Abstract: Objective To explore the related factors for hypoxemia during anesthesia recovery period after thoracoscopic radical resection of lung cancer under general anesthesia and construct a nomogram prediction model.Methods A total of 227 patients who underwent thoracoscopic radical resection of lung cancer under general anesthesia in our hospital from May 2022 to May 2025 were retrospectively enrolled and divided into hypoxemia group and non-hypoxemia group according to whether hypoxemia occurred during postoperative anesthesia recovery period.Variables were screened by LASSO regression, and independent related factors were determined by multivariate Logistic regression analysis.A nomogram model was constructed and Bootstrap method was used for internal validation.Receiver operating characteristic (ROC) curve, calibration curve, and decision curve were used to evaluate model performance and clinical applicability.SHAP model was used to explore the contribution of each independent influencing factor to the prediction model.A risk stratification system was established based on the risk score of the nomogram model.Results A total of 60 cases of hypoxemia occurred during postoperative anesthesia recovery period.Six non-zero coefficient variables were obtained through LASSO regression variable screening, namely body mass index (BMI), the ratio of forced expiratory volume in one second to forced vital capacity (FEV1/FVC), one-lung ventilation time, intraoperative dexmedetomidine use, intraoperative hypothermia, and airway aspiration frequency.The results of the multivariate Logistic regression analysis showed that BMI (OR=1.189, 95%CI: 1.040-1.360), intraoperative hypothermia (OR=2.227, 95%CI: 1.079-4.598), postoperative airway suction frequency (OR=5.326, 95%CI: 2.927-9.690), intraoperative use of dexmedetomidine (OR=2.502, 95%CI: 1.101-5.686), and single-lung ventilation time (OR=1.035, 95%CI: 1.009-1.063) were risk factors for postoperative anesthesia recovery period-related hypoxemia in patients, while preoperative FEV1/FVC (OR=0.860, 95%CI: 0.803-0.921) was a protective factor (P<0.05).A nomogram model was constructed based on the above factors.ROC curve showed that the area under the curve (AUC) of this model was 0.897 (95%CI: 0.854-0.940).The C-index after internal validation of the model obtained by Bootstrap method was 0.904.Calibration curve showed high consistency between the ideal probability curve and the actual probability curve.Clinical decision curve showed that when the threshold probability was between 0.06 and 0.88, the nomogram prediction model could produce better clinical benefit.The SHAP results showed that the order of importance of each independent related factor on the ogram model was airway aspiration frequency, preoperative FEV1/FVC, one-lung ventilation time, BMI, intraoperative hypothermia, and the use of dexmedetomidine during the operation.The risk stratification of the risk prediction model established based on the above related influencing factors was low risk ≤134 points and high risk >134 points, with risk probabilities of 60.24% and 6.94%, respectively.Conclusion The related factors of hypoxemia during anesthesia recovery period after thoracoscopic radical resection for lung cancer under general anesthesia include BMI, intraoperative hypothermia, frequency of postoperative airway suctioning, intraoperative use of dexmedetomidine, one-lung ventilation time, and preoperative FEV1/FVC.The nomogram model constructed accordingly demonstrates good predictive utility.

Key words: thoracoscopic radical resection of lung cancer; anesthesia recovery period; hypoxemia; nomogram