Accuracy of CO2 angiography in vessel diameter assessment: a comparative study of CO2 versus iodinated contrast material in an aortoiliac flow model.

Purpose: Precise vessel sizing prior to endovascular intervention is critical to achievement of technical success. Diameter measurements obtained with CO2 and iodinated contrast material in an aortoiliac flow model were compared.

Materials And Methods: Aortoiliac flow was simulated in a compliant, silicone elastomer phantom of the aortoiliac system using an autoperfusion pump (flow volume, approximately 1100 mL/min; mean arterial pressure, 70-80 mm Hg at 80-90 cycles/minute) and a glycerol solution (40% by weight; viscosity 3.7 centipoise at 20 degrees C). Digital subtraction angiography was performed with the phantom in the anteroposterior (AP) plane and in three oblique planes with both CO2 and iodinated contrast material. Five sets of images for both CO2 and iodinated contrast material were obtained for each projection. Two readers independently performed vessel diameter measurements at seven sites (distal abdominal aorta, bilateral proximal and distal common iliac, and mid-external iliac arteries). The model was then evaluated with intravascular ultrasound (IVUS) using a 20-MHz imaging catheter. Actual diameter measurements were taken from the inner wall to inner wall in orthogonal planes at the same locations within the model, as described previously. Analysis was performed to determine local difference in measurements (t tests), difference when compared to the standard AP projection with iodinated contrast material (Dunnett's test) and inter-reader variability (Pitman's test).

Results: The contralateral iliac vessel segment did not opacify when imaging with CO2 in the 45 degrees obliquities; thus, 22 of 28 sites were available for comparison. At 18 of 22 (81.8%) sites, there was significant difference in vessel measurements (P < .01), with CO2 yielding a significantly larger diameter at 17 of 22 (77.3%) of the sites. The difference in mean diameter ranged from -1.28 to 4.47 mm. With use of the AP iodinated contrast material run as the standard, there were significant differences (P < .05) in vessel diameter at 17 of 22 (77.3%) and four of 21 (19%) sites for CO2 and iodinated contrast material respectively, with CO2 tending toward greater diameter measurements. Significant differences (P < .05) in variance between the two readers were present but occurred primarily with CO2 in the AP projection and iodinated contrast material in the 45 degrees left obliquity. With use of IVUS as the standard, there were significant differences (P < .05) in the measured vessel diameters with CO2 at nine of 22 (40.9%) of the sites and with iodinated contrast material at 17 of 28 (60.7%) of the sites. Of the measurements made with CO2, seven of nine (77.8%) of the measurements were of larger diameter than those obtained with IVUS. By contrast, of the measurements made with iodinated contrast material angiography, IVUS measured larger diameters in 16 of 17 (94.1%).

Conclusion: CO2 angiography consistently yielded significantly larger vessel measurements when compared to both iodinated contrast angiography and WVUS. These results have important implications in regards to planning intervention based solely on CO2 angiography. Further evaluation is needed before recommending CO2 for vessel sizing in clinical practice.