TY - JOUR T1 - Comparison of Myocardial Perfusion <sup>82</sup>Rb PET Performed with CT- and Transmission CT–Based Attenuation Correction JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1992 LP - 1998 DO - 10.2967/jnumed.108.056580 VL - 49 IS - 12 AU - Piotr J. Slomka AU - Ludovic Le Meunier AU - Sean W. Hayes AU - Wanda Acampa AU - Muneo Oba AU - Gillian G. Haemer AU - Daniel S. Berman AU - Guido Germano Y1 - 2008/12/01 UR - http://jnm.snmjournals.org/content/49/12/1992.abstract N2 - CT-based attenuation correction (AC) for myocardial perfusion PET studies is challenging because of respiratory motion. Our study aimed to compare the transmission CT (TCT)–based and CT-based AC for myocardial perfusion PET/CT images with a direct semiquantitative approach comparing differences in segmental count distribution. Methods: Stress and rest 82Rb PET scans from 54 consecutive patients acquired on a PET/CT scanner with dual CT-based and TCT-based AC were considered. TCT- and CT-based AC images were automatically registered to each other, and direct voxel-based and American Heart Association segment–based estimation of positive and negative changes between these scans was performed. Additionally, visual quality control (QC) of CT map alignment with PET emission data was performed by 2 expert observers, and studies with significant (≥5 mm) misalignment were reprocessed with corrected CT alignment. Results: We used the 17-segment American Heart Association model for TCT-to-CT regional change analysis in all patients and found that 4 segments on rest and 4 segments on stress scans differed more than 3% between CT- and TCT-corrected images for studies without significant misalignments (&lt;5 mm); only 1 differed by more than 5%. In cases with significant misalignment of greater than or equal to 3% TCT–CT AC, changes were observed on 14 rest and 10 stress segments; after alignment, these differences were still seen in 13 rest segments and 11 stress segments. Visual QC revealed that 46% of rest and 54% of stress PET scans were misaligned by greater than or equal to 5 mm with the CT maps acquired during normal breathing. The range of the reported PET/CT misalignment was 0–15 mm in x, 0–16 mm in y, and 0–20 mm in z directions. The overall agreement in visual QC of PET/CT alignment between the observers was 72.2% Conclusion: There are significant differences between TCT and CT AC applied to cardiac PET/CT studies, which remain after alignment of CT maps to emission data. ER -