Conventional methods to quantify infarct size after myocardial infarction in mice are not ideal, requiring either tissue destruction for histology or relying on nondirect measurements such as wall motion. We therefore implemented a fast, high-resolution method to directly measure infarct size in vivo using three-dimensional (3D) late gadolinium enhancement MRI (3D-LGE). Myocardial T1 relaxation was quantified at 9.4 Tesla in five mice, and reproducibility was tested by repeat imaging after 5 days. In a separate set of healthy and infarcted mice (n = 8 of each), continuous T1 measurements were made following intravenous or intraperitoneal injection of a contrast agent (0.5 micromol/g gadolinium-diethylenetriamine pentaacetic acid). The time course of T1 contrast development between viable and nonviable myocardium was thereby determined, with optimal postinjection imaging windows and inversion times identified. Infarct sizes were quantified using 3D-LGE and compared with triphenyltetrazolium chloride histology on day 1 after infarction (n = 8). Baseline myocardial T1 was highly reproducible: the mean value was 952 +/- 41 ms. T1 contrast peaked earlier after intravenous injection than with intraperitoneal injection; however, contrast between viable and nonviable myocardium was comparable for both routes (P = 0.31), with adequate contrast remaining for at least 60 min postinjection. Excellent correlation was obtained between infarct sizes derived from 3D-LGE and histology (r = 0.91, P = 0.002), and Bland-Altman analysis indicated good agreement free from systematic bias. We have validated an improved 3D MRI method to noninvasively quantify infarct size in mice with unsurpassed spatial resolution and tissue contrast. This method is particularly suited to studies requiring early quantification of initial infarct size, for example, to measure damage before intervention with stem cells.