Effect of Different Percentages of Adaptive Statistical Iterative Reconstruction on the Image Quality of Low Dose 64 Slice Liver MDCT Exams

Lee Mitsumori, University of Washington - Seattle Campus
Orpheus Kolokythas, University of Washington
Kent M. Koprowicz, University of Washington - Seattle Campus
Alessio A, University of Washington - Seattle Campus
May Janet, University of Washington - Seattle Campus
Gunn M, University of Washington - Seattle Campus
Shuman William, University of Washington

Abstract

Objective: To evaluate the effect of different percentages (0-100%) of adaptive statistical iterative reconstruction on the image quality of liver CT. Materials and Methods: Nineteen patients were scanned with a low-dose 64 slice liver CT technique where tube current modulation was set to achieve a 40% dose reduction (noise index 40) compared with a conventionalprotocol (noise index 30). A CT image reconstruction algorithm, adaptive statistical iterative reconstruction (ASIR), was applied to the raw data of the low dose scans to create image sets with 0, 30%, 50%, 70%, and 100% ASIR reconstructions. The image noise, sharpness, and diagnostic acceptability of the ASIR images were compared with prior liver CT scans in the same patients obtained with the conventional scan protocol. Image noise was measured in the liver for each scan. Two reviewers subjectively assessed image sharpness and diagnostic acceptability (5-point scale: 1= worst, 5 = best) of each of the ASIR and conventional exams. Results: In the low-dose studies with a noise index of 40 (NI 40), the measured image noise became lower with increasing percentage ASIR reconstruction: 65.3 ± 9.7 HU, 44.4 ± 7.6 HU, 19.7 ± 3.5 HU, 14.1 ± 2.0 HU, and 8.0 ± 1.3 HU for the 0, 30%, 50%, 70%, and 100% reconstructions (p < 0.001). Image noise of the conventional scans was 25.0 ± 3.9 HU, which was between the 30 and 50% ASIR noise values. Image sharpness scores declined with increasing ASIR percentage (3.1 ± 0.7, 3.2 ± 0.4, 3.2 ± 0.4, 1.9 ± 0.5, 1.0 ± 0.0 for the 0, 30%, 50%, 70%, and 100% ASIR) and was 4.0 ± 0.0 for the conventional scans. Diagnostic acceptability scores were 2.1 ± 0.5, 3.1 ± 0.6, 3.2 ± 0.8, 1.7 ± 0.6, 1.0 ± 0.0 for the 0-100% ASIR scans and 3.6 ± 0.5 for the conventional exams. Adjusted pair-wise comparisons revealed that all percentage ASIR levels had an image sharpness that was lower than the conventional scan (p < 0.05), and that the diagnostic acceptability scores of the 30% and 50% ASIR reconstructions were similar to the conventional studies (p > 0.6). Conclusion: Higher percentage ASIR reconstruction resulted in lower image noise and a progressive decrease in perceived image sharpness. In this study, a low-dose CT technique (NI 40) combined with 30% or 50% ASIR reconstruction resulted in images with similar diagnostic acceptability as conventional exams performed with a higher dose technique (NI 30).