The study used quantitative techniques to compare image quality for filtered back projection and AIDR 3D, and a subjective evaluation was also performed to support the quantitative findings, according to the research team.
"For evaluation of iterative reconstruction techniques, it is essential to assess noise frequencies in CT images in addition to standard deviation measurements," said Joemai, who received a research grant from Toshiba Medical Systems.
The group acquired images using a Catphan phantom at six tube current levels, and images were evaluated to assess spatial resolution, noise, and low-contrast detectability.
Both objective and subjective evaluations showed superior image quality using AIDR 3D compared to filtered back projection. Low-contrast resolution improved as much as 34% with a substantial reduction in noise compared to FBP, and subjective grading of clinical images confirmed AIDR 3D's improved image quality, artifact reduction, and dose reduction.