Annals of medicine · 2025
Identification of early prognostic biomarkers in Severe Fever with Thrombocytopenia Syndrome using machine learning algorithms.
СтатьяМногоцентровое исследованиеПоддержка исследования: вне правительства СШАНазвание 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.
Zhu Jie, Zhou Jianmei, Tao Chunhui, Xia Guomei, Liu Bingyan, Zheng Xiaowei, Li Xu, Zhang Zhenhua
Аннотация
OBJECTIVE: We aimed at identifying acute phase biomarkers in Severe Fever with Thrombocytopenia Syndrome (SFTS), and to establish a model to predict mortality outcomes.
METHODS: A retrospective analysis was conducted on multicenter clinical data. Group-based trajectory modeling (GBTM) was utilized to demonstrate the overall trend of laboratory indicators and their correlation with mortality. Six different machine learning algorithms were employed to develop prognostic models based on the clinical features during the acute phase, which were reduced using Lasso regression.
RESULTS: Seven indicators (ALT, AST, BUN, LDH, a-HBDH, DD, and PLT) at 7-10 days post-onset and their change slopes were found to be crucial during disease progression. These, along with other clinical features, were reduced to 8 variables using Lasso regression for model construction. The random forest model demonstrated the best performance in both internal validation (AUC: 0.961) and external validation (AUC: 0.948). Decision Curve Analysis indicated a good balance between model benefits and risks.
CONCLUSIONS: a-HBDH and its change slope along with central nervous symptom manifestations within 7-10 days after onset accurately predicted mortality in SFTS. Various algorithms provided a comprehensive overview of disease progression and constructed more stable and efficient models.
METHODS: A retrospective analysis was conducted on multicenter clinical data. Group-based trajectory modeling (GBTM) was utilized to demonstrate the overall trend of laboratory indicators and their correlation with mortality. Six different machine learning algorithms were employed to develop prognostic models based on the clinical features during the acute phase, which were reduced using Lasso regression.
RESULTS: Seven indicators (ALT, AST, BUN, LDH, a-HBDH, DD, and PLT) at 7-10 days post-onset and their change slopes were found to be crucial during disease progression. These, along with other clinical features, were reduced to 8 variables using Lasso regression for model construction. The random forest model demonstrated the best performance in both internal validation (AUC: 0.961) and external validation (AUC: 0.948). Decision Curve Analysis indicated a good balance between model benefits and risks.
CONCLUSIONS: a-HBDH and its change slope along with central nervous symptom manifestations within 7-10 days after onset accurately predicted mortality in SFTS. Various algorithms provided a comprehensive overview of disease progression and constructed more stable and efficient models.