Fourier and factorial analysis: an objective and comparative evaluation on a cardiac phantom.

Orthogonal and oblique factor analysis represent an alternative to Fourier analysis in the evaluation of cardiac dynamic behaviour in gated blood pool studies. In order to estimate their respective places, orthogonal factor analysis (OFA), factor analysis of dynamic structures (FADS) and Fourier analysis (FA) are tested on a dynamic and periodical phantom with well known and reproducible kinetics. The phantom data are acquired under standard conditions by varying the counting rates and the temporal frequency sampling. To compare the results of the three methods with maximal objectivity, the relative contribution of each component is calculated. With standard acquisition conditions, FA and OFA give very close results. Only a minor advantage in evaluation of small phase differences is observed with OFA. FADS solutions are effectively related to the dynamic behaviour of the phantom, but their interpretation is more complicated and the quality of the oblique factors is reduced as the number of calculated factors increases. The influence of the counting statistics on FA, OFA and FADS is very similar. However, in cases of undersampling, robustness is demonstrated with the factorial technics.