A novel deep learning architecture outperforming ‘off‑the‑shelf’ transfer learning and feature‑based methods in the automated assessment of mammographic breast density

Trivizakis,Eleftherios; Ioannidis,Georgios   S.; Melissianos,Vasileios   D.; Papadakis,Georgios   Z.; Tsatsakis,Aristidis; Spandidos,Demetrios   A.; Marias,Kostas

doi:10.3892/or.2019.7312

A novel deep learning architecture outperforming ‘off‑the‑shelf’ transfer learning and feature‑based methods in the automated assessment of mammographic breast density

Authors:
- Eleftherios Trivizakis
- Georgios S. Ioannidis
- Vasileios D. Melissianos
- Georgios Z. Papadakis
- Aristidis Tsatsakis
- Demetrios A. Spandidos
- Kostas Marias
View Affiliations

Affiliations: Computational BioMedicine Laboratory (CBML), Institute of Computer Science (ICS), Foundation for Research and Technology‑Hellas (FORTH), 70013 Heraklion, Greece, Department of Forensic Sciences and Laboratory of Toxicology, Medical School, University of Crete, 70013 Heraklion, Greece, Laboratory of Clinical Virology, Medical School, University of Crete, 70013 Heraklion, Greece
Published online on: September 12, 2019 https://doi.org/10.3892/or.2019.7312
Pages: 2009-2015
Copyright: © Trivizakis et al. This is an open access article distributed under the terms of Creative Commons Attribution License.

Metrics: Total Views: 0 (Spandidos Publications: | PMC Statistics: )

Metrics: Total PDF Downloads: 0 (Spandidos Publications: | PMC Statistics: )

Cited By (CrossRef): 0 citations Loading Articles...

Abstract

Potentially suspicious breast neoplasms could be masked by high tissue density, thus increasing the probability of a false‑negative diagnosis. Furthermore, differentiating breast tissue type enables patient pre‑screening stratification and risk assessment. In this study, we propose and evaluate advanced machine learning methodologies aiming at an objective and reliable method for breast density scoring from routine mammographic images. The proposed image analysis pipeline incorporates texture [Gabor filters and local binary pattern (LBP)] and gradient‑based features [histogram of oriented gradients (HOG) as well as speeded‑up robust features (SURF)]. Additionally, transfer learning approaches with ImageNet trained weights were also used for comparison, as well as a convolutional neural network (CNN). The proposed CNN model was fully trained on two open mammography datasets and was found to be the optimal performing methodology (AUC up to 87.3%). Thus, the findings of this study indicate that automated density scoring in mammograms can aid clinical diagnosis by introducing artificial intelligence‑powered decision‑support systems and contribute to the ‘democratization’ of healthcare by overcoming limitations, such as the geographic location of patients or the lack of expert radiologists.

View Figures

View References

1	Duffy SW, Morrish OWE, Allgood PC, Black R, Gillan MGC, Willsher P, Cooke J, Duncan KA, Michell MJ, Dobson HM, et al: Mammographic density and breast cancer risk in breast screening assessment cases and women with a family history of breast cancer. Eur J Cancer. 88:48–56. 2018. View Article : Google Scholar : PubMed/NCBI
2	Titus-Ernstoff L, Tosteson AN, Kasales C, Weiss J, Goodrich M, Hatch EE and Carney PA: Breast cancer risk factors in relation to breast density (United States). Cancer Causes Control. 17:1281–1290. 2006. View Article : Google Scholar : PubMed/NCBI
3	Checka CM, Chun JE, Schnabel FR, Lee J and Toth H: The relationship of mammographic density and age: Implications for breast cancer screening. AJR Am J Roentgenol. 198:W292–W295. 2012. View Article : Google Scholar : PubMed/NCBI
4	Byrne C, Schairer C, Wolfe J, Parekh N, Salane M, Brinton LA, Hoover R and Haile R: Mammographic features and breast cancer risk: Effects with time, age, and menopause status. J Natl Cancer Inst. 87:1622–1629. 1995. View Article : Google Scholar : PubMed/NCBI
5	D'Orsi CJ, Sickles EA, Mendelson EB, Morris EA, et al: ACR BI-RADS^® Atlas, Breast Imaging Reporting and Data SystemReston, VA: American College of Radiology; 2013
6	Bovis K and Singh S: Classification of mammographic breast density using a combined classifier paradigm. Proceedings of the 4th International Workshop on Digital Mammography. 177–180. 2002.
7	Tzikopoulos S, Georgiou H, Mavroforakis M and Theodoridis S: A fully automated scheme for breast density estimation and asymmetry detection of mammograms. Eur Signal Process Conf. 1869–1873. 2009.
8	Oliver A, Freixenet J and Zwiggelaar R: Automatic classification of breast densityProceedings of the IEEE International Conference on Image Processing. IEEE; pp. II–1258. 2005
9	Fonseca P, Mendoza J, Wainer J, Ferrer J, Pinto J, Guerrero J and Castaneda B: Automatic breast density classification using a convolutional neural network architecture search procedureProceedings of the Medical Imaging 2015. Computer-Aided Diagnosis. 9414. SPIE Medical Imaging; Orlando, FL: 2015
10	Petersen K, Chernoff K, Nielsen M and Ng AY: Breast density scoring with multiscale denoising autoencoders. Proceedings of the MICCAI Workshop on Sparsity Techniques in Medical Imaging. 2012.
11	Kallenberg M, Petersen K, Nielsen M, Ng AY, Pengfei Diao, Igel C, Vachon CM, Holland K, Winkel RR, Karssemeijer N and Lillholm M: Unsupervised deep learning applied to breast density segmentation and mammographic risk scoring. IEEE Trans Med Imaging. 35:1322–1331. 2016. View Article : Google Scholar : PubMed/NCBI
12	Dalal N and Triggs B: Histograms of Oriented Gradients for Human Detection. Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05). 1. IEEE; pp. 886–893. 2005, View Article : Google Scholar
13	Bay H, Tuytelaars T and Van Gool L: SURF: Speeded Up Robust FeaturesSpringer; Berlin, Heidelberg: pp. 404–417. 2006
14	Ojala T, Pietikainen M and Harwood D: Performance evaluation of texture measures with classification based on Kullback discrimination of distributions. Proceedings of 12th International Conference on Pattern Recognition. 1. IEEE; pp. 582–585. 1994, View Article : Google Scholar
15	Fogel I and Sagi D: Gabor filters as texture discriminator. Biol Cybern. 61:103–113. 1989. View Article : Google Scholar
16	Deng J, Dong W, Socher R, Li LJ, Li K and Fei-Fei L: ImageNet: A Large-Scale Hierarchical Image DatabaseProceedings of 2009 IEEE Conference on Computer Vision and Pattern Recognition. IEEE; 2009, View Article : Google Scholar
17	Otsu N: A Threshold selection method from gray-level histograms. IEEE Trans Syst Man Cybern. 9:62–66. 1979. View Article : Google Scholar
18	Ma WY and Manjunath BS: EdgeFlow: A technique for boundary detection and image segmentation. IEEE Trans Image Process. 9:1375–1388. 2000. View Article : Google Scholar : PubMed/NCBI
19	Yang W, Wang K and Zuo W: Fast neighborhood component analysis. Neurocomput. 83:31–37. 2012. View Article : Google Scholar
20	Zhao W, Chellappa R and Phillips PJ: Subspace Linear Discriminant Analysis for Face Recognition (Technical Report CAR-TR-914)Center for Automation Research University of Maryland; College Park, MD: 1999
21	Chollet F: Keras. GitHub Repository. 2015.
22	Lehman CD, Yala A, Schuster T, Dontchos B, Bahl M, Swanson K and Barzilay R: Mammographic breast density assessment using deep learning: Clinical implementation. Radiology. 290:52–58. 2019. View Article : Google Scholar : PubMed/NCBI
23	Mohamed AA, Berg WA, Peng H, Luo Y, Jankowitz RC and Wu S: A deep learning method for classifying mammographic breast density categories. Med Phys. 45:314–321. 2018. View Article : Google Scholar : PubMed/NCBI
24	Ma X, Fisher CE, Wei J, et al: Multi-path deep learning model for automated mammographic density categorizationProceedings of the Medical Imaging 2019. Computer-Aided Diagnosis. 10950. SPIE Medical Imaging; pp. 862019