99 - Degree of Uncertainty in Reporting Imaging Findings for Necrotizing Enterocolitis: A Secondary Analysis of a Pilot Diagnostic Randomized Clinical Trial

Poster Number: 99
Publication Number: 99.237

Alain Cuna, Children's Mercy Hospitals and Clinics, Kansas City, MO, United States; Disa Rathore, Kansas City University, Miami, FL, United States; Sherwin S. Chan, Children's Mercy Hospitals and Clinics, KANSAS CITY, MO, United States; Erin Opfer, Children’s Mercy Hospital, Kansas City, KS, United States; Kira Bourret, Kansas City University, Ashburn, VA, United States

Background: Uncertainty in reporting of imaging findings for necrotizing enterocolitis (NEC) can be challenging and result in ambiguity, miscommunication, and potential diagnostic errors. The degree to which uncertainty complicates diagnostic imaging for NEC has not been characterized. 

Objective: To determine the degree of uncertainty in diagnostic imaging for NEC.

Design/Methods: We conducted a retrospective study using data from a previously completed pilot diagnostic randomized clinical trial (RCT) (Clinicaltrials.gov NCT03963011). The study population comprised of preterm infants with suspected NEC who were randomized to either standard imaging with abdominal radiographs (AXR) alone or experimental imaging with AXR + add-on bowel ultrasound (BUS). Three investigators independently reviewed radiology reports and assigned level of uncertainty scores using a 4-point Likert scale (Table 1). Our primary outcome was uncertainty scores for pneumatosis, portal venous gas, and free air. Secondary outcomes included rates for complete reporting for all three main NEC findings, use of standardized reporting templates, and inclusion of diagnostic certainty scale.

Results: A total of 113 AXR and 24 BUS were conducted on sixteen preterm infants (mean gestational age 27.2 ± 2.2 weeks, mean birth weight 1020 ± 373 grams) with concern for NEC as part of a pilot diagnostic RCT. Overall, BUS had less uncertainty for reporting each of the three main NEC findings compared to AXR (Pneumatosis: 1 [1 – 1.75) vs 3 [2 – 3], P< 0.0001; Portal venous gas: 1 [1 – 1] vs 1 [1 – 1], P=0.02; Free air: 1 [1 – 1] vs 2 [1 – 3], P< 0.0001; Figure). BUS reports also had higher rates of complete reporting and higher rates for using standardized reporting template and diagnostic certainty scales compared to AXR reports (Table 2). 

Conclusion(s): The superior technical accuracy of BUS may allow radiologists to report the presence or absence of important NEC findings with less uncertainty compared to AXR. The substantial improvement in uncertainty with BUS provides additional evidence to support the use of BUS as an adjunct to AXR for NEC diagnosis. Further studies are needed to determine whether improved uncertainty translates to improved clinical outcomes.