Patients with coronary artery disease (CAD) and severe left ventricular (LV) dysfunction have a high but variable annual mortality and some may benefit from myocardial revascularization. This study aimed to evaluate the prognostic value of positron emission tomography (PET), and its interrelation with the choice of medical therapy or revascularization for predicting survival and improvement in symptoms of heart failure in patients with CAD and LV dysfunction. Ninety-three consecutive patients with angiographic CAD and a mean LV ejection fraction of 0.25 who underwent cardiac PET studies for assessment of hypoperfused yet viable myocardium ("mismatch pattern") using N-13 ammonia and 18-F deoxyglucose were followed up for an average of 13.6 months. Fifty patients underwent medical treatment and 43 underwent revascularization. The Cox model analysis showed that the extent of mismatch had a negative effect (p = 0.02), whereas revascularization had a positive effect on survival (p = 0.04). The annual survival probability of patients with mismatch receiving medical therapy was lower than of those without mismatch (50 vs 92%, p = 0.007). Patients with mismatch who underwent revascularization had a higher survival rate than those treated medically (88 vs 50%, P = 0.03). The presence of mismatch also predicted improvement in heart failure symptoms after revascularization (p < 0.001). These results suggest that the presence of mismatch in patients with CAD and severe LV dysfunction is associated with poor annual survival with medical therapy. Revascularization in patients with PET mismatch appears to be associated with improved survival and heart failure symptoms.