DU Wenwen,LI Peng,ZHANG Yu, et al. Influencing factors of PACU hypoxemia after thoracoscopic radical resection of lung cancer: A study based on the decision tree model[J]. JOURNAL OF WEZHOU MEDICAL UNIVERSITY, 2023, 53(6): 473-481.
Abstract：Objective: To establish a decision tree model that can predict hypoxemia in postanesthesia care unit of lung cancer patients undergoing thoracoscopic radical resection, and to explore predictive value of decision tree model. Methods: A retrospective study was conducted. 5 560 patients who underwent thoracoscopic radical resection of lung cancer in the Anesthesiology Department of the First Affiliated Hospital of Wenzhou Medical University were selected as the study objects from January 2017 to June 2022. According to whether hypoxemia was complicated in the PACU, the patients were divided into two groups: non-hypoxemia group and hypoxemia group. A prediction model was established and validated based on the decision algorithm, and receiver operator characteristic curve, calibration curve, and decision curve were draw, respectively. Results: Among 5 560 patients, 1 653 patients developed hypoxemia in the PACU. The results of decision tree model analysis based on training sets showed that age, sex, intraoperative auxiliary use of dexmedetomidine, anesthesiologist’s shift, time of surgery, and intraoperative auxiliary use of diazosine were related to PACU hypoxemia after thoracoscopic radical resection of lung cancer (P<0.05). Among them, ≥64 years old, female, the shift of anesthesiologist, intraoperative auxiliary use of diazosine, dexmedetomidine, and non-morning surgery were the promoting factors for postoperative PACU hypoxemia, while male, intraoperative auxiliary use of flurbiprofen axetil, and morning surgery were the protective factors for postoperative PACU hypoxemia (P<0.05). Decision tree model was validated by validation set data. The results showed that the AUC value was 0.916, and the predicted value was in good agreement with the measured value. The model had good net benefits. Conclusion: The prediction model for PACU hypoxemia in patients undergoing thoracoscopic radical resection of lung cancer based
on decision tree algorithm has good predictive ability, high accuracy, good net benefit, and intuitive structure,which is helpful for PACU medical staff to identify and prevent hypoxemia in advance.
