Development and validation of a clinical features-based nomogram for predicting neonatal cerebral microbleeds

Area of Science:

• Neonatal Neurology
• Pediatric Radiology
• Clinical Prediction Modeling

Background:

• Neonatal cerebral microbleeds (CMBs) are infrequent and often asymptomatic, leading to diagnostic and treatment delays.
• Limited research exists on neonatal CMBs and their risk factors, hindering proactive prevention strategies.
• Early identification of risk factors is crucial for effective management of neonatal CMBs.

Purpose of the Study:

• To develop and validate a nomogram for predicting neonatal CMBs.
• To identify independent clinical risk factors associated with neonatal CMBs.
• To assess the efficacy of the developed prediction model.

Main Methods:

• Retrospective study of 230 neonates (115 with CMBs, 115 controls) using 1.5-T MRI.
• Data collection included clinical variables and CMB characteristics; data split into training (70%) and validation (30%) cohorts.
• Multivariate logistic regression used to construct a nomogram; model performance evaluated using ROC, calibration, and decision curve analysis.

Main Results:

• Independent risk factors for neonatal CMBs identified: spontaneous delivery, neonatal pneumonia, gestational hypertension, and gestational diabetes.
• The nomogram demonstrated good predictive performance with an AUC of 0.811 in the training cohort and 0.780 in the validation cohort.
• Ischemic infarction was an independent risk factor for moderate-to-severe CMBs, with an AUC of 0.731.

Conclusions:

• The developed nomogram effectively predicts neonatal CMBs based on identified independent risk factors.
• The model shows high calibration and clinical utility, potentially aiding in early diagnosis and treatment.
• Further research could refine prediction models for specific CMB severities.