To investigate the significance and mechanism of dilatation of the left ventricular cavity on dipyridamole thallium-201 imaging, we performed both dipyridamole thallium-201 imaging and dipyridamole radionuclide angiography on 83 patients with known angiograms. The dipyridamole/delayed ratio of the left ventricular dimension from the thallium-201 image was defined as the left ventricular dilatation ratio (LVDR). An LVDR greater than the mean + two standard deviations in patients without coronary artery disease was defined as abnormal. Twenty-two of 83 patients showed an abnormal LVDR, and 18 of the 22 patients (82%) had triple-vessel disease. By defect and washout analysis, the sensitivity and specificity for correctly identifying the patients as having triple-vessel disease was 72% and 76%, respectively, whereas LVDR had a sensitivity of 72% and a specificity of 93%. When LVDR was used in combination with the defect and washout criteria, sensitivity increased to 84% without a loss of specificity. In those 22 patients with abnormal LVDRs, end-diastolic volume measured by radionuclide angiography did not change after dipyridamole infusion. Dilatation of the left ventricular cavity on dipyridamole thallium-201 imaging reflected relative subendocardial hypoperfusion induced by dipyridamole rather than actual chamber enlargement. The LVDR was moderately sensitive and highly specific for triple-vessel disease and provided complementary information to dipyridamole thallium-201 imaging.