人工智能对比增强乳腺摄影诊断乳腺癌的验证:与组织病理学结果的关系
研究利用CDD-CESM数据库中326例女性的2006幅图像训练深度学习多视图对比增强乳腺摄影模型,并在37例女性288幅外部测试集上验证,以组织病理学和随访为标准。AI的AUC-ROC为0.936,最佳截断值AI评分=0.28时敏感度75%、特异度96.3%、准确率90.1%;放射科医生敏感度88.9%、特异度92.6%、准确率91.7%。作者认为模型在乳腺癌检测中结果有前景,但需进一步研究、外部训练和验证。
要点
- AI模型在外部测试集上AUC-ROC为0.936,最佳截断值为AI评分0.28。
- AI敏感度75%、特异度96.3%、准确率90.1%,PPV 87.1%、NPV 92%。
- 放射科医生敏感度88.9%、特异度92.6%、准确率91.7%,PPV 80%、NPV 96.2%。
核心结论
- 深度学习多视图对比增强乳腺摄影模型在女性队列中显示出检测乳腺癌的有前景结果,但需进一步研究、外部训练和验证。
研究设计与样本
- 研究利用已发表的CDD-CESM数据库中326例女性的2006幅全标注图像训练深度学习多视图对比增强乳腺摄影模型(MVCM),用于分类CEM低能量图和重组图。
诊断标准与关键数值
- 外部测试集为37例女性(平均年龄51.31岁±11.07)的288幅图像,AI的AUC-ROC为0.936(95%CI:0.898,0.973;P<0.001),预测恶性的最佳截断值为AI评分=0.28。
局限性
- 原文结论指出需进一步研究、外部训练和验证。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
INTRODUCTION: Contrast-enhanced mammography (CEM) is used for characterization of breast lesions with increased diagnostic accuracy compared to digital mammography (DM). Artificial intelligence (AI) approaches are emerging with accuracies equal to an average radiologist. However, most studies trained deep learning (DL) models on DM images and there is a paucity in literature for discovering the application of AI using CEM.
OBJECTIVES: To develop and test a DL model that classifies CEM images and produces corresponding highlights of lesions detected.
METHODS: Fully annotated 2006 images of 326 females available from the previously published Categorized Digital Database for Contrast Enhanced Mammography images (CDD-CESM) were used for training. We developed a DL multiview contrast mammography model (MVCM) for classification of CEM low energy and recombined images. An external test set of 288 images of 37 females not included in the training was used for validation. Correlation with histopathological results and follow-up was considered the standard reference. The study protocol was approved by the Institutional Review Board and patient informed consent was obtained.
RESULTS: Assessment was done on an external test set of 37 females (mean age, 51.31 years ± 11.07 [SD]) with AUC-ROC for AI performance 0.936; (95 % CI: 0.898, 0.973; p < 0.001) and the best cut off value for prediction of malignancy using AI score = 0.28. Findings were then correlated with histopathological results and follow up which revealed a sensitivity of 75 %, specificity 96.3 %, total accuracy of 90.1 %, positive predictive value (PPV) 87.1 %, and negative predictive value (NPV) 92 %, p-value (<0.001). Diagnostic indices of radiologists were sensitivity 88.9 %, specificity 92.6 %, total accuracy 91.7 %, PPV 80 %, and NPV 96.2 %, p-value (<0.001).
CONCLUSION: A deep learning multiview CEM model was developed and evaluated in a cohort of female participants and showed promising results in detecting breast cancer. This warrants further studies, external training, and validation.
引用
- Helal M, Khaled R, Alfarghaly O, et al. Validation of artificial intelligence contrast mammography in diagnosis of breast cancer: Relationship to histopathological results. Eur J Radiol. 2024. PMID: 38428255. doi:10.1016/j.ejrad.2024.111392
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。