Comparison of fixed and variable temporal resolution methods for creating gated cardiac blood-pool image sequences.

In each of 50 resting subjects, two gated blood-pool image sequences were created from the same LIST mode data set. One sequence was created using a sorting method that spans each individual cardiac cycle with the same number of images (the "variable temporal" or VT method), while the other (the "fixed temporal" or FT method) spans the average cardiac cycle with images of fixed temporal duration. Left ventricular time-activity curves were extracted from each sequence using identical regions-of-interest and analyzed with identical methods to obtain estimates of ejection fraction, peak ejection rate, peak filling rate, and the times of occurrence of these peak rates. Differences among certain of these parameters in kind and amount support the hypothesis that estimates of resting cardiac function are more accurately portrayed by the FT method. The magnitudes of these differences are small for systolic parameters but large for early diastolic parameters. Thus, although both methods might be used for measuring systolic function, the FT method will yield a more accurate estimate of peak filling rate in resting subjects.