1.Machine Learning Models to Identify Individuals With Imminent Suicide Risk Using a Wearable Device: A Pilot Study
Jumyung UM ; Jongsu PARK ; Dong Eun LEE ; Jae Eun AHN ; Ji Hyun BAEK
Psychiatry Investigation 2025;22(2):156-166
Objective:
We aimed to determine whether individuals at immediate risk of suicide could be identified using data from a commercially available wearable device.
Methods:
Thirty-nine participants experiencing acute depressive episodes and 20 age- and sex-matched healthy controls wore a commercially available wearable device (Galaxy Watch Active2) for two months. We collected data on activities, sleep, and physiological metrics like heart rate and heart rate variability using the wearable device. Participants rated their mood spontaneously twice daily on a Likert scale displayed on the device. Mood ratings by clinicians were performed at weeks 0, 2, 4, and 8. The suicide risk was assessed using the Hamilton Depression Rating Scale’s suicide item score (HAMD-3). We developed two predictive models using machine learning: a single-level model that processed all data simultaneously to identify those at immediate suicide risk (HAMD-3 scores ≥1) and a multilevel model. We compared the predictions of imminent suicide risk from both models.
Results:
Both the single-step and multi-step models effectively predicted imminent suicide risk. The multi-step model outperformed the single-step model in predicting imminent suicide risk with area under the curve scores of 0.89 compared to 0.88. In the multi-step model, the HAMD total score and heart rate variability were most significant, whereas in the single-step model, the HAMD total score and diagnosis were key predictors.
Conclusion
Wearable devices are a promising tool for identifying individuals at immediate risk of suicide. Future research with more refined temporal resolution is recommended.
2.Observer-Blind Randomized Control Trial for the Effectiveness of Intensive Case Management in Seoul: Clinical and Quality-of-Life Outcomes for Severe Mental Illness
Hye-Young MIN ; Seung-Hee AHN ; Jeung Suk LIM ; Hwa Yeon SEO ; Sung Joon CHO ; Seung Yeon LEE ; Dohhee KIM ; Kihoon YOU ; Hyun Seo CHOI ; Su-Jin YANG ; Jee Eun PARK ; Bong Jin HAHM ; Hae Woo LEE ; Jee Hoon SOHN
Psychiatry Investigation 2025;22(5):513-521
Objective:
In South Korea, there is a significant gap in systematic, evidence-based research on intensive case management (ICM) for individuals with severe mental illness (SMI). This study aims to evaluate the effectiveness of ICM through a randomized controlled trial (RCT) comparing ICM with standard case management (non-ICM).
Methods:
An RCT was conducted to assess the effectiveness of Seoul-intensive case management (S-ICM) vs. non-ICM in individuals with SMI in Seoul. A total of 78 participants were randomly assigned to either the S-ICM group (n=41) or the control group (n=37). Various clinical assessments, including the Brief Psychiatric Rating Scale (BPRS), Montgomery–Åsberg Depression Rating Scale, Health of the Nation Outcome Scale, and Clinical Global Impression-Improvement (CGI-I), along with quality-of-life measures such as the WHO Disability Assessment Schedule, WHO Quality of Life scale, and Multidimensional Scale of Perceived Social Support (MSPSS) were evaluated over a 3-month period. Statistical analyses, including analysis of covariance and logistic regression, were used to determine the effectiveness of S-ICM.
Results:
The S-ICM group had significantly lower odds of self-harm or suicidal attempts compared to the control group (adjusted odds ratio [aOR]=0.30, 95% confidence interval [CI]: 0.21–1.38). Psychiatric symptoms measured by the BPRS and perceived social support measured by the MSPSS significantly improved in the S-ICM group. The S-ICM group also had significantly higher odds of CGI-I compared to the control group (aOR=8.20, 95% CI: 2.66–25.32).
Conclusion
This study provides inaugural evidence on the effectiveness of S-ICM services, supporting their standardization and potential nationwide expansion.
3.Prevalence and factors influencing postpartum depression and its culture-specific cutoffs for women in Asia: a scoping review
Bora MOON ; Hyun Kyoung KIM ; Ju-Hee NHO ; Hyunkyung CHOI ; ChaeWeon CHUNG ; Sook Jung KANG ; Ju Hee KIM ; Ju-Young LEE ; Sihyun PARK ; Gisoo SHIN ; Ju-Eun SONG ; Min Hee LEE ; Sue KIM
The Ewha Medical Journal 2025;48(1):e15-
The prevalence of postpartum depression (PPD) in Asia is reported to range from 13.53% to 22.31%. However, there remains a gap in the identification of PPD, particularly regarding cultural cutoff points. Therefore, the purpose of this scoping review was to determine the prevalence and associated factors of PPD in Eastern, South-eastern, Western, and Southern Asian countries and analyze the cutoff points of the Edinburgh Postnatal Depression Scale (EPDS) used across these countries. Following Arksey and O'Malley’s five-step scoping review framework, the population was defined as mothers, the concept as the EPDS, and the context as the Asian region. A literature search was conducted using PubMed, Embase, CINAHL, PsycINFO, and Web of Science. The data analysis focused on demographic characteristics, EPDS cutoffs and features, PPD prevalence, and its associated factors. Nineteen studies were selected. Most countries used translated versions of the EPDS with demonstrated reliability and validity. The cutoff scores varied, with most using scores of 10 or higher. The prevalence of PPD ranged from 5.1% to 78.7%. Key associated factors for PPD included cultural factors such as relationships with in-laws and preferences for the newborn’s sex. To improve the accuracy of PPD screening in Asia, the EPDS should be used consistently, and appropriate cutoff criteria must be established. In addition, prevention strategies and programs that reflect the cultural characteristics and social context of Asia need to be developed for the early detection and prevention of PPD.
4.Characteristics and outcomes of portal vein thrombosis in patients with inflammatory bowel disease in Korea
Ki Jin KIM ; Su-Bin SONG ; Jung-Bin PARK ; June Hwa BAE ; Ji Eun BAEK ; Ga Hee KIM ; Min-Jun KIM ; Seung Wook HONG ; Sung Wook HWANG ; Dong-Hoon YANG ; Byong Duk YE ; Jeong-Sik BYEON ; Seung-Jae MYUNG ; Suk-Kyun YANG ; Chang Sik YU ; Yong-Sik YOON ; Jong-Lyul LEE ; Min Hyun KIM ; Ho-Su LEE ; Sang Hyoung PARK
The Korean Journal of Internal Medicine 2025;40(2):243-250
Background/Aims:
Portal vein thrombosis (PVT) frequently occurs in patients with inflammatory bowel disease (IBD), particularly when influenced by factors such as abdominal infections, IBD flare-ups, or surgical procedures. The implications of PVT range from immediate issues such as intestinal ischemia to long-term concerns including portal hypertension and its complications. However, there is a notable gap in comprehensive studies on PVT in IBD, especially with the increasing incidence of IBD in Asia. This research aimed to evaluate the clinical features and outcomes of PVT in patients with IBD at a leading hospital in South Korea.
Methods:
This retrospective analysis reviewed adult patients diagnosed with both IBD and PVT from 1989 to 2021 at a renowned South Korean medical center. The study focused on patient characteristics, specifics of PVT, administered treatments, and outcomes, all confirmed through enhanced CT scans.
Results:
A total of 78 patients met the study’s criteria. Notably, only 20.5% (16/78) were treated with oral anticoagulants; however, a vast majority (96.2%; 75/78) achieved complete radiographic resolution (CRR). When comparing patients receiving anticoagulants to those who did not, a significant preference for anticoagulant use was observed in cases where the main portal vein was affected, as opposed to just the left or right veins (p = 0.006). However, multivariable analysis indicated that neither anticoagulant use nor previous surgeries significantly impacted CRR.
Conclusions
Patients with IBD and PVT generally had favorable outcomes, regardless of anticoagulant use.
5.Key factors for successful eruption of the mandibular third molar after extraction of the mandibular second molar
Jung Jin PARK ; Yoonjeong NOH ; Yoon Jeong CHOI ; Jae Hyun PARK ; Ji Hyun LEE ; Chooryung Judi CHUNG ; Kyung-Ho KIM
The Korean Journal of Orthodontics 2025;55(2):154-163
Objective:
Extraction of the mandibular second molar (L7) and substitution by the mandibular third molar (L8) is an effective treatment option. This study aimed to evaluate spontaneously erupted L8 occlusion after L7 extraction, and identify the influencing factors.
Methods:
This study assessed 46 L8 from 28 patients using dental study models, panoramic radiographs, and lateral cephalograms obtained during L7 extraction (T1) and completion of L8 eruption (T2). At T2, samples were categorized as acceptable (A-group) or unacceptable (U-group) based on the American Board of Orthodontics index.L8 angulation and position, retromolar space, distance between the Xi point and mandibular first molar (L6), and Nolla stage were compared between the groups to identify the predictive factors for successful eruption.
Results:
At T2, 58.7% of L8 exhibited acceptable occlusion. Age at T1 was significantly higher in the U-group than that in the A-group. Angles ∠6-MnP and ∠8-MnP differed significantly between the groups at T2. Xi-L6 distance was considerably longer in the A-group than that in the U-group at T1 and T2. Younger age at extraction and Xi-L6 distance at T1 affected the acceptable occlusion.
Conclusions
Younger age at L7 extraction and adequate eruption space (Xi-L6 distance) appear to be the key factors for achieving acceptable L8 occlusion.
6.Potential of histone deacetylase 6 inhibitors in alleviating chemotherapy-induced peripheral neuropathy
Su Jung PARK ; Soung-Min LEE ; Seong Mook KANG ; Hyun-Mo YANG ; Su-Kil SEO ; Ju-Hee LEE
The Korean Journal of Pain 2025;38(2):152-162
Background:
Histone deacetylase 6 (HDAC6), belonging to class IIb of histone deacetylases, regulates theacetylation of the cytoplasmic protein α-tubulin. The overexpression of HDAC6 is linked to the development oftumors, and inhibiting HDAC6 is known to trigger apoptosis in multiple myeloma cells. In addition to its application in cancer treatment, bortezomib, a proteasome inhibitor, is widely used in managing multiple myeloma and has shown effectiveness in patients with both newly diagnosed and relapsed disease. However, the treatment regimen may be delayed or discontinued due to the risk of peripheral neuropathy, a significant non-hematologic side effect.
Methods:
Animal models of peripheral neuropathy induced by various anti-cancer drugs were established, confirming the potential of HDAC6 inhibitors as a treatment for this condition. Six- to eight-week-old male Sprague Dawley rats were utilized to create these models. Mechanical allodynia and electron microscopy served as indicators of peripheral neuropathy. The HDAC6 inhibitor CKD-011 was administered at doses of 5, 10, 20, and 40 mg/kg.
Results:
In an animal model of bortezomib-induced peripheral neuropathy, CKD-011, an HDAC6 inhibitor, effectively ameliorated peripheral neuropathy. Similarly, CKD-011 administration demonstrated recovery from peripheral neuropathy in models induced with oxaliplatin, paclitaxel, and cisplatin.
Conclusions
These findings suggest that HDAC6 inhibitors have the potential to mitigate peripheral neuropathy induced by chemotherapeutic agents.
7.Understanding of Patients with Severe COVID-19 Using Lung Ultrasound
Seo-Hee YANG ; Eun Ju PARK ; Jung-Hyun KIM ; Jin Woo SONG ; Young-Jae CHO
Tuberculosis and Respiratory Diseases 2025;88(2):380-387
Background:
Lung ultrasound (LUS) has proven valuable in the initial assessment of coronavirus disease 2019 (COVID-19), but its role in detecting pulmonary fibrosis following intensive care remains unclear. This study aims to assess the presence of pulmonary sequelae and fibrosis-like changes using LUS in survivors of severe COVID-19 pneumonia one month after discharge.
Methods:
We prospectively enrolled patients with severe COVID-19 who required mechanical ventilation in the intensive care unit (ICU) and conducted LUS assessments from admission to the outpatient visit after discharge. We tracked changes in key LUS findings and applied our proprietary LUS scoring system. To evaluate LUS accuracy, we correlated measured LUS values with computed tomography scores.
Results:
We evaluated B-line presence, pleural thickness, and consolidation in 14 eligible patients. The LUS scores exhibited minimal changes, with values of 19.1, 19.2, and 17.5 at admission, discharge, and the outpatient visit, respectively. Notably, the number of B-lines decreased significantly, from 1.92 at admission to 0.56 at the outpatient visit (p<0.05), while pleural thickness increased significantly, from 2.05 at admission to 2.48 at the outpatient visit (p≤0.05).
Conclusion
This study demonstrates that LUS can track changes in lung abnormalities in severe COVID-19 patients from ICU admission through to outpatient follow-up. While pleural thickening and B-line patterns showed significant changes, no correlation was found between LUS and high-resolution computed tomography fibrosis scores. These findings suggest that LUS may serve as a supplementary tool for assessing pulmonary recovery in severe COVID-19 cases.
8.Vitamin D Attenuates Non-Alcoholic Fatty Liver Disease in High-Fat Diet-Induced Obesity Murine Model
Sook In CHUNG ; Lin LIANG ; Heejae HAN ; Kyung Hee PARK ; Jae-Hyun LEE ; Jung-Won PARK
Yonsei Medical Journal 2025;66(2):75-86
Purpose:
Obesity and metabolic syndrome are acknowledged as key factors contributing to the development of non-alcoholic fatty liver disease (NAFLD). Vitamin D (VitD) is a multifaceted secosteroid hormone known for its anti-fibrotic and anti-inflammatory properties, with its deficiency often linked to obesity. Our study aimed to investigate whether VitD supplementation could mitigate the liver pathology associated with NAFLD.
Materials and Methods:
The NAFLD model was developed by subjecting male C57BL/6 mice to a high-fat diet (HFD) for 14 weeks.These mice were supplemented with VitD through intraperitoneal injection at a dosage of 7 μg/kg, administered three times per week for 7 weeks.
Results:
HFD resulted in VitD deficiency, insulin resistance, and increased liver weight. It elevated serum levels of liver aminotransferases and triglyceride, ultimately leading to steatohepatitis with fibrosis. This model exhibited increased levels of transforming growth factor (TGF)-β1, pro-inflammatory cytokines, HNF4α transcription factors, reactive oxygen species (ROS), renin-angiotensin system activity, and epithelial-mesenchymal transitions (EMT) within the liver. Supplementation with VitD resulted in the recovery of liver weight, improvement in histologic features associated with steatohepatitis, and reduction in alanine aminotransferases and triglyceride levels induced by the HFD. Additionally, it mitigated the HFD-induced over-expressions of TGF-β1 and fibrosis-related genes, along with pro-inflammatory cytokines and ROS. Notably, no adverse effect was found due to VitD supplementation in this model.
Conclusion
VitD ameliorates steatohepatitis within obesity-induced NAFLD through its multifaceted pathways. VitD supplementation emerges as a potentially safe, cost-effective, and direct treatment approach for NAFLD patients dealing with obesity or metabolic dysfunction.
9.Differences in Treatment Outcomes Depending on the Adjuvant Treatment Modality in Craniopharyngioma
Byung Min LEE ; Jaeho CHO ; Dong-Seok KIM ; Jong Hee CHANG ; Seok-Gu KANG ; Eui-Hyun KIM ; Ju Hyung MOON ; Sung Soo AHN ; Yae Won PARK ; Chang-Ok SUH ; Hong In YOON
Yonsei Medical Journal 2025;66(3):141-150
Purpose:
Adjuvant treatment for craniopharyngioma after surgery is controversial. Adjuvant external beam radiation therapy (EBRT) can increase the risk of long-term sequelae. Stereotactic radiosurgery (SRS) is used to reduce treatment-related toxicity.In this study, we compared the treatment outcomes and toxicities of adjuvant therapies for craniopharyngioma.
Materials and Methods:
We analyzed patients who underwent craniopharyngioma tumor removal between 2000 and 2017. Of the 153 patients, 27 and 20 received adjuvant fractionated EBRT and SRS, respectively. We compared the local control (LC), progression-free survival (PFS), and overall survival between groups that received adjuvant fractionated EBRT, SRS, and surveillance.
Results:
The median follow-up period was 77.7 months. For SRS and surveillance, the 10-year LC was 57.2% and 57.4%, respectively. No local progression was observed after adjuvant fractionated EBRT. One patient in the adjuvant fractionated EBRT group died owing to glioma 94 months after receiving radiotherapy (10-year PFS: 80%). The 10-year PFS was 43.6% and 50.7% in the SRS and surveillance groups, respectively. The treatment outcomes significantly differed according to adjuvant treatment in nongross total resection (GTR) patients. Additional treatment-related toxicity was comparable in the adjuvant fractionated EBRT and other groups.
Conclusion
Adjuvant fractionated EBRT could be effective in controlling local failure, especially in patients with non-GTR, while maintaining acceptable treatment-related toxicity.
10.Feasibility of a Machine Learning Classifier for Predicting Post-Induction Hypotension in Non-Cardiac Surgery
Insun PARK ; Jae Hyon PARK ; Young Hyun KOO ; Chang-Hoon KOO ; Bon-Wook KOO ; Jin-Hee KIM ; Ah-Young OH
Yonsei Medical Journal 2025;66(3):160-171
Purpose:
To develop a machine learning (ML) classifier for predicting post-induction hypotension (PIH) in non-cardiac surgeries.
Materials and Methods:
Preoperative data and early vital signs were obtained from 3669 cases in the VitalDB database, an opensource registry. PIH was defined as sustained mean arterial pressure (MAP) <65 mm Hg within 20 minutes since induction or from induction to incision. Six different ML algorithms were used to create binary classifiers to predict PIH. The primary outcome was the area under the receiver operating characteristic curve (AUROC) of ML classifiers.
Results:
A total of 2321 (63.3%) cases exhibited PIH. Among ML classifiers, the random forest regressor and extremely gradient boosting regressor showed the highest AUROC, both recording a value of 0.772. Excluding these models, the light gradient boosting machine regressor showed the second highest AUROC [0.769; 95% confidence interval (CI), 0.767–0.771], followed by the gradient boosting regressor (0.768; 95% CI, 0.763–0.772), AdaBoost regressor (0.752; 95% CI, 0.743–0.761), and automatic relevance determination regression (0.685; 95% CI, 0.669–0.701). The top three important features were mean diastolic blood pressure (DBP), minimum MAP, and minimum DBP from anesthetic induction to tracheal intubation, and these features were lower in cases with PIH (all p<0.001).
Conclusion
ML classifiers exhibited moderate performance in predicting PIH, and have the potential for real-time prediction.

Result Analysis
Print
Save
E-mail