Comparison of PNI and GNRI nomogram models for predicting postoperative complications in elderly patients with lung cancer: a retrospective study

Objective This study aims to clarify the correlation between the nutritional indicators PNI and GNRI and postoperative cardiopulmonary complications in elderly patients with non-small cell lung cancer (NSCLC). The study also aims to construct an early postoperative visual prediction tool to optimize the efficacy of assessing postoperative cardiopulmonary complication risk in elderly NSCLC patients. Methods This retrospective study analyzed the clinical data of elderly lung cancer surgery patients in our department. 364 Patients who met the inclusion criteria were assigned at a 7:3 ratio. Logistic regression analysis was performed to create a nomogram that predicts postoperative cardiopulmonary complications and identifies independent risk factors. We evaluated the model's performance using the C-index, the area under the curve (AUC), the calibration curve, and the decision curve analysis (DCA). We verified the model's stability using the validation set. Results This study included 364 elderly patients undergoing lung cancer surgery, with a postoperative cardiopulmonary complication rate of 26.65%.Multivariate logistic regression identified smoking history, postoperative PNI, ΔPNI, preoperative GNRI, and operative duration as independent risk factors ( P < 0.05). A nomogram was constructed based on these factors, achieving AUC values of 0.82 (95% CI: 0.76–0.88) in the training set and 0.87 (95% CI: 0.79–0.94) in the validation set. Calibration was satisfactory (H-L test, P > 0.05), and decision curve analysis confirmed clinical net benefit, with prediction accuracies of 76.0% and 80.0%, respectively. Conclusion This study confirms that smoking history, nutritional indicators (postoperative PNI and ΔPNI and preoperative GNRI), and operation time are all independent risk factors for postoperative cardiopulmonary complications in elderly patients with non-small cell lung cancer (NSCLC). The predictive model based on these factors exhibits good discrimination and calibration and provides significant clinical net benefit. This model can effectively assist in the early identification of risk and the individualized perioperative intervention for postoperative cardiopulmonary complications in these patients.