Background. A prospective study was conducted to evaluate the clinical usefulness of three-dimensional (3D) surface-shaded maps for routine practice of myocardial perfusion single-photon emission computed tomography (SPECT) by comparison with 2D slices and 2D bull's-eye qualitative analysis. Methods and results. Angiograms were performed on 201 consecutive patients, 155 with coronary artery disease (CAD) and 46 with no significant CAD. One-day 201Tl stress/rest-reinjection protocol was performed in 110 patients, and 1-day 99mTc-sestamibi or tetrofosmin stress/rest protocol was performed in 91. The stress protocol was either exercise or dipyridamole (0.56 mg/kg) infusion. Three-dimensional surface maps were obtained by using a threshold for the transaxial data at 50%, 55%, 60%, 65%, and 70% of the maximum pixel value in the first 60 patients. Interpretation of 3D maps was based on the presence of a complete (transmural-looking) perfusion hole within the myocardial wall; doubtful patterns were considered pathologic or normal. Good diagnostic values were found for the 50% to 60% thresholds, but the 60% setting showed the best concordance with multislice and bull's-eye analysis; higher values drastically degraded the specificity. Considering doubtful patterns as normal clarified interpretation and led to a small loss in sensitivity but high gain in specificity. Applied to the whole population, the 3D maps using a 60% threshold provided similar diagnostic value to detect CAD as did conventional and bull's-eye analysis. Moreover, the 3D maps showed a trend toward higher specificity and a proportionally smaller decrease in sensitivity (sensitivity: 92.9%, 90.3%, 89.7%; specificity: 45.6%, 50.0%, 58.7% for tomograms, bull's-eye analysis, and 3D maps, respectively), especially for the detection of left anterior descending and right CAD. Multivessel disease was detected in an identical manner. Three-dimensional maps might improve detection of perfusion defects in the basal regions. However, 3D maps were found to be less sensitive than slices and particularly bull's-eye analysis for the reversibility of stress defects. Conclusions. Three-dimensional surface display of myocardial perfusion is a valuable independent tool for determining presence, extent, and location of CAD. It can convey useful first-look information to the referring physician, especially through a cine-rotational motion (as done in our practice through use of a floppy disk).