Construction and validation of a risk prediction model for venous thromboembolism post-VATS in simultaneous multicentric primary lung cancers.

BACKGROUND

Synchronous multiple primary lung cancers (sMPLCs) represent 0.8% to 20% of new lung cancer diagnoses. Currently, there is a lack of risk prediction models for venous thromboembolism (VTE) after video-assisted thoracoscopic surgery (VATS) in sMPLC patients. This study seeks to create and validate a VTE risk prediction model tailored for sMPLC patients undergoing VATS.

METHODS

A retrospective cohort analysis was conducted on patients who underwent lung cancer resection from November 2017 to December 2024 using Hospital Information System (HIS), telephone follow-up, and the Questionnaire Star electronic questionnaire. Categorical variables were analyzed using χ tests and continuous variables were assessed with -tests for univariate analysis. Variables with statistical significance from the univariate analysis and the least absolute shrinkage and selection operator (LASSO) regression algorithm were included in the logistic regression analysis to identify risk factors and construct the prediction model. A nomogram was used for the visualization of the model. The discriminative ability and calibration of the model were evaluated using the area under the receiver operating characteristic (ROC) curve and calibration plots, respectively. The clinical utility of the model was assessed using decision curve analysis.

RESULTS

The occurrence of VTE post-VATS in patients with sMPLC was associated with age, smoking history, coronary artery disease, cerebrovascular disease, chronic obstructive pulmonary disease (COPD), atherosclerotic plaques in the extremities, surgical method, intraoperative transfusion, Postoperative Caprini score, and the number of primary lesions (P<0.05). The area under the ROC curve was 0.917 [95% confidence interval (CI): 0.894-0.941], with a sensitivity of 0.885 and a specificity of 0.818. The calibration curve demonstrated a good fit between the observed and predicted curves, with a mean absolute error of 0.008. The clinical decision curve analysis indicated that the model offered superior clinical benefits compared to the Caprini score.

CONCLUSIONS

The prediction model constructed in this study exhibits robust predictive performance, providing a theoretical basis for clinical medical staff to identify high-risk groups of patients with sMPLC who may develop VTE after VATS at an early stage and to facilitate timely interventions.