Ultrasound for midgut volvulus and malrotation: frequency and predictors of a non-diagnostic examination in a multi-institutional cohort.

Background: Ultrasound (US) is increasingly used as the first-line imaging modality for the diagnosis of midgut volvulus, but may be non-diagnostic in some cases.

Objective: To determine the frequency and factors associated with non-diagnostic US for each midgut volvulus and malrotation in a multi-institutional sample.

Materials and methods: We conducted a retrospective multi-institutional study of children (age 0-18 years) who underwent US to evaluate for midgut volvulus and malrotation between January 1, 2018, and June 30, 2021, and had an available reference standard of one of the following: upper GI series, CT/MRI, surgery, or, for volvulus, clinical follow-up at 30 days or greater. Blinded review of US images was performed by a single radiologist at each institution. When available, radiographs acquired ≤3 h from the US were reviewed for bowel gas pattern. After blinded review, original ultrasound reports were classified as diagnostic or non-diagnostic for midgut volvulus and malrotation. Stepwise logistic regression identified the most important predictors of non-diagnostic US.

Results: In total, 637 patients were imaged for midgut volvulus and 311 for malrotation. Based on original report review, non-diagnostic proportions of examinations for volvulus and malrotation were 13.5% (86/637) and 25.7% (80/311), respectively. Based on blinded review, non-diagnostic proportions of examinations for volvulus and malrotation were 17.3% (110/637) and 37.6% (117/311), respectively. Of the patients with US considered non-diagnostic for volvulus by original reports, 2.3% (2/86) were subsequently found to have volvulus. Among patients with non-diagnostic US for volvulus by blinded review (n=110), none was found to have volvulus. Gaseous dilation with elongation of bowel on radiography was the single best predictor of a non-diagnostic US in blinded interpretation for volvulus and malrotation (OR=8.2 and 9.2; 95%CI 3.7-19.8 and 1.7-89.4, respectively) and in original radiology reports for volvulus (OR=4.5; 95%CI 2.2-9.5).

Conclusion: A small fraction of a multi-institutional sample of US for midgut volvulus was non-diagnostic; however, assessment of malrotation without volvulus is associated with a higher frequency of non-diagnostic examinations. Dilated bowel gas pattern on radiography is the strongest predictor for a non-diagnostic US, although it does not necessarily preclude a diagnostic exam.