Abstract: Purpose To construct a venous thromboembolism(VTE) risk prediction model for postoperative patients with esophageal cancer. Methods Convenience sampling was used to select 220 patients who underwent esophageal cancer surgery in a tertiary grade-A hospital from January 2020 to April 2022 as the modeling group, with 60 in the VTE group and 160 in the non-VTE group. Logistic regression analysis was used to establish the VTE risk prediction model, which was validated by a separate validation group. Results Risk factors for postoperative VTE in patients with esophageal cancer included age ≥60 years, comorbidities, neoadjuvant chemotherapy (NACT), operative duration, postoperative use of prothrombin complex concentrates (PCC), and postoperative D-D at 12 hours. The model showed good predictive accuracy. In the modeling group, HL test showed P=0.250, area under receiver operating characteristic curve(AUC) was 0.766, the best cutoff value was 0.445, the sensitivity was 0.783, the specificity was 0.662. In the validation group, AUC was 0.738, the sensitivity was 0.682, the specificity was 0.769, the prediction accuracy was 80%. Conclusion The prediction model has demonstrated good predictive performance and can provide reference for assessing VTE risk in postoperative patients with esophageal cancer.

Key words: esophageal cancer, venous thromboembolism, risk prediction, influencing factor, nomogram