ABSTRACT
Precise detection and localization of the Endotracheal tube (ETT) is essential for patients receiving chest radiographs. A robust deep learning model based on U-Net++ architecture is presented for accurate segmentation and localization of the ETT. Different types of loss functions related to distribution and region-based loss functions are evaluated in this paper. Then, various integrations of distribution and region-based loss functions (compound loss function) have been applied to obtain the best intersection over union (IOU) for ETT segmentation. The main purpose of the presented study is to maximize IOU for ETT segmentation, and also minimize the error range that needs to be considered during calculation of distance between the real and predicted ETT by obtaining the best integration of the distribution and region loss functions (compound loss function) for training the U-Net++ model. We analyzed the performance of our model using chest radiograph from the Dalin Tzu Chi Hospital in Taiwan. The results of applying the integration of distribution-based and region-based loss functions on the Dalin Tzu Chi Hospital dataset show enhanced segmentation performance compared to other single loss functions. Moreover, according to the obtained results, the combination of Matthews Correlation Coefficient (MCC) and Tversky loss functions, which is a hybrid loss function, has shown the best performance on ETT segmentation based on its ground truth with an IOU value of 0.8683.
ABSTRACT
Objectives: Patients with rheumatoid arthritis (RA) are at a higher risk of extra-articular manifestations, especially hearing loss (HL). Although Chinese herbal medicines (CHM) are proven safe and effective treatments for inflammatory conditions, the effect of CHM use on HL in RA patients is unknown. This cohort study aims to determine the relationship between CHM use and the subsequent risk of HL among RA patients. Methods: From health insurance claims data in Taiwan, a total of 6,905 persons aged 20-80 years with newly-diagnosed RA in 2000-2009 were identified. Of these, we recruited 2,765 CHM users and randomly selected 2,765 non-CHM users who matched with the users by the propensity score. Both cohorts were followed up until the end of 2012 to estimate the incidence of HL. Cox proportional hazards regression was used to estimate the adjusted hazard ratio (HR) for HL. Results: The incidence of HL was lower in the CHM users than in the comparison cohort (8.06 vs. 10.54 per 1,000 person-years) (adjusted HR, 0.77; 95% CI, 0.63-0.94). Those who received CHM for more than 2 years had the greatest benefit against the onset of HL, with over 50% risk reduction. Prescriptions of Hai Piao Xiao, Yan Hu Suo, San-Qi, Huang Qin, Dang Shen, Jia-Wei-Xiao-Yao-San, Shu-Jing-Huo-Xue-Tang, and Dang-Gui-Nian-Tong-Tang were found to be associated with a reduced risk of HL. Conclusions: Our findings suggest that adding CHM to conventional therapy may reduce the subsequent risk of HL in RA patients. Prospective randomized trials are recommended to further clarify whether the association revealed in this study supports such a causal relationship.
