J Nucl Med. 2007; 48 (Supplement 2):42P
Instrumentation & Data Analysis: Image Generation
Motion Effects and Compensation in PET and SPECT |
Motion correction (MC) of multi-pinhole SPECT (MP-SPECT) myocardial perfusion imaging (MPI) on a breath-by-breath (BbB) basis improves accuracy in measurement of LV function
Dennis Kirch1,
John Koss1 and
Peter Steele1
1 Western Cardiology Associates, Westminster, Colorado
143
Objectives: The major component affecting cardiac movement during MPI is respiration, both in terms of BbB movement and long term changes in tidal volume (creep). The high statistical content of the data acquired by MP-SPECT technique supports MC, including BbB respiration. When properly applied MC quantitatively improves stress/rest comparability of MPI as well as the accuracy of segmental wall motion and wall thickening analysis. Methods: Acquisition of all MP-SPECT views simultaneously allows the respirational motion of the heart to be tracked by analyzing all of the views together using a correlative MC approach. A static homomorphic cardiac phantom (1.5 cm wall thickness) was imaged with background activity by MP-SPECT while undergoing linear motion of ±0.75 cm at 12 cyc/min. This data was evaluated using Circumferential Profile Analysis (CPA) to assess the uniformity in quantification of the perfusion uptake, cardiac volumetrics, and cardiac wall thickness measurements. Clinical validation in 12 patients was accomplished by tracking the phase and amplitude of the computed cardiac movement during prompted breathing. This was implimented by analyzing the ECG trigger data included in the list-mode data acquisition format in order to synchronize the correlative MC computation to include data intervals which covered integral numbers of heart beats. Results: The phantom results showed a ±10% variability in the CPA for sinusoidal motion of the phantom which was reduced to ±2% following MC correction. The error in absolute volume caused by this motion was +8.4 % and the maximum error in the FWHM wall thickness measured in the wall segments which were perpendicular to the cardiac motion was +31%. In patient studies done under controlled conditions using synchronized breathing, the systematic error in stress/rest CPA differences reduced from ±5% to ±2.2%. The average ejection fractions for these patients were 51% (w/o MC) and 57% (with MC), respectively. Conclusions: MC on a BbB basis is a prerequisite for obtaining accurate MPI diagnosis, measurements of wall motion, wall thickening and diastolic and systolic performance. Unsynchronized aplication of MC to MPI under some conditions can make the motion artifacts demonstrably worse.
