CT images of abdomen and pelvis: effect of nonlinear three-dimensional optimized reconstruction algorithm on image quality and lesion characteristics.

Abstract

The purpose of this study was to retrospectively assess the effect of a postprocessing nonlinear three-dimensional optimized reconstruction algorithm on image quality and lesion characteristics in abdominal and pelvic computed tomographic (CT) images. Institutional review board approved the HIPAA-compliant study protocol; informed consent was waived. Abdominal and pelvic CT images (40 patients; male-female ratio, 20:20; age range, 28-86 years) at 5-mm (n = 20) and 2-mm (n = 20) section thicknesses were postprocessed with the algorithm at three noise reduction levels. Image noise at the level of porta hepatis and acetabulum was evaluated with a five-point scale (1, no or minimal noise; 5, unacceptable noise), and presence and number of lesions and conspicuity were assessed. Statistical analysis was performed (Wilcoxon signed rank test, analysis of variance). Significant noise reduction was noted at all three levels with the algorithm (P < .05). Reduction in image contrast was noted with only one noise reduction level (P < .0001). The algorithm improves image noise without affecting lesion conspicuity and detection on low-dose abdominal and pelvic CT images.