Room: AAPM ePoster Library
To investigate the image quality of the half-reconstruction computed tomography (CT) images with a differently positioned X-ray tube and detector of projection dataset using filtered back projection (FBP) and model-based iterative reconstruction (MBIR) methods.
All CT images (Aquilion One, Canon Medical Systems) were acquired with scanning conditions of 120 kV, 1.0 sec/rotation, and non-helical scanning. Half- and full-reconstruction algorithms for the FBP and the MBIR methods were employed, respectively. The resolution property was evaluated by a two-dimensional (multi-directional) modulation transfer function (2D-MTF) at a 100-mm off-center position using a wire method. The noise property was evaluated by measurements of the noise power spectrum (NPS) at a 100-mm off-center area (50×50 mm²) under two different exposure levels (50 and 250 mAs). The MTFs and NPSs of half-reconstruction CT image were measured with various X-ray tube and detector positions of the projection dataset.
The MTFs of the half- and full-reconstruction algorithms with the MBIR method were higher than those of the FBP method. The MTF of the tangential direction in the X-ray tube side was higher than that in the detector side for both methods. The effect of the measurement direction of the MBIR method on the change of 2D-MTF was smaller than that of the FBP method for both the half- and full-reconstruction algorithms. The NPSs of the half-reconstruction measured at the X-ray tube side were higher than those measured at the detector side in high spatial frequencies for both the FBP and MBIR methods at each exposure level.
The effect of the measurement area on the image quality evaluation was smaller in the MBIR method than in the FBP method. The half-reconstruction algorithm indicated higher resolution and more noise at the X-ray tube side and lower resolution and lesser noise at the detector side, respectively.
Noise Power Spectrum, MTF, Image Analysis