Comparison of dual-window scatter correction and effective attenuation coefficients for quantification in dedicated breast SPECT

Abstract

Correction for scattered and attenuated photons is necessary for accurate quantification in dedicated breast SPECT. An implemented dual energy window (DEW) scatter correction method along with attenuation correction has been shown to be accurate to within 10% of true values; however, the DEW method requires multiple processing steps, and thus more time, than an effective attenuation coefficient (EAC) method. This study aims to determine an EAC for quantification accuracy comparable to the DEW method. Our dedicated, high performance CZT-based breast SPECT system is capable of novel 3D, non-traditional trajectories. A breast phantom containing lesions ranging from 0.1 to 1.6mL was filled with Tc-99m to 8:1 and 4:1 lesion-to-background activity concentrations. Multiple acquisitions were collected using vertical axis of rotation (VAOR) and projected sine wave (PROJSINE) orbits and an 8% energy window. The PROJSINE trajectory incompletely samples the volume, but offers potential clinical benefits by contouring the breast and acquiring chest wall projections. Images were reconstructed with OSEM to 20 iterations using both uniform EAC masks with values of 0.12 to 0.165cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> and the DEW method with measured 0.16 scatter coefficient and the NIST attenuation value of 0.1545cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> for water at 140keV. ROIs were seeded and automatically drawn within each lesion, and measured values were compared to known truth. Results indicate that the DEW method is superior for VAOR trajectories and a single EAC value may not adequately compensate non-uniformly shaped objects that have a spatially varying scatter distribution. However, for PROJSINE trajectories, an EAC of approximately 0.13cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">-1</sup> yields comparable quantification accuracy for background and large lesions without the need for additional image processing, indicating this may have clinical utility in image quantification.