Purpose The purpose of this study was to determine whether the measured size of active gastrointestinal hemorrhage was useful in predicting subsequent positive findings at catheter angiography. Materials and methods This was a single-institution retrospective study consisting of 32 patients with computed tomography angiography (CTA) positive for gastrointestinal bleeding who went on to receive catheter angiography. Each CTA was reviewed, with axial measurements of the anterior-posterior and transverse dimensions of the largest foci of hemorrhage recorded. Volumetric analysis was used to measure the volume of hemorrhage. These measurements were performed for both the arterial and portal venous phases. Additionally, the interval growth between the arterial and portal venous phase was also calculated. Results There was a statistically significant difference in the absolute size of the maximum transverse dimension on portal venous phase imaging (mean = 19.8 mm, p < 0.001), as well as an interval increase in transverse (mean = 8.5 mm, p < 0.001) and anteriorposterior (mean = 5.4 mm, p = 0.027) size between arterial and portal venous phases in patients with positive catheter angiography versus negative catheter angiography. There was a statistically significant difference in the volume of hemorrhage on arterial (mean = 1.72 cm(3), p = 0.020) and portal venous phases (mean = 5.89 cm(3), p = 0.016), as well as an interval change in the size of hemorrhage between the two phases (mean = 4.17 cm(3), p = 0.020) in patients with positive catheter angiography versus patients in the negative catheter angiography group. Conclusions The absolute axial size and volume of hemorrhage, as well as the interval change between the arterial and portal venous phases of CTA imaging is predictive of subsequent positive catheter angiography.
