We evaluated the safety, accuracy, and potential clinical utility of intravenous dipyridamole perfusion scintigraphy with thallium-201 in 170 patients, 78 with suspected and 92 with known unstable angina. All had coronary angiography. Noncardiac side effects (26%), induced chest discomfort (44%), and ST segment changes (12%) were similar in the two groups. No significant arrhythmias occurred. Two patients had prolonged chest pain, both with extensive reversible image abnormalities and associated creatinine kinase-MB release. Both had elective bypass surgery. Twenty-eight patients had normal coronary arteries, and 35 had single-vessel disease. Scintigraphic per patient sensitivity and specificity were 91% and 79% with a per vessel sensitivity of 74% and a per vessel specificity of 78% without between-group differences. During a brief follow-up period, 62 patients with image abnormalities had coronary revascularization, and there were seven deaths without intergroup differences. In a similar patient group that did not have angiography, scintigraphic defects were less frequent and less extensive, revascularization was not performed, and subsequent deaths occurred less often. Dipyridamole perfusion scintigraphy is an accurate alternative to exercise testing in the evaluation of patients with unstable angina pectoris. Although not without risk, the method appears relatively safe and should be considered as a guide to diagnosis, and probably to prognosis and management.