Comprehensive nomogram models for predicting checkpoint inhibitor pneumonitis

Checkpoint inhibitor pneumonitis (CIP) is a common type of immune-related adverse events (irAEs) with poor clinical prognosis. Currently, there is a lack of effective biomarkers and predictive models to predict the occurrence of CIP. This study retrospectively enrolled 547 patients who received immunotherapy. The patients were divided into CIP cohorts of any grade, or grade & GE;2 or & GE;3. Multivariate logistic regression analysis was used to determine the independent risk factors, based on which we established Nomogram A and B for respectively predicting any grade or grade & GE;2 CIP. For Nomogram A to predict any grade CIP, the C indexes in the training and validation cohorts were 0.827 (95% CI=0.772-0.881) and 0.860 (95% CI=0.741-0.918), respectively. Similarly, for Nomogram B to predict grade 2 or higher CIP, the C indexes of the training and validation cohorts were 0.873 (95% CI=0.8260.921) and 0.904 (95% CI=0.804-0.973), respectively. In conclusion, the predictive power of nomograms A and B has proven satisfactory following internal and external verification. They are promising clinical tools that are convenient, visual, and personalized for assessing the risks of developing CIP.