Effect of scintillation camera nonuniformity on ejection fraction measurements.

A rotating cardiac phantom with three possible ejection fraction (EF) values was used in conjunction with a scintillation camera employing energy correction and count skim arithmetic for uniformity correction. Studies were collected with and without any correction, with the energy window of the analyzer set properly, and with the camera properly tuned. The uniformity was then degraded in one experiment by off-setting the analyzer window both high and low with respect to the primary photopeak and in another experiment by de-tuning a selected photomultiplier tube. In both experiments studies were taken with no correction enabled, and then with each of the correction options enabled. The results of both experiments show that ejection fraction values could be in error when the differential uniformity using National Electrical Manufacturers Association (NEMA) protocols exceeds 10%. If either energy correction alone, or energy correction combined with count skim correction is used, the ejection fraction values return to more acceptable values. Asymmetric windows, improper setting of the energy window or a badly tuned photomultiplier will likely result in poor analog images before the effect on ejection fraction measurements becomes evident. Uniformity correction devices do not adversely affect the numerical results obtained from these phantom studies, but should, nevertheless, be used with caution.