In recent years, several of 99mTc-labeled myocardial perfusion imaging agents have been developed, such as 99mTc-sestamibi, 99mTc-tetrofosmin and 99mTc-furifosmin. Although images obtained with these new tracers have a general similar appearance, there are differences in the myocardial kinetics, body distribution, general quality of images and imaging protocols. The aim of this study was to quantitatively compare normal exercise planar and SPECT data files obtained with 201TI and 99mTc-labeled agents. METHODS Lower-limit-of-normal curves were generated for each specific radiopharmaceutical from normal subjects with low (< 3%) pretest likelihood of coronary artery disease using circumferential count distribution profiles from planar and SPECT exercise images. Lower-limit-of-normal curves were statistically compared using the nonparametric Kruskall-Wallis and Wilcoxon tests. RESULTS Planar and SPECT lower-limit-of-normal curves generated for each radiopharmaceutical showed general similarities. Statistically significant differences among the lower-limit-of-normal curves were found in the planar left anterior oblique view and in the planar left lateral view (p < 0.05 for each). No statistically significant differences existed between lower-limit-of-normal curves of various radiopharmaceuticals on the planar anterior view and on SPECT imaging. CONCLUSIONS For quantitative analysis of planar images, radiopharmaceutical-specific normal data files are mandatory. Although SPECT normal data files of various radiopharmaceuticals are not statistically different, they are not identical. It appears, nevertheless, prudent to use radiopharmaceutical-specific normal data files for quantitative analysis of SPECT images.