Abstract Archives of the RSNA, 2009
SSE20-01
Evaluation of the Accuracy of Contrast Enema for Diagnosing Hirschsprung Disease in Neonates, Infants, and Children
Scientific Papers
Presented on November 30, 2009
Presented as part of SSE20: Pediatrics (Gastrointestinal)
Alex Towbin MD, Presenter: Nothing to Disclose
Beverly M Schnell PhD, Abstract Co-Author: Nothing to Disclose
Shelia Salisbury PhD, Abstract Co-Author: Nothing to Disclose
Steven Jay Kraus MD, Abstract Co-Author: Nothing to Disclose
Hirschsprung disease (HD) is a congenital disorder caused by absent ganglia in affected bowel. Findings on contrast enema include a transition zone, rectosigmoid index <1, irregular contractions, spasm, cobblestone mucosal pattern, mucosal irregularity, serrations, and microcolon. The purpose of this study is to evaluate the accuracy of contrast enema in diagnosing HD in neonates, infants, and children for the various signs.
A retrospective case-control study was performed enrolling patients with biopsy-proven HD and those with a negative biopsy. Contrast enemas were independently reviewed by two pediatric radiologists for the presence or absence of nine signs of HD. A rectosigmoid ratio was calculated for each patient. Each radiologist determined the likelihood of HD using a 5-point Likert scale.
Stepwise logistic regression was performed to identify a subset of findings that were predictive of biopsy results. Sensitivity and specificity of each finding were calculated. ROC curves of the Likert scale were created for each radiologist and the inter-reviewer reliability was calculated.
There were 52 patients with biopsy-proven HD and 63 with a negative biopsy. The only signs predictive of HD were the presence of a transition zone and an abnormal rectosigmoid ratio. Using the results of a less experienced reviewer, whose experience is more typical of a general radiologist (i.e. not a specialist in pediatric fluoroscopy), the sensitivity and specificity with 95%-CI of a transition zone was 68% (49–82) and 56% (36–74) in neonates, 100% (56–100) and 69% (48–85) in infants, and 27% (7–61) and 90% (54–99) in children. For an abnormal rectosigmoid ratio, the sensitivity and specificity were 68% (49–82) and 78% (57–91) in neonates, 86% (42–99) and 62% (41–79) in infants, and 55% (25–82) and 70% (35–92) in children. The area under the curve for the Likert scale was 0.76 for the entire population. Inter-reviewer reliability was moderately high.
A transition zone and abnormal rectosigmoid ratio are the signs most predictive of HD in the pediatric population; however, their sensitivity and specificity are low. Given that biopsy provides a definitive diagnosis and has a low associated morbidity, it should be relied upon as the diagnostic method of choice.
Contrast enema is neither sensitive nor specific for a diagnosis of Hirschsprung disease.
Towbin, A,
Schnell, B,
Salisbury, S,
Kraus, S,
Evaluation of the Accuracy of Contrast Enema for Diagnosing Hirschsprung Disease in Neonates, Infants, and Children. Radiological Society of North America 2009 Scientific Assembly and Annual Meeting, November 29 - December 4, 2009 ,Chicago IL.
http://archive.rsna.org/2009/8007548.html