Development and validation of nomogram models for predicting immune-related adverse events in recurrent and metastatic nasopharyngeal carcinoma patients treated with PD-L1 inhibitors.

Objective: To predict the incidence of immune-related Adverse Events (irAEs) in patients with recurrent or metastatic Nasopharyngeal Carcinoma (NPC) treated with Programmed Death-Ligand 1 (PD-L1) inhibitors, this study developed and validated nomogram models incorporating demographic, clinical, and biological variables. Methods: Data from 153 NPC patients were analyzed, incorporating variables including age, sex, Body Mass Index (BMI), clinical stage, and biomarkers. Predictive models were constructed using multivariable logistic regression, Least Absolute Shrinkage and Selection Operator (LASSO) regression, and Ridge regression. The models' performance was evaluated using Receiver Operating Characteristic (ROC) curves, calibration curves, and Decision Curve Analysis (DCA). Internal validation was conducted through k-fold cross-validation. Results: Independent predictors of irAEs included PD-L1, Free Thyroxine (FT4), Sodium (Na), and lymphocyte counts. Of the three models, the stepwise regression model performed best, with an area under the curve (AUC) of 0.78. Calibration curves showed a strong correlation between predicted and observed outcomes, and DCA demonstrated high clinical utility. Conclusion: The nomogram models effectively predict irAEs in NPC patients treated with PD-L1 inhibitors. Early identification of patients with elevated PD-L1, abnormal FT4, Na, or irregular lymphocyte counts allows for closer monitoring and personalized treatment, potentially improving outcomes. Further research is required to confirm these findings across other cancer types and therapies.