It is obvious that the indication and clinical applications of radionuclide stress testing have been expanded and that both techniques described in this article are useful for diagnostic and prognostic purposes. The sensitivity and specificity of noninvasive stress testing have been significantly enhanced by the introduction of these radionuclide approaches for detecting ischemia in patients with undiagnosed chest pain. As mentioned previously, these tests should be ordered with a sound understanding of the principle of Bayes' theorem and of probability analysis. High-risk patients with either stable CAD or recent myocardial infarction can be identified by the severity of the abnormal response elicited. Patients with multiple thallium defects, particularly of the redistribution type, appear to be at the highest risk for subsequent cardiac events. Similarly, patients with a greater than 10 per cent fall in ejection fraction with development of multiple wall motion abnormalities and an increase in end-systolic volume seem to be in a high risk subset. Further developments with single photon emission tomography and computer quantitation of thallium or ventriculographic images should make these tests even more reliable in obtaining useful information in patients with CAD.