Development of shape-based descriptors for diagnosis of pulmonary lesions
Abstract
Lung cancer is one of the most common types of cancer and the one with the highest mortality rate in the world. The automation of diagnostic computer vision systems, through the analysis of medical images, provides an interpretation regarding the pathology. The idea is to use the features of the form from the images of lung nodules, then sort on malignant or benign. This paper presents the development of descriptors based on shape analysis for characterization of nodule. The tests have promising results with an accuracy of 92 %, specificity of 89.2%, sensitivity of 91.5 % and a área under the ROC curve of 0.920.
References
Braz Júnior, G. B.. “Detecção de regiões de massas em mamografias usando índices de diversidade geoestatística e geometria côncava”. Univ. Federal do Maranhão, Maranhão, Centro de Ciências e Tecnologias 2014.
Carvalho Filho, A. O de.. Métodos para sistemas CAD e CADx de nódulo pulmonar baseada em tomografia computadorizada usando análise de forma e textura, tese de doutorado, Área de Engenharia de Eletricidade, Universidade Federal do Maranhão, Centro de Ciências Exatas e Tecnologia, São Luís, Maranhão, 2016.
Chen, W.; Li, Z.; Bai, L.; e Lin, Y..”NF-kappaB, a mediator for lung carcinogenesis and a target for lung cancer prevention and therapy.” Frontiers in bioscience: a journal and virtual library,2011, v. 16, pp. 1172-1185.
Cook, R. J.. “Kappa and its Dependence on Marginal Rates”, Encyclopedia of BioStatistics, P. Armitage and T. Colton, eds., John Wiley &, Sons, 1998, pp. 2166&ndash,2168.
Dandil, E.; Cakiroglu, M.; Eksi, Z.; Ozkan, M.; Kurt, O.; Canan, A., "Artificial neural network-based classification system for lung nodules on computed tomography scans", Soft Computing and Pattern Recognition (SoCPaR) 2014 6th International Conference of, pp. 382-386, pp. 11–14 Aug. 2014.
Dhara, A. K., Mukhopadhyay, S., Dutta, A., Garg, M., & Khandelwal, N. A combination of shape and texture features for classification of pulmonary nodules in lung ct images. Journal of digital imaging, 2016, vol. 29, n. 4, pp. 466-475.
Frank, E. “Fully supervised training of Gaussian radial basis function networks in WEKA”. Department of Computer Science, The University of Waikato, 2014.
Frazão, A., "Que tipo de nódulo pode ser câncer," Tua saúde, [link] (acesso em Nov. 15, 2016).
Froner, A. P. P.. Caracterização de nódulos pulmonares em imagens de tomografia computadorizada para fins de auxílio ao diagnóstico, Dissertação de mestrado, Àrea de Engenharia Elétrica, Pontifícia Universidade Católica do Rio Grande Do Sul, Faculdade de Engenharia Elétrica, Troy, Porto Alegre, 2015.
Giger, M. L.. “Computer-aided diagnosis”. RSNA Categorial Course in Breast Imaging, 1999, pp. 783-792.
International Agency for Research on Cancer - IARC. “About ANCERModial”. Disponível em: [link]. Acessado em 15 Fev. 2017).
Klein, P. N.; Tirthapura, S.; Sharvit, D.; e Kimia, B. B.. A tree-edit-distance algorithm for comparing simple, closed shapes. In Symposium on Discrete Algorithms, pp. 696–704, 2000.
Landwehr, N.; Hall, M. ; e Frank, E.. “Logistic model trees,” in Proc. 14th Eur. Conf. Machine Learning, 2003, vol. 2837, pp. 241–252.
Manzano-Mancho, D.; e Gómez-Pérez, A..” An overview of methods and tools forontology learning from texts”, The Knowledge Engineering Review, vol 19, n.3, pp. 187-212, 2005.
Martinez, E. Z.. Louzada, N. F.; e Pereira, B. B.. “A curva ROC para tests diagnósticos”. Cadernos Saúde Coletiva 11, 2003, pp. 7-31.
Naticional Cancer Institute - NCI. “Lung Cancer. U.S. National Institute of Health”. Disponível em: [link] -- What Is Cancer?r. Acessado em 15 Fev. 2017).
Orozco, H.M.; Villegas, O. O. V.; Sanchez, V. G. C.; Dominguez, M. D. J. N. “Automated system for lung nodules classification basead on wavelet feature descritor and support vector machine”, Biomedical Engineering Online, v.14, no. 9 ,2015.
Provost, F. e Domingos, P.. “Well-Trained Pets: Improving Probability Estimation Trees”, 2000, Stern School of Business, New york Univ.
Sampaio, W. B. . “Mass Detection in Mammography Images using a Methodology Adapted to Breast Density”, Univ. Federal of the Maranhão, Maranhão, Center for Science and Technology, 2015.
Sebastian, T. B.; Klein, P. N.; e Kimia, B. B.. "Recognition of shapes by editing shock graphs," in IEEE ICCV, 2001.
Silva, A. C.. “Algoritmos para Dianóstico Assistido de Nódulos Pulmonares Solitários em Imagens de Tomografia Computadorizada”. Tese de Doutorado do Programa de Pós-gradução em Informática, Departament de Informática, Pontifícia Universidade Católica do Rio de Janeiro, 2004.
Silva, A. C.. “Medidas globais em 3d para diagnóstico de nódulo pulmonar”. Disse, Instituto de Matemática Pura e Aplicada–IMPA, 2007.
Sousa, J. R. F. S.. “Metodologia para detecção automática de nódulos pulmonares”. Dissertação de Mestrado do Programa de Pós-gradução em Engenharia da Eletricidade, Universidade Federal do Maranhão 2007.
Sousa, J. R. F. S.; Silva, A. C.; e Paiva, A. C.. “Lung structure classification using 3D geometric measurements and SVM. Progress in Pattern Recognition, Image Analysis and Applications.”. Lecture Notes in Computer Science.,2008, v. 4756, pp. 783-792.
Wang, Jun, et al. "Prediction of malignant and benign of lung tumor using a quantitative radiomic method." Engineering in Medicine and Biology Society (EMBC), 2016 IEEE 38th Annual International Conference of the. IEEE, 2016.
WEKA – Machine Learning Group at the University of Waikato. “Weka 3: Data Mining Software in Java”. Disponível em: [link]. Acessado 20 Mar. 2017)
Zhu, W.; Zeng, N.; e Wang, N.. “Sensitivity, specificity, accuracy associated confidence interval and roc analysis with practical sas implementations” in In Proceedings of the NorthEast SAS Users Group. Conference on Evolutionary Programming, San Diego, CA, 1998, pp. 201-208.
