To determine the incidence of incomplete redistribution on conventional delayed thallium images, 41 patients with persistent perfusion defects on myocardial images recorded 3 to 4 hours after thallium injection during exercise were studied. At the conclusion of their delayed images the patients were reinjected at rest with approximately 1 mCi of thallium-201 and a third set of images was recorded. The images were presented at random in pairs (initial:delayed, initial:reinjection) to 2 experienced observers for qualitative scoring of 9 segments/patient. Of the 360 segments analyzed, concordance between the delayed and reinjected images occurred in 307 (85%). Of 141 segments that demonstrated a persistent perfusion abnormality on 3- to 4-hour delayed images, 44 (31%) were reassigned to a redistribution score after reinjection. In 9 patients, reinjection images provided the only evidence of ischemia from the scintigraphic data. In 13 of 14 vascular territories that demonstrated redistribution after reinjection, intact perfusion (either anterograde or via collaterals) was detected at coronary angiography. These data suggest that rest reinjection imaging may provide a means of detecting viable myocardium in segments that demonstrate a fixed perfusion abnormality on conventional 3- to 4-hour delayed thallium images.