231. Adaptive iterative dose reduction in 3D (AIDR 3D) algorithm: Quality image and dose to the patient

Abstract

Purpose The iterative reconstruction techniques have been proposed for decades to improve the quality of CT images and to reduce the quantum noise, dose and artefacts. This work evaluated the variability of the image quality by varying the strength of the adaptive iterative dose reduction in 3D algorithm (AIDR 3D) and how this was reflected in the dose delivered to patient. Materials and Methods 3D AIDR is an iterative algorithm, integrated in the acquisition procedure to ensure dose reduction. We assessed image noise, linearity and accuracy of CT numbers, the spatial resolution with dedicated phantom (Catphan503) varying the strength of the algorithm. Introducing the 3D AIDR algorithm in the acquisition phase, the dosimetric index, CTDIvol, was monitored. For some clinical cases, the images were reconstructed retrospectively with AIDR 3D and the diagnostic quality of the image was assessed by an expert radiologist. Results We acquired images of Catphan503 without the iterative algorithm. These images were reconstructed with different algorithm strength. The images were analyzed by ImageJ to extract the physical parameters. The accuracy of the CT number of the different materials included in the Catphan503 does not vary with the strength of the iterative algorithm. It has been verified that the image noise is reduced by about 40% by introducing the 3D AIDR Standard and a further 8% by AIDR 3D Strong. MTF value was reduced compared to images reconstructed by FBP algorithm of about 3%, 4.5% and 7% in images reconstructed respectively by AIDR Mild, AIDR Standard and AIDR Strong iterative algorithm. The dosimetric index was reduced of approximately 50% when the images were acquired by the iterative algorithm integrated in the acquisition phase. SNR of a uniform region of clinical images was increased compared to images reconstructed by FBP algorithm of about 60%, 67% and 73% in images reconstructed respectively by AIDR Mild, AIDR Standard and AIDR Strong iterative algorithm. Conclusions The algorithm preserved the physical parameters of the image. The radiologist involved in this study has not found a significant loss of quality diagnostic image.