C-TIRADS联合S-Detect技术在甲状腺结节诊断中的初步应用  

Preliminary Application of Chinese Version of TI-RADS Combined with S-Detect Technology in Thyroid Nodules

作　　者：方明娣[1] 彭梅[1] 毕玉[1] Fang Mingdi;Peng Mei;Bi Yu(Department of Ultrasound,the Second Affiliated Hospital of Anhui Medical University,Hefei,Anhui 230601,China)

机构地区：[1]安徽医科大学第二附属医院超声诊断科,合肥市230601

出　　处：《中国超声医学杂志》2021年第11期1212-1215,共4页Chinese Journal of Ultrasound in Medicine

基　　金：安徽省重点研究与开发计划项目(No.201904a07020068);安徽医科大学科研基金(No.2019xkj045)。

摘　　要：目的探讨C-TIRADS分类方法联合人工智能S-Detect技术对甲状腺结节的诊断价值。方法采用C-TIRADS分类方法和S-Detect技术对98例患者(共98个甲状腺结节)进行评估,以术后病理结果作为金标准,对比分析采用C-TIRADS分类方法、S-Detect技术以及两者联合对甲状腺结节的诊断效能。结果经手术病理证实,98例患者共98个甲状腺结节中,良性40个、恶性58个。超声医师采用C-TIRADS分类方法联合S-Detect技术诊断甲状腺结节良恶性的灵敏度为95.2%、特异度为82.5%、准确率为91.8%,其诊断效能优于单独应用常规超声和S-Detect技术(P<0.05)。结论 C-TIRADS分类方法兼顾分类的准确性及临床易用性,适应目前我国甲状腺结节诊治流程,与S-Detect技术联合使用可提高甲状腺结节的诊断效能。Objective To investigate the diagnostic value of Chinese TI-RADS classification method combined with artificial intelligence S-Detect technology in thyroid nodules. Methods A total of 98 patients(98 thyroid nodules) were evaluated using the Chinese version of TI-RADS classification and S-Detect technique. The postoperative pathological results were taken as the gold standard, and the diagnostic efficacy of C-TIRADS classification, S-Detect and the combination of the two methods was analyzed. Results As confirmed by surgical pathology, 98 patients with 98 thyroid nodules, among which 40 were benign and 58 were malignant. The sensitivity, specificity and accuracy of C-TIRADS combined with S-Detect in the diagnosis of benign and malignant thyroid nodules by sonographers were 95.2%, 82.5% and 91.8%,which was better than that of routine ultrasound and S-Detect alone(P<0.05). Conclusions The Chinese version of TI-RADS classification can be used in combination with the S-Detect technique to improve the diagnostic efficiency of the thyroid nodules, which takes into account the accuracy of the classification and the ease of clinical use. Combined with S-Detect technology, it can improve the diagnostic efficiency of thyroid nodule.

关 键 词：人工智能 计算机辅助诊断 甲状腺结节 超声检查 

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]