Influence of different reconstruction parameters in the visualization of intracranial stents using C-arm flat panel CT angiography: experience in an animal model.

Abstract

C-arm flat panel computed tomography angiography (CA-CTA) is a relatively new imaging modality. Consequently, knowledge about postprocessing parameters and their influence on image quality is still limited, especially for the visualization of implanted microstents. To optimize reconstruction parameters by evaluating the influence of these different parameters for CA-CTA visualization of microstents in an animal model. Eleven microstents were implanted within the left common carotid artery of 11 New Zealand white rabbits. Both CA-CTA, using intra-venous delivery of contrast material, and conventional digital subtraction angiography (DSA) was performed. CA-CTA datasets were reconstructed using three different image characteristics (normal, sharp, smooth). Two experienced neuroradiologists evaluated the image quality and performed measurements of inner and outer stent diameters as well as measurements of the lumen area. Stent deployment was performed successfully in all animals. Inter-observer correlation coefficient for all measurements was high (r = 0.87-0.92). Lumen area and inner stent diameter were significantly smaller in image characteristic "smooth" (P < 0.01) than in "sharp" and "normal". Outer stent diameter was larger in "smooth" than in "sharp" and "normal" (P < 0.01). Stent strut size was significantly wider using image characteristic "smooth". "Sharp" and "normal" compared best to DSA, with "sharp" providing the closest match to DSA measurements, with the trade-off of significantly more noise than in the "normal" reconstructions. The use of different image characteristics in the postprocessing of CA-CTA datasets has an influence on the visualization of implanted stents. Image characteristic "sharp" and "normal" compared best to DSA.