人工智能在乳腺钼靶X射线摄影诊断乳腺疾病中的应用价值分析  

Analysis of the Application Value of Artificial Intelligence in Mammography Diagnosis of Breast Diseases

作　　者：徐佳 XU Jia(Jiangyin Traditional Chinese Medicine Hospital,Jiangsu Jiangyin 214400)

机构地区：[1]江阴市中医院,江苏江阴214400

出　　处：《中国医疗器械信息》2025年第5期89-91,共3页China Medical Device Information

摘　　要：目的:探讨在乳腺钼靶X射线摄影诊断乳腺疾病中应用人工智能的效果。方法:回顾性分析2022年1月~2024年10月本院收治的行乳腺钼靶X射线摄影检查的乳腺疾病患者215例的临床资料,所有患者均分别进行人工智能诊断与放射科主治医师诊断,将穿刺活检病理诊断结果作为“金标准”,比较人工智能诊断、放射科主治医师诊断与联合诊断乳腺疾病的灵敏度、特异度与准确度,并评价人工智能诊断与放射科主治医师诊断对病变特征的一致性。结果:经穿刺活检病理诊断显示,215例患者中恶性病变102例,良性病变113例;联合诊断的灵敏度、特异度、准确度高于人工智能诊断(P<0.05);人工智能诊断特异度与主治医师诊断比较,差异无统计学意义(P>0.05);经Kappa系数计算,人工智能诊断与放射科主治医师诊断对病变特征Kappa系数为0.637,95%CI为0.131~0.795(P<0.05)。结论:人工智能在乳腺钼靶X射线摄影中的应用具有重要价值,联合诊断可最大化地发挥人工智能与临床医师的互补作用。Objective:To explore the value of artificial intelligence in the diagnosis of breast diseases using mammography.Methods:A retrospective analysis was conducted on the clinical data of 215 breast disease patients who underwent mammography in our hospital from January 2022 to October 2024.All patients were diagnosed by artificial intelligence and radiologist respectively,and the pathological diagnosis results of biopsy were used as the“gold standard”.The sensitivity,specificity,and accuracy of artificial intelligence diagnosis,radiologist diagnosis,and combined diagnosis of breast disease were compared,and the consistency of lesion characteristics between artificial intelligence diagnosis and radiologist diagnosis was evaluated.Results:Pathological diagnosis through biopsy showed that out of 215 patients,102 had malignant lesions and 113 had benign lesions;The sensitivity,specificity,and accuracy of joint diagnosis are higher than those of artificial intelligence diagnosis(P<0.05);There was no statistically significant difference(P>0.05)between the diagnostic specificity of artificial intelligence and that of the attending physician;According to the Kappa coefficient calculation,the Kappa coefficient of artificial intelligence diagnosis and radiology attending physician diagnosis for lesion characteristics is 0.637,with a 95%CI of 0.131-0.795 and P<0.05.Conclusion:The application of artificial intelligence in mammography has significant value,and combined diagnosis can maximize the complementary role of artificial intelligence and clinical physicians.

关 键 词：人工智能 乳腺钼靶 X射线摄影 乳腺疾病 诊断 

分 类 号：R737.9[医药卫生—肿瘤]