Performance Assessment of Dynamic Spiral Scan Modes With Variable Pitch for Quantitative Perfusion Computed Tomography

Purpose: Perfusion computed tomography is increasingly being used in diagnostic radiology. Axial coverage of the traditional approach is limited to the width of the detector. Using continuous periodic table movement coverage can be increased beyond this limit. In this study, we compared tissue flow values determined from scans with a periodic spiral implementation with variable pitch with ones determined from standard dynamic scan modes. Methods: A flow phantom (preserved porcine kidney) was scanned with 2 settings of a periodic spiral (Adaptive 4D Spiral) with a range of 100 and 148 mm and a temporal sampling of 1.5 seconds. Additionally, the whole phantom was scanned with the standard dynamic mode (detector width 38.4 mm, temporal sampling 1.0 seconds) at various overlapping positions as a reference. Scan parameters (80 kV, 140 mAs, 40s scan time) were selected similar to a typical brain perfusion study. All scans were repeated 5 times. Tissue flow was calculated with a dedicated deconvolution algorithm. In a center slice and 3 additional slices at various off center positions flow values were recorded in a total of 126 regions of interest (ROI). Reproducibility was determined from the variation of the repeat scans. Agreement between periodic spirals and standard mode was determined by Bland Altman plots and correlation analysis. Results: The reproducibility of the tissue flow determination ranged from 2.7 to 4.4 mL/100 mL/min and was similar for all scan modes. The coefficient of variation ranged from 3.9% to 6.1%. Mean tissue flow in the 126 ROIs ranged from 35 to 121 mL/100 mL/min. There was excellent correlation between both periodic spiral ranges and the standard dynamic mode with a Pearson correlation coefficient of r = 0.97. The regression slope (intercept 0) for the 100 mm range was 1.01, for the 148 mm range it was 0.97. The absolute differences per ROI varied between 1.5 and 4.1 mL/100 mL/min, the relative differences between 1.9% and 6.5%. Differences did not depend on the slice location. Conclusions: Periodic spiral scan modes with variable pitch and a sampling rate of 1.5 seconds can be used for the quantitative determination of tissue flow. Their performance is equivalent to equidistant sampling with standard dynamic scan modes. The ranges of 100 and 148 mm investigated allow coverage