Differential diagnosis between intussusception and gastroenteritis by plain film.

BACKGROUND: The objective of this study is to reassess the validity of plain film diagnosis for pediatric intussusception with consideration of the combinations of radiologic findings. METHODS: Sixty-six cases of intussusception and 81 controls of gastroenteritis were collected. Their films were read blindly by three radiologists together, with emphasis on nine radiologic findings: sparse colon gas (F1), sparse fecaloid content in colon (F2), gas-filled small bowel loops in right hypochondrium (F3), small bowel obstruction (F4), difficulty in assessing cecum position (F5), pneumoperitoneum (F6), discernible mass lesion (F7), target sign (F8) and crescent sign (F9). The sensitivity and specificity of individual findings and combinations of findings were analyzed. RESULTS: No cases showed F6. Listed in decreasing order of sensitivity, the other eight findings were F1, F5, F2, F7, F3, F4, F9 and F8. The sensitivity and specificity of these eight findings were inversely proportional with significant p-values (F8: < 0.005; the other seven findings: < 0.001). Sixty-three films (95%) in case group displayed combination of at least three radiologic findings. If the diagnostic criteria must consist of at least three radiologic findings, 60 cases (74%) of control group can be excluded. CONCLUSIONS: After eliminating the confounding effect of small bowel obstruction on the percentage of gas-filled small bowel loops in right hypochondrium and excluding the data agreed by only one radiologist from the study by Ratcliffe et al, the results of plain film findings were consistent with those of previous reports. The plain abdominal film can play an active role in the diagnosis of pediatric intussusception. Its validity increases when combinations of radiologic findings rather than individual signs are emphasized.