Combined Fourier amplitude and phase imaging in patients with coronary artery disease.

Fourier amplitude and phase image analysis were evaluated in 66 patients. Thirty patients served as a control group and 36 had coronary artery disease (CAD). Each patient in our control group had a rest and exercise radionuclide cineangiogram (RNCA) study (60 total). The amplitude and phase images for controls were uniform. No statistically significant difference in the histogram distribution of amplitude or phase occurred between rest and exercise. Twenty-five patients with CAD had a prior myocardial infarction (MI). Fourier analysis, when compared to the left ventricular ejection fraction (LVEF) and visual cine wall motion analysis, improved the sensitivity of the rest RNCA study to detect CAD from 68% to 92%. Regional Fourier amplitude and phase image analysis demonstrated an 85% sensitivity in localizing regions of previous infarction. Thirteen of 25 patients with prior MI and an additional 11 patients with recent onset of angina but no previous infarction, had exercise RNCA studies. When compared to cardiac catheterization data, regional Fourier amplitude and phase image analysis demonstrated sensitivities of 87% and 80%, respectively, in correctly identifying clinically significant major coronary artery involvement in these two subsets of patients. Fourier amplitude and phase image analysis were also able to distinguish normals from patients with previous MI, and patients with CAD but no prior infarction. All comparisons were statistically significant. We conclude that the semiquantitative image analysis of Fourier amplitude and phase data increases the clinical utility of the RNCA study in patients with CAD.