Noninvasive stress testing is generally recommended to detect patients who are at increased risk for cardiac death and myocardial infarction. Such tests depend on the presence of a physiologically significant coronary stenosis to detect disease. The low prevalence of events in patients who are able to exercise, however, results in a poor positive predictive value. Also, recent data suggest that a significant number of morbid events result from rapid progression of disease in segments of the coronary artery that initially had only minimal obstruction. Further, from a therapeutic standpoint, only catheterization has been shown in randomized trials to predict which patients are candidates for bypass surgery. Thus, noninvasive testing as an intermediate step to select those patients who require invasive study remains an attractive but unproven hypothesis.