Performance of simultaneous emission-transmission systems for attenuation-corrected SPECT: a method for validation applied to two camera systems

Several commercially available systems for attenuation correction in single photon emission computed tomography (SPECT) based on a transmission scan h ave been introduced that vary in performance. A test procedure for attenuat ion correction in SPECT is described and applied to two principally differe nt gamma camera systems (the Siemens Multispect 3 triple-headed system [3HS ] and the ADAC Genesys Vertex double-headed system [2HS]). The test procedu re was based on geometrically well-defined phantoms. A torso phantom was us ed to illustrate the attenuation correction methods. The test procedure can be used without detailed knowledge of or access to the algorithms used for attenuation correction. The influence on the transmission measurement of r adioactivity in a phantom was higher for the 2HS than for the 3HS. The 3HS produced satisfactory attenuation maps and corrected emission count rates t o a constant value independent of phantom density and size. With the 2HS, t here was a progressive decrease in the correction of emission count rates w ith increasing phantom density, and about 30% lower corrected count rates i n the large compared with the small phantom. A decrease in measured attenua tion coefficients in the vicinity of an emission source was demonstrated in large but not small phantoms. A likely explanation is erroneous correction of downscatter into the transmission energy window. This study demonstrate s the need for independent evaluation of systems for attenuation correction in SPECT. ((C) 2001 Lippincott Williams & Wilkins).