Beat-by beat validation of ECG gating

Ejection fraction, normalized peak ejection and filling rates, and the time of occurrence of these events relative to the R-wave were determined in each of 512 consecutive individual cardiac cycles in each of 30 patients using an ultra-high-efficiency nonimaging detector system. For a given patient the 512 measurements of each quantity were averaged and compared with the value of this same quantity as determined from an R-wave-gated left-ventricular (LV) time-activity curve (TAC) derived from the same 512 cycles. We conclude (a) that a small but detectable systematic underestimate occurs in some LV function parameters when they are derived from gated LV TACs; (b) that the magnitude of this underestimate is smaller and less variable for systolic than for diastolic measurements; (c) that the magnitude of the underestimate is not greater than 20% in any single patient for diastolic parameters, nor greater than 8% in any individual patient for systolic parameters, and is substantially less for most patients; and (d) that a small subset of patients may require beat-length windowing if the gated values of diastolic parameters are to fall within these limits. Thus LV function measurements obtained from gated TACs adequately reflect the true average of such values during the measurement interval.