Objective: We evaluated the capability of CT to depict findings that allowed differentiation of small-bowel ischemia from intramural hemorrhage.
Materials and methods: Findings of 35 CT examinations (19 patients with small-bowel ischemia and 16 patients with intramural hemorrhage) were analyzed by two abdominal radiologists for the degree of wall thickening, location and length of involvement (short, <or = 15 cm; medium, 16-30 cm; or long, >30 cm), presence of hemoperitoneum, and pattern of attenuation. Patency and caliber of the superior mesenteric artery and vein were noted. Diagnosis was confirmed by laboratory findings, clinical parameters, and follow-up examinations, or at surgery. A Mann-Whitney U or Fisher's exact test was used to compare the two conditions for the following features: wall thickening, location and length of involvement, presence of hemoperitoneum, and appearance of the target sign.
Results: Among the 35 examinations, 18 abnormal segments with intramural hemorrhage and 19 abnormal segments with ischemia were identified. (Two patients with intramural hemorrhage each had two segments involved.) Mean bowel wall thickness was 11.7 mm (range, 4-25 mm) in patients with intramural hemorrhage and 4.0 mm (range, 1-9 mm) in patients with ischemia. Length of involvement was short in 14 segments with intramural hemorrhage and medium in four segments with intramural hemorrhage; none of the segments with intramural hemorrhage had long involvement. Among the segments with ischemia, length of involvement was medium in three and long in 16; none of the ischemic segments had short involvement. Fifteen (94%) of 16 segments with intramural hemorrhage and six (32%) of 19 segments with ischemia had hemoperitoneum. Seven of the 18 segments with intramural hemorrhage and nine of the 19 with ischemia had a target sign. Segments with intramural hemorrhage exhibited a higher statistically significant degree of wall thickening (p < 0.001), a shorter length of involvement (p < 0.0001), and a higher incidence of hemoperitoneum (p < 0.001) than did segments with ischemia. The two groups were not statistically different in location of involvement (p = 0.12) or in the incidence of the target sign (p = 0.18).
Conclusion: Although some of the CT features overlap, a short segment involvement with wall thickening of 1 cm or greater is typical of intramural hemorrhage; a long segment involvement with wall thickening of less than 1 cm is typical of ischemia.