RESUMO
Purpose: To evaluate the classification performance of structured report features, radiomics, and machine learning (ML) models to differentiate between Coronavirus Disease 2019 (COVID-19) and other types of pneumonia using chest computed tomography (CT) scans. Methods: Sixty-four COVID-19 subjects and 64 subjects with non-COVID-19 pneumonia were selected. The data was split into two independent cohorts: one for the structured report, radiomic feature selection and model building (n = 73), and another for model validation (n = 55). Physicians performed readings with and without machine learning support. The model's sensitivity and specificity were calculated, and inter-rater reliability was assessed using Cohen's Kappa agreement coefficient. Results: Physicians performed with mean sensitivity and specificity of 83.4 and 64.3%, respectively. When assisted with machine learning, the mean sensitivity and specificity increased to 87.1 and 91.1%, respectively. In addition, machine learning improved the inter-rater reliability from moderate to substantial. Conclusion: Integrating structured reports and radiomics promises assisted classification of COVID-19 in CT chest scans.
RESUMO
BACKGROUND: Domestic dogs and cats are very well known to develop chronic hepatic diseases, including hepatic lipidosis and cirrhosis. Ultrasonographic examination is extensively used to detect them. However, there are still few reports on the use of the ultrasound B-mode scan in correlation with histological findings to evaluate diffuse hepatic changes in rodents, which represent the most important animal group used in experimental models of liver diseases. The purpose of this study was to determine the reliability of ultrasound findings in the assessment of fatty liver disease and cirrhosis when compared to histological results in Wistar rats by following up a murine model of chronic hepatic disease. RESULTS: Forty Wistar rats (30 treated, 10 controls) were included. Liver injury was induced by dual exposure to CCl4 and ethanol for 4, 8 and 15 weeks. Liver echogenicity, its correlation to the right renal cortex echogenicity, measurement of portal vein diameter (PVD) and the presence of ascites were evaluated and compared to histological findings of hepatic steatosis and cirrhosis. Liver echogenicity correlated to hepatic steatosis when it was greater or equal to the right renal cortex echogenicity, with a sensitivity of 90%, specificity of 100%, positive and negative predictive values of 100% and 76.9% respectively, and accuracy of 92.5%. Findings of heterogeneous liver echogenicity and irregular surface correlated to liver cirrhosis with a sensitivity of 70.6%, specificity of 100%, positive and negative predictive values of 100% and 82.1% respectively, and accuracy of 87.5%. PVD was significantly increased in both steatotic and cirrhotic rats; however, the later had greater diameters. PVD cut-off point separating steatosis from cirrhosis was 2.1 mm (sensitivity of 100% and specificity of 90.5%). One third of cirrhotic rats presented with ascites. CONCLUSION: The use of ultrasound imaging in the follow-up of murine diffuse liver disease models is feasible and efficient, especially when the studied parameters are used in combination. The potential implication of this study is to provide a non-invasive method that allows follow-up studies of fatty liver disease and cirrhosis of individual rats for pre-clinical drug or cell based therapies.