RESUMEN
Objective@#To establish a CT image radiomics-based prediction model for the differential diagnosis of silicosis and tuberculosis nodules.@*Methods@#A total of 53 patients with silicosis and 89 patients with tuberculosis who underwent routine CT scans in Suzhou Fifth People's Hospital from January to August, 2018 were enrolled in this study. AK/ITK software was used to segment the images to obtain 139 silicosis lesions and 119 tuberculosis lesions. For each lesion image, 396 features were extracted, and feature dimension reduction was applied to select the most characteristic feature subset. Support vector machine (SVM) , feedforward back propagation neural network (FNN-BP) , and random forest (RF) were implemented using R software (Rstudio V1.1.463) , and the algorithm that achieved the largest area under of the receiver operating characteristic (ROC) curve (AUC) was selected as the final prediction model.@*Results@#RF was the best prediction model for the differential diagnosis of silicosis and tuberculosis nodules, with an accuracy of 83.1%, a sensitivity of 0.76, a specificity of 0.9, and an AUC of 0.917 (95% confidence interval: 0.8431-0.9758) . RF had a significantly larger AUC than SVM and FNN-BP (P<0.05) .@*Conclusion@#CT image-based RF prediction model can be used to differentially diagnose silicosis and tuberculosis nodules.
RESUMEN
Objective To evaluate the roles of high-resolution computer tomography (HRCT) and bron- choalveolar lavage (BALF) in the diagnosis of connective tissue diseases associated with interstitial lung disease (CTD-ILD). Methods Clinical data of chest HRCT and BALF of patients with CTD-ILD from January 1997 to December 2007 in in-patient department of Peking University First Hospital, were retrospectively analyzed. Results ①Among 46 cases with the picture of chest HRCT, 19 (41.3%)showed usual interstitial pneumonia(UIP) -like pattern and 18 (39.1%) showed lobular and interlobular septa thickening. 8 (8/17) of ANCA vasculitis (AASV) and 5 (5/9) of rheumatoid arthritis (RA) manifested as UIP-like patterns respectively. In polymyositis/dermatomyositis(PM/DM) and Sjogren's syndrome (KS) patients, the organizing pneumonia(OP)-like pattern and lymphocytic interstitial pneumonia(LIP)-like pattern were 2/4 and 2/4 respectively. ②Among 32 cases undergoing BAL, 10/10 patients with AASV-ILD all showed that neutrophils were dominant in BALF, while, the other 22 patients showed that the ratio of neutrophils elevation (14/22, 63.6%) and the ratio of lymphocytes elevation (18/22, 81.8%) were comparable, and there were 12/22(54.5%)patients with both types of cell elevation. Among 13 cases with iymphocytes elevation in BALF who performed analysis of sub-type lymphocytes, 10/13 cases showed decreased CD4/CD8 ratio, 3/13 cases showed increased CD4/CD8 ratio which were all related with SS. ③Among 15 patients undertaken HRCT and BALF detection together, 7/7 UIP-like cases showed the ratio of neutrephils elevation in BALK While in non-UIP-like cases, 5/8 showed the ratio of lymphocytes elevation. Conclusion ①UIP-like patterns and patterns of lobular and interlobular septa thickening are the most common imaging features of HRCT in CTD-ILD, the former are mostly seen in AASV and RA. OP-like patterns and LIP-like patterns are commonly seen in PM/DM and SS respectively. ②Tbe increased neutrophil percentage is dominant in BAL fluid of patients with AASV-ILD, while the others show that the ratio of neutrophil and lymphocyte elevation are comparable, lymphocytes subtype analysis shows decreased CD4/CD8 ratio is dominant in CTD-ILD patients with lymphocytes increased. There is a significant relationship between increased CD4/CD8 ratio and SS-LIP. ③All of the cases with UIP-like patterns show the ratio of neutrophils elevation in BALF. While the ratio of lymphocytes elevation is dominant in non-UIP-like cases.