Optimizing Image Quality in Cerebral 3D Rotational Angiography: A Study on the Impact of Voxel Size

Secondary reconstruction with 2 voxel sizes (0.3 mm and 0.05 mm). A, FOV sizes for secondary reconstruction: 15.36 cm (green square, 0.3 mm × 512 matrix) and 2.56 cm (yellow square, 0.05 mm × 512 matrix). B and C, Fusion 3D images of the right (orange) and left (gray) internal carotid arteriograms (posterior view) are presented. The aneurysm at the anterior communicating artery origin of the right anterior cerebral artery is presented. A small subcallosal artery (arrow) arises from the anterior communicating artery. While the artery is depicted more clearly in the smaller voxel reconstruction (C), more noise is noted in the overall vessels. The window setting (W: window width, C: window center) was maintained across both reconstructed images to ensure a fair and unbiased comparative analysis.

Quantitative measurement of vessel intensity. A, Multiplanar reconstruction images of a vessel by using voxel sizes of 0.30 mm (right) and 0.05 mm (left). The line intensity profile was derived from a crossing line (indicated by the red line) perpendicular to the vessel. Regions of interest (circled) were designated to determine the maximum intensity and background noise. B. Line intensity profiles corresponding to each voxel size. C, Intensity profile with Gaussian curve fitting (black line). The slope of the curve was determined between the 20% and 80% intensity thresholds.

Quantitative and qualitative analyses of image quality across voxel sizes. Upper row, Quantitative metrics for image quality at varying voxel sizes (0.05 mm to 0.30 mm): A, Maximum intensity; B, Slope of grayscale intensity at the vessel margin; and C, Background noise levels. Lower row: Qualitative assessments of image quality by using a 5-point Likert scale, capturing cumulative scores from 3 raters: D, Image sharpness; E, Noise; and F, Overall image quality. Significant differences between adjacent voxel sizes are denoted with corresponding P values.