European journal of radiology · 2026
Who benefits from a filter? Developing clinical prediction models for thrombus dislodgement in hospitalized patients with deep vein thrombosis: A multicenter retrospective study.
СтатьяМногоцентровое исследованиеНазвание 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.
Gong Maofeng, Kong Jie, Gu Jianping, Sun Yanting, He Xu
Аннотация
PURPOSE: Patients with deep vein thrombosis (DVT) who are most likely to benefit from inferior vena cava filter (IVCF) placement remain poorly elucidated. This study aimed to develop and validate predictive models to assess the risk of thrombus dislodgement (TD) in hospitalized patients with DVT.
METHODS: 252 patients from a multicenter and a single-center database were included. The least absolute shrinkage and selection operator (LASSO)-minimum, LASSO-1 standard error (SE), and logistic regression (LR) were used to identify potential predictors for model development. A multivariate LR was then conducted to construct a nomogram. Model discriminatory was evaluated using the area under the curve (AUC) with a 95% confidence interval (CI). Calibration analysis and decision curve analysis (DCA) were performed to assess model accuracy and clinical utility.
RESULTS: Three models were developed based on predictors selected by LASSO-min (age, sex, thrombus limbs, pulmonary embolism [PE], diabetes mellitus, immobilization, previous venous thromboembolism, and percutaneous transluminal angioplasty [PTA]), LASSO-1SE (sex, thrombus limbs, PE, and immobilization), and LR (sex, thrombus limbs, PE, immobilization, and PTA). The AUCs were 0.765 (95% CI: 0.686-0.844), 0.711 (95% CI: 0.626-0.796), and 0.743 (95% CI: 0.663-0.823) in the training cohort, and 0.707 (95% CI, 0.587-0.827), 0.692 (95% CI, 0.574-0.810), and 0.695 (95% CI, 0.572-0.817) in the validation cohort. DeLong's test showed no statistically significant differences among the three models (all p > 0.05). Sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were comparable. The LR model, incorporating a broader set of statistically significant predictors, demonstrated moderately parsimonious and stable performance. Calibration curves showed strong concordances, and DCA indicated a notable net clinical benefit within a threshold probability range of 10%-95%.
CONCLUSION: A clinically interpretable and statistically robust model was developed and internally validated to predict TD in hospitalized patients with DVT. The model showed good calibration and discriminative ability, suggesting its potential value as a practical tool for early risk stratification of TD.
METHODS: 252 patients from a multicenter and a single-center database were included. The least absolute shrinkage and selection operator (LASSO)-minimum, LASSO-1 standard error (SE), and logistic regression (LR) were used to identify potential predictors for model development. A multivariate LR was then conducted to construct a nomogram. Model discriminatory was evaluated using the area under the curve (AUC) with a 95% confidence interval (CI). Calibration analysis and decision curve analysis (DCA) were performed to assess model accuracy and clinical utility.
RESULTS: Three models were developed based on predictors selected by LASSO-min (age, sex, thrombus limbs, pulmonary embolism [PE], diabetes mellitus, immobilization, previous venous thromboembolism, and percutaneous transluminal angioplasty [PTA]), LASSO-1SE (sex, thrombus limbs, PE, and immobilization), and LR (sex, thrombus limbs, PE, immobilization, and PTA). The AUCs were 0.765 (95% CI: 0.686-0.844), 0.711 (95% CI: 0.626-0.796), and 0.743 (95% CI: 0.663-0.823) in the training cohort, and 0.707 (95% CI, 0.587-0.827), 0.692 (95% CI, 0.574-0.810), and 0.695 (95% CI, 0.572-0.817) in the validation cohort. DeLong's test showed no statistically significant differences among the three models (all p > 0.05). Sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were comparable. The LR model, incorporating a broader set of statistically significant predictors, demonstrated moderately parsimonious and stable performance. Calibration curves showed strong concordances, and DCA indicated a notable net clinical benefit within a threshold probability range of 10%-95%.
CONCLUSION: A clinically interpretable and statistically robust model was developed and internally validated to predict TD in hospitalized patients with DVT. The model showed good calibration and discriminative ability, suggesting its potential value as a practical tool for early risk stratification of TD.