OBJECTIVE The purpose of the present study was to compare single-photon emission computed tomographic (SPECT) myocardial images of technetium-99m (Tc-99m) sestamibi and thallium-201 (Tl-201) isotopes in the same dog undergoing partial coronary occlusion during pharmacologic vasodilation. BACKGROUND To date, no controlled study has been reported comparing SPECT Tc-99m sestamibi with SPECT Tl-201 imaging during stress with anatomic and physiologic standards. METHODS Mongrel dogs were anesthetized with chloralose and instrumented to record left anterior descending coronary blood flow and aortic pressure. Partial coronary occlusion with a hydraulic cuff reduced coronary vascular conductance, which is equal to the coronary blood flow normalized to aortic pressure during peak vasodilation with intravenous adenosine. Each dog received 5 mCi of Tl-201, then 30 mCi of Tc-99m sestamibi during partial coronary occlusion at peak vasodilation. Tomographic myocardial imaging was performed in a 180 degrees anterior arc scan for 33.5 min, first with Tl-201, and later, without moving the dog, for 33.5 min with Tc-99m sestamibi. Postmortem staining defined the region underperfused because of its dependence on the artery that was partially occluded. RESULTS In seven dogs with moderate reduction in coronary blood flow, coronary vascular conductance decreased with partial coronary occlusion (47 +/- 12%) during Tl-201 imaging and (47 +/- 8%, p = NS) during Tc-99m sestamibi imaging. The underperfused region was 23.9 +/- 6.4% of total left ventricular mass. Counts in the defects were 39% higher (0.86 +/- 0.08 of normal counts) for Tc-99m sestamibi than for Tl-201 (0.64 +/- 0.09 of normal counts, p < 0.001), and the defect on SPECT Tc-99m sestamibi images occupied only a fraction (0.37 +/- 0.30) of the area of the defect on the Tl-201 images of the same dog. Bull's-eye displays constructed from the pathologic slices showed that the Tl-201 defect size was closer to the underperfused region of the left ventricular mass determined pathologically than was the Tc-99m sestamibi defect size. In four additional dogs a severe, near total coronary occlusion was created during Tl-201 and Tc-99m sestamibi administration. In these dogs, similar defect contrast (0.55 +/- 0.12 vs. 0.62 +/- 0.09, p = NS) and areas (0.18 +/- 0.07 vs. 0.18 +/- 0.11, p = NS) were observed with Tl-201 and Tc-99m sestamibi, respectively. CONCLUSIONS Tomographic myocardial imaging with Tc-99m sestamibi during moderately severe partial coronary occlusion underestimated the area of the defect relative to Tl-201 or to the pathologic reference standard in dogs. Defect contrast was sharper with tomographic myocardial Tl-201 than with tomographic myocardial Tc-99m sestamibi during moderately severe partial coronary occlusion.