The partition coefficient of Gd-DTPA was thought to vary with the amount of cellular membrane damage after an acute myocardial infarction. The relationship between the partition coefficient of Gd-DTPA (lambda) and the uptake of 201Tl (as a marker of tissue viability) was studied 2 h to 3 weeks after reperfusion of a 2-h occlusion to the left anterior descending coronary artery in a canine model. Gd-DTPA was infused as a bolus followed by a prolonged constant infusion, and this infusion protocol was optimized such that the concentration of Gd-DTPA was directly related to lambda. After this infusion, MR images of excised hearts showed regions of increased signal intensity corresponding to increased Gd-DTPA concentration. At all time points, lambda and 201Tl uptake were strongly negatively correlated indicating that lambda is an accurate indicator of myocardial viability. Furthermore, lambda in the infarcted regions was increased relative to normal regions after 2 h of reperfusion and stayed elevated up to 3 weeks. At all time points, lambda in the infarcted and normal regions were significantly different. As well, this data showed a trend that lambda in infarcted regions decreased monotonically from 1 day to 3 weeks. This trend was confirmed with MR imaging by examining the change in signal intensity of in vivo images from 4 days to 3 weeks in two animals. These results suggest that MRI with Gd-DTPA could be used to measure the extent of myocardial damage after an acute myocardial infarction.