Development and Validation of a Dynamic Nomogram to Predict the Risk of Neonatal White Matter Damage

Two risk assessment models for predicting white matter injury in extremely preterm infants

BACKGROUND

Extremely preterm infants (EPIs) are at high-risk of white matter injury (WMI), leading to long-term neurodevelopmental impairments. We aimed to develop nomograms for WMI.

METHODS

The study included patients from 31 provinces, spanning ten years. 6074 patients before 2018 were randomly divided into a training and internal validation group (7:3). The external validation group comprised 1492 patients from 2019. Predictors were identified using the least absolute shrinkage and selection operator (LASSO) and multivariable logistic regression and nomograms were constructed. Models' performance was evaluated using receiver operating characteristic (ROC), decision curve analysis (DCA) and calibration curves.

RESULTS

The prenatal nomogram included multiple gestation, premature rupture of membranes (PROM), chorioamnionitis, prenatal glucocorticoids, hypertensive disorder complicating pregnancy (HDCP) and Apgar 1 min, with area under the curve (AUC) of 0.805, 0.816 and 0.799 in the training, internal validation and external validation group, respectively. Days of mechanical ventilation (MV), shock, patent ductus arteriosus (PDA) ligation, intraventricular hemorrhage (IVH) grade III-IV, septicemia, hypothermia and necrotizing enterocolitis (NEC) stage II-III were identified as postpartum predictors. The AUCs were 0.791, 0.813 and 0.823 in the three groups, respectively. DCA and calibration curves showed good clinical utility and consistency.

CONCLUSION

The two nomograms provide clinicians with precise and efficient tools for prediction of WMI.

IMPACT

This study is a large-sample multicenter study, spanning 10 years. The two nomograms are convenient for identifying high-risk infants early, allowing for reducing poor prognosis.