Risk prediction models for immune-related adverse events in patients with non-small cell lung cancer: a systematic review and meta-analysis

Immune checkpoint inhibitors have significantly improved survival outcomes in patients with non-small cell lung cancer, but they are frequently associated with immune-related adverse events that can compromise treatment continuity and patient quality of life. A high-quality prediction model could facilitate early identification and prevention. However, the quality and applicability of existing models in clinical practice and future research remain unknown. To systematically review published studies on risk prediction models for immune-related adverse events in patients with non-small cell lung cancer. PubMed, Web of Science, Embase, China National Knowledge Infrastructure, Chinese Biomedical Literature Database, and Wanfang Data Knowledge Service Platform were searched from inception to May 7, 2025. Two reviewers independently screened the literature, extracted data, and used the Prediction Model Risk of Bias Assessment Tool to assess the risk of bias and applicability. Stata 16.0 software was used to perform a meta-analysis of the area under the curve restricted to externally validated models predicting checkpoint inhibitor pneumonitis and to pool predictors reported in at least two studies. A total of 32 studies comprising 9462 patients were included, the incidence of immune-related adverse events in patients with non-small cell lung cancer ranged from 5.00% to 47.83%. 48 prediction models for immune-related adverse events in non-small cell lung cancer patients were included, with the area under the curve ranging from 0.57 to 1.00. 31 studies were found to have a high risk of bias, while one study was rated as low risk; all studies were found to have a good applicability. The pooled area under the curve of externally validated checkpoint inhibitor pneumonitis models was 0.83 ( 95% CI : 0.76 ~ 0.91). Meta-analysis showed: Eastern Cooperative Oncology Group performance status, treatment regimen, history of chronic obstructive pulmonary disease, interstitial lung disease, emphysema, any lung disease, hypertension, age, smoking history, radiotherapy history, body mass index, programmed death-ligand 1 expression, platelet-to-lymphocyte ratio, and targeted therapy history were identified as candidate predictors associated with the occurrence of immune-related adverse events. Current risk prediction models for immune-related adverse events in non-small cell lung cancer patients are still in an exploratory phase. While the models demonstrate good discriminative ability, the overall risk of bias is high. Future research should prioritize optimized study designs and improved reporting standards to ensure the development of prediction models with robust clinical utility.