Optimising Neurosonography in the NICU: An Audit of 130 Cases and Proposal of a Referral and Reporting Protocol

Introduction Neurosonography is a first-line, safe bedside modality for evaluating neonates at risk of intracranial pathology. International guidelines recommend routine screening in every preterm infant and targeted scans in neonates with systemic risk factors. However, in practice, it is frequently both underutilised (missed scans in at-risk neonates) and overutilised (repeated scans in low-risk infants without clinical indication). This results in either delayed detection of critical conditions or unnecessary imaging with minimal clinical yield. We conducted an audit of 130 neonatal neurosonograms at our tertiary care centre to evaluate current utilisation patterns, assess clinical yield, and analyse the impact on management. Based on the findings, we propose a referral and reporting protocol as an exploratory instrument to improve neonatal outcomes. However, this hypothesis needs to be validated in further prospective studies. Methods We retrospectively analysed 130 consecutive neurosonograms performed over 12 months in neonates admitted to the neonatal intensive care unit (NICU) of Frank Institute of Medical Sciences, India. Data collected included clinical indication for scan, birth history, neurosonographic diagnosis, and outcome. Based on the current American Academy of Pediatrics (AAP)/American Institute of Ultrasound in Medicine (AIUM) guidelines and clinical judgement, referrals were classified as: Appropriate, possibly appropriate, inappropriate and indeterminate. To assess underutilization, delivery records during the same period were reviewed. Infants were also stratified into high-risk and low-risk groups based on international criteria. The proportion of high-risk neonates who did not undergo neurosonography was calculated. Statistical analysis compared the yield and management impact between appropriate and inappropriate referrals using chi-square analysis. Results Germinal matrix haemorrhage (GMH) was detected in 18 cases (13.8%). Periventricular leukomalacia (PVL) was identified in 19 cases (14.6%). Other findings included hydrocephalus or VP shunt-related changes in five cases (3.8%), cephalohematoma or caput succedaneum in five cases (3.8%), meningitis in one case (0.7%), and cerebral abscess in one case (0.7%). The majority of scans were normal. Diagnostic yield was significantly higher in appropriate referrals compared to the inappropriate group (chi-square p = 0.038). Twenty-four scans (18%) directly influenced clinical management, representing significant clinical implications. Of 24 neonates meeting the high-risk criteria, nine (36%) did not receive neurosonography, suggesting underutilisation. Discussion Our audit shows a wide variation in neurosonography utilisation, with underuse in high-risk neonates and overuse where the yield is low. The diagnostic yield was significantly higher in appropriate referrals, suggesting an association between referral quality and imaging value. We have developed a referral checklist and reporting template. The checklist has been developed to be used at the point of clinical decision-making. The reporting template includes disease-specific action triggers to maximise clinical effect. However, both tools are proposed as exploratory instruments derived from a single-centre retrospective audit. Conclusion Neurosonography is an essential tool for NICU care. Its clinical impact may improve when guided by structured referral protocols and standardised action-oriented reporting.