Search

• A tracheal sound database is established and conditionally open to the public. • Lung sound and tracheal sound analyses must be treated as two distinct problems. • Domain adaptation and mixed-set training benefit deep learning in respiratory sound. • The models trained on the mixed set are more flexible in respiratory sound analysis. Many deep learning–based computerized respiratory sound analysis methods have previously been developed. However, these studies focus on either lung sound only or tracheal sound only. The effectiveness of using a lung sound analysis algorithm on tracheal sound and vice versa is rarely reported. Furthermore, no one knows whether using lung and tracheal sounds together in training a deep learning-based respiratory sound analysis model is beneficial. In this study, we first constructed a tracheal sound database, HF_Tracheal_V1, containing 10,448 15-s tracheal sound recordings, 21,741 inhalation labels, 15,858 exhalation labels, and 6414 continuous adventitious sound (CAS) labels collected from 227 participants undergoing a diagnostic/surgical procedure under monitored anesthesia care. HF_Tracheal_V1 and our previously built lung sound database, HF_Lung_V2, were either combined (mixed set), used one after the other (domain adaptation), or used alone to train convolutional neural network bidirectional gate recurrent unit models for inhalation, exhalation, and CAS detection in lung and tracheal sounds. The results revealed that the models trained using lung sound alone performed poorly in tracheal sound analysis and vice versa. However, mixed set training or domain adaptation improved the performance for 1) inhalation and exhalation detection in lung sounds and 2) inhalation, exhalation, and CAS detection in tracheal sounds compared to positive controls (the models trained using lung sound alone and used in lung sound analysis and vice versa). In particular, the model trained on the mixed set had great flexibility to serve two purposes, lung and tracheal sound analyses, at the same time. [ABSTRACT FROM AUTHOR]

Background Total gastrectomy (TG) is potentially curative for upper gastric cancer (GC) and includes both stomach removal and enteral bypass reconstruction. Therefore, similar to bariatric surgery, TG may contribute to a change in glucose metabolism. Objectives The aim of this population-based study was to determine if there are any changes in glucose metabolism after TG in patients with GC. Setting Nationwide population database. Methods We conducted a nationwide cohort study using data from Taiwan’s National Health Insurance Research Database collected between 2000 and 2011. Two cohorts were sampled for further analysis of changes in glucose metabolism after TG, including a diabetes group and non-diabetes group. Results A total of 8593 patients with TG were identified. Of the 579 patients with diabetes with GC, 178 (30.7%) achieved diabetes remission after TG. On Cox multivariate analysis, patients aged 50 to 64 years (odds ratio [OR], .48; 95% confidence interval [CI], .31–.75; P <.01) and those who used insulin (OR, .31; 95% CI, .19–.51; P <.01) had lower rates of diabetes remission, whereas liver cirrhosis patients had higher rates of diabetes remission (OR, 1.77; 95% CI, 1.16–2.69; P <.01). On the other hand, patients without diabetes but who had GC (n = 1565) had lower rates of newly diagnosed diabetes after TG compared with the general population as assessed by control-to-case analysis (OR, .56; 95% CI, .47–.66; P < .01). Conclusion Our data showed that TG contributes to improved glucose metabolism in patients with GC. [ABSTRACT FROM AUTHOR]

Accurate assessment of the percentage of total body surface area (%TBSA) burned is crucial in managing burn injuries. It is difficult to estimate the size of an irregular shape by inspection. Many articles reported the discrepancy of estimating %TBSA burned by different doctors. We set up a system with multiple deep learning (DL) models for %TBSA estimation, as well as the segmentation of possibly poor-perfused deep burn regions from the entire wound. We proposed boundary-based labeling for datasets of total burn wound and palm, whereas region-based labeling for the dataset of deep burn wound. Several powerful DL models (U-Net, PSPNet, DeeplabV3+, Mask R-CNN) with encoders ResNet101 had been trained and tested from the above datasets. With the subject distances, the %TBSA burned could be calculated by the segmentation of total burn wound area with respect to the palm size. The percentage of deep burn area could be obtained from the segmentation of deep burn area from the entire wound. A total of 4991 images of early burn wounds and 1050 images of palms were boundary-based labeled. 1565 out of 4994 images with deep burn were preprocessed with superpixel segmentation into small regions before labeling. DeeplabV3+ had slightly better performance in three tasks with precision: 0.90767, recall: 0.90065 for total burn wound segmentation; precision: 0.98987, recall: 0.99036 for palm segmentation; and precision: 0.90152, recall: 0.90219 for deep burn segmentation. Combining the segmentation results and clinical data, %TBSA burned, the volume of fluid for resuscitation, and the percentage of deep burn area can be automatically diagnosed by DL models with a pixel-to-pixel method. Artificial intelligence provides consistent, accurate and rapid assessments of burn wounds. • The %TBSA burned could be calculated from the segmentation results of patients' burn wounds and palms. • Superpixel segmentation preprocessing is a helpful method for images that are difficult to be labeled. • Deep learning has the potential to be a tool providing instant and precise diagnoses for acute burn wound management. [ABSTRACT FROM AUTHOR]

Background Morbidly obese patients display both an autonomic nervous imbalance and impaired glucose metabolism, and both of these conditions can be partially reversed after bariatric surgery. The aim of the present study was to investigate changes in heart rate variability (HRV) and glucose metabolism in patients after laparoscopic sleeve gastrectomy (SG). Methods Eighteen morbidly obese patients who underwent SG were examined before surgery and at 7, 30, 90, and 180 days after surgery. Indices of HRV included time-domain, frequency-domain, and nonlinear parameters. Glucose metabolism was evaluated by the measuring levels of insulin resistance, glycated hemoglobin (HbA 1c ), and gut hormones. Results The study included 9 men and 9 women with a mean age of 34 years. In the HRV study, the average R-R interval, median R-R interval, standard deviation of the R-R intervals, root mean squared successive difference of the R-R intervals (RMSSD), and the number of pairs of successive normal-to-normal beat intervals that differed by>50 ms significantly increased at 180 days after surgery. Regarding the frequency-domain indices, the low frequency (LF)/high frequency (HF) ratio was more balanced at 90 days after SG compared with baseline, and increases in the total power, LF band, and HF band were observed at 180 days. The assessments of insulin resistance, glucose metabolism, and gut hormones revealed not only improvements in the homeostasis model assessment of insulin resistance and HbA 1c levels but also increases in the levels of glucagon-like peptide-1 at 90 and 180 days after surgery compared with baseline. A multivariable regression model revealed significantly negative associations between the perioperative changes in HOMA-IR and changes in both the RMSSD and HF band. Conclusions SG leads to early improvements in insulin resistance and glucose metabolism that are followed by improvements in HRV indices. Improvements in insulin resistance were associated with increases in the RMSSD and HF band index, but the mechanism of these changes require further study. [ABSTRACT FROM AUTHOR]

Cancer has been ranked first in the causes of death for 31 consecutive years in Taiwan. Radiofrequency ablation (RFA) is a treatment for hepatocellular carcinoma (HCC) and it becomes one of the important therapies for HCC these years. For those who had HCC and were treated by RFA, their clinical data are collected to build predictive models which can be used in predicting the recurrence or not of liver cancer after RAF treatment. Clinical data with multiple measurements are merged based on different time periods and these data are further transformed based on temporal abstraction (TA). Data processed by TA reveal variations of clinical data with different time points. The goal of this study is to evaluate whether clinical data handled by TA could facilitate performance of predictive models. Different data sets are used in developing predictive models, including clinical data which are not processed by TA called the original data set, clinical data which are processed by TA called the TA data set, and combination of the original data set and the TA data set called the TA+original data set. Support vector machine (SVM) was selected as a classifier to develop predictive models. The results demonstrate data sets processed by TA provide benefit for predictive models. [ABSTRACT FROM AUTHOR]

Obesity has become an important health issue in Taiwan due to its increasing prevalence among population and the patients are getting younger. In this research, we build a health care information system, named the Pauian Health, for people who were on diet and to control their weight, or others to keep in shape. The Pauian Health is designed based on model–view–controller (MVC) pattern with a proposed architecture framework and has been successfully implemented in Pauian Archiland Company. The architecture centralizes the functions of accessing databases, user identification, and the validation of inherited clinical knowledge in a single component. Under this architecture, the system ensures its scalability, extensibility, data consistency, and the ability of cross-platform. This architecture has also been implemented in different devices and platforms, including iPhone, iPad and website. The architecture also supports unstable network environment, which uses Token as session identifier and user validation. The proposed framework has the potential in duplicating the service for different targeted users or deployed as a cloud service, which is fully capable in extending the service in the future. [ABSTRACT FROM AUTHOR]

This project proposes a pulse analysis system for doctors to diagnose their patients using their pulse signals. A novice and easy time-domain algorithm is applied to pulse signal for periodic function examination. Monitoring and modeling our body as a system with digitalized pulse signals is one of the linkages that we can find among Traditional Chinese Medicine (TCM), hemodynamics and modern Information Communication Technology (ICT). With the labeling of diseases or some modern Bio-markers, we are able to link pulse analysis, ICT and the latest medical science together. The proposed method is capable of calculating the instability of the pulse wave of subjects. After finding the starting point of each period in a periodic wave, we use set theory as the constraint to detect stable periodic wave. With normal heart rate checking and the variability of each period checking, our algorithm can detect whether the input signal is normal, stable and periodic. A coefficient that represents the instability is calculated by the average Standard Deviation (SD) of each period in the waveform. The proposed system helps the automation of pulse examination to select a proper segment for harmonic analysis. With the periodic function examination method, non-survival sepsis patients in Continuous Mandatory Ventilation (CMV) mode or Continuous Positive Airway Pressure (CPAP) mode can be significantly separated. [ABSTRACT FROM AUTHOR]

Abstract: In many network applications, including distant learning, audio webcasting, video streaming, and online gaming, often a source has to send data to many receivers. IP multicasts and application-layer multicasts provide efficient and scalable one-to-many or many-to-many communications. A common secret key, the group key, shared by multiple users can be used to secure the information transmitted in the multicast communication channel. In this paper, a new group key management protocol is proposed to reduce the communication and computation overhead of group key rekeying caused by membership changes. With shared key derivation, new keys derivable by members themselves do not have to be encrypted or delivered by the server, and the performance of synchronous and asynchronous rekeying operations, including single join, single leave, and batch update, is thus improved. The proposed protocol is shown to be secure and immune to collusion attacks, and it outperforms the other comparable protocols from our analysis and simulation. The protocol is particularly efficient with binary key trees and asynchronous rekeying, and it can be tuned to meet different rekeying delay or key size requirements. [Copyright &y& Elsevier]

Purpose Clinical pathways fall under the process perspective of health care quality. For care providers, clinical pathways can be compared to improve health care quality. The objective of this study was to design a convenient physician order set comparison system based on claim records from the National Health Insurance Research Database (NHIRD) of Taiwan. Methods Data were retrieved from the NHIRD for the period of 2003–2007 for frequent physician order sets found in hospital surgical hernia repair inpatient claim records. The derived frequent physician order sets were divided into five frequency thresholds: 80%, 85%, 90%, 95% and 100%. A consistency index was defined and calculated to understand each care providers’ adherence to clinical pathways. In addition, the average count of physician orders, average amount of cost, Charlson comorbidity index, and recurrence rate were calculated; these variables were considered in frequent physician order sets comparison. Results Records for 3262 patients from 257 hospitals were retrieved. The frequent physician order sets of various frequency thresholds, Charlson comorbidities, and recurrence rates were extracted and computed for comparison among hospitals. A recurrence rate threshold of 2% was established to separate low and high quality of herniorrhaphy at each hospital. Univariable analysis showed that low recurrence rate was associated with high consistency index (70.99 ± 23.88 vs. 52.60 ± 20.30; P < .001), few surgeons at each hospital (3.50 ± 4.41 vs. 7.09 ± 6.57; P < .001), and non-medical center facility type ( P = .042). A multivariable Cox regression analysis indicated an association of low recurrence rates with consistency index only (one percentage increased: OR = 0.973; CI: 0.957–0.990; P = .002). Conclusions The proposed system leveraged the claim records to generate frequent physician order sets at hospitals, thus solving the difficulty in obtaining clinical pathway data. This allows medical professionals and management to conveniently and effectively compare and query similarities and differences in clinical pathways among hospitals. [ABSTRACT FROM AUTHOR]

Background Gastric cancer is among the most common gastrointestinal cancers worldwide. Patients who have undergone surgery for gastric cancer may suffer from malnutrition and potential consequences such as gastrointestinal complications, surgical stress, and cancer cachexia. A tablet PC-based intervention via a mobile application might enhance the early recovery of postgastrectomy patients. Objective The aim of this study was to develop and test a tablet personal computer (PC)-assisted intervention to hasten the recovery of postgastrectomy cancer patients with respect to nutritional status. Methods This single-arm pilot study investigated a tablet PC application developed to serve the functions of nutritional monitoring, medical information management, drainage follow-up, and wound care. All services were delivered by medical professionals. Results Twenty consecutive gastrectomy patients at the National Taiwan University Hospital received perioperative care via this application (App group). During the study period, we retrospectively collected an additional 20 demographically matched gastrectomy cases as a control group. The App group had a lower body weight loss percentage relative to the control group during a 6-month follow-up period (4.8 ± 1.2% vs. 8.7 ± 2.4%; p < 0.01). However, the patients in the App group had more outpatient clinic (OPC) visits than did those in the control group (9.8 ± 0.9 vs. 5.6 ± 0.8; p < 0.01). Conclusions This study supported the feasibility of a tablet PC-based application for the perioperative care of gastric cancer subjects to promote a lower body weight loss and the collection of comprehensive surgical records. [ABSTRACT FROM AUTHOR]

Background and objective Recurrence of hepatocellular carcinoma (HCC) is an important issue despite effective treatments with tumor eradication. Identification of patients who are at high risk for recurrence may provide more efficacious screening and detection of tumor recurrence. The aim of this study was to develop recurrence predictive models for HCC patients who received radiofrequency ablation (RFA) treatment. Methods From January 2007 to December 2009, 83 newly diagnosed HCC patients receiving RFA as their first treatment were enrolled. Five feature selection methods including genetic algorithm (GA), simulated annealing (SA) algorithm, random forests (RF) and hybrid methods (GA+RF and SA+RF) were utilized for selecting an important subset of features from a total of 16 clinical features. These feature selection methods were combined with support vector machine (SVM) for developing predictive models with better performance. Five-fold cross-validation was used to train and test SVM models. Results The developed SVM-based predictive models with hybrid feature selection methods and 5-fold cross-validation had averages of the sensitivity, specificity, accuracy, positive predictive value, negative predictive value, and area under the ROC curve as 67%, 86%, 82%, 69%, 90%, and 0.69, respectively. Conclusions The SVM derived predictive model can provide suggestive high-risk recurrent patients, who should be closely followed up after complete RFA treatment. [ABSTRACT FROM AUTHOR]

Abstract: Background: Pancreaticoduodenectomy (PD) is a major operation with high complication rate. Thereafter, patients may develop morbidity because of the complex reconstruction and loss of pancreatic parenchyma. A well-designed database is very important to address both the short-term and long-term outcomes after PD. Objective: The objective of this research was to build an international PD database implemented with security and clinical rule supporting functions, which made the data-sharing easier and improve the accuracy of data. Methods: The proposed system is a cloud-based application. To fulfill its requirements, the system comprises four subsystems: a data management subsystem, a clinical rule supporting subsystem, a short message notification subsystem, and an information security subsystem. After completing the surgery, the physicians input the data retrospectively, which are analyzed to study factors associated with post-PD common complications (delayed gastric emptying and pancreatic fistula) to validate the clinical value of this system. Results: Currently, this database contains data from nearly 500 subjects. Five medical centers in Taiwan and two cancer centers in Mongolia are participating in this study. A data mining model of the decision tree analysis showed that elderly patients (>76 years) with pylorus-preserving PD (PPPD) have higher proportion of delayed gastric emptying. About the pancreatic fistula, the data mining model of the decision tree analysis revealed that cases with non-pancreaticogastrostomy (PG) reconstruction – body mass index (BMI)>29.65 or PG reconstruction – BMI>23.7 – non-classic PD have higher proportion of pancreatic fistula after PD. Conclusions: The proposed system allows medical staff to collect and store clinical data in a cloud, sharing the data with other physicians in a secure manner to achieve collaboration in research. [Copyright &y& Elsevier]

Abstract: The electrocardiogram (ECG) analysis is one of the most important approaches to cardiac arrhythmia detection. Many algorithms have been proposed, however, the recognition rate is still unsatisfactory due to unreliable feature extraction in signal characteristic analysis or poor generalization capability of the classifier. In this paper, we propose a system for cardiac arrhythmia detection in ECGs with adaptive feature selection and modified support vector machines (SVMs). Wavelet transform-based coefficients and signal amplitude/interval parameters are first enumerated as candidates, but only a few specific ones are adaptively selected for the classification of each class pair. A new classifier, which integrates k-means clustering, one-against-one SVMs, and a modified majority voting mechanism, is proposed to further improve the recognition rate for extremely similar classes. The experimental results show that the proposed ECG analysis approach can obtain a higher recognition rate than the published approaches. By testing the system with more than 100,000 samples in MIT-BIH arrhythmia database, the average recognition rate is 98.92%, and the recognition rate for each class is kept above 92%. [Copyright &y& Elsevier]