Quantitative aspect of two-factor analysis applied to radionuclide gated cardiac studies

Abstract

The quantitative aspect of two factor analysis (TFA) applied to radionuclide gated cardiac studies was investigated using clinical data and a computer based simulation method. In the clinical investigation, the ejection fraction (EF) calculated from the extracted ventricular pattern (EVP) significantly correlated with the left ventricular EF(LVEF) determined from multiple gated cardiac blood pool scans (MUGA), while the regression line deviated considerably from the line of identity. The contribution of the EVP(CEVP) to the total information inversely correlated significantly with the LVEF. In the computer simulation, a model simulating the atrial, ventricular, and background regions was studied. In the case with no statistical noise and constant background, the EF calculated from the EVP was nearly equal to the true EF, and the CEVP decreased as the EF increased. The statistical noise increased the EF and decreased the CEVP, this tendency was most marked when the true EF was less than about 50%. When background varied largely with time, the quantitative aspect deteriorated considerably. In conclusion, the clinical investigation results suggest that TFA has some possible use in quantitative analysis, however, the stimulation results suggest that it should be used carefully because its quantitative aspect is affected by various factors.