Critical physical parameters in cardiac scintillation camera images

Experimental studies at the New York Hospital-Cornell Medical Center using canine hearts as a model for myocardial infarction raised serious doubts about how much information could be reliably extracted from multiple two-dimensional scintillation camera projections. The principal radiotracer used was Tc-99m-glucoheptonate. A central issue was whether the doughnut sign is quantitatively correlated with infarct size. To resolve these questions, a computer model of the canine thorax was developed. The calculation includes the effects of thoracic geometry, infarct geometry, tracer distributions in four tissue types, inverse square law, tissue attenuation, tissue scatter, finite collimation, and statistical noise. Model calculations indicate that the doughnut sign is correlated with infarct size, but it is also strongly dependent on the relative angular orientation of the infarct and the detector. For any given infarct orientation, there is an optimal viewing angle