Precise identification of infarcted myocardial tissue is of importance in diagnostic and interventional cardiology. A three-dimensional, catheter-based endocardial electromechanical mapping technique was used to assess the ability of local endocardial impedance in delineating the exact location, size, and border of canine myocardial infarction. Electromechanical mapping of the left ventricle was performed in a control group (n = 10) and 4 wk after left anterior descending coronary artery ligation (n = 10). Impedance, bipolar electrogram amplitude, and endocardial local shortening (LS) were quantified. The infarcted area was compared with the corresponding regions in controls, revealing a significant reduction in impedance values [infarcted vs. controls: 168.8 ± 11.7 and 240.7 ± 22.3 , respectively (means ± SE), P < 0.05] bipolar electrogram amplitude (1.8 ± 0.2 mV, 4.4 ± 0.7 mV, P < 0.05), and LS (2.36 ± 1.6%, 11.9 ± 0.9%, P < 0.05). The accuracy of the impedance maps in delineating the location and extent of the infarcted region was demonstrated by the high correlation with the infarct area (Pearson's correlation coefficient = 0.942) and the accurate identification of the infarct borders in pathology. By accurately defining myocardial infarction and its borders, endocardial impedance mapping may become a clinically useful tool in differentiating healthy from necrotic myocardial tissue.