Development and validation of a multi-parameter nomogram for venous thromboembolism in patients undergoing surgery for brain tumors: a retrospective analysis.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
Venous thromboembolism (VTE) is a serious complication in patients undergoing brain tumor surgery, particularly those requiring neurological intensive care unit (NICU) monitoring. Risk factors for VTE vary among tumor types, and effective early warning tools remain scarce. This study aimed to develop a nomogram-based clinical prediction model to identify independent risk factors for postoperative VTE in brain tumor patients. A total of 400 NICU patients (aged 18-82 years; mean age 52.5 ± 15.6 years; 46.5% male) from January 2018 to July 2023 who underwent brain tumor surgery were included, among whom 96 developed postoperative VTE (24%). Univariate, LASSO regression, and multivariate logistic regression analyses identified independent predictors, and a nomogram prediction model was constructed. The C-index and calibration curve evaluated the model's discriminative ability and stability, while decision curve analysis assessed its clinical utility. The Linear Support Vector Machine (LSVM) method validated the model's robustness. Seven independent predictors were identified and incorporated into a nomogram: elevated preoperative D-dimer levels, elevated HbA1c levels post-admission, longer surgery duration, increased TT4 levels upon admission, increased BMI, older age and hyperlipidemia history. These factors were incorporated into a nomogram risk prediction model with an AUC of 0.937 using multivariate logistic regression. The calibration curve verified the model's stability, the C-index demonstrated strong discriminative ability, and decision curve analysis confirmed its clinical value. A linear support vector machine (LSVM) was further used to validate the predictive efficacy of the same set of variables achieved an overall accuracy of 85%. The nomogram integrates multiple clinical parameters to enhance the prediction and identification of VTE risk following brain tumor surgery. It demonstrated excellent discriminative ability, stability, and clinical applicability.