Objective  To investigate the risk factors for postoperative venous thromboembolism (VTE) in patients with non-small cell lung cancer (NSCLC), and establish a nomogram model for the accurate prediction of high-risk individuals.

Methods  A total of 472 patients with NSCLC who underwent radical surgical resection in The First Affiliated Hospital of Anhui Medical University from June 2019 to December 2023 were included in the study. All patients were randomly assigned to the modeling group (n=332) or the internal validation group (n=140) at a ratio of 7∶3. In addition, 200 patients with NSCLC admitted to Fuyang Hospital Affiliated with Anhui Medical University during the same period were randomly selected as the external validation group. To analyze the risk factors for post-operative VTE, patients in the modeling group were further assigned to the VTE group (n=58) or the non-VTE group (n=274), and the demographic data, clinicopathological features, and laboratory test results of the two groups were compared. Multivariate Logistic regression analysis was used to identify independent risk factors for VTE and to construct a nomogram model to predict VTE risk. The predictive ability of the model was evaluated using receiver operating characteristic (ROC) and calibration curves.

Results The incidence of post-operative VTE in patients with NSCLC was 16.9%. Patients in the VTE group were older (P=0.006), had a more advanced TNM stage (P<0.001), had more frequent vascular invasion (P=0.001), and had a longer duration of surgery (P=0.033) than patients in the non-VTE group. In addition, there were significant differences between patients in the VTE and non-VTE groups for pre-operative activated partial thromboplastin time (APTT) (P=0.003), D-dimer level (P<0.001), and serum carcinoembryonic antigen (CEA) level (P=0.029). Age, TNM stage, and pre-operative D-dimer level were independent risk factors for VTE in patients with NSCLC. Based on these four variables, a nomogram model was developed to predict the risk of post-operative VTE. The areas under the ROC curves for the modeling, internal validation, and external validation groups were 0.836, 0.871, and 0.864, respectively. The calibration curve indicates a high degree of consistency between the predicted risks of the model and the actual risks that occur.

Conclusions The nomogram model based on age, TNM stage, operative time, and pre-operative D-dimer level can effectively identify individuals at risk of VTE, and it promises to be a valuable tool for risk assessment.