Ability of specific and nonspecific signs of multidetector computed tomography (MDCT) in the diagnosis of blunt surgically important bowel and mesenteric injuries.

PURPOSE: To determine the accuracy of MDCT in the evaluation of blunt surgically relevant bowel and/or mesenteric injuries (BMIs) using single specific CT signs together with specific pairs of nonspecific signs.
METHODS: Fifty-four patients examined with MDCT were divided into two groups: a 'surgical' group of 20 patients-which underwent surgery for blunt BMIs-and a control group of 34 'nonsurgical' trauma patients. Two radiologists with different experience performed a double-blind retrospective evaluation of the images, classifying the patients in the two groups by using only single specific signs; then, the images were reviewed in consensus with a third radiologist and sensitivity and specificity were calculated. Subsequently, the frequency of every single sign and of every possible combination of nonspecific signs in the two groups was registered, to find combinations present only in the surgical group; sensitivity and specificity were calculated by using even those specific combinations.
RESULTS: At the first consensual evaluation, sensitivity and specificity were 75 and 100%, respectively. Two combinations of nonspecific signs (focal wall thickening + extraluminal air; focal wall thickening + seat belt sign) were found only in surgical patients that did not present any single specific sign: Sensitivity calculated adding those two combinations was 95%, without a decrease in specificity.
CONCLUSIONS: MDCT is an accurate technique in the evaluation of blunt surgically relevant BMIs. The single specific CT signs were sufficient for the diagnosis in only 75% of the cases; adding the two specific combinations allowed an increase in sensitivity of 20%.