Background. There is no radiological measurement to estimate the severity of pediatrics juvenile dermatomyositis (JDM) with interstitial lung disease (ILD). We validated the effectiveness of CT scoring assessment in JDM patients with ILD. Aim. To establish a CT scoring system and calculate CT scores in JDM patients with ILD and to determine its reliability and the correlation with Krebs von den Lungen-6 (KL-6). Methods. The study totally enrolled 46 JDM-ILD patients and 16 JDM without ILD (non-ILD, NILD) patients. The chest CT images (7.0 ± 3.6 years; 32 male and 30 female) were all analyzed. CT scores of six lung zones were retrospectively calculated, included image pattern score and distribution range score. Image pattern score was defined as follows: increased broncho-vascular bundle (1 point); ground glass opacity (GGO) (2 points); consolidation (3 points); GGO with bronchiectasis (4 points); consolidation with bronchiectasis (5 points); and honeycomb lung (6 points). Distribution range score was defined as no infiltrate (0 point); <30% (1 point); 30%–60% (2 points); and ≥60% (3 points). Two pediatric radiologists reviewed all CT images independently. The ROC curve was established, and the optimal cutoff score for severity discrimination was set. Results. The agreement between two observers was excellent, and the ICC was 0.930 (95% CI 0.882–0.959, ). CT score and KL-6 level had a positive linear correlation (r = 0.784, ). However, the correlation between CT scores of different lung zone and KL-6 level was different. The KL-6 cut off level suggested for JDM with ILD was 209.0 U/ml, with 73.9% sensitivity and 87.5% specificity, and the area under curve was (AUC) 0.864 ( ). Conclusion. The CT scoring system we established, as a semiquantitative method, can effectively evaluate ILD in JDM-PM patients and provide reliable evidence for treatment.
Background: The shortage of pediatric medical resources and overcrowding in children’s hospitals are severe issues in China. Accurately predicting waiting times can help optimize hospital operational efficiency. Objective: This study aims to develop machine learning models to predict waiting times for various laboratory and radiology examinations at a pediatric hospital. Methods: Time stamp data from laboratory and radiology examinations were retrospectively collected from the pediatric hospital information system between November 1, 2024, and March 13, 2025. Two queue-related and 4 time-based features were extracted using queue theory. Linear regression and 8 machine learning models were trained to predict waiting times for each medical task. Hyperparameters were fine-tuned using randomized search and 10-fold cross-validation, and the bootstrap method was used for model evaluation. Mean absolute error, mean square error, root mean square error, and the coefficient of determination (R²) were used as evaluation metrics. Shapley additive explanations values were used to assess feature importance. Results: A total of 230,864 time-stamped records were included after data preprocessing. The median waiting time was 4.817 (IQR 1.867-12.050) minutes for all medical tasks. Waiting times for radiology examinations were generally longer than those for laboratory tests. Tree-based algorithms, such as random forest and classification and regression trees, performed best in predicting laboratory test waiting times, with R² values ranging from mean 0.880(SD 0.003) to mean 0.934 (SD 0.003). However, the machine learning models did not perform well in predicting radiology examination waiting times, with R2 ranging from 0.114 (SD 0.005) to 0.719 (SD 0.004). Feature importance analysis revealed that queue-related predictors, especially the number of queuing patients, were the most important in predicting waiting times. Conclusions: Task-specific prediction models are more appropriate for accurately predicting waiting times across various medical tasks. Guided by queue theory principles, we developed machine learning models for the waiting time prediction of each medical task and highlighted the importance of queue-related predictors.