Unencrypted login | home

Program Information

An Experimental Study of the Effects of Different Beam-Hardening Filters On CTDIvol and Low-Contrast Image Detectability in a CT Scanner


A Maghsoodpour

A Maghsoodpour*, S Aldoohan, Univeristy of Oklahoma Health Science Center, Oklahoma City, OK

TU-C-103-12 Tuesday 10:30AM - 12:30PM Room: 103

Purpose: To study the effects of different beam-hardening filters on CTDIvol and low-contrast image detectability in a CT scanner

Methods: The default Molybdenum beam-hardening filter of GE HiSpeed CT/i scanner was replaced with three different types of filters, having K-edges in low, middle, and high diagnostic energy range. The thickness of each filter was chosen such that it approximated Molybdenum energy-equivalent thickness for the simulated mean energy of X-ray spectrum generated at 120 kVp. The scanner calibration routine was run for each filter prior to image acquisition to ensure optimal image quality (optimized signal-to-noise ratio, optimized contrast-to-noise ratio, and no noticeable ring artifacts). An ACR CT phantom was imaged under a routine axial head protocol to evaluate the CNR for a wide range of mAs. Next, the head dose phantom was scanned to obtain CTDIvol for the same mAs range. The beam quality was also recorded. Filters were compared according to measured CNR, and CTDIvol for the smallest mAs providing CNR > 1 (= 200 mAs).

Results: Examination of CNRs indicates no significant difference among filters. Evaluation of NPS for raw images acquired from different filters will address the image noise. CTDIvol measurements indicate that the filters attenuate the beam differently due to their K-edge energies, and copper, having a low-energy K-edge (= 8.98 keV), provides the minimum CTDIvol.

Conclusion: CNR is a function of CTDIvol and better addressed when accompanied by CTDIvol. The current project has the potential for selecting a new beam-hardening filter to possibly reduce patient dose without compromising low-contrast detectability. This work is in progress to optimize CNR by investigating other available kVps, filters, measured energy spectra, NPS, and K-edge effect.

Contact Email: