Abstract
OBJECTIVE: Severe intraventricular hemorrhage (IVH) remains a major complication in extremely preterm infants, with significant clinical implications. We aimed to develop and internally validate a nomogram for forecasting the likelihood of early onset of severe IVH in extremely preterm neonates. METHODS: In this study, a retrospective review of clinical data was conducted among premature infants born before 32 weeks' gestation who were treated at the pediatric unit of the First Affiliated Hospital of the Army Medical University in Chongqing, China, from January 2017 through December 2023. The group of infants was split randomly into two segments-a training group consisting of 230 individuals and an internal validation group with 98-essentially a 7:3 split. According to the Volpe classification of IVH, the training group was divided into a severe IVH group (Volpe grades III-IV, n = 46) and a mild/no IVH group (Volpe grades I-II and no IVH, n = 184). Key predictive variables were identified through least absolute shrinkage and selection operator (LASSO) regression. The predictive model's performance was assessed using multiple metrics: receiver operating characteristic (ROC) curve analysis to measure discrimination, calibration plots to evaluate accuracy, and decision curve analysis (DCA) to determine clinical utility. RESULTS: Six predictors were identified in the training cohort: gestational age, 5-min Apgar score, septic shock, pulmonary hemorrhage, hemoglobin count, and thrombocytes count. The nomogram showed very good performance, yielding an area under the ROC curve (AUC) of 0.877 (95% CI, 0.815-0.939) in the training set and 0.838 (95% CI, 0.712-0.964) in the validation set. Calibration plots showed close agreement with the ideal line, and DCA indicated a substantial net clinical benefit. CONCLUSION: This nomogram offers a precise, personalized method for early detection of severe IVH risk in extremely preterm infants, aiding prompt clinical decisions.