To examine the value of rest and redistribution thallium-201 imaging in predicting improvement in left ventricular (LV) ejection fraction (EF) after coronary artery bypass grafting (CABG), 26 patients with coronary artery disease (CAD) and abnormal LV function were studied. Nineteen patients had pathologic Q waves preoperatively. Rest and redistribution thallium-201 images and radionuclide ventriculograms were obtained before and after CABG, and the thallium scintigrams were evaluated both quantitatively and qualitatively. The patients were divided according to the preoperative thallium scintigrams into 2 groups: Group I (16 patients) had either normal resting thallium-201 images or reversible resting perfusion defects, and Group II (10 patients) had fixed resting perfusion defects. The resting EF was less than 50% preoperatively in all patients. Fourteen patients (54%) showed improvement in EF postoperatively. Three patients (2 in Group I and 1 in Group II) showed new postoperative perfusion defects, and none of the 3 showed improvement in LV function. Of the remaining 14 patients in Group I, 12 (86%) showed improvement in LV function, compared with 2 of 9 patients in Group II (p less than 0.01). Improvement in LV function was observed in 8 of the 19 patients (42%) with abnormal Q waves. Nitroglycerin intervention radionuclide ventriculograms were obtained in 20 patients before CABG. Of the 6 patients who showed improvement in LV function with nitroglycerin, 4 also showed improvement postoperatively. Postoperative improvement in LV function was also observed in 6 of the 14 patients who did not improve with nitroglycerin. Thus, rest and redistribution thallium imaging is useful in identifying patients whose LV function will improve after CABG. Normal rest thallium-201 images or reversible resting defects correctly identified 12 of 14 patients (86%) who showed improvement in LV function postoperatively. Nitroglycerin-intervention ventriculography and abnormal Q waves were less useful in this differentiation.