TY - JOUR T1 - Problems Created in Attenuation-Corrected SPECT Images by Artifacts in Attenuation Maps: A Simulation Study JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 335 LP - 343 VL - 46 IS - 2 AU - Anna Celler AU - Katherine L. Dixon AU - Zheng Chang AU - Stephan Blinder AU - John Powe AU - Ronald Harrop Y1 - 2005/02/01 UR - http://jnm.snmjournals.org/content/46/2/335.abstract N2 - The importance of accurate attenuation correction, especially for imaging of the thorax region, is widely acknowledged. Appropriate compensation methods have been developed and introduced into clinical practice. Most of these methods use attenuation maps obtained using various transmission scanning systems. However, when maps are inaccurate, the correction procedure may introduce artifacts into the final images that can be difficult to identify and might inadvertently alter diagnosis and study outcome. As a result, attenuation correction is often avoided in clinical practice. Our objective was to examine issues related to the quality of attenuation maps and the effects that map artifacts may have on attenuation-corrected emission images. Methods: The topics that are investigated include the problem of low transmission counts, cross-talk contributions from the emission isotope, truncation of the transmission data, and methods of map reconstruction and segmentation. Examples of patient studies displaying specific problems guided our investigations, but, because truth in these studies is seldom known, analytic and Monte Carlo-simulated data were used in the analysis. Attenuation maps and final emission images were visually checked for artifacts and for the presence of perfusion defects. In addition, quantitative evaluation of map uniformity, defect visibility, and size variation was performed. Results: The statistical paired-sample t test showed significant (P < 0.05) improvement of relative SD for attenuation maps reconstructed with iterative methods as compared with filtered backprojection and for maps created with higher photon fluxes. When maps with artifacts were used to correct emission data, an increase in myocardial infarct size and creation of false heart defects were observed. Conclusion: Our study strongly recommends that at least a visual inspection of the quality of attenuation maps be performed before their use in compensation procedures. To improve image quality, remove artifacts, and increase diagnostic confidence, attenuation maps used in the correction procedure must be accurate and free of artifacts. ER -