CT for evaluation of urolithiasis: image quality of ultralow-dose (Sub mSv) CT with knowledge-based iterative reconstruction and diagnostic performance of low-dose CT with statistical iterative reconstruction

Abstract

To compare radiation dose and image quality in regular, low, and ultralow-dose CT protocols, and to evaluate diagnostic performance of low-dose CT for urolithiasis. Sixty-five patients with suspected urolithiasis underwent three different scans under the regular, low, and ultralow-dose protocols. The regular dose scans were reconstructed using filtered back projection and the low-dose scans were reconstructed using a statistical iterative reconstruction. The ultralow-dose scans were reconstructed using both techniques in addition to a knowledge-based IR. Effective radiation doses were compared. Objective image noise was assessed by measuring standard deviation of HU and subjective image assessment was performed with a 3- or 5-point scale. Diagnostic performance of the low-dose image was evaluated, using the regular dose image as a standard reference and the interobserver agreement between two reviewers with different levels of experience was calculated. The effective radiation dose was significantly different in each protocol (p < 0.001) and estimated dose reduction of the low-dose and ultralow-dose protocols was 76.4% and 89.8%, respectively. The knowledge-based iterative reconstruction algorithm showed poorer subjective image quality than the regular and low-dose protocols, but it also had the least objective image noise. Overall, the low-dose image set showed a greater than 84% concordance rate and 100% in ureter stones larger than 3 mm. Interobserver agreement was substantial (kappa value = 0.61). The knowledge-based IR can provide a better quality image while reducing radiation exposure under the same protocol. Furthermore, the diagnostic performance of the low-dose CT protocol is comparable to the regular dose scan.