Purpose: To compare a novel method, percent-infarct-mapping (PIM), with conventional delayed enhancement (DE) of contrast for accurate myocardial viability assessment. Contrary to signal intensity (SI), the longitudinal relaxation-rate enhancement (ΔR1) is an intrinsic parameter linearly proportional to the concentration of contrast agent (CA). Determining ΔR1 voxel-by-voxel, after administering an infarct-avid CA, allows determination of pervoxel percentage of infarcted tissue. The feasibility of generating PIM is demonstrated in canine reperfused infarction using an infarct-avid, persistent-CA (PCA), (Gd)(ABE-DTTA). PIM is compared to the DE method using Gd(DTPA), and both to triphenyltetrazolium chloride (TTC) staining histochemistry. Materials and Methods: In six dogs, 48 hours following closed-chest, 180-minute coronary occlusion, DE imaging was carried out. PCA was administered immediately thereafter. Pixel-by-pixel R1 maps of the entire left ventricle (LV) were generated 48 hours after PCA using inversion-recovery with multiple inversion times. R1, ΔR1, and percent-infarct values were calculated voxel-by-voxel. Results: Significant correlations (P < 0.01, R ≥ 0.8) were obtained for percent-infarct-per-slice (PIS) determined by DE or PIM vs. those from corresponding TTC-stained slices. Compared to TTC, DE overestimated PIS by 32 ± 18%. PIM underestimated PIS by 2.5 ± 4.9%. Infarction fraction overestimation was 29 ± 6% of LV by DE, but only 1.0 ± 2.3% by PIM. Errors with PIM were significantly smaller than those with DE. Infarct area determined by signal intensity was similarly overestimated whether using Gd(DTPA) or Gd(ABE-DTTA). Conclusion: The ΔR1-based PIM method is highly reproducible and more accurate than DE for quantifying myocardial viability in vivo. © 2008 Wiley-Liss, Inc.