Multiple-hospital survey of ejection-fraction variability using a cardiac phantom.

A dynamic cardiac phantom was used to provide identical input data at 11 different nuclear medicine laboratories throughout the Philadelphia area, and the variability in the resulting calculations of ejection fraction (EF) was assessed. The variability observed between different operators using the same computer system averaged 3 EF units, which was similar to that between different observers using different types of computers. In the range of low ejection fractions, however, there was a suggestion that EFs calculated with an MDS computer were slightly lower than those from a DEC computer.