Use of Tin Filters in Cone Beam CT of Head Region for Streak Artifact Reduction: Phantom Study

Abstract

<h3>Purpose/Objective(s)</h3> This study attempts to reduce streak artifacts in cone-beam computed tomography (CBCT), which are primarily caused by high-density materials, by mounting tin filters. <h3>Materials/Methods</h3> We utilized four thicknesses of tin filters, such as default (no tins), 0.2 mm, 0.4 mm, and 0.6 mm, in front of the designated full fan bow-tie filter. The scanning protocol included both standard dose head (SDH) and high-quality head (HQH) protocols to evaluate the compensation for the reduction of low-energy photons in SDH protocol. To evaluate the degree of streak artifact reductions, an anthropomorphic phantom with insertions of aluminum and titanium plugs was scanned with above eight protocols. The percent integrity uniformity (PIU) and coefficient of fast Fourier transform (FFT) of a circumference that was twice the rod radius were evaluated. Images of a Catphan 504 phantom were analyzed in terms of noise levels, integral non-uniformities, contrast-to-noise ratios, and low-contrast visibilities. To evaluate the accompanying imaging doses, weighted CT dose indices and point doses using optically stimulated luminescence dosimeters were obtained. <h3>Results</h3> When 0.2 mm, 0.4 mm, and 0.6 mm thickness tin filters were mounted, the average PIUs increased by 13.4%, 18.9%, and 17.8% and by 35.7%, 53.0%, and 72.5% in the vicinity of aluminum and titanium, respectively. Similarly, the coefficients of lower-quarter FFTs changed by -5.2%, -11.5%, and -13.5% and by -0.7%, 2.7%, and -3.6% in the vicinity of aluminum and of titanium, respectively. The other image qualities such as noise levels, integral non-uniformities, contrast-to-noise ratios, and low-contrast visibilities were degraded as thicker tin filters were mounted, although the degree varied. By mounting 0.2 mm, 0.4 mm, and 0.6 mm tin filters, 69.4%, 84.7%, and 92.0% reductions in CT dose index and 52.7%, 70.4%, and 78.2% reductions in average point doses were obtained. <h3>Conclusion</h3> In this phantom study, attachment of a 0.4 mm tin filter with HQH protocols could effectively reduce streak artifacts without degradation of other image quality indices. In addition, the accompanied imaging doses did not increase.