Watanabe Alyssa T, Dib Valerie, Wang Junhao, Mantey Richard, Daughton William, Chim Chi Yung, Eckel Gregory, Moss Caroline, Goel Vinay, Nerlekar Nitesh
Department of Radiology, University of Southern California Keck School of Medicine, Los Angeles, CA, USA.
CureMetrix Inc, San Diego, CA, USA.
J Breast Imaging. 2025 Mar 18;7(2):168-176. doi: 10.1093/jbi/wbae064.
The performance of a commercially available artificial intelligence (AI)-based software that detects breast arterial calcifications (BACs) on mammograms is presented.
This retrospective study was exempt from IRB approval and adhered to the HIPAA regulations. Breast arterial calcification detection using AI was assessed in 253 patients who underwent 314 digital mammography (DM) examinations and 143 patients who underwent 277 digital breast tomosynthesis (DBT) examinations between October 2004 and September 2022. Artificial intelligence performance for binary BAC detection was compared with ground truth (GT) determined by the majority consensus of breast imaging radiologists. Area under the receiver operating curve (AUC), sensitivity, specificity, positive predictive value and negative predictive value (NPV), accuracy, and BAC prevalence rates of the AI algorithm were compared.
The case-level AUCs of AI were 0.96 (0.93-0.98) for DM and 0.95 (0.92-0.98) for DBT. Sensitivity, specificity, and accuracy were 87% (79%-93%), 92% (88%-96%), and 91% (87%-94%) for DM and 88% (80%-94%), 90% (84%-94%), and 89% (85%-92%) for DBT. Positive predictive value and NPV were 82% (72%-89%) and 95% (92%-97%) for DM and 84% (76%-90%) and 92% (88%-96%) for DBT, respectively. Results are 95% confidence intervals. Breast arterial calcification prevalence was similar for both AI and GT assessments.
Breast AI software for detection of BAC presence on mammograms showed promising performance for both DM and DBT examinations. Artificial intelligence has potential to aid radiologists in detection and reporting of BAC on mammograms, which is a known cardiovascular risk marker specific to women.
介绍一种基于人工智能（AI）的市售软件在乳腺钼靶片上检测乳腺动脉钙化（BAC）的性能。
本回顾性研究无需获得机构审查委员会（IRB）批准，并遵循了《健康保险流通与责任法案》（HIPAA）的规定。对2004年10月至2022年9月期间接受314次数字乳腺钼靶（DM）检查的253例患者以及接受277次数字乳腺断层合成（DBT）检查的143例患者，评估使用AI检测乳腺动脉钙化的情况。将AI对二元BAC检测的性能与由乳腺影像放射科医生的多数共识确定的真实情况（GT）进行比较。比较AI算法的受试者操作特征曲线下面积（AUC）、灵敏度、特异度、阳性预测值和阴性预测值（NPV）、准确度以及BAC患病率。
AI在DM中的病例级AUC为0.96（0.93 - 0.98），在DBT中为0.95（0.92 - 0.98）。DM的灵敏度、特异度和准确度分别为87%（79% - 93%）、92%（88% - 96%）和91%（87% - 94%），DBT的分别为88%（80% - 94%）、90%（84% - 94%）和89%（85% - 92%）。DM的阳性预测值和NPV分别为82%（72% - 89%）和95%（92% - 97%），DBT的分别为84%（76% - 90%）和92%（88% - 96%）。结果为95%置信区间。AI和GT评估的乳腺动脉钙化患病率相似。
用于在乳腺钼靶片上检测BAC存在的乳腺AI软件在DM和DBT检查中均表现出良好的性能。人工智能有潜力帮助放射科医生在乳腺钼靶片上检测和报告BAC，BAC是一种已知的女性特有的心血管风险标志物。
