A prediction nomogram for perioperative deep vein thrombosis risk in breast cancer patients: A retrospective cohort study with prospective clinical application.

Abstract

e13541 Background: Deep vein thrombosis (DVT) represents a prevalent postoperative complication among breast cancer perioperative patients. However, there remains a paucity of validated predictive models specifically designed for perioperative DVT risk assessment in this population. This study aims to analyze the risk factors and develop a nomogram model for predicting perioperative DVT in breast cancer patients, subsequently prospect validate its risk stratification utility, and assess the preventive effects of dextran 40. Methods: This retrospective study analyzed clinical data from 594 breast cancer surgery patients who underwent surgery and did not receive prophylactic anticoagulation between May and November 2024 as the modeling cohort, which was randomly divided into a training cohort (n=416) and a validation cohort (n=178) at a 7:3 ratio. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors, and a nomogram prediction model was constructed. Model performance was evaluated using receiver operating characteristic (ROC) curves, decision curve analysis (DCA), and calibration curves. A clinical validation cohort was prospectively collected from April to August 2025, which was stratified into high-risk and low-risk groups based on the risk score derived from the nomogram model. DVT occurrence between the two groups was compared to evaluate the model's discrimination ability, and the preventive effect of dextran 40 on DVT was analyzed. Results: Multivariate regression analysis revealed that elevated D-dimer levels, age >50 years, BMI ≥28 kg/m², hypertension, and diabetes were independent risk factors for perioperative lower extremity DVT in breast cancer patients. The nomogram model demonstrated good predictive performance in both the training cohort (AUC=0.790) and internal validation cohort (AUC=0.819). In the clinical validation cohort, the model classified 281 patients (64.0%) into the high-risk group, which exhibited a significantly higher DVT incidence than the low-risk group (14.6% vs. 1.2%, P<0.001). Further analysis revealed that among the high-risk group (n=281), patients treated by dextran 40 had a DVT incidence of 10.0% (16/160), significantly lower than that of control group (21.5%, 26/121). But in the low-risk group (n=157), the DVT incidence rates were 1.2% (1/84) in the intervention group and 1.4% (1/73) in the control group, with no statistically significant difference (P=0.920). Conclusions: D-dimer, age >50 years, BMI ≥28 kg/m², hypertension, and diabetes are the risk factors for perioperative DVT in breast cancer patients and successfully established nomogram prediction model. In clinical practice, dextran 40, not affecting wound healing, provides a preventive benefit in DVT high-risk patients.