RESUMO
Purpose: Artificial intelligence (AI)-based image analysis tools to quantify the brain have become commercialized. However, insufficient data for learning and scanner specificity is a limitation for achieving high quality. In the present study, the performance of personalized brain segmentation software when applied to multicenter data using an AI model trained on data from a single institution was improved. Methods: Preindicators of brain white matter (WM) information from the training dataset were utilized for preprocessing. During learning, data of cognitively normal (CN) individuals from a single center were utilized, and data of CN individuals and Alzheimer disease (AD) patients enrolled in multiple centers were considered the test set. Results: The preprocessing based on the preindicator (dice similarity coefficient [DSC], 0.8567) resulted in a better performance than without (DSC, 0.7921). The standard deviation (SD) of the WM region intensity (DSC, 0.8303) had a more substantial influence on the performance than the average intensity (DSC, 0.6591). When the SD of the test data WM intensity was smaller than the learning data, the performance improved (0.03 increase in lower SD, 0.05 decrease in higher SD). Furthermore, preindicator-based pretreatment increased the correlation of mean cortical thickness of the entire gray matter between Atroscan and FreeSurfer, and data augmentation without preprocessing did not.Both preindicator processing and data augmentation improved the correlation coefficient from 0.7584 to 0.8165. Conclusion: Data augmentation and preindicator-based preprocessing of training data can improve the performance of AI-based brain segmentation software, both increasing the generalizability and stability of brain segmentation software.
RESUMO
BACKGROUND: Health problems in shift workers vary including obesity acting as a risk factor in cerebrovascular diseases. Recent studies have commonly determined the prevalence of obesity in shift workers on the basis of body mass index. The accuracy of BMI for diagnosing obesity are still limited apparently. Consequently, this study aimed to determine the relationship between shift work and obesity according to the total body fat percentage in Korean wage workers. METHODS: From the Fourth and the Fifth Korea National Health and Nutrition Examination Survey (2008-2011), after military personnel were excluded, a total of 2952 wage workers (20 ≤ age ≤ 65) whose current jobs were their longest jobs were selected as subjects of the study. The total body fat percentage was used to determine the obesity standards (≥25.7 % in males and ≥36.0 % in females). The subjects were divided into groups by gender and work type (manual vs non-manual), and chi-squared test was used to evaluate the relationship between socio-economic, health behavior, and work-related factors, on the one hand, and obesity, on the other. In addition, multivariate logistic regression analysis was performed to examine the effects of shift work on obesity. RESULTS: When other factors were controlled for, the risk of obesity in shift work showed a statistically significant increase (odds ratio = 1.779, 95 % confidence interval = 1.050-3.015) in the male manual worker group. However, there were no significant results in the male non-manual and female worker groups. CONCLUSION: Shift work was related to a higher risk of obesity in the Korean male manual worker group.
