Early Prediction of Cardiac Arrest in the Intensive Care Unit Using Explainable Machine Learning: Retrospective Study.

BACKGROUND: Cardiac arrest (CA) is one of the leading causes of death among patients in the intensive care unit (ICU). Although many CA prediction models with high sensitivity have been developed to anticipate CA, their practical application has been challenging due to a lack of generalization and validation. Additionally, the heterogeneity among patients in different ICU subtypes has not been adequately addressed. OBJECTIVE: This study aims to propose a clinically interpretable ensemble approach for the timely and accurate prediction of CA within 24 hours, regardless of patient heterogeneity, including variations across different populations and ICU subtypes. Additionally, we conducted patient-independent evaluations to emphasize the model's generalization performance and analyzed interpretable results that can be readily adopted by clinicians in real-time. METHODS: Patients were retrospectively analyzed using data from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) and the eICU-Collaborative Research Database (eICU-CRD). To address the problem of underperformance, we constructed our framework using feature sets based on vital signs, multiresolution statistical analysis, and the Gini index, with a 12-hour window to capture the unique characteristics of CA. We extracted 3 types of features from each database to compare the performance of CA prediction between high-risk patient groups from MIMIC-IV and patients without CA from eICU-CRD. After feature extraction, we developed a tabular network (TabNet) model using feature screening with cost-sensitive learning. To assess real-time CA prediction performance, we used 10-fold leave-one-patient-out cross-validation and a cross-data set method. We evaluated MIMIC-IV and eICU-CRD across different cohort populations and subtypes of ICU within each database. Finally, external validation using the eICU-CRD and MIMIC-IV databases was conducted to assess the model's generalization ability. The decision mask of the proposed method was used to capture the interpretability of the model. RESULTS: The proposed method outperformed conventional approaches across different cohort populations in both MIMIC-IV and eICU-CRD. Additionally, it achieved higher accuracy than baseline models for various ICU subtypes within both databases. The interpretable prediction results can enhance clinicians' understanding of CA prediction by serving as a statistical comparison between non-CA and CA groups. Next, we tested the eICU-CRD and MIMIC-IV data sets using models trained on MIMIC-IV and eICU-CRD, respectively, to evaluate generalization ability. The results demonstrated superior performance compared with baseline models. CONCLUSIONS: Our novel framework for learning unique features provides stable predictive power across different ICU environments. Most of the interpretable global information reveals statistical differences between CA and non-CA groups, demonstrating its utility as an indicator for clinical decisions. Consequently, the proposed CA prediction system is a clinically validated algorithm that enables clinicians to intervene early based on CA prediction information and can be applied to clinical trials in digital health.

Identification of Atrial Fibrillation With Single-Lead Mobile ECG During Normal Sinus Rhythm Using Deep Learning.

BACKGROUND: The acquisition of single-lead electrocardiogram (ECG) from mobile devices offers a more practical approach to arrhythmia detection. Using artificial intelligence for atrial fibrillation (AF) identification enhances screening efficiency. However, the potential of single-lead ECG for AF identification during normal sinus rhythm (NSR) remains under-explored. This study introduces a method to identify AF using single-lead mobile ECG during NSR. METHODS: We employed three deep learning models: recurrent neural network (RNN), long short-term memory (LSTM), and residual neural networks (ResNet50). From a dataset comprising 13,509 ECGs from 6,719 patients, 10,287 NSR ECGs from 5,170 patients were selected. Single-lead mobile ECGs underwent noise filtering and segmentation into 10-second intervals. A random under-sampling was applied to reduce bias from data imbalance. The final analysis involved 31,767 ECG segments, including 15,157 labeled as masked AF and 16,610 as Healthy. RESULTS: ResNet50 outperformed the other models, achieving a recall of 79.3%, precision of 65.8%, F1-score of 71.9%, accuracy of 70.5%, and an area under the receiver operating characteristic curve (AUC) of 0.79 in identifying AF from NSR ECGs. Comparative performance scores for RNN and LSTM were 0.75 and 0.74, respectively. In an external validation set, ResNet50 attained an F1-score of 64.1%, recall of 68.9%, precision of 60.0%, accuracy of 63.4%, and AUC of 0.68. CONCLUSION: The deep learning model using single-lead mobile ECG during NSR effectively identified AF at risk in future. However, further research is needed to enhance the performance of deep learning models for clinical application.

Explainable Artificial Intelligence Warning Model Using an Ensemble Approach for In-Hospital Cardiac Arrest Prediction: Retrospective Cohort Study.

BACKGROUND: Cardiac arrest (CA) is the leading cause of death in critically ill patients. Clinical research has shown that early identification of CA reduces mortality. Algorithms capable of predicting CA with high sensitivity have been developed using multivariate time series data. However, these algorithms suffer from a high rate of false alarms, and their results are not clinically interpretable. OBJECTIVE: We propose an ensemble approach using multiresolution statistical features and cosine similarity-based features for the timely prediction of CA. Furthermore, this approach provides clinically interpretable results that can be adopted by clinicians. METHODS: Patients were retrospectively analyzed using data from the Medical Information Mart for Intensive Care-IV database and the eICU Collaborative Research Database. Based on the multivariate vital signs of a 24-hour time window for adults diagnosed with heart failure, we extracted multiresolution statistical and cosine similarity-based features. These features were used to construct and develop gradient boosting decision trees. Therefore, we adopted cost-sensitive learning as a solution. Then, 10-fold cross-validation was performed to check the consistency of the model performance, and the Shapley additive explanation algorithm was used to capture the overall interpretability of the proposed model. Next, external validation using the eICU Collaborative Research Database was performed to check the generalization ability. RESULTS: The proposed method yielded an overall area under the receiver operating characteristic curve (AUROC) of 0.86 and area under the precision-recall curve (AUPRC) of 0.58. In terms of the timely prediction of CA, the proposed model achieved an AUROC above 0.80 for predicting CA events up to 6 hours in advance. The proposed method simultaneously improved precision and sensitivity to increase the AUPRC, which reduced the number of false alarms while maintaining high sensitivity. This result indicates that the predictive performance of the proposed model is superior to the performances of the models reported in previous studies. Next, we demonstrated the effect of feature importance on the clinical interpretability of the proposed method and inferred the effect between the non-CA and CA groups. Finally, external validation was performed using the eICU Collaborative Research Database, and an AUROC of 0.74 and AUPRC of 0.44 were obtained in a general intensive care unit population. CONCLUSIONS: The proposed framework can provide clinicians with more accurate CA prediction results and reduce false alarm rates through internal and external validation. In addition, clinically interpretable prediction results can facilitate clinician understanding. Furthermore, the similarity of vital sign changes can provide insights into temporal pattern changes in CA prediction in patients with heart failure-related diagnoses. Therefore, our system is sufficiently feasible for routine clinical use. In addition, regarding the proposed CA prediction system, a clinically mature application has been developed and verified in the future digital health field.

Coinfections with multiple sexually transmitted pathogens in Republic of Korea, 2018-2020.

BACKGROUND: Sexually transmitted infections (STIs) can have serious consequences, and the global STI incidence remains high. However, there is little information on the frequency of STIs with multiple pathogens according to age. Accordingly, we conducted a study to determine the trends of coinfection with sexually transmitted pathogens according to age in the Republic of Korea from 2018 to 2020. METHODS: From January 2018 to December 2020, 65,191 samples of swab, urine, and other types submitted for STI screening were obtained from U2Bio Co. Ltd. (Seoul, Republic of Korea). Multiplex polymerase chain reaction, a sensitive and rapid method for simultaneous detection of STIs caused by multiple different pathogens, was performed using an AccuPower STI4C-Plex Real-Time PCR kit, AccuPower STI8A-Plex Real-Time PCR kit, and AccuPower STI8B-Plex Real-Time PCR kit with an Exicycler 96 Real-Time Quantitative Thermal Block. RESULTS: Of the 65,191 samples tested, 35,366 (54.3%) tested positive for one or more sexually transmitted pathogens. The prevalence of coinfections with two or more sexually transmitted pathogens was inversely proportional to age. Furthermore, the rates of coinfection with sexually transmitted pathogens and age distribution differed according to sex and the sexually transmitted pathogen type. CONCLUSION: This study confirmed that a significant proportion of patients with STIs are coinfected with multiple pathogens. Public health managers could use these results to recognize and prevent STIs according to age.

Prediction of the Cause of Glaucoma Disease Identified by Glaucoma Optical Coherence Tomography Test in Relation to Diabetes and Hypertension at a National Hospital in Seoul: A Retrospective Study.

Glaucoma remains the primary cause of long-term blindness. While diabetes mellitus (DM) and hypertension (HTN) are known to influence glaucoma, other factors such as age and sex may be involved. In this retrospective study, we aimed to investigate the associations between age, sex, DM, HTN, and glaucoma risk. We employed optical coherence tomography (OCT) conducted using a 200 × 200-pixel optic cube (Cirrus HD OCT 6000, version 10.0; Carl Zeiss Meditec, Dublin, CA, USA). Effects obscured by low-test signals were disregarded. Data were amassed from 1337 patients. Among them, 218 and 402 patients had DM and HTN, respectively, with 133 (10%) exhibiting both. A sex-based comparison revealed slightly greater retinal nerve fiber layer (RNFL) and ganglion cell-inner plexiform layer (GCIPL) thickness in females. Patients without DM and HTN were predominantly in their 50 s and 60 s, whereas DM and HTN were most prevalent in those in their 60 s and 70 s. Both RNFL and GCIPL thicknesses decreased with advancing age in most patients. The study revealed that older individuals were more prone to glaucoma than younger individuals, with a higher incidence among patients with DM and HTN and reduced RNFL and GCIPL thicknesses. Furthermore, early detection before advancing age could furnish valuable preventive insights.

Ensemble Learning Method for In-Hospital Cardiac Arrest Prediction.

Cardiac arrest prediction for multivariate time series data have been developed and obtained high precision performance. However, these algorithms still did not achieved high sensitivity and suffer from a high false-alarm. Therefore, we propose a ensemble approach for prediction satisfying precision-recall result compared than other machine learning methods. As a result, our proposed method obtained an overall area under precision-recall curve of 46.7%. It is possible to more accurately respond rapidly cardiac arrest event.

The biphasic impact of apolipoprotein E ε4 allele on age-related hearing loss.

Both the ε4 variant of the apolipoprotein E (APOE) gene and hearing loss are well-known risk factors for Alzheimer's disease. However, previous studies have produced inconsistent findings regarding the association between APOE genotypes and hearing levels, necessitating further investigation. The aim of this study was to investigate the relationship between APOE genotypes and hearing levels. This retrospective study analyzed clinical data from a clinical data warehouse of seven affiliated Catholic Medical Center hospitals. The study included 1,162 participants with records of APOE genotypes, audiometric tests, and cognitive function tests. In Generalized linear mixed model analysis, ε4 carriers exhibited lower pure tone audiometry thresholds with an estimate of -0.353 (SE = 0.126, p = 0.005). However, the interaction term for age and APOE ε4 had a coefficient of 0.577 (SE = 0.214 p = 0.006), suggesting that the APOE ε4 gene may accelerate hearing deterioration with age. Subgroup analysis based on an age cut-off of 75 revealed that ε4 carriers had better hearing at younger ages, but showed no significant difference at older ages. These results indicate that the ε4 allele may have a biphasic effect on hearing levels depending on age.

A retrospective study on the status of sexually transmitted co-infections in university hospitals in Korea from 2017 to 2021.

Background: The prevalence of sexually transmitted infections (STIs) remains high worldwide. Despite the worldwide increase in the incidence of STIs every year, there are few reports on the frequency of STIs with different pathogens according to age and gender. Accordingly, a study was conducted to determine trends in co-infection with STIs by age and gender in Cheonan, South Korea from 2017 to 2021. Objectives: To identify trends by age or sex in co-infection of STIs in this region. Design: A retrospective study was conducted on clinical samples examined at Dankook University Hospital from January 2017 to November 2021. A total of 3297 specimens were collected from patients visiting Dankook University Hospital (Cheonan, Korea), and statistical analysis was performed on patients ranging in age from 1 day to 93 years. Methods: Multiplex polymerase chain reaction, the most efficient method to diagnose a bacterial infection, was performed using an MJ Research PTC-200 Thermal Cycler (Marshall Scientific, Richmond, VA, USA) and a Seeplex STD Detection Kit (Seegene, Seoul, Republic of Korea). The co-infection rate with STI pathogens was analyzed according to age and sex. Results: Of the 3297 clinical samples, 1017 (30.9%) tested positive for sexually transmitted pathogens, ranging from one to six co-infections. Analysis of the co-infection rate by age revealed that the average age gradually decreased as the total number of co-infection pathogens increased. The co-infection percentage and age distribution of STIs differed according to sex. Co-infection was more prevalent in female patients. Furthermore, co-infection in male patients occurred frequently in the 30-39-year-old group, while those in female patients occurred in the 20-29- and 30-39-year-old groups. Conclusion: Our statistical analysis showed that STI co-infections were more common among younger than older people. Therefore, it helps in recognizing STIs at a young age and provides possible indicator data to prevent STIs at a young age. In addition, further research is needed on co-infection in other regions.

Association between HbA1c and hearing loss: a tertiary care center-based study.

The purpose of this study was to investigate the correlation between glycated hemoglobin (HbA1c) levels and hearing loss (HL) using data from a tertiary hospital. Our hypothesis regarding the relationship between HL and HbA1c levels was that elevated HbA1c levels are associated with an increased risk of HL. We retrospectively reviewed the medical charts of patients diagnosed with sensorineural HL or diabetes between 2006 and 2021 at the Catholic Medical Center (CMC). Data were collected from the CMC's Clinical Data Warehouse. Participants were selected from patients who were prescribed pure-tone audiometry and an HbA1c blood test. The survey was completed for 5287 participants. The better ear pure-tone audiometry (PTA) for air conduction thresholds at 500, 1000, 2000, and 4000 Hz was calculated. Sensorineural HL was defined as a better ear PTA of 25 dB or higher. We used the HbA1c level as a diagnostic criterion for diabetes. The following criteria were used to define the HbA1c level: normal, HbA1c level below 5.6%; prediabetes, level between 5.6 and 6.4%; and diabetes, level of 6.5% or more. Among 5287 participants, 1129 were categorized as normal, 2119 as prediabetic, and 2039 as diabetic. The diabetic group was significantly older (p < 0.05). The PTA also significantly deteriorated in the diabetes group (p < 0.05). We analyzed the effects of age, sex, and HbA1c level on frequency-specific hearing using multiple regression. The hearing thresholds at all frequencies deteriorated significantly with increasing age and HbA1c level (p < 0.05). A case-control study was also performed to facilitate a comprehensive comparison between distinct groups. The participants were categorized into two groups: a case (PTA > 25 dB) and control group (PTA ≤ 25 dB), based on their PTA threshold of four frequencies. After adjusting for age and sex, we found no significant odds ratio (OR) of HL between the prediabetes group and the normal group. Notably, the OR of HL was significantly higher in the diabetes group with each PTA threshold and frequency. The 6.3% HbA1c level cutoff value was determined by analyzing the receiver operating characteristic curve for predicting hearing impairment > 25 dB. Diabetes was associated with hearing loss in all frequency ranges, particularly at high frequencies. Screening for HL is strongly recommended for patients with elevated HbA1c levels.

Development and evaluation of smartphone usage management system for preventing problematic smartphone use.

Objective: The increased use of smartphones has led to several problems, including excessive smartphone use and the decreased self-ability to control smartphone use. To prevent these problems, the MindsCare app was developed as a method of self-management and intervention based on an evaluation of smartphone usage. We designed the MindsCare app to manage smartphone usage and prevent problematic smartphone use by providing personalized interventions. Methods: We recruited 342 Korean participants over the age of 20 and asked them to use MindsCare for 13 weeks. Subsequently, we evaluated the changes in average smartphone usage time and the usability of the app. We designed a usability evaluation questionnaire based on the Technology Acceptance Model and conducted factor and reliability analyses on the participants' responses. In the eighth week of the study, participants responded to a survey on the usability of the app. We ultimately collected data from 190 participants. Results: The average score for the usability of the system was 3.61 on a five-point Likert scale, and approximately 58% of the participants responded positively to the evaluation items. In addition, our analysis of MindsCare data revealed a significant reduction in average smartphone use time in the eighth week compared to the baseline (t = 3.47, p = 0.001). Structural equation model analysis revealed that effort expectancy and performance expectancy had a positive relation with behavior intention for the app. Conclusions: Through this study, we confirmed the MindsCare app's smartphone usage time reduction effect and proved its good usability. As a result, MindsCare may contribute to achieving users' goals of reducing problematic smartphone use.

Transforming Thyroid Cancer Diagnosis and Staging Information from Unstructured Reports to the Observational Medical Outcome Partnership Common Data Model.

BACKGROUND: Cancer staging information is an essential component of cancer research. However, the information is primarily stored as either a full or semistructured free-text clinical document which is limiting the data use. By transforming the cancer-specific data to the Observational Medical Outcome Partnership Common Data Model (OMOP CDM), the information can contribute to establish multicenter observational cancer studies. To the best of our knowledge, there have been no studies on OMOP CDM transformation and natural language processing (NLP) for thyroid cancer to date. OBJECTIVE: We aimed to demonstrate the applicability of the OMOP CDM oncology extension module for thyroid cancer diagnosis and cancer stage information by processing free-text medical reports. METHODS: Thyroid cancer diagnosis and stage-related modifiers were extracted with rule-based NLP from 63,795 thyroid cancer pathology reports and 56,239 Iodine whole-body scan reports from three medical institutions in the Observational Health Data Sciences and Informatics data network. The data were converted into the OMOP CDM v6.0 according to the OMOP CDM oncology extension module. The cancer staging group was derived and populated using the transformed CDM data. RESULTS: The extracted thyroid cancer data were completely converted into the OMOP CDM. The distributions of histopathological types of thyroid cancer were approximately 95.3 to 98.8% of papillary carcinoma, 0.9 to 3.7% of follicular carcinoma, 0.04 to 0.54% of adenocarcinoma, 0.17 to 0.81% of medullary carcinoma, and 0 to 0.3% of anaplastic carcinoma. Regarding cancer staging, stage-I thyroid cancer accounted for 55 to 64% of the cases, while stage III accounted for 24 to 26% of the cases. Stage-II and -IV thyroid cancers were detected at a low rate of 2 to 6%. CONCLUSION: As a first study on OMOP CDM transformation and NLP for thyroid cancer, this study will help other institutions to standardize thyroid cancer-specific data for retrospective observational research and participate in multicenter studies.

Differential effects of ziprasidone and haloperidol on immobilization-stress-induced CRF mRNA expression in the hypothalamic paraventricular nucleus of rats.

OBJECTIVES: Corticotropin-releasing factor (CRF) plays a prominent role in mediating the effect of stressors on the hypothalamic-pituitary-adrenal axis. In this study, we examined the effects of chronic administration of second-generation antipsychotic drug ziprasidone on CRF mRNA expression in the hypothalamic paraventricular nucleus (PVN) of rats with or without immobilization stress. METHODS: The rats were subjected to immobilization stress 2 h/day for 3 weeks. The effect of ziprasidone (2.5 mg/kg, 21 days) on CRF mRNA expression was determined using in situ hybridization of tissue sections from the rat hypothalamic PVN. Haloperidol (1.0 mg/kg, 21 days) was used for comparison. RESULTS: Haloperidol increased the expression of CRF mRNA in the PVN under basal conditions, whereas ziprasidone had no effect. Chronic immobilization stress increased CRF expression. The chronic administration of ziprasidone prevented the increase in CRF mRNA expression caused by immobilization stress. CONCLUSIONS: These results suggest that ziprasidone may have a regulatory effect on the stress-induced CRF mRNA expression and a role in the treatment of depressive mood symptom.

TGF-beta promotes Th17 cell development through inhibition of SOCS3.

TGF-beta, together with IL-6 and IL-21, promotes Th17 cell development. IL-6 and IL-21 induce activation of STAT3, which is crucial for Th17 cell differentiation, as well as the expression of suppressor of cytokine signaling (SOCS)3, a major negative feedback regulator of STAT3-activating cytokines that negatively regulates Th17 cells. However, it is still largely unclear how TGF-beta regulates Th17 cell development and which TGF-beta signaling pathway is involved in Th17 cell development. In this report, we demonstrate that TGF-beta inhibits IL-6- and IL-21-induced SOCS3 expression, thus enhancing as well as prolonging STAT3 activation in naive CD4(+)CD25(-) T cells. TGF-beta inhibits IL-6-induced SOCS3 promoter activity in T cells. Also, SOCS3 small interfering RNA knockdown partially compensates for the action of TGF-beta on Th17 cell development. In mice with a dominant-negative form of TGF-beta receptor II and impaired TGF-beta signaling, IL-6-induced CD4(+) T cell expression of SOCS3 is higher whereas STAT3 activation is lower compared with wild-type B6 CD4(+) T cells. The addition of a TGF-beta receptor I kinase inhibitor that blocks Smad-dependent TGF-beta signaling greatly, but not completely, abrogates the effect of TGF-beta on Th17 cell differentiation. Our data indicate that inhibition of SOCS3 and, thus, enhancement of STAT3 activation is at least one of the mechanisms of TGF-beta promotion of Th17 cell development.

Machine Learning Approach to Predict the Probability of Recurrence of Renal Cell Carcinoma After Surgery: Prediction Model Development Study.

BACKGROUND: Renal cell carcinoma (RCC) has a high recurrence rate of 20% to 30% after nephrectomy for clinically localized disease, and more than 40% of patients eventually die of the disease, making regular monitoring and constant management of utmost importance. OBJECTIVE: The objective of this study was to develop an algorithm that predicts the probability of recurrence of RCC within 5 and 10 years of surgery. METHODS: Data from 6849 Korean patients with RCC were collected from eight tertiary care hospitals listed in the KOrean Renal Cell Carcinoma (KORCC) web-based database. To predict RCC recurrence, analytical data from 2814 patients were extracted from the database. Eight machine learning algorithms were used to predict the probability of RCC recurrence, and the results were compared. RESULTS: Within 5 years of surgery, the highest area under the receiver operating characteristic curve (AUROC) was obtained from the naïve Bayes (NB) model, with a value of 0.836. Within 10 years of surgery, the highest AUROC was obtained from the NB model, with a value of 0.784. CONCLUSIONS: An algorithm was developed that predicts the probability of RCC recurrence within 5 and 10 years using the KORCC database, a large-scale RCC cohort in Korea. It is expected that the developed algorithm will help clinicians manage prognosis and establish customized treatment strategies for patients with RCC after surgery.

Behavioral Intention to Use a Smartphone Usage Management Application Between a Non-Problematic Smartphone Use Group and a Problematic Use Group.

Despite the many advantages of smartphone in daily life, there are significant concerns regarding their problematic use. Therefore, several smartphone usage management applications have been developed to prevent problematic smartphone use. The purpose of this study is to investigate the factors of users' behavioral intention to use smartphone usage management applications. Participants were divided into a smartphone use control group and a problematic use group to find significant intergroup path differences. The research model of this study is fundamentally based on the Technology Acceptance Model and Expectation-Confirmation Theory. Based on this theorem, models were modified to best suit the case of problematic smartphone use intervention by smartphone application. We conducted online surveys on 511 randomly selected smartphone users aged 20-60 in South Korea, in 2018. The Smartphone Addiction Proneness Scale was used to measure participants' smartphone dependency. Descriptive statistics were used for the demographic analysis and collected data were analyzed using IBM SPSS Statistics 24.0 and Amos 24.0. We found that in both non-problematic smartphone use group and problematic smartphone use group, facilitating factors and perceived security positively affect the intentions of users to use the application. One distinct difference between the groups was that the latter attributed a lower importance to perceived security than the former. Some of our highlighted unique points are envisioned to provide intensive insights for broadening knowledge about technology acceptance in the field of e-Addictology.

Machine learning prediction of dropping out of outpatients with alcohol use disorders.

BACKGROUND: Alcohol use disorder (AUD) is a chronic disease with a higher recurrence rate than that of other mental illnesses. Moreover, it requires continuous outpatient treatment for the patient to maintain abstinence. However, with a low probability of these patients to continue outpatient treatment, predicting and managing patients who might discontinue treatment becomes necessary. Accordingly, we developed a machine learning (ML) algorithm to predict which the risk of patients dropping out of outpatient treatment schemes. METHODS: A total of 839 patients were selected out of 2,206 patients admitted for AUD in three hospitals under the Catholic Central Medical Center in Korea. We implemented six ML models-logistic regression, support vector machine, k-nearest neighbor, random forest, neural network, and AdaBoost-and compared the prediction performances thereof. RESULTS: Among the six models, AdaBoost was selected as the final model for recommended use owing to its area under the receiver operating characteristic curve (AUROC) of 0.72. The four variables affecting the prediction based on feature importance were the length of hospitalization, age, residential area, and diabetes. CONCLUSION: An ML algorithm was developed herein to predict the risk of patients with AUD in Korea discontinuing outpatient treatment. By testing and validating various machine learning models, we determined the best performing model, AdaBoost, as the final model for recommended use. Using this model, clinicians can manage patients with high risks of discontinuing treatment and establish patient-specific treatment strategies. Therefore, our model can potentially enable patients with AUD to successfully complete their treatments by identifying them before they can drop out.

PD-1 restraint of regulatory T cell suppressive activity is critical for immune tolerance.

Inhibitory signals through the PD-1 pathway regulate T cell activation, T cell tolerance, and T cell exhaustion. Studies of PD-1 function have focused primarily on effector T cells. Far less is known about PD-1 function in regulatory T (T reg) cells. To study the role of PD-1 in T reg cells, we generated mice that selectively lack PD-1 in T reg cells. PD-1-deficient T reg cells exhibit an activated phenotype and enhanced immunosuppressive function. The in vivo significance of the potent suppressive capacity of PD-1-deficient T reg cells is illustrated by ameliorated experimental autoimmune encephalomyelitis (EAE) and protection from diabetes in nonobese diabetic (NOD) mice lacking PD-1 selectively in T reg cells. We identified reduced signaling through the PI3K-AKT pathway as a mechanism underlying the enhanced suppressive capacity of PD-1-deficient T reg cells. Our findings demonstrate that cell-intrinsic PD-1 restraint of T reg cells is a significant mechanism by which PD-1 inhibitory signals regulate T cell tolerance and autoimmunity.

Expression and functional significance of SOCS-1 and SOCS-3 in astrocytes.

Astrocytes play a number of important physiological roles in CNS homeostasis. Inflammation stimulates astrocytes to secrete cytokines and chemokines that guide macrophages/microglia and T cells to sites of injury/inflammation. Herein, we describe how these processes are controlled by the suppressor of cytokine signaling (SOCS) proteins, a family of proteins that negatively regulate adaptive and innate immune responses. In this study, we describe that the immunomodulatory cytokine IFN-beta induces SOCS-1 and SOCS-3 expression in primary astrocytes at the transcriptional level. SOCS-1 and SOCS-3 transcriptional activity is induced by IFN-beta through IFN-gamma activation site (GAS) elements within their promoters. Studies in STAT-1alpha-deficient astrocytes indicate that STAT-1alpha is required for IFN-beta-induced SOCS-1 expression, while STAT-3 small interfering RNA studies demonstrate that IFN-beta-induced SOCS-3 expression relies on STAT-3 activation. Specific small interfering RNA inhibition of IFN-beta-inducible SOCS-1 and SOCS-3 in astrocytes enhances their proinflammatory responses to IFN-beta stimulation, such as heightened expression of the chemokines CCL2 (MCP-1), CCL3 (MIP-1alpha), CCL4 (MIP-1beta), CCL5 (RANTES), and CXCL10 (IP-10), and promoting chemotaxis of macrophages and CD4(+) T cells. These results indicate that IFN-beta induces SOCS-1 and SOCS-3 in primary astrocytes to attenuate its own chemokine-related inflammation in the CNS.

Novel Antihypertension Mechanism of ß-Glucan by Corin and ANP-Mediated Natriuresis in Mice.

Many of the ß-glucans are known to have antihypertensive activities, but, except for angiotensin-converting enzyme II inhibition, the underlying mechanisms remain unclear. Corin is an atrial natriuretic peptide (ANP)-converting enzyme. Activated corin cleaves pro-ANP to ANP, which regulates water-sodium balance and lowers blood pressure. Here, we reported a novel antihypertensive mechanism of ß-glucans, involved with corin and ANP in mice. We showed that multiple oral administrations of ß-glucan induced the expression of corin and ANP, and also increased natriuresis in mice. Microarray analysis showed that corin gene expression was only upregulated in mice liver by multiple, not single, oral administrations of the ß-glucan fraction of Phellinus baumii (BGF). Corin was induced in liver and kidney tissues by the ß-glucans from zymosan and barley, as well as by BGF. In addition to P. baumii, ß-glucans from two other mushrooms, Phellinus linteus and Ganoderma lucidum, also induced corin mRNA expression in mouse liver. ELISA immunoassays showed that ANP production was increased in liver tissue by all the ß-glucans tested, but not in the heart and kidney. Urinary sodium excretion was significantly increased by treatment with ß-glucans in the order of BGF, zymosan, and barley, both in 1% normal and 10% high-sodium diets. In conclusion, we found that the oral administration of ß-glucans could induce corin expression, ANP production, and sodium excretion in mice. Our findings will be helpful for investigations of ß-glucans in corin and ANP-related fields, including blood pressure, salt-water balance, and circulation.

Differential effects of ziprasidone and haloperidol on immobilization stress-induced mRNA BDNF expression in the hippocampus and neocortex of rats.

Recent in vivo and in vitro experiments have demonstrated that second-generation antipsychotic drugs (SGAs) might have neuroprotective effects. Ziprasidone is a SGA that is efficacious in the treatment of schizophrenia. In this study, we sought to analyze the effects of ziprasidone on the expression of the neuroprotective protein brain-derived neurotrophic factor (BDNF) in the rat hippocampus and neocortex, with or without immobilization stress. The effect of ziprasidone (2.5mg/kg) on the expression of BDNF mRNA was determined by in situ hybridization in tissue sections from the rat hippocampus and neocortex. Haloperidol (1.0mg/kg) was used for comparison. Haloperidol strongly decreased the expression of BDNF mRNA in both the hippocampal and cortical regions, with or without immobilization stress (p<0.01). In contrast, the administration of ziprasidone significantly attenuated the immobilization stress-induced decrease in BDNF mRNA expression in the rat hippocampus and neocortex (p<0.01). Ziprasidone exhibited differential effects on BDNF mRNA expression in the rat hippocampus and neocortex. These results suggest that ziprasidone might have a neuroprotective effect by recovering stress-induced decreases in BDNF mRNA expression.