Construction of a Nomogram Model for Risk Assessment of Immune Checkpoint Inhibitor- Related Pneumonitis: A Study Based on Clinical Data

Background Immune checkpoint inhibitor (ICI) therapy-related pneumonitis (CIP) is a rare and potentially lethal side effect associated with ICIs. This study aimed to develop and validate a noninvasive nomogram for the identification of independent risk factors for CIP in advanced non-small cell lung cancer (NSCLC) patients treated with ICIs. Methods This study retrospectively enrolled 73 patients with advanced NSCLC and CIP and 583 healthy controls. All patients were randomized into a training set (n = 420) and a test set (n = 181) at a ratio of 7:3. Univariate and binary logistic regression analyses were used to determine independent risk factors and to construct a prediction model. Internal validation was evaluated using the area under the curve (AUC), a calibration curve and decision curve analysis (DCA). Results Binary logistic regression analysis revealed that the risk factors for CIP were red blood cell distribution width (RDW) (odds ratio [OR], 6.242; 95% CI: 2.661–14.645), absolute eosinophils count (EOS) (OR, 5.453; 95% CI: 1.732–17.170), lactate dehydrogenase (LDH) (OR, 14.032; 95% CI: 5.562–35.395), fibrinogen (Fib) (OR, 4.951; 95% CI: 2.213–11.073) and the use of antibiotics (OR, 6.449; 95% CI: 2.746–15.145). A nomogram model was constructed for CIP based on these risk factors, and the AUC was 0.918 (95% CI: 0.877–0.958). Therefore, the model shows good differentiation, calibration and clinical value. Conclusions A noninvasive predictive nomogram was developed and validated to help clinicians predict the risk of CIP in patients treated with ICIs. Trial registration Not applicable.