Application of Hilbert transform to radionuclide-gated cardiac studies: analysis of asynchronous emptying and filling in various heart diseases.

Fourier phase analysis has generally been used to investigate asynchronous emptying and filling in various heart diseases. A potential limitation of this form of analysis is curve fitting error, since a truncated Fourier series may not adequately describe the shape of a time-volume curve (TVC) and thus may produce errors in indices calculated from the fit. To overcome this problem, we developed a new method using Hilbert transform. Using Hilbert transform, the instantaneous phase (IP) curve was calculated directly from the TVC obtained from multigated cardiac blood pool images. Four parameters [time to maximum IP [T(max)], time to 0 in IP[T(0)], time to minimum IP[T(min)], and time from 0 to minimum IP [T(min-0)]] were extracted from the IP curves for each pixel, and functional images were constructed in 40 patients with ischemic heart disease (IHD), 16 with hypertrophic cardiomyopathy (HCM), 3 with dilated cardiomyopathy (DCM), and 7 normal controls (N). The standard deviations (SD) of these parameters were then calculated for the left ventricle. In IHD patients with a left ventricle ejection fraction (LVEF) of less than 50% and in DCM, the SDs of all parameters were significantly higher than in group N. In IHD patients with an LVEF of greater than 50%, the SDs of T(min), T(0), and T(min-0) were significantly higher than in group N, but there was no significant difference in the SD of T(max). In HCM patients, the SDs of T(min) and T(min-0) were significantly higher than in group N, suggesting the presence of asynchronous filling. In conclusion, this method appears to be promising for the quantitative analysis of asynchronous emptying and filling in various heart diseases.