Evaluation of the accuracy of steady-state krypton-81m method for calculating right ventricular ejection fraction.

The accuracy of a steady-state 81mKr method for calculating the right ventricular ejection fraction (RVEF) has been examined in this study. Causes of errors using this method and their effects on the calculated RVEF were evaluated. The results suggested that mixing in the right ventricle during continuous infusion of 81mKr was homogenous, allowing for the calculation of ejection fraction using the count rate ratio. Lung activity was quite important and could not be neglected in computing RVEF, but the use of [99mTc]MAA lung perfusion scanning seemed to allow a correct subtraction of this background activity. The delineation of right ventricular regions of interest (ROIs) was complicated by the translation movements of the right ventricle during heart contraction. These ROIs should be drawn carefully on the count density distribution images and data shown by parametric images; such as first and second harmonic phase, amplitude images, and composite stroke volume image should be considered. Furthermore, this study demonstrates the superiority of the 81mKr technique compared with 99mTc methods for computing RVEF. In conclusion, even if the true accuracy of the 81mKr method for calculating RVEF cannot be proven due to the lack of reference methods, strong, suggestive evidence that the technique should be accurate is shown here.