The purpose of this study was to compare the accuracy of post-reperfusion cardiac magnetic resonance (CMR) and pre-reperfusion multidetector computed tomography (MDCT) imaging to measure the size of the area at risk (AAR), using pathology as a reference technique in a porcine acute myocardial infarction model. Fifteen pigs underwent balloon-induced coronary artery occlusion for 40 min followed by reperfusion. The AAR was assessed with arterial enhanced MDCT performed during occlusion, while two different T2 weighted (T2W) CMR imaging sequences and the contrast-enhanced (ce-) CMR endocardial surface length (ESL) were performed after 90 min of reperfusion. Animals were euthanized and the AAR was assessed by pathology. Additional measurements of the myocardial water content in the AAR, remote and the AAR border zones were performed. AAR by pathology best correlated with measurements made by MDCT (R ² = 0.88; p < 0.001) with little bias on Bland–Altman plots (bias 2.5%, SD 6.1% LV area). AAR measurements obtained by T2W STIR, T2W ACUTE sequences or the ESL on ce-CMR showed a fair correlation with pathology (R ² = 0.72, R ² = 0.65 and R ² = 0.69, respectively; all p ≤ 0.001), but significantly overestimated the size of the AAR with important bias (17.4 ± 10.8% LV area; 11.7 ± 11.0% LV area; 13.0 ± 10.3% LV area, respectively). The myocardial water content in the AAR border zones was significantly higher than the remote (82.8 vs. 78.8%; p < 0.001). Our data suggest that post-reperfusion imaging methods overestimated the AAR likely due to the presence of edema outside of the boundaries of the AAR. Pre-reperfusion arterial enhanced MDCT showed the greatest accuracy for the assessment of the AAR.