Combining abdominal ultrasound and radiography for surgical risk stratification in necrotising enterocolitis: a prospective cohort pilot study.

Abstract

Objective: To evaluate whether combining abdominal ultrasound with radiography improves diagnostic accuracy and surgical risk stratification in neonates with suspected necrotising enterocolitis (NEC) compared with radiography alone.

Design, setting and patients: Prospective cohort pilot study conducted in two tertiary neonatal intensive care units. Sixty-seven neonates with suspected NEC underwent concurrent abdominal radiography and ultrasound assessments. Imaging studies were independently reviewed by masked investigators using pre-specified criteria to classify each study as reassuring or non-reassuring.

Main outcome measures: The main outcome measure was the need for surgical intervention. Imaging data were analysed using unsupervised k-means clustering (k=2): logistic regression-testing associations with surgery and principal component analysis (PCA)-to identify imaging features most contributing to group separation.

Results: Ultrasounds were reassuring in all cases subsequently diagnosed with non-NEC, that is, feeding intolerance, whereas most radiographs in this group were non-reassuring. Clustering based on radiographs alone did not significantly discriminate surgical risk (58.8% vs 39.4%; p=0.11). Combined model (radiograph+ultrasound) produced two distinct clusters with significantly different surgical rates (78.3% vs 34.1%; OR 6.96, 95% CI 2.29 to 24.58). PCA highlighted complex ascites, absent peristalsis and abnormal bowel perfusion as key discriminating features.

Conclusion: Combining abdominal ultrasound with radiography improved the identification of neonates at high surgical risk from NEC, in our pilot study. A reassuring ultrasound reliably identified infants with feeding intolerance, suggesting potential to reduce unnecessary transfers and treatments. Larger multicentre studies are needed to validate these findings and inform development of a unified multimodal imaging score for NEC diagnosis.