,the First Affiliated Hospital of Zhengzhou University,Zhengzhou,Henan 450052,China)

机构地区：[1]郑州大学第一附属医院超声科,郑州市450052

出　　处：《中国超声医学杂志》2021年第4期367-371,共5页Chinese Journal of Ultrasound in Medicine

摘　　要：目的比较AI TI-RADS与ACRTI-RADS对甲状腺结节的诊断价值和非必要穿刺率。方法回顾性分析2583例患者共3833个甲状腺结节的超声图像,用AITI-RADS和ACRTI-RADS对所有结节进行分类,构建ROC曲线,比较两者的诊断价值和非必要穿刺率。结果①AI TI-RADS和ACR TI-RADS的ROC曲线下面积分别为0.882、0.867(Z=4.682,P<0.0001)。两者之间净重新分类指数为5.98%(Z=6.282,P=0.000)。AI TI-RADS的灵敏度、准确度高于ACR TI-RADS(P<0.05),特异度的差异无统计学意义(χ^(2)=0.019,P=0.891)。②AI TI-RADS的非必要穿刺率(11.4%)低于ACR TI-RADS(15.2%)(χ^(2)=24.457,P=0.000)。结论AITI-RADS可提高准确度和灵敏度而不降低特异度,同时降低良性结节的非必要穿刺率。Objective To compare the diagnostic value and unnecessary fine-needle aspiration(FNA)rate of Artificial Intelligence Thyroid Imaging Reporting and Data System(AI TI-RADS)and American College of Radiology TI-RADS(ACR TI-RADS)for thyroid nodules. Methods The ultrasound images of 3833 thyroid nodules of 2583 patients were retrospectively analysed. All nodules were categorized by AI TI-RADS and ACR TI-RADS. The ROC curves were plotted to compare the diagnostic value and unnecessary FNA rate of the two. Results(1)The area under ROC curves of AI TI-RADS and ACR TI-RADS were 0.882,0.867,respectively(Z=4.682,P<0.000 1).The net reclassification index(NRI)between the two was 5.98%(Z=6.282,P=0.000).The sensitivity and accuracy of AI TI-RADS were higher than that of ACR TI-RADS(P<0.05), with no significant difference in specificity(χ^(2)=0.019,P=0.891).(2)The unnecessary FNA rate of AI TI-RADS(11.4%) was lower than that of ACR TI-RADS(15.2%)(χ^(2)=24.457,P=0.000). Conclusions AI TI-RADS can improve accuracy and sensitivity without reducing specificity, while reducing the unnecessary FNA rate of benign nodules.

关 键 词：超声风险分层系统 ACR TI-RADS AI TI-RADS 

分 类 号：R445.1[医药卫生—影像医学与核医学] R581[医药卫生—诊断学] R736.1[医药卫生—临床医学]