Técnica de análise de textura em imagens de tomografia computadorizada de feixe cônico como ferramenta para o aprimoramento do diagnóstico diferencial entre sinusites odontogênicas e não odontogênicas
Carregando...
Data
2022-06-29
Autores
Ribeiro, Isabela Teixeira
Título da Revista
ISSN da Revista
Título de Volume
Editor
Universidade Estadual Paulista (Unesp)
Resumo
O objetivo deste estudo foi avaliar a técnica de análise de textura (AT), em
imagens por tomografia computadorizada de feixe cônico (TCFC), como ferramenta
quantitativa para o diagnóstico diferencial entre sinusites odontogênicas (SO) e não
odontogênicas (SNO). Material e Métodos: Foram selecionados exames de TCFC
(20 SO e 20 SNO), classificados seguindo os critérios de Maillet et al. (2011). Estes
foram exportados para o software OnDemand3D, para produzir cortes transversais.
Dos exames exportados, foram selecionados 03 cortes consecutivos, abrangendo
diferentes regiões da opacificação sinusal, os quais foram exportados para o
software MaZda 4.60 para realização do cálculo da AT, na região de interesse
delimitada. Os 07 parâmetros de textura foram calculados por meio da matriz de
coocorrência. Foi realizada análise exploratória dos dados pelo cálculo da mediana,
mínimo e máximo. Os grupos foram comparados pelo teste de Mann-Whitney, e a
homogeneidade dos dados entre os grupos pelo teste de Levene. Nível de
significância de 5%. Resultados: Três (03) parâmetros apresentaram diferenças
significativas: contraste, grupo SNO apresentou maiores valores; correlação e
momento da diferença inversa, grupo SO com maiores valores. Constatou-se que o
grupo SO apresentou maior homogeneidade, havendo diferenças estatísticas nos
valores de desvio padrão dos parâmetros: correlação, soma dos quadrados, soma
da entropia e entropia. Conclusão: a AT possibilitou uma diferenciação quantitativa
pelos parâmetros contraste, correlação e momento da diferença inversa entre a SNO
e SO.
The aim of this study was to evaluate the texture analysis (TA) technique in cone beam computed tomography (CBCT) images as a quantitative tool for the differential diagnosis between odontogenic (SO) and non-odontogenic (NOS) sinusitis. Material and Methods: CBCT scans (20 SO and 20 SNO) were selected, classified according to the criteria of Maillet et al. (2011). These were exported to the OnDemand3D software to produce cross-sections. Of these, 03 consecutive sections were selected, covering different regions of sinus opacification. These sections were exported to the MaZda 4.60 software to perform the AT calculation, in the delimited region of interest. The 07 texture parameters were calculated using the cooccurrence matrix. Exploratory data analysis was performed by calculating the median, minimum and maximum. The groups were compared by the Mann-Whitney test, and the homogeneity of the data between the groups by the Levene test. 5% significance level. Results: Three (03) parameters showed significant differences: contrast, the SNO group showed higher values; correlation and inverse moment of difference, SO group with higher values. It was found that the SO group presented greater homogeneity, with statistical differences in the standard deviation values of the parameters: correlation, sum of squares, sum of entropy and entropy. Conclusion: the AT allowed a quantitative differentiation by the parameters contrast, correlation and inverse moment of difference between ONS and OS
The aim of this study was to evaluate the texture analysis (TA) technique in cone beam computed tomography (CBCT) images as a quantitative tool for the differential diagnosis between odontogenic (SO) and non-odontogenic (NOS) sinusitis. Material and Methods: CBCT scans (20 SO and 20 SNO) were selected, classified according to the criteria of Maillet et al. (2011). These were exported to the OnDemand3D software to produce cross-sections. Of these, 03 consecutive sections were selected, covering different regions of sinus opacification. These sections were exported to the MaZda 4.60 software to perform the AT calculation, in the delimited region of interest. The 07 texture parameters were calculated using the cooccurrence matrix. Exploratory data analysis was performed by calculating the median, minimum and maximum. The groups were compared by the Mann-Whitney test, and the homogeneity of the data between the groups by the Levene test. 5% significance level. Results: Three (03) parameters showed significant differences: contrast, the SNO group showed higher values; correlation and inverse moment of difference, SO group with higher values. It was found that the SO group presented greater homogeneity, with statistical differences in the standard deviation values of the parameters: correlation, sum of squares, sum of entropy and entropy. Conclusion: the AT allowed a quantitative differentiation by the parameters contrast, correlation and inverse moment of difference between ONS and OS
Descrição
Palavras-chave
análise de textura, sinusite maxilar, sinusite odontogênica, tomografia computadorizada de feixe cônico, cone-Beam computed tomography;, maxillary sinusitis, odontogenic sinusitis, texture analysis