We investigated the ability of coronary artery calcium (CAC) to predict a normal adenosine stress rubidium-82 (Rb-82) myocardial perfusion positron emission tomography (PET) in patients admitted to the chest pain unit. Eighty-four consecutive patients (33 men; mean age 62 +/- 14.8 years) with low to intermediate likelihood of coronary artery disease were included. A single noncontrast computed tomogram under shallow breathing was obtained for attenuation correction and to assess the presence of CAC. This was followed by a rest and adenosine stress dynamic Rb-82 emission PET. Computed tomography and PET images were interpreted independently. There was a high prevalence of risk factors (80% hypertension, 30% diabetes, 38% hypercholesterolemia, 13% smoking); prior coronary revascularization and myocardial infarction were present in 21% and 15% of the patients, respectively. The absence of CAC was associated with a normal adenosine stress Rb-82 myocardial perfusion PET in 34 of 34 patients, yielding a negative predictive value of 100%. The presence of CAC (50 of 84) was associated with a higher incidence of myocardial perfusion defects (13 of 50), yielding a positive predictive value of 26%. Sensitivity was 100% (13 of 13) and specificity was 48% (34 of 71). In conclusion, the absence of CAC is predictive of a normal adenosine stress Rb-82 myocardial perfusion PET in patients admitted to the chest pain unit. If these results are confirmed, myocardial perfusion imaging probably can be safely avoided in chest pain patients with negative CAC with low to intermediate pretest likelihood of disease. This approach may decrease overall radiation exposure and hospital time and prove to be cost effective.