Análise da influência do tamanho da região de interesse no diagnóstico auxiliado por computador de fraturas por fragilidade óssea
Resumo
O diagnóstico de fraturas lombares devido a osteoporose e osteopenia é uma tarefa desafiadora devido a semelhança entre os dois casos. A princípio, ambas saõ assintomáticas, podendo posteriormente levar a fratura por fragilidade óssea. Na literatura há estudos promissores na diferenciação entre corpos vertebrais com e sem fraturas. Nestes estudos saõ usados a abordagem radiômica e redes neurais com transferência de aprendizado em corpos vertebrais segmentados. Contudo, a segmentação manual ou semi-automática ainda é um processo trabalhoso. Por este motivo, neste artigo é realizada a comparação da influência do tamanho da regiaõ de interesse sobre a classificação final. Assim, nosso principal objetivo consiste o quanto de detalhe é necessário na segmentação de forma a se obter resultados aceitáveis. Desta forma, é possível que o processo de segmentação seja facilitado ou automatizado em trabalhos futuros. Os resultados experimentais e a validação estatística obtidos com uma base de imagens representativa contendo 2400 casos, divididos meio a meio entre classes com e sem fraturas, demonstram que é possível reduzir o tamanho das regiões de interesse em até 45% sem diferença significativa na acurácia. Com o modelo Xception apresentando resultados mais próximos quando comparado com as imagens originais.
Referências
E. López, E. Ibarz, A. Herrera, S. Puértolas, S. Gabarre, Y. Más, J. Mateo, J. Gil-Albarova, and L. Gracia, “A predictive mechanical model for evaluating vertebral fracture probability in lumbar spine under different osteoporotic drug therapies,” Computer methods and programs in biomedicine, vol. 131, pp. 37–50, 2016.
J. S. Ramos, E. J. de Aguiar, I. V. Belizario, M. V. Costa, J. G. Maciel, M. T. Cazzolato, C. Traina, M. H. Nogueira-Barbosa, and A. J. Traina, “Analysis of vertebrae without fracture on spine mri to assess bone fragility: A comparison of traditional machine learning and deep learning,” in 2022 IEEE 35th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2022, pp. 78–83.
K. Hammernik, T. Ebner, D. Stern, M. Urschler, and T. Pock, “Vertebrae segmentation in 3d ct images based on a variational framework,” in Recent advances in computational methods and clinical applications for spine imaging. Springer, 2015, pp. 227–233.
H.-Z. Wu, X.-F. Zhang, S.-M. Han, L. Cao, J.-X. Wen, W.-J. Wu, and B.-L. Gao, “Correlation of bone mineral density with mri t2* values in quantitative analysis of lumbar osteoporosis,” Archives of Osteoporosis, vol. 15, no. 1, pp. 1–7, 2020.
N. Binkley, S. Morin, P. Martineau, L. Lix, D. Hans, and W. Leslie, “Frequency of normal bone measurement in postmenopausal women 220 with fracture: a registry-based cohort study,” Osteoporosis International, vol. 31, no. 12, pp. 2337–2344, 2020. 222
M. Tellache, M. Pithioux, P. Chabrand, and C. Hochard, “Femoral neck fracture prediction by anisotropic yield criteria,” European Journal of Computational Mechanics/Revue Européenne de Mécanique Numérique, vol. 18, no. 1, pp. 33–41, 2009.
J. G. Maciel, I. M. d. Araújo, L. C. Trazzi, P. M. d. Azevedo-Marques, C. E. G. Salmon, F. J. A. d. Paula, and M. H. Nogueira-Barbosa, “Association of bone mineral density with bone texture attributes extracted using routine magnetic resonance imaging,” Clinics, vol. 75, 2020.
L. He, Z. Liu, C. Liu, Z. Gao, Q. Ren, L. Lei, and J. Ren, “Radiomics based on lumbar spine magnetic resonance imaging to detect osteoporosis,” Academic radiology, vol. 28, no. 6, pp. e165–e171, 2021. 233
A. Y. Kim, M. A. Yoon, S. J. Ham, Y. C. Cho, Y. Ko, B. Park, S. Kim, E. Lee, R.W. Lee, C. G. Chee et al., “Prediction of the acuity of vertebral compression fractures on ct using radiologic and radiomic features,” Academic Radiology, 2022.
J. S. Ramos, J. G. Maciel, M. T. Cazzolato, C. Traina, M. H. Nogueira-Barbosa, and A. J. Traina, “Beaut: a radiomic approach to identify potential lumbar fractures in magnetic resonance imaging,” in 2021 IEEE 34th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2021, pp. 389–394.
J. d. S. Ramos, J. G. Maciel, A. J. M. Traina, P. M. d. Azevedo-Marques, and M. H. Nogueira-Barbosa, “Spine mri texture analysis and prediction of osteoporotic vertebral fracture,” International Journal of Computer Assisted Radiology and Surgery, vol. 16, pp. S115–S116, 2021.
J. S. Ramos, C. Y. Watanabe, M. H. Nogueira-Barbosa, and A. J. Traina, “Bgrowth: an efficient approach for the segmentation of vertebral compression fractures in magnetic resonance imaging,” in Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing, 2019, pp. 220–227.
J. S. Ramos, M. T. Cazzolato, B. S. Faiçal, M. H. Nogueira-Barbosa, C. Traina, and A. J. Traina, “3dbgrowth: volumetric vertebrae segmentation and reconstruction in magnetic resonance imaging,” in 2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2019, pp. 435–440.
J. S. Ramos, M. T. Cazzolato, B. S. Faiçal, O. A. Linares, M. H. Nogueira-Barbosa, C. Traina, and A. J. Traina, “Fast and smart segmentation of paraspinal muscles in magnetic resonance imaging with cleverseg,” in 2019 32nd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI). IEEE, 2019, pp. 76–83.
J. S. Ramos, M. T. Cazzolato, M. H. Nogueira-Barbosa, and A. J. Traina, “Fine: improving time and precision of segmentation techniques for vertebral compression fractures in mri,” in Proceedings of the 35th Annual ACM Symposium on Applied Computing, 2020, pp. 198–201.
M. Roberts, D. Driggs, M. Thorpe, J. Gilbey, M. Yeung, S. Ursprung, A. I. Aviles-Rivero, C. Etmann, C. McCague, L. Beer et al., “Common pitfalls and recommendations for using machine learning to detect and prognosticate for covid-19 using chest radiographs and ct scans,” Nature Machine Intelligence, vol. 3, no. 3, pp. 199–217, 2021.
M. Sokolova and G. Lapalme, “A systematic analysis of performance measures for classification tasks,” Information processing & management, vol. 45, no. 4, pp. 427–437, 2009.
J. S. Ramos, E. J. de Aguiar, I. V. Belizario, M. V. Costa, J. G. Maciel, M. T. Cazzolato, C. Traina, M. H. Nogueira-Barbosa, and A. J. Traina, “Analysis of vertebrae without fracture on spine mri to assess bone fragility: A comparison of traditional machine learning and deep learning,” in 2022 IEEE 35th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2022, pp. 78–83.
K. Hammernik, T. Ebner, D. Stern, M. Urschler, and T. Pock, “Vertebrae segmentation in 3d ct images based on a variational framework,” in Recent advances in computational methods and clinical applications for spine imaging. Springer, 2015, pp. 227–233.
H.-Z. Wu, X.-F. Zhang, S.-M. Han, L. Cao, J.-X. Wen, W.-J. Wu, and B.-L. Gao, “Correlation of bone mineral density with mri t2* values in quantitative analysis of lumbar osteoporosis,” Archives of Osteoporosis, vol. 15, no. 1, pp. 1–7, 2020.
N. Binkley, S. Morin, P. Martineau, L. Lix, D. Hans, and W. Leslie, “Frequency of normal bone measurement in postmenopausal women 220 with fracture: a registry-based cohort study,” Osteoporosis International, vol. 31, no. 12, pp. 2337–2344, 2020. 222
M. Tellache, M. Pithioux, P. Chabrand, and C. Hochard, “Femoral neck fracture prediction by anisotropic yield criteria,” European Journal of Computational Mechanics/Revue Européenne de Mécanique Numérique, vol. 18, no. 1, pp. 33–41, 2009.
J. G. Maciel, I. M. d. Araújo, L. C. Trazzi, P. M. d. Azevedo-Marques, C. E. G. Salmon, F. J. A. d. Paula, and M. H. Nogueira-Barbosa, “Association of bone mineral density with bone texture attributes extracted using routine magnetic resonance imaging,” Clinics, vol. 75, 2020.
L. He, Z. Liu, C. Liu, Z. Gao, Q. Ren, L. Lei, and J. Ren, “Radiomics based on lumbar spine magnetic resonance imaging to detect osteoporosis,” Academic radiology, vol. 28, no. 6, pp. e165–e171, 2021. 233
A. Y. Kim, M. A. Yoon, S. J. Ham, Y. C. Cho, Y. Ko, B. Park, S. Kim, E. Lee, R.W. Lee, C. G. Chee et al., “Prediction of the acuity of vertebral compression fractures on ct using radiologic and radiomic features,” Academic Radiology, 2022.
J. S. Ramos, J. G. Maciel, M. T. Cazzolato, C. Traina, M. H. Nogueira-Barbosa, and A. J. Traina, “Beaut: a radiomic approach to identify potential lumbar fractures in magnetic resonance imaging,” in 2021 IEEE 34th International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2021, pp. 389–394.
J. d. S. Ramos, J. G. Maciel, A. J. M. Traina, P. M. d. Azevedo-Marques, and M. H. Nogueira-Barbosa, “Spine mri texture analysis and prediction of osteoporotic vertebral fracture,” International Journal of Computer Assisted Radiology and Surgery, vol. 16, pp. S115–S116, 2021.
J. S. Ramos, C. Y. Watanabe, M. H. Nogueira-Barbosa, and A. J. Traina, “Bgrowth: an efficient approach for the segmentation of vertebral compression fractures in magnetic resonance imaging,” in Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing, 2019, pp. 220–227.
J. S. Ramos, M. T. Cazzolato, B. S. Faiçal, M. H. Nogueira-Barbosa, C. Traina, and A. J. Traina, “3dbgrowth: volumetric vertebrae segmentation and reconstruction in magnetic resonance imaging,” in 2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS). IEEE, 2019, pp. 435–440.
J. S. Ramos, M. T. Cazzolato, B. S. Faiçal, O. A. Linares, M. H. Nogueira-Barbosa, C. Traina, and A. J. Traina, “Fast and smart segmentation of paraspinal muscles in magnetic resonance imaging with cleverseg,” in 2019 32nd SIBGRAPI Conference on Graphics, Patterns and Images (SIBGRAPI). IEEE, 2019, pp. 76–83.
J. S. Ramos, M. T. Cazzolato, M. H. Nogueira-Barbosa, and A. J. Traina, “Fine: improving time and precision of segmentation techniques for vertebral compression fractures in mri,” in Proceedings of the 35th Annual ACM Symposium on Applied Computing, 2020, pp. 198–201.
M. Roberts, D. Driggs, M. Thorpe, J. Gilbey, M. Yeung, S. Ursprung, A. I. Aviles-Rivero, C. Etmann, C. McCague, L. Beer et al., “Common pitfalls and recommendations for using machine learning to detect and prognosticate for covid-19 using chest radiographs and ct scans,” Nature Machine Intelligence, vol. 3, no. 3, pp. 199–217, 2021.
M. Sokolova and G. Lapalme, “A systematic analysis of performance measures for classification tasks,” Information processing & management, vol. 45, no. 4, pp. 427–437, 2009.
Publicado
06/11/2023
Como Citar
FONSECA, Thiago F. C. da; WATANABE, Carolina Y. V.; MACIEL, Jamilly G.; NOGUEIRA-BARBOSA, Marcello H.; TRAINA, Caetano Jr.; TRAINA, Agma J. M.; RAMOS, Jonathan S..
Análise da influência do tamanho da região de interesse no diagnóstico auxiliado por computador de fraturas por fragilidade óssea. In: WORKSHOP DE TRABALHOS DA GRADUAÇÃO - CONFERENCE ON GRAPHICS, PATTERNS AND IMAGES (SIBGRAPI), 36. , 2023, Rio Grande/RS.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2023
.
p. 168-171.
DOI: https://doi.org/10.5753/sibgrapi.est.2023.27474.
