Abstract:Objective To construct a risk model of postoperative delirium (POD) after radical resection of lung cancer in elderly patients and evaluate its efficiency. Methods A total of 618 elderly patients with radical resection of lung cancer in the Department of Intensive Care Medicine of the Tumor Hospital Affiliated to Xinjiang Medical University were selected from March 2022 to August 2022 as the research subjects. The patients were divided into a training set (n=433) and a validation set (n=185) in a 7∶3 ratio using simple random sampling. At 1-3 days after surgery, the occurrence of POD was evaluated by the qualified staff using confusion assessment method. General demographic data, preoperative, intraoperative and postoperative indicators were collected. SPSS 26.0 was used for data processing. According to the data type, t test or Chi-square test was performed. Factors affecting the occurrence of POD in the training set were screened by binary logistic regression model. A POD prediction model was constructed, and receiver operating characteristic (ROC) curve was drawn. The validation set was used to verify the efficiency of the prediction model. Results Of 618 elderly patients with lung cancer, 119 (19.26%) developed POD, with 88 (20.32%) in the training set and 31 (16.76%) in the validation set, and there was no statistically significant difference between the two groups (χ2=1.061, P=0.303). With presence or absence of POD in the training group as the dependent variable, binary logistic regression analysis indicated that age > 79 years (OR=2.221, 95%CI 1.642-3.004), Charlson comorbidity index (CCI) score ≥ 2 points (OR=1.540, 95%CI 1.011-2.348), thoracotomy (OR=1.902, 95%CI 1.177-3.075), and propofol dosage (OR=1.413, 95%CI 1.133-1.764) were independent influencing factors of POD occurrence in elderly patients undergoing radical resection of lung cancer. ROC curve showed an area under the curve (AUC) of 0.832 (95%CI 0.746-0.917) for the prediction model constructed on the above results in predicting POD risk in the validation set, with a sensitivity of 0.710 and a specificity of 0.838. Conclusion Age, CCI score, surgical method and propofol dosage are independent risk factors for POD in elderly patients after radical resection of lung cancer. The risk prediction model based on these factors demonstrates good performance in predicting POD in patients. In order to ensure the reliability of the conclusion, external verification is necessary with increased samples.