Abstract: Objective: To establish a predictive model based on Perioperative Hypothermia (PH) and clinicopathological features to predict the occurrence of Surgical Site Infection (SSI) in gastric cancer patients. Methods: The clinical data and anesthesia records of patients who underwent radical surgery in a hospital in Xinjiang from January 2020 to April 2021 and were pathologically confirmed as gastric cancer were collected for retrospective analysis. The Logistic regression was used to analyze the independent risk factors affecting the occurrence of SSI, and a nomogram prediction model for the occurrence of SSI was established. Results: A total of 355 patients were enrolled, PH occurred in 99 patients (27.9%) and SSI occurred in 66 patients (18.6%). The Logistic regression analysis results showed that age, diabetes mellitus, BMI, body temperature entering the operating room, heat preservation interruption time, intraoperative PH, operative modalities and operation time were independent influencing factors of SSI in gastric cancer patients (P<0.05), based on which the nomogram prediction model was established with its AUC and Brier score being 0.826 (95%CI: 0.770-0.883) and 1.14, respectively. In addition, the fitting degree of the calibration curve was good, and the decision curve showed that patients could benefit when the threshold was between 0.10 and 0.73. Conclusion: The nomogram model formed by combining intraoperative PH and clinical characteristics can more accurately predict the risk of postoperative SSI in gastric cancer patients.

Key words: gastric cancer; Perioperative Hypothermia; Surgical Site Infections; nomogram; prediction