Deep transfer learning for detection of breast arterial calcifications on mammograms: a comparative study.

IF 3.7 Q1 RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING
Nazanin Mobini, Davide Capra, Anna Colarieti, Moreno Zanardo, Giuseppe Baselli, Francesco Sardanelli
{"title":"Deep transfer learning for detection of breast arterial calcifications on mammograms: a comparative study.","authors":"Nazanin Mobini, Davide Capra, Anna Colarieti, Moreno Zanardo, Giuseppe Baselli, Francesco Sardanelli","doi":"10.1186/s41747-024-00478-6","DOIUrl":null,"url":null,"abstract":"<p><strong>Introduction: </strong>Breast arterial calcifications (BAC) are common incidental findings on routine mammograms, which have been suggested as a sex-specific biomarker of cardiovascular disease (CVD) risk. Previous work showed the efficacy of a pretrained convolutional network (CNN), VCG16, for automatic BAC detection. In this study, we further tested the method by a comparative analysis with other ten CNNs.</p><p><strong>Material and methods: </strong>Four-view standard mammography exams from 1,493 women were included in this retrospective study and labeled as BAC or non-BAC by experts. The comparative study was conducted using eleven pretrained convolutional networks (CNNs) with varying depths from five architectures including Xception, VGG, ResNetV2, MobileNet, and DenseNet, fine-tuned for the binary BAC classification task. Performance evaluation involved area under the receiver operating characteristics curve (AUC-ROC) analysis, F<sub>1</sub>-score (harmonic mean of precision and recall), and generalized gradient-weighted class activation mapping (Grad-CAM++) for visual explanations.</p><p><strong>Results: </strong>The dataset exhibited a BAC prevalence of 194/1,493 women (13.0%) and 581/5,972 images (9.7%). Among the retrained models, VGG, MobileNet, and DenseNet demonstrated the most promising results, achieving AUC-ROCs > 0.70 in both training and independent testing subsets. In terms of testing F<sub>1</sub>-score, VGG16 ranked first, higher than MobileNet (0.51) and VGG19 (0.46). Qualitative analysis showed that the Grad-CAM++ heatmaps generated by VGG16 consistently outperformed those produced by others, offering a finer-grained and discriminative localization of calcified regions within images.</p><p><strong>Conclusion: </strong>Deep transfer learning showed promise in automated BAC detection on mammograms, where relatively shallow networks demonstrated superior performances requiring shorter training times and reduced resources.</p><p><strong>Relevance statement: </strong>Deep transfer learning is a promising approach to enhance reporting BAC on mammograms and facilitate developing efficient tools for cardiovascular risk stratification in women, leveraging large-scale mammographic screening programs.</p><p><strong>Key points: </strong>• We tested different pretrained convolutional networks (CNNs) for BAC detection on mammograms. • VGG and MobileNet demonstrated promising performances, outperforming their deeper, more complex counterparts. • Visual explanations using Grad-CAM++ highlighted VGG16's superior performance in localizing BAC.</p>","PeriodicalId":36926,"journal":{"name":"European Radiology Experimental","volume":"8 1","pages":"80"},"PeriodicalIF":3.7000,"publicationDate":"2024-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11247067/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"European Radiology Experimental","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1186/s41747-024-00478-6","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"RADIOLOGY, NUCLEAR MEDICINE & MEDICAL IMAGING","Score":null,"Total":0}
引用次数: 0

Abstract

Introduction: Breast arterial calcifications (BAC) are common incidental findings on routine mammograms, which have been suggested as a sex-specific biomarker of cardiovascular disease (CVD) risk. Previous work showed the efficacy of a pretrained convolutional network (CNN), VCG16, for automatic BAC detection. In this study, we further tested the method by a comparative analysis with other ten CNNs.

Material and methods: Four-view standard mammography exams from 1,493 women were included in this retrospective study and labeled as BAC or non-BAC by experts. The comparative study was conducted using eleven pretrained convolutional networks (CNNs) with varying depths from five architectures including Xception, VGG, ResNetV2, MobileNet, and DenseNet, fine-tuned for the binary BAC classification task. Performance evaluation involved area under the receiver operating characteristics curve (AUC-ROC) analysis, F1-score (harmonic mean of precision and recall), and generalized gradient-weighted class activation mapping (Grad-CAM++) for visual explanations.

Results: The dataset exhibited a BAC prevalence of 194/1,493 women (13.0%) and 581/5,972 images (9.7%). Among the retrained models, VGG, MobileNet, and DenseNet demonstrated the most promising results, achieving AUC-ROCs > 0.70 in both training and independent testing subsets. In terms of testing F1-score, VGG16 ranked first, higher than MobileNet (0.51) and VGG19 (0.46). Qualitative analysis showed that the Grad-CAM++ heatmaps generated by VGG16 consistently outperformed those produced by others, offering a finer-grained and discriminative localization of calcified regions within images.

Conclusion: Deep transfer learning showed promise in automated BAC detection on mammograms, where relatively shallow networks demonstrated superior performances requiring shorter training times and reduced resources.

Relevance statement: Deep transfer learning is a promising approach to enhance reporting BAC on mammograms and facilitate developing efficient tools for cardiovascular risk stratification in women, leveraging large-scale mammographic screening programs.

Key points: • We tested different pretrained convolutional networks (CNNs) for BAC detection on mammograms. • VGG and MobileNet demonstrated promising performances, outperforming their deeper, more complex counterparts. • Visual explanations using Grad-CAM++ highlighted VGG16's superior performance in localizing BAC.

Abstract Image

用于检测乳房 X 光照片上乳腺动脉钙化的深度传输学习:一项比较研究。
简介乳房动脉钙化(BAC)是常规乳房 X 光检查中常见的偶然发现,被认为是心血管疾病(CVD)风险的性别特异性生物标志物。之前的研究表明,预训练卷积网络(CNN)VCG16 对自动检测 BAC 非常有效。在本研究中,我们通过与其他十种 CNN 的比较分析进一步测试了该方法:本回顾性研究纳入了 1,493 名女性的四视角标准乳腺 X 光检查结果,并由专家将其标记为 BAC 或非 BAC。比较研究使用了 11 个经过预训练的卷积网络(CNN),这些网络来自 Xception、VGG、ResNetV2、MobileNet 和 DenseNet 等五种架构,深度各不相同,并针对二元 BAC 分类任务进行了微调。性能评估包括接受者操作特征曲线下面积(AUC-ROC)分析、F1-分数(精确度和召回率的调和平均值)以及用于视觉解释的广义梯度加权类激活映射(Grad-CAM++):数据集显示,BAC 发生率为 194/1,493 名女性(13.0%)和 581/5,972 幅图像(9.7%)。在重新训练的模型中,VGG、MobileNet 和 DenseNet 的结果最有希望,在训练和独立测试子集中的 AUC-ROC 均大于 0.70。在测试 F1 分数方面,VGG16 排名第一,高于 MobileNet(0.51)和 VGG19(0.46)。定性分析显示,VGG16 生成的 Grad-CAM++ 热图始终优于其他生成的热图,能对图像中的钙化区域进行更精细、更有辨别力的定位:深度迁移学习在乳房 X 光照片的 BAC 自动检测中大有可为,其中相对较浅的网络表现出了卓越的性能,需要更短的训练时间和更少的资源:深度迁移学习是一种很有前途的方法,它能提高乳房 X 光照片上 BAC 的报告率,并有助于开发高效的工具,利用大规模乳房 X 光照片筛查计划对女性进行心血管风险分层:- 我们测试了不同的预训练卷积网络 (CNN),以检测乳房 X 光照片上的 BAC。- VGG和MobileNet表现出了良好的性能,超过了更深、更复杂的同类产品。- 使用 Grad-CAM++ 进行的可视化解释凸显了 VGG16 在定位 BAC 方面的卓越性能。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
European Radiology Experimental
European Radiology Experimental Medicine-Radiology, Nuclear Medicine and Imaging
CiteScore
6.70
自引率
2.60%
发文量
56
审稿时长
18 weeks
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信