CT of small-bowel obstruction in children: sensitivity and specificity

OBJECTIVE.:The aim of this study was to determine the sensitivity, specificity, and accuracy of CT in the diagnosis of small-bowel obstruction in children.

The CT scans of 30 children with surgically proven small-bowel obstruction, 22 children with ileus, and 29 children who served as controls were retrospectively reviewed by two of four interpreters who were unaware of the children's final diagnoses. Causes of obstruction in the patients included 19 adhesions, six cases of volvulus, five intussusceptions, four strictures, and two cases each of internal hernia and abscess. Eight obstructions had multiple causes. The CT scans were evaluated for the presence of small-bowel obstruction using a scale with five degrees of confidence. In cases of discrepancy of more than one level of certainty, a third interpreter was consulted. Criteria for small-bowel obstruction included a discrepancy in caliber between the proximal dilated and the more distal small bowels or generalized small-bowel dilatation (>2.5 cm) in the presence of a collapsed colon. An interpreter's rating that an obstruction was either present or probable was considered a positive finding; a rating indicating that the interpreter was not sure whether an obstruction was present or believed that an obstruction was not probable or saw normal anatomic structures was considered a negative finding for small-bowel obstruction. The cause and level of obstruction also were recorded.

There were 26 true-positive (87%) and four false-negative (13%) interpretations for small-bowel obstruction. Among the interpretations of scans of patients with ileus, 68% were true-negative and 32% were false-positive interpretations for small-bowel obstruction. Among the control group, there were no false-positive readings. Sensitivity of CT was 87%, specificity was 86%, and accuracy was 86%. In the scans of children 2 years and younger, CT had a sensitivity of 100% and specificity of 0%. Of the patients with surgically confirmed levels of obstruction, the correct level of obstruction was described by both interpreters in 12 (86%) of 14 scans. The causes of obstruction were correctly identified in 14 (47%) of 30 scans.

CT is both sensitive and specific for use in diagnosing small-bowel obstruction in children, especially in children older than 2 years.