Search

Purpose: The purpose of this study was to verify the maintenance of low-contrast detectability at different CT dose reduction levels, in patients of different sizes, as a consequence of the application of iterative reconstruction at different strengths combined with tube current modulation. Methods: Anthropomorphic abdominal phantoms of two sizes (small and large) were imaged at a fixed noise with iterative algorithm ASIR-V percentages in the range between 0 and 70% and corresponding dose reductions in the range of 0–83%. A total of 1400 images with and without liver low-contrast simulated lesions were evaluated by five radiologists, using the receiver operating characteristics (ROC) paradigm and evaluating the area under the ROC curve (AUC). The human observer results were then compared with AUC obtained with a channelized Hotelling observer (CHO). CNR values were also calculated. Results: For the small phantom, the AUC values lie between 0.90 and 0.93 for human evaluations of images acquired without iterative reconstruction, with 30% ASIR-V and with 50% ASIR-V. The AUC decreased significantly to 0.81 (p = 0.0001) at 70% ASIR-V. The CHO results were in coherence with human observer scores. Also, similar results were observed for the large size phantom. CNR values were stable for the different ASIR-V percentages. Conclusions: The iterative algorithm maintained the low-contrast detectability up to a dose reduction of about 70%, following application of a 50% ASIR-V combined with automatic tube current modulation, regardless of the phantom size. At further dose reductions using greater iterative percentages, a significant decrease in detectability was observed. [ABSTRACT FROM AUTHOR]

Objective: The purpose of this study was to optimize an aorta angiographic CT protocol, by investigating the best combination of tube current modulation, iterative algorithm strength and kV reduction. Materials and methods: Anthropomorphic phantoms of three sizes were imaged by CT with different values of noise index, of iterative algorithm ASIR percentages and kV in the range 80-120. Quantitative noise and contrast noise ratios were evaluated at different phantom locations. Three radiologists assessed the subjective image quality by comparing the image series with the one acquired with the reference protocol (120 kV, slice thickness 0.625, noise index 28, ASIR 40 %). Results: Although the highest CNR values were obtained for the 80 kV acquisitions, qualitative scores were higher for 100 and 120 kV at the same noise index. An optimized protocol was established with a NI of 39.2, ASIR 60 %, 100 kV for small- and medium-sized patients and 120 kV for large-sized patients, with a dose reduction of 47 %. Conclusion: When different dose reduction parameters are available, anthropomorphic phantoms of different sizes help to find the optimal combination. For aorta studies, 100 kV with relative high values of noise indexes and iterative levels provides the best balance between dose reduction and image quality. [ABSTRACT FROM AUTHOR]