Sintigrafik Görüntülerden Tiroid Nodülleri için Bilgisayar Destekli Tanı Sistemi

Öz

Modern tıpta, anatomik bölgelerin segmentasyonu yoluyla görüntü tanıma ve tıbbi görüntüler kullanılarak hastalıkların otomatik olarak sınıflandırılması, çeşitli hastalıkların teşhisinde artan bir potansiyel role sahiptir. Tiroid sintigrafisi, tiroid bezi bozukluklarının teşhisi için kullanılan görüntüleme yöntemlerinden biridir. Çalışmamızda optimize edilmiş Bayesian yerel olmayan ortalama filtresi ile sintigrafi görüntülerinde benek gürültüsü azaltılmıştır. Tiroid bezi lokal bazlı aktif kontur yöntemi ile otomatik olarak segmentlere ayrıldı ve tiroid bezi patolojileri konvolüsyonel sinir ağları (CNN) ile sınıflandırıldı. Önerilen bilgisayar destekli tanı (CAD) sistemi, Histogramlar Piramidi Oryantasyon Gradyanları (PHOG), Gri Düzey Ortak Oluşum Matrisi (GLCM), Yerel Yapılandırma Modeli (LCP) ve Özellik Çantası (BoF) yöntemleriyle karşılaştırıldı. Tiroid bezinin sintigrafik görüntülerinin ortak patolojik paternleri, CNN tarafından %91.19 ile başarıyla sınıflandırıldı. Karşılaştırmalı yöntemler sırasıyla %7.61, %86.04, %88.91 ve %85.72 genel başarı oranları sağlayan PHOG, GLCM, LCP ve BoF yöntemleriydi. Önerilen CNN tabanlı otomatik teşhis sistemi, el yapımı yöntemlere kıyasla umut verici sonuçlar vermiştir.

Anahtar Kelimeler

• Derin öğrenme
• görüntü sınıflama
• Aktif kontur
• gürültü giderme
• tiroid nodül
• sintigrafi

Computer Aided Diagnosis System of Thyroid Nodules from Scintigraphic Images

Öz

In modern medicine, image recognition via segmentation of anatomical regions and automatic classification of diseases using medical images has a growing potential role in diagnosis of various diseases. Scintigraphy of thyroid is one of the established imaging modalities for diagnosis of thyroid gland disorders. In our study, the speckle noise was reduced in the scintigraphy images with the optimized Bayesian nonlocal mean filter. The thyroid gland was automatically segmented by local based active contour method and the thyroid gland pathologies were classified with convolutional neural networks (CNN). The proposed computer aided diagnosis (CAD) system was compared with Pyramid of Histograms of Orientation Gradients (PHOG), Gray Level Co occurrence Matrix (GLCM), Local Configuration Pattern (LCP) and Bag of Feature (BoF) methods. The common pathological patterns of scintigraphic images of the thyroid gland were successfully classified by CNN with an overall success rate of 91.19%. The comparative methods were PHOG, GLCM, LCP and BoF methods which provided overall success rates of 7.61%, 86.04%, 88.91% and 85.72% respectively. The proposed CNN based automatic diagnosis system provided promising results compared to handcrafted methods.

Anahtar Kelimeler

• Deep Learning
• image classification
• Active contour
• noise reduction
• thyroid nodules
• Scintigraphic images

Kaynakça

1. Referans1 Vorländer, C., Wolff, J., Saalabian, S., Lienenlüke, R. H., Wahl, R. A. 2010. Real-time ultrasound elastography -a noninvasive diagnostic procedure for evaluating dominant thyroid nodules. Langenbeck's archives of surgery, Cilt. 395(7), s. 865-871. DOI: 10.1007/s00423-010-0685-3
2. Referans2 Zhu, C., Zheng, T., Kilfoy, B. A., Han, X., Ma, S., Ba, Y., Bai, Y., Wang, R., Zhu, Y., Zhang, Y. 2009. A birth cohort analysis of the incidence of papillary thyroid cancer in the United States, 1973–2004. Thyroid, Cilt.19(10), s. 1061-1066. DOI: 10.1089/thy.2008.0342
3. Referans3 Koundal, D., Gupta, S., Singh, S. 2016. Automated delineation of thyroid nodules in ultrasound images using spatial neutrosophic clustering and level set. Applied Soft Computing, Cilt. 40, s. 86-97. DOI: 10.1016/j.asoc.2015.11.035
4. Referans4 Savelonas, M. A., Iakovidis, D. K., Legakis, I., Maroulis, D. 2008. Active contours guided by echogenicity and texture for delineation of thyroid nodules in ultrasound images. IEEE Transactions on Information Technology in Biomedicine, Cilt. 13(4), s. 519-527. DOI: 10.1109/TITB.2008.2007192
5. Referans5 Maroulis, D. E., Savelonas, M. A., Iakovidis, D. K., Karkanis, S. A., Dimitropoulos, N. 2007. Variable background active contour model for computer-aided delineation of nodules in thyroid ultrasound images. IEEE Transactions on Information Technology in Biomedicine, Cilt. 11(5), s. 537-543. DOI: 10.1109/TITB.2006.890018
6. Referans6 Tsantis, S., Dimitropoulos, N., Cavouras, D., Nikiforidis, G. 2006. A hybrid multi-scale model for thyroid nodule boundary detection on ultrasound images. Computer methods and programs in biomedicine, Cilt. 84(2-3), s. 86-98. DOI: 10.1016/j.cmpb.2006.09.006
7. Referans7 Iakovidis, D. K., Savelonas, M. A., Karkanis, S. A., Maroulis, D. E. 2007. A genetically optimized level set approach to segmentation of thyroid ultrasound images. Applied Intelligence, Cilt. 27(3), s.193-203. DOI: 10.1007/s10489-007-0066-y
8. Referans8 Keramidas, E. G., Iakovidis, D. K., Maroulis, D., Karkanis, S. 2007. Efficient and effective ultrasound image analysis scheme for thyroid nodule detection. In International Conference Image Analysis and Recognition, August 22-24, Montreal, 1052-1060

9. Referans9 Keramidas, E. G., Maroulis, D., Iakovidis, D. K. 2012. ΤND: a thyroid nodule detection system for analysis of ultrasound images and videos. Journal of medical systems, Cilt. 36(3), s. 1271-1281. DOI: 10.1007/s10916-010-9588-7
10. Referans10 Ma, J., Luo, S., Dighe, M., Lim, D. J., Kim, Y. 2010. Differential diagnosis of thyroid nodules with ultrasound elastography based on support vector machines. IEEE International Ultrasonics Symposium 11-14 October, 1372-1375.
11. Referans11 Ding, J., Cheng, H., Ning, C., Huang, J., Zhang, Y. 2011. Quantitative measurement for thyroid cancer characterization based on elastography. Journal of Ultrasound in Medicine, Cilt. 30(9), s. 1259-1266. DOI: 10.7863/jum.2011.30.9.1259
12. Referans12 Singh, N., Jindal, A. 2012. Ultra sonogram images for thyroid segmentation and texture classification in diagnosis of malignant (cancerous) or benign (non-cancerous) nodules. Int. J. Eng. Innov. Technol, Cilt. 1, s. 202-206.
13. Referans13 Acharya, U. R., Sree, S. V., Krishnan, M. M. R., Molinari, F., Garberoglio, R., Suri, J. S. 2012. Non-invasive automated 3D thyroid lesion classification in ultrasound: a class of ThyroScan™ systems. Ultrasonics, Cilt. 52(4), s. 508-520. DOI: 10.1016/j.ultras.2011.11.003
14. Referans14 Ma, J., Wu, F., Zhu, J., Xu, D., Kong, D. 2017. A pre-trained convolutional neural network based method for thyroid nodule diagnosis. Ultrasonics, Cilt. 73, s. 221-230. DOI: 10.1016/j.ultras.2016.09.011
15. Referans15 Coupé, P., Hellier, P., Kervrann, C., Barillot, C. 2009. Nonlocal means-based speckle filtering for ultrasound images. IEEE transactions on image processing, Cilt. 18(10), s. 2221-2229. DOI: 10.1109/TIP.2009.2024064
16. Referans16 Zhang, K., Zhang, L., Lam, K. M., Zhang, D. 2015. A level set approach to image segmentation with intensity inhomogeneity. IEEE transactions on cybernetics, Cilt. 46(2),s. 546-557. DOI: 10.1109/TCYB.2015.2409119
17. Referans17 Ma, J., Wu, F., Jiang, T. A., Zhu, J., Kong, D. 2017. Cascade convolutional neural networks for automatic detection of thyroid nodules in ultrasound images. Medical physics, Cilt. 44(5), s.1678-1691. DOI: 10.1002/mp.12134
18. Referans18 Arevalo, J., González, F. A., Ramos-Pollán, R., Oliveira, J. L., Lopez, M. A. G. 2016. Representation learning for mammography mass lesion classification with convolutional neural networks. Computer methods and programs in biomedicine, Cilt. 127, s. 248-257.DOI :10.1016/j.cmpb.2015.12.014
19. Referans19 Yu, L., Guo, Y., Wang, Y., Yu, J., Chen, P. 2016. Segmentation of fetal left ventricle in echocardiographic sequences based on dynamic convolutional neural networks. IEEE Transactions on Biomedical Engineering, Cilt. 64(8), s. 1886-1895. DOI: 10.1109/TBME.2016.2628401
20. Referans20 Chen, H., Ni, D., Qin, J., Li, S., Yang, X., Wang, T., Heng, P. A. 2015. Standard plane localization in fetal ultrasound via domain transferred deep neural networks. IEEE journal of biomedical and health informatics, Cilt. 19(5), s. 1627-1636. DOI: 10.1109/JBHI.2015.2425041
21. Referans21 Corvilain, B., Van Sande, J., Dumont, J. E., Bourdoux, P., Ermans, A. M. 1998. Autonomy in endemic goiter. Thyroid, Cilt. 8(1), s. 107-113. DOI: 10.1089/thy.1998.8.107
22. Referans22 Hegedüs, L. 2004. The thyroid nodule. New England Journal of Medicine, Cilt. 351(17), s.1764-1771. DOI: 10.1056/NEJMcp031436
23. Referans23 Sarkar, S. D. 2006. Benign thyroid disease: what is the role of nuclear medicine?. In Seminars in nuclear medicine. Cilt. 36-3, s 185-193). DOI: 10.1053/j.semnuclmed.2006.03.006
24. Referans24 Zhang, J., Wang, C., Cheng, Y. 2015. Comparison of despeckle filters for breast ultrasound images. Circuits, Systems, and Signal Processing, Cilt. 34(1), s.185-208. DOI: 10.1007/s00034-014-9829-y
25. Referans25 Zhang, K., Zhang, L., Zhang, S. 2010. A variational multiphase level set approach to simultaneous segmentation and bias correction. IEEE International Conference on Image Processing, Hong Kong, 26-29 September, 4105-4108
26. Referans26 Bosch, A., Zisserman, A., Munoz, X. 2007. Representing shape with a spatial pyramid kernel. In Proceedings of the 6th ACM international conference on Image and video retrieval, Amsterdam, July 5-7 ,401-408.
27. Referans27 Soh, L. K., Tsatsoulis, C. 1999. Texture analysis of SAR sea ice imagery using gray level co-occurrence matrices. IEEE Transactions on geoscience and remote sensing, Cilt. 37(2), s. 780-795. DOI: 10.1109/36.752194
28. Referans28 Guo, Y., Zhao, G.,Pietikäinen, M. 2011. Texture classification using a linear configuration model based descriptor. BMVC. s. 1-10. DOI: 10.5244/C.25.119
29. Referans29 Deselaers, T., Pimenidis, L., Ney, H. 2008. Bag-of-visual-words models for adult image classification and filtering. In 2008 19th International Conference on Pattern Recognition . Tampa, 08-11 December,1-4