Attenuation correction of SPECT using X-ray CT on an emission-transmission CT system: Myocardial perfusion assessment

We present experimental results for attenuation correction of emission tomography through use of an attenuation map derived from X-ray CT data The X-ray and gamma-ray data were obtained from a third-generation X-ray CT scanner and a single-headed SPECT scanner juxtaposed to enable systematic registration of SPECT and X-ray CT images. CT measurements were obtained from known materials at a given X-ray technique, then related to attenuation coefficients at 140 keV. The resulting calibration data were used to generate attenuation maps for the coregistered SPECT data. In addition, phantom studies provided recovery coefficients as a function of object size to compensate for partial volume effects in the coregistered radionuclide data. This technique has been applied to a torso phantom, yielding attenuation maps with an average accuracy error < 1%. Reconstruction of the SPECT image without attenuation correction underestimated the regional myocardial activity errors with an average accuracy error of -90%. Reconstruction with attenuation correction reduced the average accuracy error to -37%, but addition of a partial volume correction overestimated the regional myocardial activity concentration by an average error of 94%. The technique has also been applied to small pigs in myocardial perfusion studies by correlating the image-derived myocardial activity concentrations with those obtained directly from excised tissue samples. Correcting for attenuation increased the slope of the regression line from 0.16 to 0.36 (compared with an ideal slope of 1.00.) Adding an additional correction for partial volume effects improved the slope of the regression line to 0.93. The absolute precision error (standard error of the estimate) obtained without correction (2.14 mu Ci/ml) was degraded by attenuation correction (4.23 mu Ci/ml) and was also degraded by adding a correction for partial volume errors (10.89 mu Ci/ml), although the relative precision error without correction (31.2%) was improved by attenuation correction to 27.6% when the images were reconstructed with attenuation correction alone or with a combination of attenuation correction and compensation for partial volume effects.