Dose-saving in hepatic CT perfusion by shortening the unenhanced stage of the examinations

Abstract: Purpose: The high radiation dose delivered to patients in CT perfusion (CTp) examinations often prevents the use of CTp. This study investigates the unenhanced phase of hepatic CTp to find the minimum baseline needed to compute reliable perfusion values, thus minimizing the radiation dose delivered. Methods and Materials: 64 patients with colorectal cancer, with liver free of known diseases, underwent hepatic axial CTp examination, with acquisition (every 1s during the first 30s, every 3s after, 66s scan duration, 42 samples) starting with contrast agent injection. Considering the single compartment dual-input model, regions of interest (ROIs) were drawn for each patient on aorta, portal vein, and liver. Voxel-based blood flow (BF) values computed during first-pass phase using the entire sequence were compared (repeated ANOVA, p-value≤0.05) with those attained progressively disregarding the first unenhanced samples. Mean BF and percentage variations, dose length product (DLP) and effective dose (ED) were computed for each dataset. Results: Mean BF variations of maximum 1, 2, and 5mL/min/100g were observed in at least one examination by excluding 1, 7 (p=0.96), and 9 (p=0.1) data points, respectively. BF variations became statistically significant after that 10, or more, unenhanced samples were excluded, although perfusion maps remained highly correlated. For each sampling saved, the overall DLP and ED decreased by 2.38%. Conclusion: Starting the acquisition after 7 or 9 seconds from contrast medium injection yields a dose reduction of about 17% and 21%, respectively. This achievement would reduce the dose, without losing information, and thus could contribute to develop CTp indications.