Influence of arrhythmias on gated SPECT myocardial perfusion and function quantification

Despite the importance of R-wave gating myocardial perfusion tomography for ventricular function assessment, neither prevalence of gating errors nor their influence on quantified cardiac parameters has been studied.

Arrhythmia-induced anomalies in curves of counts versus projection angle for each R-wave segment were detected visually and algorithmically. Arrhythmia prevalence was tabulated for 379 patients (group 1) with prospective coronary artery disease (mean age 63+/-13 y, 47% male). Myocardial counts were analyzed from all reconstructed cinematic midventricular slices to assess arrhythmia effects on percentage of systolic count increase, generally assumed to equal percentage of wall thickening. In a separate retrospective analysis of 41 patients (group 2), with coronary artery disease (mean age 64+/-12 y, 68% male) having no significant arrhythmias, 36 of whom also underwent equilibrium radionuclide angiography, original projection data were altered to simulate arrhythmia-induced aberrant count patterns to evaluate effects on ventricular function and perfusion measurements.

Group 1 patients consisted of 26% without gating errors, 32% with count losses only in the last R-wave interval due to inconsistent transient increase of heart rate, 24% with count decreases in several late intervals due to consistently variable rates, 8% with early interval count increases paired with late interval count decreases due to ectopic beats and 9% with erratic count changes due to atrial fibrillation. Observed count patterns were strongly associated (P < 10(-3)) with arrhythmias detected by electrocardiogram monitoring. In group 2 simulations, ventricular volumes changed by only 2%+/-9% and ejection fraction (EF) by only 1%+/-4% from control values and correlated linearly (r> or = 0.96) with control values for all simulated arrhythmias. SPECT and equilibrium radionuclide angiography EFs correlated similarly (r = 0.85-0.89) for control and all simulations. Percentage changes from control in perfusion defect extent and severity were larger than processing reproducibility limits, the largest change being for atrial fibrillation. Control wall thickening was 38%+/-17%, significantly lower (P < 10(-6)) than for simulated arrhythmias, reflecting similar observations for group 1 patients.

Even though ventricular volumes and EFs were affected minimally by arrhythmias, both perfusion analysis and wall thickening were compromised. Consequently, quality assurance of gating may be critically important for obtaining accurate quantified parameters.