Objectives. Because climatic and air-pollution factors are known to influence the occurrence of respiratory diseases, we used these factors to develop machine learning models for predicting the occurrence of respiratory diseases.Methods. We obtained the daily number of respiratory disease patients in Seoul. We used climatic and air-pollution factors to predict the daily number of patients treated for respiratory diseases per 10,000 inhabitants. We applied the reliefbased feature selection algorithm to evaluate the importance of feature selection. We used the gradient boosting and Gaussian process regression (GPR) methods, respectively, to develop two different prediction models. We also employed the holdout cross-validation method, in which 75% of the data was used to train the model, and the remaining 25% was used to test the trained model. We determined the estimated number of respiratory disease patients by applying the developed prediction models to the test set. To evaluate the performance of each model, we calculated the coefficient of determination (R 2 ) and the root mean square error (RMSE) between the original and estimated numbers of respiratory disease patients. We used the Shapley Additive exPlanations (SHAP) approach to interpret the estimated output of each machine learning model. Results.Features with negative weights in the relief-based algorithm were excluded. When applying gradient boosting to unseen test data, R 2 and RMSE were 0.68 and 13.8, respectively. For GPR, the R 2 and RMSE were 0.67 and 13.9, respectively. SHAP analysis showed that reductions in average temperature, daylight duration, average humidity, sulfur dioxide (SO2), total solar insolation amount, and temperature difference increased the number of respiratory disease patients, whereas increases in atmospheric pressure, carbon monoxide (CO), and particulate matter ≤2.5 µm in aerodynamic diameter (PM2.5) increased the number of respiratory disease patients.Conclusions. We successfully developed models for predicting the occurrence of respiratory diseases using climatic and airpollution factors. These models could evolve into public warning systems.

Background: An association between low muscle mass and nonalcoholic fatty liver disease (NAFLD) has been suggested. We investigated this relationship using controlled attenuation parameter (CAP). Methods: A retrospective cohort of subjects had liver FibroScan® (Echosens, Paris, France) and bioelectrical impedance analyses during health screening exams. Low muscle mass was defined based on appendicular skeletal muscle mass/body weight ratios of one (class I) or two (class II) standard deviations below the sex-specific mean for healthy young adults. Results: Among 960 subjects (58.1 years; 67.4% male), 344 (45.8%, class I) and 110 (11.5%, class II) had low muscle mass. After adjusting for traditional metabolic risk factors, hepatic steatosis, defined as a CAP ≥ 248 dB/m, was associated with low muscle mass (class I, odds ratio (OR): 1.96, 95% confidence interval (CI): 1.38–2.78; class II, OR: 3.33, 95% CI: 1.77–6.26). A dose-dependent association between the grade of steatosis and low muscle mass was also found (class I, OR: 1.88, for CAP ≥ 248, <302; OR: 2.19, in CAP ≥ 302; class II, OR: 2.33, for CAP ≥ 248, <302; OR: 6.17, in CAP ≥ 302). High liver stiffness was also significantly associated with an increased risk of low muscle mass (class I, OR: 1.97, 95% CI: 1.31–2.95; class II, OR: 2.96, 95% CI: 1.51–5.78). Conclusion: Hepatic steatosis is independently associated with low muscle mass in a dose-dependent manner. The association between hepatic steatosis and low muscle mass suggests that particular attention should be given to subjects with NAFLD for an adequate assessment of muscle mass.

To investigate the associations of weekend catch-up sleep (WCS) and high-sensitivity C-reactive protein (hs-CRP) levels according to bedtime inconsistency in the Korean population. In this cross-sectional study using the Korea National Health and Nutrition Examination Survey (2016–2018) with 17,665 participants, four groups were defined: no-WCS (WCS within ± 1 h of weekday sleep time), moderate WCS (1 ≤ , < 3 h), severe WCS (≥ 3 h), and inverse WCS (≤ − 1 h). An inconsistent bedtime was defined as a > 2 h difference between weekend and weekday bedtimes. Outcomes were divided into quartiles based on the hs-CRP level: Lowest (< 0.34), Middle-low (≥ 0.34, < 0.55), Middle-high (≥ 0.55, < 1.10), Highest (≥ 1.10). Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were calculated using multinomial logistic regression, controlling for relevant covariates. Moderate WCS was associated with a lower risk for the highest hs-CRP levels than no WCS (aOR = 0.87, 95% CI 0.78–0.97), and a similar association was observed only in participants with consistent bedtimes (aOR = 0.88, 95% CI 0.78–0.99). Significant interactions of those associations of WCS and hs-CRP levels with bedtime inconsistency were found. These findings provide evidence that people with inconsistent bedtimes would have limited protective effect of WCS on hs-CRP.

No abstract

Background To evaluate the correlation between single- and multi-slice cross-sectional thoracolumbar and whole-body compositions. Methods We retrospectively included patients who underwent whole-body PET–CT scans from January 2016 to December 2019 at multiple institutions. A priori-developed, deep learning-based commercially available 3D U-Net segmentation provided whole-body 3D reference volumes and 2D areas of muscle, visceral fat, and subcutaneous fat at the upper, middle, and lower endplate of the individual T1–L5 vertebrae. In the derivation set, we analyzed the Pearson correlation coefficients of single-slice and multi-slice averaged 2D areas (waist and T12–L1) with the reference values. We then built prediction models using the top three correlated levels and tested the models in the validation set. Results The derivation and validation datasets included 203 (mean age 58.2 years; 101 men) and 239 patients (mean age 57.8 years; 80 men). The coefficients were distributed bimodally, with the first peak at T4 (coefficient, 0.78) and the second peak at L2-3 (coefficient 0.90). The top three correlations in the abdominal scan range were found for multi-slice waist averaging (0.92) and single-slice L3 and L2 (0.90, each), while those in the chest scan range were multi-slice T12–L1 averaging (0.89), single-slice L1 (0.89), and T12 (0.86). The model performance at the top three levels for estimating whole-body composition was similar in the derivation and validation datasets. Conclusions Single-slice L2–3 (abdominal CT range) and L1 (chest CT range) analysis best correlated with whole-body composition around 0.90 (coefficient). Multi-slice waist averaging provided a slightly higher correlation of 0.92.