To detect coronary artery disease (CAD) noninvasively and to predict the occurrence of future cardiac events, 671 patients were evaluated using dipyridamole perfusion scintigraphy. 1. Although chest pain and ST depression were induced by the administration of dipyridamole in 34% and 22% of the patients, respectively, and additional intravenous aminophylline was needed in 19% of the patients, dipyridamole perfusion scintigraphy could be completed in nearly all patients. In contrast, treadmill exercise test was not accomplished in 24% of the patients. 2. The patients were classified in three groups by scintigraphic perfusion defects; i.e., group I (322 patients) with fixed defects, group II (107 patients) with reversible defects, and group III (242 patients) without perfusion defects. The patients in Group I were subclassified three groups according to three high risk parameters (extensive fixed defect, partial redistribution and diffuse slow washout)--group Ia (69 patients) with two or more high risk parameters, group Ib (144 patients) with one high risk parameter and group Ic (109 patients) without high risk parameters. 3. Coronary angiography performed in 377 patients revealed significant CAD (luminal narrowing greater than or equal to 50%) in 96%, 89%, 56%, 90% and 8% of the patients in groups Ia, Ib, Ic, II and III, respectively. Multi-vessel CAD was present in 87%, 32%, 11%, 51% and 2% of the patients in each group, respectively.