Clinical Data Evaluation of C-arm-based Motion Compensated Coronary Artery Reconstruction

Angiographic C-arm systems are capable of performing computed tomography (CT) imaging for assisting in therapy planning, performance and assessment. Due to the long acquisition time, C-arm CT of dynamic structures is challenging. Cardiac motion has to be estimated and compensated in the reconstruction step. The quality of the motion estimation algorithm mainly dictates the resulting image quality. A common strategy to reduce the requirement of motion estimation is to exclude problematic heart phases by ECG-gating. A small ECG window improves the temporal resolution, but the usage of fewer data leads to undersampling artifacts. In contrast, larger ECG windows yield better image quality at the account of stronger cardiac motion artifacts. The bootstrapping approach presented here allows increasing the size of the ECG window in an iterative manner. The technique was evaluated on a clinical data set of 58 cases. Image quality was assessed by a human observer. Vessel sharpness and diameter were determined by a semi-automatic evaluation tool. The vessel diameter in 3-D was compared to a gold standard measurement in the 2-D projection images. Good image quality was achieved. The diameter of the arteries was determined reliably. The evaluation study clearly shows the benefit of using more projection data for dynamic image reconstruction. Besides avoiding undersampling artifacts a sharper reconstruction filter kernel can be applied. There is no clear choice in using either an 80% or 100% width of the ECG gating window. While using (almost) all acquired projection data, the technique appears efficient in dose and contrast agent. Keywords—C-arm computed tomography, cardiac imaging, motion compensation, image reconstruction, clinical evaluation