J Nucl Med. 2007; 48 (Supplement 2):99P
Instrumentation & Data Analysis: Image Generation
SPECT Reconstruction and Compensation |
Dual-peak attenuation compensation (DPAC) using the upper and lower Tl-201 peaks during simultaneous stress/rest multi-pinhole SPECT (MP-SPECT) for myocardial perfusion imaging (MPI)
Dennis Kirch1,
John Koss1 and
Peter Steele1
1 Western Cardiology Associates, Westminster, Colorado
337
Objectives: The ability to simultaneously image Tc-99m (rest) and Tl-201 (stress) is clearly advantageous in application to MPI. MP-SPECT images can be uitlized for this purpose because of their high statistical quality which is needed to perform scatter and crosstalk corrections. We demonstrate here that this statistical content also supports the use of both Tl peaks for performing DPAC of the stress images to achieve better Tl-stress/Tc-rest diagnostic comparability. This is in spite of the close proximity of the Tc peak (140 keV) to the upper Tl peak (168 keV). Methods: The methodology for weighted combination of the upper and lower (80 keV) peaks was validated using a homomorphic phantom loaded to simulate patient imaging conditions. This indicated that equal weighted additive combination of the data from both peaks worked best. Then a group of 26 pts with documented CAD, underwent simultaneous Tl-stress/Tc-rest dual-isotope MPI by MP-SPECT. The Tl-stress images where reconstructed using upper peak data alone, lower peak alone and the DPAC combination of both peaks. Circumferential Profile Analysis CPA provided precise Tl-stress/Tc-rest comparison of the data from these dual-isotope MP-SPECT studies processed with and without DPAC. Results: Use of equal weights to combine the image data from the upper and lower peaks of Tl was supported by the fact that reconstruction of the upper peak data alone produced images of near diagnostic quality. The superiority of the images produced by equally weighted combination of the two peaks was confirmed by CPA analysis of the homomorphic phantom data which showed the systematic error between the Tl-stress/Tc-rest curves to be reduced from ± 5% using the lower Tl peak alone to ± 3% for the combined DPAC reconstructions. Review of the visual and CPA Tl-stress/Tc-rest comparisons revealed improvement in the level of confidence for clinical interpretation in 8 out of the 26 pts. Conclusions: The improved statistical quality of MP-SPECT which supports DPAC leads to reduced systematic error as demonstrated by CPA results for both phantom and patient MPI. The reduction in systematic error confirms the fact that DPAC improves the level of confidence in the clinical interpretation of patient studies.
