Abstract
The development of democratized, generalizable deep learning applications for health care systems is challenging as potential biases could easily emerge. This paper provides an overview of the potential biases that appear in image analysis datasets that affect the development and performance of artificial intelligence algorithms. Especially, an exhaustive analysis of mammography data has been carried out at the patient, image and source of origin levels. Furthermore, we summarize some techniques to alleviate these biases for the development of fair deep learning models. We present a learning task to classify negative and positive screening mammographies and analyze the influence of biases in the performance of the algorithm.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Hammer, G.P., du Prel, J.B., Blettner, M.: Avoiding bias in observational studies: part 8 in a series of articles on evaluation of scientific publications. Dtsch Arztebl Int. 106, 664 (2009)
Yu, A.C., Eng, J.: One algorithm may not fit all: how selection bias affects machine learning performance. Radiographics 40, 1932–1937 (2020)
Varoquaux, G., Cheplygina, V.: How I failed machine learning in medical imaging - shortcomings and recommendations. Electr. Eng. Syst. Sci. (2021)
Tong, S., Kagal, L.: Investigating bias in image classification using model explanations. Comput. Sci. Comput. Vis. Pattern Recogn. (2020)
Oakden-Rayner, L., Dunnmon, J., Carneiro, G., Re, C.: Hidden stratification causes clinically meaningful failures in machine learning for medical imaging. Comput. Sci. Mach. Learn. (2019)
K. Winkler, et al.: Association between surgical skin markings in dermoscopic images and diagnostic performance of a deep learning convolutional neural network for melanoma recognition. JAMA Dermatol (2019)
Pot, M., Kieusseyan, N., Prainsack, B.: Not all biases are bad: equitable and inequitable biases in machine learning and radiology. Insights Imaging 12, 1–10 (2021)
Larrazabal, A.J., Nieto, N., Peterson, V., Milone, D.H., Ferrante, E. Gender imbalance in medical imaging datasets produces biased classifiers for computer-aided diagnosis. In: Proc. Natl. Acad. Sci. USA , 117, 12592–12594 (2020)
Park, S.H., Han, K.: Methodologic guide for evaluating clinical performance and effect of artificial intelligence technology for medical diagnosis and prediction. Radiology 286, 800–809 (2003)
Zhao, Q., Adeli, E., Pohl, K.M.: Training confounder-free deep learning models for medical applications. Nat. Commun. 11, 1–9 (2020)
Li, X., Cui, Z., Wu, Y., Gu, L., Harada, T.: Stimating and improving fairness with adversarial learning. Comput. Sci. Comput. Vis. Pattern Recogn. (2021)
Seyyed-Kalantari, L., Zhang, H., McDermott, M., Chen, I.Y., Ghassemi, M.: Underdiagnosis bias of artificial intelligence algorithms applied to chest radiographs in under-served patient populations. Nat. Med. 27, 2176–2182 (2021)
Catala, O.D.T., et al.: Bias analysis on public x-ray image datasets of pneumonia and Covid-19 patients. IEEE Access. 9, 42370–42383 (2021)
E. H. P. Pooch, P. L. Ballester, R. C. Barros: Can we trust deep learning based diagnosis? the impact of domain shift in chest radiograph classification. Electr. Eng. Syst. Sci. Image Video Process. (2020)
Zech, J.R., Badgeley, M.A., Liu, M., Costa, A.B., Titano, J.J., Oermann, E.K.: Variable generalization performance of a deep learning model to detect pneumonia in chest radiographs: a cross-sectional study. PLoS Med. 15, e1002683 (2018)
Yala, A., et al.: Toward robust mammography-based models for breast cancer risk. Sci Transl. Med. 13, eaba4373 (2021)
Mayer McKinney, S., et al.: International evaluation of an AI system for breast cancer screening. Nature 577, 89–94 (2020)
Wu, N., et al.: Deep neural networks improve radiologists’ performance in breast cancer screening. IEEE Trans. Med. Imaging 39, 1184–1194 (2020)
Ganin, Y., et al.: Domain-adversarial training of neural networks. Stat. Mach. Learn. 17, 2096–2130 (2016)
Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., Batra, D.: Visual explanations from deep networks via gradient-based localization. In: IEEE International Conference on Computer Vision (ICCV). (2017)
Acknowledgments
This work has been partially funded by FEDER “Una manera de hacer Europa”. This research has been done within the project CADIA - Sistema de Detecciónn de Diversas Patologíss Basado en el Analisis de Imagen con Inteligencia Artificial (DG-SER1-19-003) under the Codigo100 Public Procurement and Innovation Programme by the Galician Health Service - Servizo Galego de Saude (SERGAS) cofunded by the European Regional Development Fund (ERDF).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Zufiria, B. et al. (2022). Analysis of Potential Biases on Mammography Datasets for Deep Learning Model Development. In: Wu, S., Shabestari, B., Xing, L. (eds) Applications of Medical Artificial Intelligence. AMAI 2022. Lecture Notes in Computer Science, vol 13540. Springer, Cham. https://doi.org/10.1007/978-3-031-17721-7_7
Download citation
DOI: https://doi.org/10.1007/978-3-031-17721-7_7
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-17720-0
Online ISBN: 978-3-031-17721-7
eBook Packages: Computer ScienceComputer Science (R0)
