Using a convolutional neural network to predict readers' estimates of mammographic density for breast cancer risk assessment

Background: Mammographic density is an important risk factor for breast cancer. Recent research demonstrated that percentage density assessed visually using Visual Analogue Scales (VAS) showed stronger risk prediction than existing automated density measures, suggesting readers may recognise relevant image features not yet captured by automated methods. Method: We have built convolutional neural networks (CNN) to predict VAS scores from full-field digital mammograms. The CNNs are trained using whole-image mammograms, each labelled with the average VAS score of two independent readers. They learn a mapping between mammographic appearance and VAS score so that at test time, they can predict VAS score for an unseen image. Networks were trained using 67520 mammographic images from 16968 women, and tested on a large dataset of 73128 images and case-control sets of contralateral mammograms of screen detected cancers and prior images of women with cancers detected subsequently, matched to controls on age, menopausal status, parity, HRT and BMI. Results: Pearson's correlation coefficient between readers' and predicted VAS in the large dataset was 0.79 per mammogram and 0.83 per woman (averaging over all views). In the case-control sets, odds ratios of cancer in the highest vs lowest quintile of percentage density were 3.07 (95%CI: 1.97 - 4.77) for the screen detected cancers and 3.52 (2.22 - 5.58) for the priors, with matched concordance indices of 0.59 (0.55 - 0.64) and 0.61 (0.58 - 0.65) respectively. Conclusion: Our fully automated method demonstrated encouraging results which compare well with existing methods, including VAS.

[1]  Trevor Darrell,et al.  Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation , 2013, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[2]  C. D'Orsi.  Breast Imaging Reporting and Data System (BI-RADS) , 2018 .

[3]  D. Shen,et al.  Computer-Aided Diagnosis with Deep Learning Architecture: Applications to Breast Lesions in US Images and Pulmonary Nodules in CT Scans , 2016, Scientific Reports.

[4]  Jingmei Li,et al.  Digital mammographic density and breast cancer risk: a case–control study of six alternative density assessment methods , 2014, Breast Cancer Research.

[5]  Karolin Baecker,et al.  Two Dimensional Signal And Image Processing , 2016 .

[6]  N. Boyd,et al.  The quantitative analysis of mammographic densities. , 1994, Physics in medicine and biology.

[7]  Douglas G. Altman,et al.  Measurement in Medicine: The Analysis of Method Comparison Studies , 1983 .

[8]  Susan M. Astley,et al.  Volumetric and Area-Based Breast Density Measurement in the Predicting Risk of Cancer at Screening (PROCAS) Study , 2012, Digital Mammography / IWDM.

[9]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[10]  Shivani Pahwa,et al.  Evaluation of breast parenchymal density with QUANTRA software , 2015, Indian Journal of Radiology and Imaging.

[11]  P. Narula.  MAMMOGRAPHIC DENSITY AND THE RISK AND DETECTION OF BREAST CANCER , 2016 .

[12]  A. Ng,et al.  Breast Density Scoring with Multiscale Denoising Autoencoders , 2012 .

[13]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[14]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[15]  Susan M. Astley,et al.  Same task, same observers, different values: the problem with visual assessment of breast density , 2013, Medical Imaging.

[16]  J. Hopper,et al.  Mammographic density—a review on the current understanding of its association with breast cancer , 2014, Breast Cancer Research and Treatment.

[17]  Bram van Ginneken,et al.  A survey on deep learning in medical image analysis , 2017, Medical Image Anal..

[18]  Hongmin Cai,et al.  Discrimination of Breast Cancer with Microcalcifications on Mammography by Deep Learning , 2016, Scientific Reports.

[19]  Gustavo Carneiro,et al.  Deep Learning and Structured Prediction for the Segmentation of Mass in Mammograms , 2015, MICCAI.

[20]  Jack Cuzick,et al.  A novel and fully automated mammographic texture analysis for risk prediction: results from two case-control studies , 2017, Breast Cancer Research.

[21]  Gustavo Carneiro,et al.  Automated Mass Detection in Mammograms Using Cascaded Deep Learning and Random Forests , 2015, 2015 International Conference on Digital Image Computing: Techniques and Applications (DICTA).

[22]  Mary Wilson,et al.  Visually assessed breast density, breast cancer risk and the importance of the craniocaudal view , 2008, Breast Cancer Research.

[23]  Ron Kimmel,et al.  Computational mammography using deep neural networks , 2018, Comput. methods Biomech. Biomed. Eng. Imaging Vis..

[24]  Mary Wilson,et al.  Assessing Individual Breast Cancer Risk within the U.K. National Health Service Breast Screening Program: A New Paradigm for Cancer Prevention , 2012, Cancer Prevention Research.

[25]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[26]  Yahong Luo,et al.  Understanding Clinical Mammographic Breast Density Assessment: a Deep Learning Perspective , 2018, Journal of Digital Imaging.

[27]  Stephen W Duffy,et al.  A concordance index for matched case–control studies with applications in cancer risk , 2015, Statistics in medicine.

[28]  Susan M. Astley,et al.  A comparison of five methods of measuring mammographic density: a case-control study , 2018, Breast Cancer Research.

[29]  Yahong Luo,et al.  A deep learning method for classifying mammographic breast density categories , 2018, Medical physics.

[30]  Susan M. Astley,et al.  Mammographic density adds accuracy to both the Tyrer-Cuzick and Gail breast cancer risk models in a prospective UK screening cohort , 2015, Breast Cancer Research.

[31]  Nico Karssemeijer,et al.  Robust Breast Composition Measurement - VolparaTM , 2010, Digital Mammography / IWDM.

[32]  Geoffrey E. Hinton,et al.  Rectified Linear Units Improve Restricted Boltzmann Machines , 2010, ICML.

[33]  Nico Karssemeijer,et al.  Unsupervised Deep Learning Applied to Breast Density Segmentation and Mammographic Risk Scoring , 2016, IEEE Transactions on Medical Imaging.