人工智能S-Detect技术联合BI-RADS分类及Adler分级法在乳腺肿块诊断中的价值  

Value of Artificial Intelligence S-Detect Technique Combined with BI-RADS Classification and Adler Classification in Diagnosis of Breast Masses

作　　者：李如冰 彭梅[1] 毕玉[1] 詹韵韵[1] 姜凡[1] 方明娣[1] LI Ru-bing;PENG Mei;BI Yu;ZHAN Yun-yun;JIANG Fan;FANG Ming-di(Department of Ultrasound Diagnosis,the Second Hospital of Anhui Medical University,Hefei 230601,Anhui,China)

机构地区：[1]安徽医科大学第二附属医院超声诊断科,安徽合肥230601

出　　处：《医学信息》2022年第20期32-36,共5页Journal of Medical Information

基　　金：安徽医科大学校科研基金项目(编号:2021xkj035)。

摘　　要：目的探讨人工智能S-Detect技术联合BI-RADS分类及Adler分级法诊断乳腺肿块中的价值。方法收集2019年9月-2021年7月在安徽医科大学第二附属医院超声诊断科行超声检查的121例乳腺肿块患者的超声图像资料,分析BI-RADS分类、S-Detect及Adler分级法诊断结果,根据S-Detect及Adler分级法诊断结果将每个肿块的分类进行联合诊断,以病理结果为金标准,比较BI-RADS分类、BI-RADS联合S-Detect技术、BI-RADS联合S-Detect技术及Adler分级法诊断乳腺肿块的灵敏度、特异度、准确度,采用受试者操作曲线(ROC)分析三种诊断方法的价值。结果病理结果显示,121例乳腺肿块患者恶性65例,良性56例。BI-RADS分类、BI-RADS联合S-Detect技术、BI-RADS联合S-Detect技术及Adler分级法的AUC分别为0.644、0.663、0.823;BI-RADS分类诊断乳腺肿块的灵敏度为98.46%、特异度为30.36%、准确度为66.94%;BI-RADS联合S-Detect诊断乳腺肿块的灵敏度为96.92%,特异度为35.71%,准确度为68.60%;BI-RADS联合S-Detect及Adler分级法诊断乳腺肿块的灵敏度为98.46%,特异度为66.07%,准确度为83.47%。结论人工智能S-Detect技术联合BI-RADS分类及Adler分级法可进一步提升对乳腺肿块的诊断效能,对临床干预与治疗具有重要的指导意义。Objective To explore the value of artificial intelligence S-Detect technology combined with BI-RADS classification and Adler classification in the diagnosis of breast masses.Methods The ultrasound image data of 121 patients with breast masses who underwent ultrasound examination in the Department of Ultrasound Diagnosis,the Second Hospital of Anhui Medical University from September 2019 to July 2021 were collected.The diagnostic results of BI-RADS classification,S-Detect and Adler classification were analyzed.According to the diagnostic results of S-Detect and Adler classification,the classification of each mass was jointly diagnosed,and the pathological results were used as the gold standard.The sensitivity,specificity and accuracy of BI-RADS classification,BI-RADS combined with S-Detect technology,BI-RADS combined with S-Detect technology and Adler grading method in the diagnosis of breast masses were compared.The value of the three diagnostic methods was analyzed by receiver operating curve(ROC).Results The pathological results showed that 65 cases were malignant and 56 cases were benign in 121 patients with breast masses.The AUC of BI-RADS classification,BI-RADS combined with S-Detect technology,BI-RADS combined with S-Detect technology and Adler grading method were 0.644,0.663 and 0.823,respectively.The sensitivity,specificity and accuracy of BI-RADS classification in the diagnosis of breast masses were 98.46%,30.36%and 66.94%,respectively.The sensitivity,specificity and accuracy of BI-RADS combined with S-Detect in the diagnosis of breast masses were 96.92%,35.71%and 68.60%,respectively.The sensitivity,specificity and accuracy of BI-RADS combined with S-Detect and Adler grading in the diagnosis of breast masses were 98.46%,66.07%and 83.47%,respectively.Conclusion Artificial intelligence S-Detect technology combined with BI-RADS classification and Adler classification can further improve the diagnostic efficiency of breast masses,and has important guiding significance for clinical intervention and treatment.

关 键 词：S-Detect技术 彩色多普勒血流显像 乳腺影像报告数据系统 乳腺肿块 

分 类 号：R445.1[医药卫生—影像医学与核医学]