Thallium-201 reinjection imaging and positron emission tomography provide concordant information regarding myocardial viability in many patients with coronary artery disease and left ventricular (LV) dysfunction. It is unclear whether this concordance applies to patients with severe, as well as those with moderate, LV dysfunction. We studied 44 patients with chronic coronary artery disease and LV dysfunction, subgrouped on the basis of severity of dysfunction: 23 patients had moderate and 21 had severe dysfunction (ejection fractions 34 +/- 6% and 19 +/- 6%). Patients underwent exercise thallium single-photon emission computed tomography (SPECT) with 3- to 4-hour redistribution and reinjection imaging, as well as positron emission tomography (PET) imaging with 18fluorodeoxyglucose and 15O-water. Data were analyzed quantitatively in aligned transaxial PET and SPECT tomograms. A myocardial region was considered nonviable by PET if 18fluorodeoxyglucose activity was <50% of that in a normal region, associated with proportional reduction in blood flow. Similarly, regions were considered nonviable by thallium if activity was <50% of activity in normal regions on redistribution and reinjection studies. Thallium SPECT and PET data were concordant regarding viability in 98% and 93% of myocardial regions, respectively, in patients with moderate and with severe LV dysfunction. Lower concordance was observed only when regions with severe irreversible thallium perfusion defects on redistribution images were considered in both groups: 86% and 78%, respectively (p <0.01). Thus, thallium SPECT with reinjection yields information regarding regional myocardial viability that is similar to that provided by PET in patients with severe as well as moderate LV dysfunction. However, there is discordance in >20% of regions manifesting severe irreversible thallium defects in patients with severely reduced LV function.