Journal of affective disorders · 2025
Machine learning-enhanced mapping of suicide risk in Bipolar Disorder: A multi-modal analysis.
СтатьяНазвание и аннотация: Перевод пока не подготовлен
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Saeed Saboor, Wang Huaizhi, Kong Lingzhuo, Geng Yimeng, Zhang Jinyu, Pan Yanmeng, Le Xu, Zhang Xuhong, Liu Ting Ting, Hu Shaohua
Аннотация
BACKGROUND: Bipolar disorder (BD) is associated with a high risk of suicide, but the complex interplay of factors contributing to this risk remains poorly understood. This study aimed to comprehensively analyze demographic, clinical, and biological factors associated with suicide risk in BD patients and develop a novel suicide risk assessment model integrating these factors.
METHODS: We conducted a cross-sectional study of 152 patients with BD, classified into four suicide-risk groups: no risk (n = 19), low risk (n = 45), moderate risk (n = 38), and high risk (n = 50). Participants underwent assessments using the Mini-International Neuropsychiatric Interview (M.I.N·I.), Hamilton Depression Rating Scale-24 items (HAMD-24), Young Mania Rating Scale (YMRS), Montgomery-Åsberg Depression Rating Scale (MADRS), and Beck Scale for Suicide Ideation (BSSI). We evaluated thyroid function, inflammatory markers, and lymphocyte subsets. Univariate and multivariate analyses were performed to identify factors associated with suicide risk.
FINDINGS: Depressive symptoms were significantly associated with increased odds of medium (odds ratio (OR) = 1.452, 95 % confidence interval (CI): 1.122-1.878, P = 0.005) and high (OR = 1.405, 95 % CI: 1.091-1.810, P = 0.009) suicide risk. Lower free thyroxine 4 (FT4) levels were associated with higher odds of low (OR = 0.581, 95 % CI: 0.404-0.835, P = 0.003) and medium (OR = 0.694, 95 % CI: 0.486-0.992, P = 0.045) risk. The no-risk group exhibited higher levels of thyroid hormones and autoantibodies. CD3+ T-cell percentages varied significantly across risk groups, with the lowest mean percentage in the no-risk group (57.59 ± 14.64 %). Our machine learning models achieved 87.1 % accuracy in predicting suicide risk. Patient Health Questionnaire-9 items, Hamilton Depression Rating Scale-24 items, and Montgomery-Åsberg Depression Rating Scale scores were identified as the strongest predictors of suicide risk by a Random-Forest model with 100 decision trees. In addition, FT4 and interferon-γ emerged as notable contributors to the model's predictions.
CONCLUSION: Depressive symptoms and thyroid function are crucial factors in assessing suicide risk in BD. Thyroid autoimmunity and T cell-mediated immunity emerge as potential biomarkers for risk stratification and therapeutic targets, offering new avenues for personalized intervention strategies.
METHODS: We conducted a cross-sectional study of 152 patients with BD, classified into four suicide-risk groups: no risk (n = 19), low risk (n = 45), moderate risk (n = 38), and high risk (n = 50). Participants underwent assessments using the Mini-International Neuropsychiatric Interview (M.I.N·I.), Hamilton Depression Rating Scale-24 items (HAMD-24), Young Mania Rating Scale (YMRS), Montgomery-Åsberg Depression Rating Scale (MADRS), and Beck Scale for Suicide Ideation (BSSI). We evaluated thyroid function, inflammatory markers, and lymphocyte subsets. Univariate and multivariate analyses were performed to identify factors associated with suicide risk.
FINDINGS: Depressive symptoms were significantly associated with increased odds of medium (odds ratio (OR) = 1.452, 95 % confidence interval (CI): 1.122-1.878, P = 0.005) and high (OR = 1.405, 95 % CI: 1.091-1.810, P = 0.009) suicide risk. Lower free thyroxine 4 (FT4) levels were associated with higher odds of low (OR = 0.581, 95 % CI: 0.404-0.835, P = 0.003) and medium (OR = 0.694, 95 % CI: 0.486-0.992, P = 0.045) risk. The no-risk group exhibited higher levels of thyroid hormones and autoantibodies. CD3+ T-cell percentages varied significantly across risk groups, with the lowest mean percentage in the no-risk group (57.59 ± 14.64 %). Our machine learning models achieved 87.1 % accuracy in predicting suicide risk. Patient Health Questionnaire-9 items, Hamilton Depression Rating Scale-24 items, and Montgomery-Åsberg Depression Rating Scale scores were identified as the strongest predictors of suicide risk by a Random-Forest model with 100 decision trees. In addition, FT4 and interferon-γ emerged as notable contributors to the model's predictions.
CONCLUSION: Depressive symptoms and thyroid function are crucial factors in assessing suicide risk in BD. Thyroid autoimmunity and T cell-mediated immunity emerge as potential biomarkers for risk stratification and therapeutic targets, offering new avenues for personalized intervention strategies.