The objective of this study was to determine whether preoperative estimates of regional myocardial uptake of <sup>18</sup>-F-fluorodeoxyglucose (FDG) could predict postoperative improvement in ejection fraction in patients undergoing coronary artery bypass grafting (CABG) for ischemic cardiomyopathy. 20 consecutive patients [left ventricular ejection fraction (LVEF) ≤27%] were referred for preoperative positron emission tomography (PET) to measure regional myocardial FDG uptake and were deemed candidates for CABG. All individuals had technically adequate vessels and significant viability. The left anterior descending (LAD) coronary artery was severely diseased in all patients. Because of the importance of the anterior wall to overall ventricular performance, FDG uptake was semiquantitated in those regions and related to postoperative LVEF. 17 patients underwent CABG with an uneventful recovery. LVEF increased from 22 ± 4 to 26 ± 7% 2 weeks postsurgery (p < 0.05), with ≥5% change noted in 7 patients (group 1) and <5% noted in 10 patients (group 2). No preoperative or perioperative clinical variable could predict those with improved ventricular function. The relative amount of FDG uptake in the anterior wall was higher in group 1 compared with group 2 (93 ± 9 vs. 81 ± 13%; p < 0.05) and correlated with the change in LVEF post-CABG (r = 0.50; p < 0.05). >88% of FDG uptake in the LAD region had a positive predictive accuracy of 67% and negative predictive accuracy of 88% for improved LVEF postbypass. Late follow-up estimates of LVEF (median of 10 months) showed that early changes in function were sustained. In summary, among patients with severe coronary artery disease and depressed LVEF, ventricular function may improve early postrevascularization. PET estimates of relative FDG uptake in the anterior wall help predict those individuals who are likely to have the greatest increment in LVEF.