Deep convolutional neural networks for classifying breast cancer using infrared thermography

ABSTRACT

Infrared thermography is a technique that can detect anomalies in temperature patterns which can indicate some breast pathologies including breast cancer. One limitation of the method is the absence of standardised thermography interpretation procedures. Deep learning models have been used for pattern recognition and classification of objects and have been adopted as an adjunct methodology in medical imaging diagnosis. In this paper, the use of a deep convolutional neural network (CNN) with transfer learning is proposed to automatically classify thermograms into two classes (normal and abnormal). A population of 311 female subjects was considered analysing two approaches to test the CNN’s performance: one with a balanced class distribution and the second study in a typical screening cohort, with a low prevalence of abnormal thermograms. Results showed that the transfer-learned ResNet-101 model had a sensitivity of 92.3% and a specificity of 53.8%, while with an unbalanced distribution the values were 84.6% and 65.3%, respectively. These results suggest that the model presented in this work can classify abnormal thermograms with high sensitivity which validates the use of infrared thermography as an adjunct method for breast cancer screening.

KEYWORDS:

• Deep convolutional neural networks
• transfer learning
• breast cancer screening
• infrared thermography

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability

The public database that supports the findings of this study is taken from http://visual.ic.uff.br/dmi/. The dataset acquired at the ‘Instituto Jaliciense de Cancerología’ (IJC) and at the ‘Instituto de Seguridad y Servicios Sociales de los Trabajadores del Estado’ (ISSSTE) did not agree for their data to be shared publicly, so this two last database are not available.

Additional information

Funding

This work was funded by “Consejo Nacional de Ciencia y Tecnología” (CONACYT) doctoral grant 461548, Laboratorio Nacional de Ciencia y Tecnología de Terahertz (LANCYTT), “Cátedras CONACYT” project 528, and the project 278291 SRE-CONACYT-UASLP.

Notes on contributors

Juan Carlos Torres-Galván

Juan Carlos Torres-Galván Received a bachelor’s degree as a mechanical and electrical engineer, and an M.S. degree in applied sciences from the Autonomous University of San Luis Potosí (UASLP), Mexico, in 2012, and 2016. He is currently pursuing a Ph.D. degree in applied sciences in the UASLP. His research interest includes optical biomedical, image processing and machine learning.

Edgar Guevara

Edgar Guevara is a CONACYT research fellow at the Universidad Autónoma de San Luis Potosí. He is a member of the National Research System (SNI) level 1. He received his Ph.D. degree in biomedical engineering from the École Polytechnique de Montréal, Canada. His current research interests include noninvasive medical diagnostics using optical imaging, functional connectivity, spectroscopy, and biomedical signal processing.

Eleazar Samuel Kolosovas-Machuca

Eleazar Samuel Kolosovas-Machuca was born in Venezuela in 1981. He received the Electronic Engineer degree, the M.S., and Ph.D. degrees in applied sciences from the Autonomous University of San Luis Potosí (UASLP), Mexico, in 2006, 2009, and 2014. Since 2017, he has been a full professor at the UASLP. He is involved in the research of infrared thermography and biosensors for medical applications. He is the author or co-author of more than 25 scientific articles.

Antonio Oceguera-Villanueva

Antonio Oceguera-Villanueva Received a bachelor’s degree as a surgeon, and a specialty in oncology from the Guadalajara University (UDG), Mexico, in 1981, and 1999. He has more than 26 years working in the “Instituto Jaliscience de Cancerología” (IJC). He is the author or co-author of more than 25 scientific articles.

Jorge L. Flores

Jorge L. Flores Received a bachelor's degree in 1994 as an Engineer in communications and electronics from the University of Guadalajara, obtaining his Ph.D. in 2001. Since 2002 is a full-time professor at the electronic department of CUCEI at the University of Guadalajara. His research interest is digital image processing, tridimensional reconstruction objects, and optic sensors development. He has more than 30 scientific articles and is a member of the Mexican National Research System (SNI) Level 2.

Francisco Javier González

Francisco Javier González is a Full Professor at the Autonomous University of San Luis Potosi and holds a Courtesy Appointment at the Physics department of the University of Central Florida (Orlando, FL). His areas of expertise are nanophotonics and biophotonics, areas in which he has contributed more than 150 journal publications, which have been cited over 3000 times (h-index: 28). Dr. González has also published 2 Book Chapters, presented more than 120 papers at conferences, and has 7 patents granted or pending in the US, Mexico, and Spain. Dr. González is a member of the Mexican National Research System (SNI) Level 3 (highest level) and is the 2012 recipient of the Mexican National Research Award from the Mexican Academy of Sciences. Dr. González is a Senior Member of IEEE, OSA, and SPIE and a Member of the Mexican Academy of Sciences.