BACKGROUND Relative myocardial perfusion imaging may underestimate severity of coronary disease (CAD), particularly in cases of balanced ischemia. Can quantification of peak left (LV) and right (RV) ventricular Rb-82 uptake measurements identify patients with left main or 3 vessel disease? METHODS Patients (N = 169) who underwent Rb-82 PET MPI and coronary angiography were categorized as having no significant coronary stenosis (n = 60), 1 or 2 vessel disease (n = 81), or left main disease/3 vessel disease (n = 28), based on angiography. Maximal LV and RV ventricular myocardial Rb-82 uptake was measured during stress and rest. RESULTS Failure to augment LV uptake by >or= 8500 Bq/cc at stress, predicted left main or 3 vessel disease with a sensitivity of 93% and specificity of 61% (area under curve = 0.83). A >or=10% increase in RV: LV uptake ratios with stress over rest was 93% specific (area under curve = 0.74) for left main or 3 vessel disease. These indices incrementally predicted left main or 3 vessel disease compared to models including age, gender, cardiac risk factors, and summed stress and difference scores. CONCLUSION Quantifying maximal rest and stress LV and RV uptake with PET myocardial perfusion imaging may independently and incrementally identify patients with left main or 3 vessel disease.