Preliminary nomogram model for predicting respiratory failure of patients with idiopathic inflammatory myopathies.

levels below 60 mmHg for more than 1 month. Predictive factors were initially screened using LASSO regression, and independent predictors were subsequently identified using multivariate Cox regression. A nomogram was developed using key predictors. Model performance was assessed through time-dependent ROC curves and decision curve analysis (DCA). Multivariate Cox analysis identified multiple independent risk factors for respiratory failure (p < 0.05). The final predictive nomogram incorporated six variables: cough, carcinoembryonic antigen, creatine kinase, lymphocyte count, myositis-specific autoantibodies status, and osteopontin (OPN) level. The model demonstrated excellent discrimination in both cohorts, with time-dependent AUCs of 0.926, 0.934, and 0.935 at 6 months, 12 months, and 60 months in the development group, and 0.942, 0.893, and 0.972 at corresponding time points in the validation group. DCA confirmed the model's clinical utility. We successfully developed and validated a novel nomogram that integrates essential clinical and laboratory indicators. This tool, incorporating the research biomarker OPN, provides a framework for personalized risk assessment in research settings and may aid future risk stratification strategies pending further validation.