Desarrollo de una red neuronal convolucional para la detección del cáncer de mama mediante la clasificación de imágenes mamográficas

Sergio Augusto Celis Esteban; Jhoan Felipe Sarmiento Ortiz; Liliana Calderón-Benavides

doi:10.24054/rcta.v1i39.1378

Authors

Sergio Augusto Celis Esteban Universidad Autónoma de Bucaramanga
Jhoan Felipe Sarmiento Ortiz Universidad Autónoma de Bucaramanga https://orcid.org/0000-0003-3132-0237
Liliana Calderón-Benavides Universidad Autónoma de Bucaramanga https://orcid.org/0000-0001-8658-9036

DOI:

https://doi.org/10.24054/rcta.v1i39.1378

Keywords:

Artificial intelligence, deep learning, convolutional neural network, transfer learning, image classification, breast cancer, early detection

Abstract

Artificial intelligence (AI) has been growing in recent years in the health area with the development of support systems for clinical decision making. With this work, it was possible to develop a deep learning algorithm capable of classifying mammographic images into five categories (normal, benign microcalcification, benign nodule, malignant microcalcification and malignant nodule) with a priority focus on the early detection of breast cancer, applying the technique of learning transfer. The DDSM and CBIS-DDSM data sets, available on the web, were used for the training and validation of the convolutional neural network obtaining an AUC of 0.9838 and 0.9773 respectively. These results demonstrate the great potential that AI brings to the health area, and the benefits it generates in this and other pathologies by reducing the percentage of false positives and false negatives that are important elements in the diagnosis.

Downloads

Download data is not yet available.

References

Ahmad, A. M., Muhammad Khan, G., & Mahmud, S. A. (2014). Classification of Mammograms Using Cartesian Genetic Programming Evolved Artificial Neural Networks. Artificial Intelligence Applications and Innovations. AIAI. IFIP Advances in Information and Communication Technology. 436, págs. 203-213. Berlin: Springer. doi: https://doi.org/10.1007/978-3-662-44654-6_20

American Cancer Society. (2019). Cancer.org. Obtenido de La densidad de los senos y el informe de su mamograma: https://www.cancer.org/es/cancer/cancer-de-seno/pruebas-de-deteccion-y-deteccion-temprana-del-cancer-de-seno/mamogramas/la-densidad-de-los-senos-y-el-informe-de-su-mamograma.html

Araque Volk, N. (20 de marzo de 2019). Regularizando nuestra red: DropOut. Obtenido de MC.AI: https://mc.ai/regularizando-nuestra-red-dropout/

Cruz-Roa, A., Gilmore, H., Basavanhally, A., Feldman, M., Ganesan, S., Shih, N., . . . Madabhushi, A. (18 de abril de 2017). Accurate and reproducible invasive breast cancer detection in whole-slide images: A Deep Learning approach for quantifying tumor extent. Scientific Reports, 7(46450). Recuperado el 30 de octubre de 2019, de https://doi.org/10.1038/srep46450

Chacón G, Rodríguez JE, Bermúdez V, Vera M, Hernández JD, Vargas S, Pardo A, Lameda C, Madriz D, Bravo AJ. Computational assessment of stomach tumor volume from multi-slice computerized tomography images in presence of type 2 cancer. F1000Res. 2018 Jul 17;7:1098. doi: 10.12688/f1000research.14491.2. PMID: 30473775; PMCID: PMC6234734.

Global Cancer Observatory. (2018). Colombia. World Health Organization, Switzerland. Geneva: Internacional Agency for Research on Cancer. Obtenido de https://gco.iarc.fr/today/data/factsheets/populations/170-colombia-fact-sheets.pdf

Jalandhar, D., & Sharma, A. (2016). DDSM Utility v3.1. Obtenido de https://github.com/trane293/DDSMUtility/blob/mamast/Tutorial.pdf

Jean-Baptiste Lamy, Boomadevi Sekar, Gilles Guezennec, Jacques Bouaud, & Brigitte Séroussi. (2019). Explainable artificial intelligence for breast cancer: A visual case-based reasoning approach. Artificial Intelligence in Medicine, 94, 42-53. doi: https://doi.org/10.1016/j.artmed.2019.01.001

Keras. (2020). Layer activation functions. Obtenido de Keras: https://keras.io/api/layers/activations/

Kontos, K., & Maragoudakis, M. (2013). Breast Cancer Detection in Mammogram Medical Images with Data Mining Techniques. 9th Artificial Intelligence Applications and Innovations (AIAI). 412, págs. 336-347. Paphos: Springer. doi: https://dx.doi.org/10.1007/978-3-642-41142-7_35

Pita Fernández, S., & Pértegas Díaz, S. (07 de diciembre de 2010). Fisterra. Recuperado el 03 de septiembre de 2019, de Pruebas diagnósticas: Sensibilidad y especificidad: https://www.fisterra.com/mbe/investiga/pruebas_diagnosticas/pruebas_diagnosticas.Asp

PLOS. (13 de marzo de 2019). Convolutional neural network for cell classification using microscope images of intracellular actin networks. Obtenido de PLOS ONE: https://journals.plos.org/plosone/article/figure?id=10.1371/journal.pone.0213626.g002

Prabhu. (04 de marzo de 2018). Obtenido de https://medium.com/@RaghavPrabhu/understanding-of-convolutional-neural-network-cnn-deep-learning-99760835f148

Renuka Joshi. (09 de septiembre de 2016). EXSILIO SOLUTIONS. Obtenido de Accuracy, Precision, Recall & F1 Score: Interpretation of Performance Measures: https://blog.exsilio.com/all/accuracy-precision-recall-f1-score-interpretation-of-performance-measures/

Rios, Y., García-Rodríguez, J., Sanchez, E., Alanis, A., Ruiz-Velázquez, E., & Pardo, A. (2020). Neuro-fuzzy control for artificial pancreas: In silico development and validation. [Control neuro-fuzzy para páncreas artificial: Desarrollo y validación in-silico] RIAI - Revista Iberoamericana De Automatica e Informatica Industrial, 17(4), 390-400. doi:10.7326/0003-4819-157-5-201209040-00508

Rose, C. (2006). DDSM: Digital Database for Screening Mammography. Obtenido de University of South Florida: http://www.eng.usf.edu/cvprg/Mammography/Database.html

Nolan, T. (14 de mayo de 2020). CBIS-DDSM. Obtenido de Cancer Imaging Archive: https://wiki.cancerimagingarchive.net/display/Public/CBIS-DDSM

Wang, J., Yang, X., Cai, H., Tan, W., Jin, C., & Li, L. (07 de junio de 2016). Discrimination of Breast Cancer with Microcalcifications on Mammography by Deep Learning. Scientific Reports, 6(27327). doi:10.1038/srep27327