Subtraction CT colonography: Feasibility in an animal model

AIM: To determine, in an in vitro animal model, the feasibility of subtracting residual faecal material, during virtual colonoscopy, by uniform mixing of the colonic contents with barium. MATERIALS AND METHODS: Segments of animal colon containing simulated soft-tissue polyps, measuring 3-10 mm, were filled with a faecal substitute consisting of a barium and peanut butter mix that had an attenuation value greater than 200 Hounsfield units. The colonic segments were then examined by CT using 3-10 mm beam collimations, and at pitch factors of 1 and 2. Using subtraction cut-off levels varying from 100-250 Hounsfield units, the barium and peanut butter mix was then subtracted to reveal the underlying polyps. RESULTS: Our optimal technical parameters required 3-mm beam collimation, a helical pitch factor of 2, reconstruction intervals of 3 mm, and a subtraction cut-off level of 150 Hounsfield units. The subtraction technique, in this animal model, had a sensitivity of 94% and a specificity of 80% for polyps measuring greater than 3 mm. CONCLUSION: In this limited animal model, the subtraction technique utilized was technically feasible and had an acceptable sensitivity.