This study is the first to evaluate a standardized protocol for obtaining and reporting abdominal ultrasound images following the implementation of a clinical guideline for universal adjunct abdominal ultrasound screening during the initial diagnostic evaluation for NEC. Before standardization, the feasibility and benefit of using abdominal ultrasound as an adjunct to abdominal radiography for diagnosing NEC were uncertain. Following implementation, adherence to the protocols was high in patients diagnosed with NEC, and no significant adverse events were associated with sonography for the diagnosis of NEC.

While prior studies have evaluated the benefits of adjunct abdominal ultrasound, a lack of standardization for its universal use has made it less apparent in which specific clinical scenarios the ultrasounds were obtained. The average length of time for performing the ultrasounds following protocol implementation was relatively brief: 13 min, and 23% of the ultrasounds were conducted between the hours of 19:00 and 07:00. These findings suggest that this protocol is feasible for adjunct abdominal ultrasounds to be obtained consistently during initial diagnostic evaluations, regardless of the time at which they occur.

Our findings are consistent with previous studies which postulated that ultrasonography is more sensitive than radiography in detecting pneumatosis intestinalis, portal venous gas, and possibly pneumoperitoneum [2, 5, 15,16,17]. A recent study by May et al. demonstrated the sensitivity and specificity of abbreviated abdominal ultrasonography for identifying high-risk findings of NEC, described as pneumoperitoneum, complex free fluid, and fluid collections, at 100% and 95%, respectively [2]. Although our study was unable to calculate the specificity of adjunct abdominal ultrasonography as it did not include negative cases of NEC, the sensitivity was high at 0.91. This was more sensitive than the initial abdominal radiograph for NEC (0.68). This is reassuring in the context of routine use, given that a 2018 meta-analysis reported sensitivities ranging from 0.27 to 0.48 for classic signs of NEC [18]. We speculate that universal practice with the abdominal ultrasound evaluation for NEC, especially to “rule in” highly suspected cases, resulted in improved sensitivity. This study highlights a significant portion of NEC cases (32%) that demonstrate positive ultrasound findings but no findings on initial abdominal radiographs. Additionally, another 23% of NEC cases had only indeterminate findings of possible pneumatosis as the sole finding on radiographs. In addition, 59% of cases resulted in a higher modified Bell’s stage from abdominal ultrasound findings when compared to the initial abdominal radiograph findings. These instances suggest that ultrasound is more sensitive than abdominal radiography alone in staging NEC. Both infants with pneumoperitoneum and bowel perforation by complex fluid collection on abdominal ultrasounds had negative radiographs for these findings. However, no cases of NEC in the abdominal ultrasound cohort had free air on radiography, so further studies are needed to investigate the sensitivity of universal adjunct abdominal ultrasonography for this finding.

In both cases, with false-negative initial abdominal radiographs and positive findings on abdominal ultrasound, universal screening with adjunctive abdominal ultrasound may have aided the earlier or more accurate staging of NEC. Precise diagnosis and staging of NEC may influence the choice of antibiotic therapy, duration of antibiotic treatment, or duration of bowel rest, depending on the center’s guidelines.

It is important to emphasize that, within this study, most of the ultrasound findings in NEC were not singularly present (e.g., only pneumatosis) but were present as a constellation of findings that informed the overall diagnosis and stage of NEC (e.g., pneumatosis, portal venous gas, and bowel wall thickening). Most clinicians agree that an isolated finding of pneumatosis in the absence of other signs and symptoms is insufficient to diagnose NEC. Pneumatosis can be present in non-NEC conditions (e.g., volvulus or benign pneumatosis coli) or it can be transient. While the presence of pneumatosis intestinalis can aid in staging NEC, it does not definitively correlate with disease severity. Therefore, multiple findings are crucial for achieving a high level of confidence in a diagnosis of NEC. Ultrasound was able to confirm the primary abdominal radiograph finding, but it often revealed multiple other findings consistent with NEC, thereby increasing provider confidence. A recent study by Le Cacheux et al. described an increased association with abdominal ultrasound findings, including thickening of the mesentery, increased echogenicity of intraluminal intestinal content, abnormalities of the abdominal wall, and poor definition of the intestinal wall, in infants with surgical NEC compared to those with medical NEC [20]. It may be valuable to incorporate these additional findings into adjunctive abdominal ultrasound protocols as part of a future feasibility study.

These results suggest that adjunctive abdominal ultrasound can aid in detecting findings during the initial evaluation of NEC and can be routinely applied in a relatively large academic neonatal care unit. Standardizing adjunct abdominal ultrasound as part of a multidisciplinary team may improve the reporting of findings in NEC compared to initial abdominal radiography, potentially leading to more accurate modified Bell’s staging of NEC or prognostication of medical versus surgical outcomes on initial evaluation.

We note that the pediatric radiology expertise at our center significantly aided in the design and implementation of the adjunct abdominal ultrasonography protocol for evaluating NEC. The Pediatric Radiology department at our center consists of six attending physicians who consistently use the created dictation template to report findings from adjunct ultrasounds, following the implementation of the protocol.

This study has several limitations, mainly due to its small sample size and retrospective design. The assessment of ultrasound sensitivity and specificity is limited because there is no definitive gold standard diagnostic test for NEC. Nevertheless, we can make relative comparisons between modalities. Diagnosis and staging of NEC can be subjective, but this is partly mitigated by a multidisciplinary team consensus on accurate staging using modified Bell’s criteria, determined by an independent panel of reviewers. Given the promising feasibility of establishing a guideline for universal adjunct abdominal ultrasonography for NEC at our single center, a multicenter trial would be helpful to further validate these findings. Typically, the number of NEC diagnoses made by ultrasound was relatively high and may reflect verification bias, as providers used ultrasound findings to guide their clinical diagnosis. We cannot calculate the specificity of abdominal ultrasound because of our unit guidelines. Since our center does not currently use abdominal ultrasound to “rule out” NEC, such as in cases of low suspicion, but mainly to confirm and stage cases with high suspicion where clinicians are already committed to an evaluation, therefore negative ultrasound results for NEC are unavailable. A separate study would be valuable for formally assessing this modality’s ability to rule out NEC in truly negative cases and for measuring changes in antibiotic use or NPO days, which comprise NEC treatment. Nonetheless, the rate of NEC at our center remains comparable to other academic centers nationwide and did not change during the study period. A larger sample size, especially from a population with a significantly lower NEC rate, might result in a higher false-positive rate and a lower positive predictive value. An extended period of follow-up after guideline implementation could provide additional insights.