甲状腺结节的超声人工智能诊断  

Artificial Intelligence Diagnosis of Thyroid Nodules by Ultrasound

作　　者：王丹[1] 花瞻[2] 武敬平[3] 邓周录[2] 赵玉珍[3] 臧树良[3] 李艳 李程[3] Wang Dan;Hua Zhan;Wu Jingping;Deng Zhoulu;Zhao Yuzhen;Zang Shuliang;Li Yan;Li Cheng(College of Basic Medicine,Beijing University of Chinese Medicine,Beijing 100029,China;Department of Surgical,China-Japan Friendship Hospital,Beijing 100029,China;Department of Nuclear Medicine,Xiyuan Hospital,Beijing 100091,China)

机构地区：[1]北京中医药大学基础医学院,北京市100029 [2]中日友好医院普外科 [3]中日友好医院超声诊断科 [4]西苑医院核医学科

出　　处：《中国超声医学杂志》2019年第12期1070-1072,共3页Chinese Journal of Ultrasound in Medicine

摘　　要：目的检测甲状腺结节超声人工智能设备的灵敏度与特异度,评价甲状腺结节的超声人工智能诊断的价值。方法从研究数据库中随机抽取600张甲状腺结节超声图像,其中良性结节300张,恶性结节300张,运用甲状腺结节超声人工智能设备对其进行诊断,对照相应的病理结果,进行诊断结果进行统计分析。结果甲状腺结节的超声人工智能设备正确识别恶性结节超声图像数据库中图像256张,识别错误44张;正确识别良性结节超声图像数据库中图像259张,识别错误41张,灵敏度为86.20%、特异度为85.48%,准确性较高。结论本研究的超声人工智能设备,对于甲状腺结节的临床诊断具有重要价值,超声人工智能技术值得深入研究和推广。Objective To evaluate the sensitivity and specificity of ultrasonic artificial intelligence device for thyroid nodule,and to evaluate the value of ultrasonic artificial intelligence diagnosis of thyroid nodules.Methods 600 ultrasound images of thyroid nodules were randomly selected from the research database,including 300 benign nodules and 300 malignant nodules.The ultrasonic artificial intelligence device for thyroid nodule was used to diagnose it,and the diagnostic results were statistically analyzed according to the corresponding pathological results.Results Ultrasound artificial intelligence device for thyroid nodules correctly identified 256 images in the malignant nodule ultrasound image database,identifying 44 errors;correctly identifying 259 images in the benign nodule ultrasound image database,identifying 41 errors,the sensitivity reached 86.20%,The specificity was 85.48%,and the accuracy was high.Conclusions The ultrasonic artificial intelligence device of this research was of great value for the clinical diagnosis of thyroid nodules.Ultrasonic artificial intelligence technology deserves in-depth research and promotion.

关 键 词：甲状腺结节 超声人工智能 诊断 

分 类 号：R445.1[医药卫生—影像医学与核医学] R736.1[医药卫生—诊断学] TP18[医药卫生—临床医学]