Predicting the Risk of Postoperative Delirium in Patients Undergoing Lobectomy: Development and Assessment of a Novel Nomogram.

Abstract

To construct and internally validate a nomogram predicting postoperative delirium (POD) in patients with pulmonary malignancies undergoing lobectomy.Clinical electronic medical record data were retrospectively collected from 1,066 patients who underwent lobectomy, divided into a training cohort (746) and a validation cohort (320) using a 7:3 temporal split. A nomogram for POD was developed using Lasso regression and multivariable logistic regression analysis according to the TRIPOD statement. Performance was assessed through receiver operating characteristic curves (ROC) and calibration plots.POD occurred in 203 patients (19.04%). The nomogram incorporated predictors such as age, body mass index (BMI), education level, history of diabetes, history of cerebrovascular disease, surgical approach, duration of surgery, and time to recovery from anesthesia. The area under the ROC curve (AUC) was 0.871 (95% confidence interval [CI]: 0.841-0.901) for the training cohort and 0.914 (95% CI: 0.877-0.951) for the validation cohort. Calibration curves demonstrated good agreement between predicted and actual probabilities in both cohorts.This novel nomogram can help clinicians and patients' families predict the likelihood of developing delirium following lobectomy, enabling the implementation of targeted prevention strategies.