Abstract
Digital Breast Tomosynthesis (DBT) is an emerging imaging technique for breast cancer screening aiming to overcome certain limitations of traditional mammography, such as the superimposition of tissues. On the downside, DBT increases the radiologists’ workload as it generates stacks of high-resolution images, which are time-consuming to review and annotate. In this work, we propose a deep- multiple-instance-based method for DBT volume classification that relies on the local summarization of DBT slices (referred to as slabbing) and only requires volume-wise labels for training. Slabbing offers several advantages: i) it reduces the classifier’s computational complexity across the depth, letting it focus on the higher transversal resolution. Thanks to this strategy, we are the first to train a method at almost full-resolution (as high as 120 \(\times \) 2500 \(\times \) 2000); ii) it produces slabs that are closer to standard mammography, favoring an efficient transfer from classifiers trained on larger mammography databases; and iii) the slabs combined with a Multiple-Instance Learning (MIL) classifier result in localized information favoring interpretability. The proposed slabbing MIL approach is also novel for the automatic classification of DBTs. Moreover, we propose a trainable alternative to the handcrafted slabbing algorithms based on slice-wise attention that improves performance. We perform an experimental validation on a subset of the public BCS-DBT dataset and achieve an AUC of 0.73 with five-fold cross-validation. On a private multi-vendor dataset we obtain a similar AUC of 0.73, demonstrating an excellent performance consistency.
Supported by Hera-MI SAS, Nantes, France, European Regional Development Fund, Pays de la Loire and Nantes Métropole (Connect Talent MILCOM).
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
Volumes can exceed \(\approx \) \(120\times 2500\times 2000\) pixels, i.e., \(\approx \) \(600M\) pixels.
- 2.
Hereafter we omit the index i from \(V_i\) and \(y_i\) to simplify the notation.
References
Antropova, N., Abe, H., Giger, M.L.: Use of clinical MRI maximum intensity projections for improved breast lesion classification with deep convolutional neural networks. J. Med. Imaging 5(01), 1 (2018). https://doi.org/10.1117/1.jmi.5.1.014503
Balleyguier, C., Ayadi, S., Nguyen, K.V., Vanel, D., Dromain, C., Sigal, R.: BIRADS™ classification in mammography. Eur. J. Radiol. 61(2), 192–194 (2007). https://doi.org/10.1016/j.ejrad.2006.08.033
Buda, M., et al.: Detection of masses and architectural distortions in digital breast tomosynthesis: a publicly available dataset of 5,060 patients and a deep learning model. arXiv:eess.IV/2011.07995 (2021)
Conant, E.F., et al.: Improving accuracy and efficiency with concurrent use of artificial intelligence for digital breast tomosynthesis. Radiol. Artif. Intell. 1(4), e180096 (2019). https://doi.org/10.1148/ryai.2019180096
DeLong, E.R., DeLong, D.M., Clarke-Pearson, D.L.: Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44(3), 837–845 (1988)
Diekmann, F., et al.: Thick slices from tomosynthesis data sets: phantom study for the evaluation of different algorithms. J. Digital Imaging 22(5), 519–526 (2009). https://doi.org/10.1007/s10278-007-9075-y
Doganay, E., Li, P., Luo, Y., Chai, R., Guo, Y., Wu, S.: Breast cancer classification from digital breast tomosynthesis using 3D multi-subvolume approach. In: Deserno, T.M., Chen, P.H. (eds.) Medical Imaging 2020: Imaging Informatics for Healthcare, Research, and Applications, vol. 11318, p. 12. SPIE (2020). https://doi.org/10.1117/12.2551376
Fisher, B., et al.: Twenty-year follow-up of a randomized trial comparing total mastectomy, lumpectomy, and lumpectomy plus irradiation for the treatment of invasive breast cancer. N. Engl. J. Med. 347(16), 1233–1241 (2002). https://doi.org/10.1056/NEJMoa022152
Geras, K.J., Mann, R.M., Moy, L.: Artificial intelligence for mammography and digital breast tomosynthesis: current concepts and future perspectives. Radiology 293(2), 246–259 (2019). https://doi.org/10.1148/radiol.2019182627
Hans Kleinknecht, J., Ileana Ciurea, A., Augusta Ciortea, C.: Pros and cons for breast cancer screening with tomosynthesis - a review of the literature. Med. Pharm. Rep. 93(4), 335–341 (2020). https://doi.org/10.15386/mpr-1698
Houssami, N., Skaane, P.: Overview of the evidence on digital breast tomosynthesis in breast cancer detection. Breast (Edinburgh, Scotland) 22(2), 101–108 (2013). https://doi.org/10.1016/j.breast.2013.01.017
Lotter, W., et al.: Robust breast cancer detection in mammography and digital breast tomosynthesis using an annotation-efficient deep learning approach. Nature Med. 1–6 (2021). https://doi.org/10.1038/s41591-020-01174-9
Mendel, K., Li, H., Sheth, D., Giger, M.: Transfer learning from convolutional neural networks for computer-aided diagnosis: a comparison of digital breast tomosynthesis and full-field digital mammography. Acad. Radiol. 26(6), 735–743 (2019). https://doi.org/10.1016/j.acra.2018.06.019
Murakami, R., Uchiyama, N., Tani, H., Yoshida, T., Kumita, S.: Comparative analysis between synthetic mammography reconstructed from digital breast tomosynthesis and full-field digital mammography for breast cancer detection and visibility. Eur. J. Radiol. Open 7, 100207 (2020). https://doi.org/10.1016/j.ejro.2019.12.001
Murphy, M.C., Coffey, L., O’Neill, A.C., Quinn, C., Prichard, R., McNally, S.: Can the synthetic C view images be used in isolation for diagnosing breast malignancy without reviewing the entire digital breast tomosynthesis data set? Irish J. Med. Sci. 187(4), 1077–1081 (2018). https://doi.org/10.1007/s11845-018-1748-7
Nguyen, T., et al.: Overview of digital breast tomosynthesis: clinical cases, benefits and disadvantages. Diagn. Intervent. Imaging 96(9), 843–859 (2015). https://doi.org/10.1016/j.diii.2015.03.003
Ribli, D., Horváth, A., Unger, Z., Pollner, P., Csabai, I.: Detecting and classifying lesions in mammograms with deep learning. Sci. Rep. 8(1), 1–7 (2018). https://doi.org/10.1038/s41598-018-22437-z
Siegel, R.L., Miller, K.D., Fuchs, H.E., Jemal, A.: Cancer statistics, 2021. CA Cancer J. Clin. 71(1), 7–33 (2021). https://doi.org/10.3322/caac.21654
Tardy, M., Mateus, D.: Looking for abnormalities in mammograms with self-and weakly supervised reconstruction. IEEE Trans. Med. Imaging 1 (2021). https://doi.org/10.1109/TMI.2021.3050040
Vedantham, S., Karellas, A., Vijayaraghavan, G.R., Kopans, D.B.: Digital breast tomosynthesis: state of the art. Radiology 277(3), 663–684 (2015). https://doi.org/10.1148/radiol.2015141303
Wang, X., Liang, G., Zhang, Y., Blanton, H., Bessinger, Z., Jacobs, N.: Inconsistent performance of deep learning models on mammogram classification. J. Am. Coll. Radiol. 17(6), 796–803 (2020). https://doi.org/10.1016/j.jacr.2020.01.006
Wu, N., et al.: Deep neural networks improve radiologists’ performance in breast cancer screening. IEEE Trans. Med. Imaging 39(4), 1184–1194 (2020). https://doi.org/10.1109/TMI.2019.2945514
Zhang, Y., Wang, X., Blanton, H., Liang, G., Xing, X., Jacobs, N.: 2D convolutional neural networks for 3D digital breast tomosynthesis classification. In: Proceedings - 2019 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2019, pp. 1013–1017. Institute of Electrical and Electronics Engineers Inc., November 2019. https://doi.org/10.1109/BIBM47256.2019.8983097
Zheng, S., Guo, J., Cui, X., Veldhuis, R.N., Oudkerk, M., Van Ooijen, P.M.: Automatic pulmonary nodule detection in CT scans using convolutional neural networks based on maximum intensity projection. IEEE Trans. Med. Imaging 39(3), 797–805 (2020). https://doi.org/10.1109/TMI.2019.2935553
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Tardy, M., Mateus, D. (2021). Trainable Summarization to Improve Breast Tomosynthesis Classification. In: de Bruijne, M., et al. Medical Image Computing and Computer Assisted Intervention – MICCAI 2021. MICCAI 2021. Lecture Notes in Computer Science(), vol 12907. Springer, Cham. https://doi.org/10.1007/978-3-030-87234-2_14
Download citation
DOI: https://doi.org/10.1007/978-3-030-87234-2_14
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-87233-5
Online ISBN: 978-3-030-87234-2
eBook Packages: Computer ScienceComputer Science (R0)