Análise textural a partir do histograma de densidades da imagem tomográfica computadorizada do nódulo pulmonar solitário: estudo da curtose e assimetria

• Autor(a) principal: Borguezan, Bruno Max
• Data de Publicação: 2018
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UERJ
• Texto Completo: http://www.bdtd.uerj.br/handle/1/8681
• topic: Pulmonary nodules; Diagnostic imaging; Computed tomography; Kurtosis; Skewness; Nódulos pulmonares; Diagnóstico por imagem; Curtose; Assimetria; Pulmões - Doenças - Imagem; Nódulo pulmonar solitário - Diagnóstico por imagem; Tomografia computadorizada por raios x; Estatística médica; CNPQ::CIENCIAS DA SAUDE::MEDICINA::RADIOLOGIA MEDICA

Incidental findings of pulmonary nodules in imaging methods used to diagnose other thoracic or extrathoracic conditions have been increasing, suggesting the need for more in-depth analysis of radiological images to identify the nodule type and avoid unnecessary invasive procedures. Objectives: The purpose of the present study was to evaluate the solid indeterminate solitary pulmonary nodules through texture analysis of CT images. Methods: One hundred computed tomography scans of the chest were evaluated: 50 benign and 50 malignant solid nodules. In these tests, the kurtosis and skewness values were determined for the whole volume of each nodule, and the graphical representation of the histogram was classified into two basic patterns peak and plateau. Results: The mean kurtosis values of the benign and malignant nodules were 3.37 ± 3.88 and 5.88 ± 5.11, respectively. The receiver operating characteristic (ROC) curve showed that for kurtosis above 6, the sensitivity and specificity for distinguishing between benign and malignant nodules were 65% and 66%, respectively, with an area under the curve (AUC) of 0.709 (p<0.0001). The mean skewness values of benign and malignant nodules were 1.73 ± 0.94 and 2.07 ± 1.01, respectively. The ROC curve showed that for skewness above 3.1, the sensitivity and specificity for distinguishing between malignant and benign nodules were 65% and 66%, respectively, with an AUC of 0.709 (p<0.0001). Histogram analysis between the peak and plateau patterns revealed a sensitivity of 84% and a specificity of 74%, with an accuracy of 79%. Conclusions: The kurtosis and skewness values, as well as the histogram shape, can help to non-invasively diagnose solid indeterminate solitary pulmonary nodules but should not be used alone or without also considering the clinical data.