Temporal resolution of multiharmonic phase analysis of ECG-gated myocardial perfusion SPECT studies

Background. Multiharmonic phase analysis (MHPA) was developed to assess left-ventricular dyssynchrony from gated myocardial perfusion single-photon emission computed tomography (GSPECT) studies. This study was intended to determine the temporal resolution of MHPA. Methods. A reference normal GSPECT study with 128 frames/cycle was simulated using NCAT, a nonuniform rational B-splines-based cardiac torso phantom. It was shifted in the time domain to insert phase delays. Realistic GSPECT studies (8 or 16 frames/cycle) were then obtained by down-sampling the reference and shifted studies. All GSPECT projections were generated with attenuation, scatter, collimator blurring, and Poisson noise. Seventeen regional phases were calculated from the GSPECT reconstructions (filtered back-projection without compensation for physical factors), using linear interpolation for the reference study, and MHPA for the realistic studies. Comparing the regional phases between the realistic studies without and with shifts determined whether MHPA could identify certain phase delays. Results. When there were enough counts/pixel (>10 counts/pixel), MHPA with either 1, 2, or 3 harmonics could resolve a phase difference of 5.6 degrees, corresponding to 1/64 of the cardiac cycle. Conclusions. With clinically equivalent counts, the temporal resolution of MHPA is 1/64 of a cardiac cycle. Achieving this high temporal resolution from data with low temporal resolution demonstrates the benefit of replacing discrete points with continuous harmonic functions.