DU-Net: Convolutional Network for the Detection of Arterial Calcifications in Mammograms

Breast arterial calcifications (BACs) are part of several benign findings present on some mammograms. Previous studies have indicated that BAC may provide evidence of general atherosclerotic vascular disease, and potentially be a useful marker of cardiovascular disease (CVD). Currently, there is no technique in use for the automatic detection of BAC in mammograms. Since a majority of women over the age of 40 already undergo breast cancer screening with mammography, detecting BAC may offer a method to screen women for CVD in a way that is effective, efficient, and broad reaching, at no additional cost or radiation. In this paper, we present a deep learning approach for detecting BACs in mammograms. Inspired by the promising results achieved using the U-Net model in many biomedical segmentation problems and the DenseNet in semantic segmentation, we extend the U-Net model with dense connectivity to automatically detect BACs in mammograms. The presented model helps to facilitate the reuse of computation and improve the flow of gradients, leading to better accuracy and easier training of the model. We evaluate the performance using a set of full-field digital mammograms collected and prepared for this task from a publicly available dataset. Experimental results demonstrate that the presented model outperforms human experts as well as the other related deep learning models. This confirms the effectiveness of our model in the BACs detection task, which is a promising step in providing a cost-effective risk assessment tool for CVD.

[1]  Patrick van der Smagt,et al.  CNN-based Segmentation of Medical Imaging Data , 2017, ArXiv.

[2]  Héctor D. Menéndez,et al.  Mimicking Anti-Viruses with Machine Learning and Entropy Profiles , 2019, Entropy.

[3]  Yoshua Bengio,et al.  FitNets: Hints for Thin Deep Nets , 2014, ICLR.

[4]  Carlos Iribarren,et al.  Breast vascular calcification and risk of coronary heart disease, stroke, and heart failure. , 2004, Journal of women's health.

[5]  Christopher Joseph Pal,et al.  Brain tumor segmentation with Deep Neural Networks , 2015, Medical Image Anal..

[6]  Dinggang Shen,et al.  Automated Delineation of Calcified Vessels in Mammography by Tracking With Uncertainty and Graphical Linking Techniques , 2012, IEEE Transactions on Medical Imaging.

[7]  Ross B. Girshick,et al.  Focal Loss for Dense Object Detection , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  N. Karssemeijer,et al.  Reducing false positives of microcalcification detection systems by removal of breast arterial calcifications. , 2016, Medical physics.

[9]  Simon K. Warfield,et al.  Asymmetric Loss Functions and Deep Densely-Connected Networks for Highly-Imbalanced Medical Image Segmentation: Application to Multiple Sclerosis Lesion Detection , 2018, IEEE Access.

[10]  Dumitru Erhan,et al.  Going deeper with convolutions , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[11]  Daniel L Rubin,et al.  A curated mammography data set for use in computer-aided detection and diagnosis research , 2017, Scientific Data.

[12]  Steven Guan,et al.  Fully Dense UNet for 2-D Sparse Photoacoustic Tomography Artifact Removal , 2018, IEEE Journal of Biomedical and Health Informatics.

[13]  Liang Chen,et al.  DRINet for Medical Image Segmentation , 2018, IEEE Transactions on Medical Imaging.

[14]  Christopher Joseph Pal,et al.  The Importance of Skip Connections in Biomedical Image Segmentation , 2016, LABELS/DLMIA@MICCAI.

[15]  C. Comstock,et al.  Intramammary arterial calcifications associated with diabetes. , 1980, Radiology.

[16]  David M. W. Powers,et al.  Evaluation: from precision, recall and F-measure to ROC, informedness, markedness and correlation , 2011, ArXiv.

[17]  P. Slanetz,et al.  Do breast arterial calcifications on mammography predict elevated risk of coronary artery disease? , 2016, European journal of radiology.

[18]  Léon Bottou,et al.  Stochastic Gradient Descent Tricks , 2012, Neural Networks: Tricks of the Trade.

[19]  Q. Bui,et al.  A Review of the Role of Breast Arterial Calcification for Cardiovascular Risk Stratification in Women , 2019, Circulation.

[20]  Femke Atsma,et al.  Breast arterial calcifications are correlated with subsequent development of coronary artery calcifications, but their aetiology is predominantly different. , 2007, European journal of radiology.

[21]  Wei Li,et al.  Multi-scale Feature Fusion Single Shot Object Detector Based on DenseNet , 2019, ICIRA.

[22]  J. Han,et al.  Prediction of Subclinical Coronary Artery Disease With Breast Arterial Calcification and Low Bone Mass in Asymptomatic Women: Registry for the Women Health Cohort for the BBC Study. , 2018, JACC. Cardiovascular imaging.

[23]  Vijayan K. Asari,et al.  Recurrent Residual Convolutional Neural Network based on U-Net (R2U-Net) for Medical Image Segmentation , 2018, ArXiv.

[24]  Dinggang Shen,et al.  Identification of breast vascular calcium deposition in digital mammography by linear structure analysis , 2012, 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI).

[25]  Nima Tajbakhsh,et al.  UNet++: A Nested U-Net Architecture for Medical Image Segmentation , 2018, DLMIA/ML-CDS@MICCAI.

[26]  Trevor Darrell,et al.  Fully Convolutional Networks for Semantic Segmentation , 2017, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[27]  D. Chakraborty,et al.  Free-response methodology: alternate analysis and a new observer-performance experiment. , 1990, Radiology.

[28]  Sabee Molloi,et al.  Detecting Cardiovascular Disease from Mammograms With Deep Learning , 2017, IEEE Transactions on Medical Imaging.

[29]  Fabrizio Angiulli,et al.  Prototype-Based Domain Description for One-Class Classification , 2012, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[30]  Dinggang Shen,et al.  Detection of Arterial Calcification in Mammograms by Random Walks , 2009, IPMI.

[31]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[32]  Andrew Zisserman,et al.  Convolutional Two-Stream Network Fusion for Video Action Recognition , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[33]  C. Henschke,et al.  Digital Mammography and Screening for Coronary Artery Disease. , 2016, JACC. Cardiovascular imaging.

[34]  Mingqi Li,et al.  Semi-supervised Transfer Learning for Convolutional Neural Networks for Glaucoma Detection , 2019, ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[35]  Guoyan Zheng,et al.  3D U-net with Multi-level Deep Supervision: Fully Automatic Segmentation of Proximal Femur in 3D MR Images , 2017, MLMI@MICCAI.

[36]  J. Szejnfeld,et al.  Correlation between intramammary arterial calcifications and CAD. , 2007, Academic radiology.

[37]  Richard H. Moore,et al.  THE DIGITAL DATABASE FOR SCREENING MAMMOGRAPHY , 2007 .

[38]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[39]  Yoshua Bengio,et al.  The One Hundred Layers Tiramisu: Fully Convolutional DenseNets for Semantic Segmentation , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW).

[40]  Berkman Sahiner,et al.  Automated detection of breast vascular calcification on full-field digital mammograms , 2008, SPIE Medical Imaging.

[41]  B. Matthews.  Comparison of the predicted and observed secondary structure of T4 phage lysozyme. , 1975, Biochimica et biophysica acta.

[42]  Yoshua Bengio,et al.  Understanding the difficulty of training deep feedforward neural networks , 2010, AISTATS.

[43]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[44]  R. Vliegenthart,et al.  Mammographic detection of breast arterial calcification as an independent predictor of coronary atherosclerotic disease in a single ethnic cohort of African American women. , 2015, Atherosclerosis.

[45]  Yanchun Zhang,et al.  MDU-Net: multi-scale densely connected U-Net for biomedical image segmentation , 2018, Health Information Science and Systems.

[46]  Chi-Wing Fu,et al.  H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tumor Segmentation From CT Volumes , 2018, IEEE Transactions on Medical Imaging.