Construction of a Risk Stratification Model for Severe Mycoplasma pneumoniae Pneumonia in Children Based on PRDM1, LDH, CRP, and the PRESS Score

Abstract Background Severe Mycoplasma pneumoniae pneumonia (SMPP) leads to poor outcomes in children, while reliable early risk stratification tools are currently insufficient. This study constructed and internally evaluated an SMPP risk stratification model combining novel immune biomarkers and routine laboratory indicators to support early risk identification and individualized clinical management of pediatric SMPP. Methods This retrospective single-center study enrolled 235 hospitalized children with MPP (164 mild and 71 severe cases) and 40 healthy controls from October 2023 to January 2026. Candidate immune biomarkers were screened based on published single-cell transcriptomic datasets and validated via RT-qPCR and ELISA. We applied LASSO-logistic regression and multivariable analysis to screen independent SMPP risk determinants, using validated immune biomarkers, clinical data, laboratory parameters and disease severity scores as candidate predictors. A nomogram-based risk stratification model was subsequently constructed. The MPP cohort was randomly split into training and test sets at a 7:3 ratio. Model performance was comprehensively assessed via ROC curve, Bootstrap calibration curves, and decision curve analysis (DCA). Results PBMC-derived PRDM1 mRNA and serum PRDM1 protein levels were significantly elevated in children with SMPP ( P < 0.05), indicating aberrant PRDM1 expression in severe pediatric MPP. Four independent risk determinants were screened, including the PRESS score, LDH, CRP, and serum PRDM1. The nomogram showed excellent discriminatory ability, with AUCs of 0.924 (95% CI: 0.881–0.967) for the training set and 0.856 (95% CI: 0.751–0.962) for the test set. The model exhibited good calibration (training MAE = 0.050; testing MAE = 0.036), and DCA showed sustained clinical net benefit across threshold probability range of 0.02–0.85. Conclusions This nomogram-based risk stratification model integrating PRDM1, PRESS score, LDH and CRP, possesses good discrimination, favorable calibration and reliable clinical utility. It facilitates early risk assessment and standardized stratified management for pediatric SMPP and aids precise clinical decision-making.