Soongsil University

Although heterogeneity of treatment effects (HTEs) is commonly observed in clinical trials, it has received little attention in studies on transcranial direct current stimulation (tDCS). This study identified the presence of HTE in tDCS treatment among participants with symptomatic knee osteoarthritis (KOA) and explored participant characteristics associated with this heterogeneity. This secondary analysis of a randomized clinical trial included 120 participants with symptomatic KOA who received 15 daily sessions of home-based 2-mA active or sham tDCS over 3 weeks. We used a multitrajectory analysis to identify distinct subgroups based on the longitudinal trajectories of KOA pain and symptoms from baseline to 3 months postintervention, capturing differential responses to tDCS. We then performed bivariate analyses to examine associations between trajectory groups and baseline demographic, clinical, and quantitative sensory testing characteristics. In the active tDCS group, 2 distinct trajectories emerged: "low initial symptoms with significant improvement" (high responders; n = 28) and "high initial symptoms with minimal improvement" (low responders; n = 32). Compared to high responders, low responders had a higher body mass index, lower educational attainment, and greater pain catastrophizing (all P < 0.05). Low responders also exhibited lower pressure pain thresholds at both the medial knee and trapezius, higher punctate mechanical pain at both the patella and hand, lower conditioned pain modulation, and higher cold pain intensity at baseline (all P < 0.05). No notable HTE was observed in the sham tDCS group. Participants exhibited varying responses to active tDCS. The characteristics associated with HTE may inform the development of personalized stimulation protocols.

Hereditary cancers frequently arise from germline pathogenic variants, yet only a small proportion of reported variants have been clinically classified, leaving most missense variants unresolved as variants of uncertain significance (VUS). Although recent machine-learning approaches have explored disease-specific or gene-specific contexts to improve pathogenicity prediction, these models remain fundamentally limited by the scarcity of labeled data and the underutilization of abundant VUS. We propose a deep-learning framework that integrates autoencoder pretraining with a deep ensemble strategy to improve variant pathogenicity prediction, effectively leverage unlabeled VUS during pretraining, and reduce uncertainty arising from limited training samples. To validate each component of our framework, we evaluated its performance under both disease-specific and gene-specific training setups. Experiments on ClinVar variants from BRCA1, BRCA2, MLH1, and MSH2 showed that our framework achieved the best performance in the gene-specific setup for BRCA1—likely because BRCA1 contains substantially more gene-specific training data than the other genes—whereas the disease-specific setup yielded superior results for the remaining genes, which had comparatively limited gene-specific samples. Overall, our method significantly outperformed existing approaches. We also introduce an interpretability approach that provides variant-level importance profiles across pathogenicity classes, thereby enhancing transparency and clinical applicability. Moreover, by projecting feature-level importance scores into a two-dimensional space, we demonstrate that pretraining enables the model to learn distinctly different feature representations, illustrating how pretraining and ensemble learning synergistically contribute to improved predictive performance. Our framework preserves the specificity of disease- and gene-specific approaches, overcomes data scarcity through VUS-guided pretraining and ensembling, and offers interpretable outcomes that may be helpful for clinical decision support. Moreover, our results suggest a promising direction for pathogenicity prediction of rare missense variants and indicate that the proposed framework may be extendable to additional genes under appropriate data and modeling conditions.

Background The growing burden of obesity has profoundly influenced the epidemiology and phenotype of diabetes. This study aimed to compare the epidemiology and clinical features between obese and non-obese diabetes in Korean adults using nationwide database.Methods We analyzed data from the Korea National Health and Nutrition Examination Survey (2012–2023) to evaluate the prevalence and management of diabetes, as well as associated comorbidities. Data from the Korean National Health Insurance Service were used to assess antidiabetic medication use, metabolic surgery trends, and cancer outcomes.Results Diabetes prevalence was nearly twice as high in adults with obesity compared with those without (17.6% vs. 9.5%), with the larger difference observed in individuals aged 30 to 59 years. Obese diabetes was associated with higher rates of hypertension and dyslipidemia and lower rates of achieving glycemic, blood pressure, and lipid targets; only 21.0% achieved all three goals. Although sodium-glucose cotransporter 2 inhibitors and thiazolidinediones were more frequently prescribed in obese diabetes, overall use remained low. Metabolic surgery was less common in individuals with diabetes than in those without; sleeve gastrectomy predominated, while Roux-en-Y gastric bypass was performed more often in those with diabetes. Higher body mass index was associated with increased incidence of thyroid, breast, prostate, and kidney cancers.Conclusion Obese diabetes represents a distinct, high-risk phenotype in Korea, characterized by a greater cardiometabolic burden and suboptimal risk-factor control. Comprehensive management strategies integrating weight reduction with metabolic and cardiovascular risk control are essential to improve outcomes in this population.

This study aimed to evaluate the associations of baseline income, cumulative income exposure, and income volatility with the incidence of pancreatic and biliary tract cancers in a nationwide Korean cohort. We analyzed 3,361,091 adults aged 30-65 years who underwent the 2012 National Health Insurance Service (NHIS) health screening. Income level was derived from insurance premium data assessed over the five years preceding baseline (2008-2012) and categorized into baseline income quartiles, cumulative exposure to low or high income, and income volatility based on annual percentage changes. Incident pancreatic and biliary tract cancers were identified using diagnostic codes and the copayment reduction registry. Associations were evaluated using Cox proportional hazards models with adjustment for demographic, lifestyle, and clinical covariates, and cumulative incidence was compared using Kaplan-Meier curves. During a median follow-up of 9.6 years, 14,469 pancreatic cancers and 6,647 biliary tract cancers were newly diagnosed. Lower baseline income was associated with a higher risk of pancreatic and biliary tract cancers, whereas sustained high-income exposure was associated with reduced risk. Cumulative low-income exposure showed a positive linear trend with pancreatic cancer incidence. Income volatility was modestly associated with pancreatic cancer and was positively associated with biliary tract cancer in the fully adjusted model. These associations were generally consistent across subgroups, with a stronger inverse association between prolonged high-income exposure and pancreatic cancer among individuals without diabetes. Income level and income stability were significantly associated with the incidence of pancreatic and biliary tract cancers. Lower baseline income was associated with higher risk, whereas sustained high-income exposure was protective. Income volatility was associated with increased cancer risk, particularly for biliary tract cancer. These findings highlight the importance of incorporating income dynamics into cancer prevention strategies and addressing socioeconomic instability among vulnerable populations.

Accurate prediction of the state of charge (SoC) of batteries is essential for ensuring the safe, reliable, and uninterrupted operation of electric vehicles (EVs). The prediction fundamentally depends on the ability to accurately predict power consumption. This study investigates the use of GPS-derived information to support SoC prediction, with a particular focus on repeated loop routes such as campus shuttles and closed-circuit EV operations. Real-world driving data are collected using a self-built electric vehicle equipped with a custom battery management system (BMS). These data are used to train three deep learning models, namely gated recurrent unit (GRU), long short-term memory (LSTM), and Transformer, to predict the future SoC of the EV. Experimental results show that the GPS-assisted model consistently outperforms the non-GPS baseline, achieving up to a 23% improvement in prediction accuracy for one-minute-ahead predictions and up to a 76% improvement for ten-minute-ahead predictions. These results demonstrate that GPS-assisted SoC prediction can be effective for forward-looking energy management in practical electric mobility applications.