Frontiers in medicine · 2026
An explainable machine learning model for predicting osteoporotic fragility fractures: a retrospective study in South China.
СтатьяНазвание и аннотация: Перевод в очереди
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.
Si Zebing, Hu Konghe, Wang Huajun, Zheng Xiaofei
Аннотация
BACKGROUND: Osteoporotic fragility fractures often lead to hospitalization and impose a significant economic burden. To investigate the clinical characteristics of hospitalized major osteoporotic fractures in South China.
METHODS: In this study, we identified 1125 hospitalized patients with osteoporotic fractures who underwent DXA examination at the Yuebei People's Hospital between January 1, 2019, and December 31, 2025. From these patients, we collected 73 features, including clinical, laboratory, and radiological features. Recursive Feature Elimination (RFE) was applied to filter these features. Then we evaluated six machine learning algorithms (XGBoost, Random Forest, Support Vector Machine, CatBoost, LightGBM, and Logistic Regression) to construct the best predictive model. In order to explain the model, SHAP (SHapley Additive exPlanations) was used to assess the importance of each attribute.
RESULTS: LightGBM achieved the highest training-set ROC-AUC (0.903) and the highest independent test-set ROC-AUC (0.840) among the six models. Through SHAP analysis, we observed that age, Hepatitis C virus IgG antibody, and serum sodium were identified as the top factors contributing to increased fracture risk.
CONCLUSION: Our model has a good ability of predicting osteoporotic fragility fractures. This paves the way for earlier intervention in patients who are susceptible to osteoporotic fragility fractures.
METHODS: In this study, we identified 1125 hospitalized patients with osteoporotic fractures who underwent DXA examination at the Yuebei People's Hospital between January 1, 2019, and December 31, 2025. From these patients, we collected 73 features, including clinical, laboratory, and radiological features. Recursive Feature Elimination (RFE) was applied to filter these features. Then we evaluated six machine learning algorithms (XGBoost, Random Forest, Support Vector Machine, CatBoost, LightGBM, and Logistic Regression) to construct the best predictive model. In order to explain the model, SHAP (SHapley Additive exPlanations) was used to assess the importance of each attribute.
RESULTS: LightGBM achieved the highest training-set ROC-AUC (0.903) and the highest independent test-set ROC-AUC (0.840) among the six models. Through SHAP analysis, we observed that age, Hepatitis C virus IgG antibody, and serum sodium were identified as the top factors contributing to increased fracture risk.
CONCLUSION: Our model has a good ability of predicting osteoporotic fragility fractures. This paves the way for earlier intervention in patients who are susceptible to osteoporotic fragility fractures.