DEVELOPING A SUPPORT SYSTEM FOR BREAST CANCER X-RAY DIAGNOSIS BASED ON ARTIFICIAL INTELLIGENCE
Main Article Content
Abstract
In recent years breast cancer has become one of the most popular cancers for women in many countries in the world, especially, in Vietnam. In order to early detect breast cancer in early stages, the mammograms classification for Vietnamese patients in breast cancer screening based on X-ray is considered as an important task. In this paper, we present a method for classifying mammograms of Vietnamese patients into three categories of BI-RADS: BI-RADS 1: breast is normal, BI-RADS 23: (according to BI-RADS 2 to BI-RADS 3): breast is abnormal with benign mass and BI-RADS 045: cannot evaluate a hurt level in breast cancer X-ray image, additional examinations are required or the breast has a hurt with high malignant rate.
Our mammograms classification system for breast cancer screening is developed based on convolutional neural networks (CNNs) with ResNet50 deep-learning method as the basic framework.
The system is trained and tested on an X-ray dataset of Vietnamese patients containing 7,912 mammograms provided by radiologists of the Hanoi Medical University Hospital. The system accuracy using the testing set achieved a combined area under the curve (mac AUC) of 0.75. To evaluate our model, we conducted a reader study with radiologists of the Department of Radiology, Hanoi Medical University Hospital. Clinical evaluation of the mammograms classification system with about 500 random mammograms of the test set shows that the system performance is more accurate compared to that of one random radiologist and it has the approximate accuracy as “committee of radiologists” when presented with the same images. With this result, our system can be considered a “second radiologist” which can help radiologists in mammograms classification of Vietnamese patients for breast cancer screening in Vietnam.
Article Details
Keywords
Ảnh X-quang ung thư vú, Loại BI-RADS, Mạng nơ ron tích chập ResNet, Học sâu
References
[2] Curated Breast Imaging Subset of DDSM, https:// wiki.cancerimagingarchive.net/display/ Public/CBIS-DDSM. (truy nhập: 19/07/2021).
[3] Dhungel, N., Carneiro, G., and Bradley, A., (2017), Fully automated classification of mammograms using deep residual neural networks, Proceedings of the 2017 IEEE 14th International Symposium on Biomedical Imaging (ISBI 2017), IEEE, United States of America, 310–314.
[4] Expert Panel on Breast, I., L. Moy, S.L. Heller, et al., (2017), ACR Appropriateness Criteria® Palpable Breast Masses, J Am Coll Radiol, 14(5S): S203-S224.
[5] Ferlay, J., Soerjomataram, I., Dikshit, R. et al., (2015), Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer, 136(5): E359-86.
[6] Geras, K.J. et al., (2018), High-Resolution Breast Cancer Screening with Multi-View Deep Convolutional Networks. arXirv: 1703.07047v3 [cs.CV] 28 Jun 2018. (truy nhập: 19/07/2021).
[7] Goodfellow, I., Bengio, Y., and Courville, A., (2016), Deep Learning, MIT Press.
[8] He, K., Zhang, X., Ren, S., Sun, J., (2016), Deep residual learning for image recognition, In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 770-778.
[9] http://icy.bioimageanalysis.org/download/ (truy nhập: 19/07/2021).
[10] https://www.californiaprotons.com/vi/ breast-cancer/stages-grades/ (truy nhập: 19/07/2021).
[11] https://radiologyhanoi.com/khong-phan- loai/1812-chup-xquang-vu-quy-trinh- chup-xquang-vu-phan-do-nghi-ngo-ac- tinh-hinh-anh-xquang-vu-theo-birads.html. (truy nhập: 19/07/2021).
[12] https://suckhoedoisong.vn/hon-21500-phu- nu-viet-mac-ung-thu-vu-nam-2020-n185530. html. (truy nhập: 19/07/2021).
[13] https://www.most.gov.vn/vn/tag/1517/chuong- trinh-kc-4-0.aspx. (truy nhập: 19/07/2021).
[14] http://www.eng.usf.edu/cvprg/Mammography/ Database.html. (DDSM: Digital Database for Screening Mammography).
[15] http://icy.bioimageanalysis.org/download/ (truy nhập: 19/07/2021).
[16] Levy, D., Jain A., (2016), Breast Mass Classification from Mammograms using Deep Convolutional Neural Networks, arXiv:1612.00542v1[cs.CV] 2 Dec 2016. (truy nhập: 19/07/2021).
[17] Litjens, G. et al., (2017), A Survey on Deep Learning in Medical Image Analysis. arXiv:1702.0574v1 [cs.CV] 19 Feb 2017.
[18] Mammographic image analysis homepage, http://www.mammoimage.org/databases/ (truy nhập: 19/07/2021).
[19] Moreira, I.C., Amaral, I., Domingues, I. et al., (2012), INbreast: toward a full-field digital mammographic database,Academic Radiology, 19(2), 236-248.
[20] Nguyen Duc Thang, Nguyen Viet Dung, Tran Vinh Duc, Anh Nguyen, Quang H. Nguyen, Nguyen Tu Anh, Nguyen Ngoc Cuong, Le Tuan Linh, Bui My Hanh, Phan Huy Phu, Nguyen Hoang Phuong, (2020), Building a X-ray Database for Mammography on Vietnamese patients and automatic Detecting ROI Using Mask- RCNN, In Book: V. Kreinovich, Nguyen Hoang Phuong (Eds.), Soft Computing for Biomedical and Related Topics, Studies in Computational Intelligence 899, Springer, 2020. 315-329.
[21] Nguyễn Hoàng Phương, (2020), Tin học và Học sâu trong Y học, Nhà xuất bản Đại học quốc gia Hà Nội.
[22] Sardanelli, F., Fallenberg, E.M., Clauser, P., et al., (2017), Mammography: an update of the EUSOBI recommendations on information for women, Insights Imaging, 8(1): 11-18.
[23] Shen, L., (2017), End-to-end Training for Whole Image Breast Cancer Diagnosis using An All Convolutional Design, Poster at NIPS 2017 Workshop on Machine Learning for Health.
[24] Shimoda, Y., Osanai, T., Nakayama, N. et al., (2016), De novo arteriovenous malformation in a patient with hereditary hemorrhagic telangiectasia, J Neurosurg Pediatr. 17(3): 330-5.
[25] Tabar, L., Yen, A.M., Wu, W.Y., et al., (2015), Insights from the breast cancer screening trials: how screening affects the natural history of breast cancer and implications for evaluating service screening programs, Breast J. 21(1): 13-20.
[26] Xi, P., Shu, C., Goubran, R., (2018), Abnormality Detection in Mammography using Deep Convolutional Neural Networks, ArXiv: 1803.01906v1 [cs.CV] 5 Mar 2018. (truy nhập: 19/07/2021).
[27] Wing, P. and Langelier, M.H., (2009), Workforce shortages in breast imaging: impact on mammography utilization, AJR Am J Roentgenol, 192(2): 370-8.
[28] Wu, N., Phang, J., Park, J., Shen, Y., Huang, Z., Zorin, M., Jastrzebski, S., Fevry, T., Katsnelson, J., Kim, E., Wolfson, S., Parikh, U., Gaddam, S., Lin, L.L.Y., Ho, K., Weinstein, J.D., Reig, B., Gao, Y. Toth, H., Pysarenko, K., Lewin, A., Lee, J., Airola, K. Mema, E., Gao, Y., Toth, H., Pysarenko, K., Lewin, A., Lee, J., Airola, K., Mema, E., Chung, S., Hwang, E., Samreen, N., Kim, S.G., Heacock, L., Moy, L., Cho, K., and Geras, K.J., (2020), Deep Neural Networks Improve Radiologists’ Performance in Breast Cancer Screening, IEEE Transactions on Medical Imaging, Vol. 39, No. 4, April 2020. Pp. 1184-1194.