Construction and Validation of Nomogram Model for Predicting Lower Extremity Venous Thrombosis in Patients with Gastric Cancer
نویسندگان
1 Department of General Surgery, Shengzhou People’s Hospital (Shengzhou Branch of the First Affiliated Hospital of Zhejiang University School of Medicine), Shengzhou City, Zhejiang Province 32400, China
2 Department of General Surgery, Shengzhou People’s Hospital (Shengzhou Branch of the First Affiliated Hospital of Zhejiang University School of Medicine), Shengzhou City, Zhejiang Province 32400, China
3 Department of General Surgery, Shengzhou People’s Hospital (Shengzhou Branch of the First Affiliated Hospital of Zhejiang University School of Medicine), Shengzhou City, Zhejiang Province 32400, China
doi
10.22034/ircmj.2025.525290.2179چکیده
Background and Objectives: Lower extremity venous thrombosis (LEVT) constitutes a severe complication that exerts negative impacts on the prognosis and functional recovery of gastric cancer patients. Few comprehensively incorporate key modifiable factors such as dietary diversity and nutritional status, which may contribute to thrombosis risk via mechanisms including endothelial dysfunction and hypercoagulability. This study develops a prognostic nomogram for predicting LEVT by integrating clinical characteristics, dietary diversity, and nutritional indicators. Methods: A retrospective analysis was conducted on 486 patients diagnosed with gastric cancer who received treatment at our hospital from January 2020 to December 2023. Data encompassing demographics, surgical recovery, dietary diversity scores, and nutritional indices were collected. Univariate analysis and multivariate logistic regression were utilized to identify the predictive factors of LEVT, and these factors were subsequently incorporated into the construction of a nomogram. The performance of the model was evaluated by means of calibration curves, receiver operating characteristic (ROC) analysis and decision curve analysis (DCA). Results: LEVT incidence was 13.99% (68/486). Multivariate analysis identified six independent predictors (P<0.05): Postoperative bed rest time (OR=1.375), Meal diversity score (OR=0.531), NRS2002 score (OR=2.234), D-dimer (OR=5.686), fibrinogen (OR=3.170), Albumin (OR=0.802). The nomogram exhibited favorable discriminatory ability (training C-index=0.901; validation C-index=0.870) and calibration performance (mean absolute error=0.075 for both datasets; Hosmer-Lemeshow P=0.174/0.403). ROC analysis revealed the area under the curve values of 0.896 (95% CI:0.845–0.947) and 0.863 (95% CI:0.745–0.980) in the training and validation groups, with corresponding sensitivities of 0.926/0.943 and specificities of 0.676/0.571, respectively. DCA confirmed superior clinical benefit across threshold probabilities 0.05–0.95 versus alternative strategies. Conclusion: The nomogram synergizes clinical, nutritional, and dietary predictors to enhance LEVT risk stratification in gastric cancer, offering a pragmatic tool for guiding thromboprophylaxis and resource allocation.
