Abstract

The jeopardised ischaemic area-at-risk (AAR) is a key prognostic determinant in acute myocardial infarction. Myocardial oedema imaging with T2-weighted cardiac magnetic resonance CMR is validated for imaging the AAR and T2 ‘mapping’ is a new method for AAR imaging with clinical and research potential. We aimed to develop an automated post-processing method that would enable straightforward volumetric quantification of AAR with T2 maps. Our approach retains user input (i.e. clinical judgement) to confirm the presence of oedema on an image which is then subjected to an automated analysis. The new method was tested on 12 acute MI patients who had a CMR within 48 hours of hospital admission. Manual segmentation of the left ventricular wall and oedema were available for comparison. Left ventricular wall boundaries were delineated automatically by variational level set methods followed by automated detection of myocardial oedema by fitting a Gaussian-Gaussian mixture statistical model. The mean perpendicular distances between automatically detected left ventricular boundaries and corresponding manual delineated boundaries were 1.8±0.2mm for endocardial boundaries and 2.3±0.3mm for endocardial boundaries. Dice similarity coefficients for agreement (0=no agreement, 1=perfect agreement) between manual delineation and automated segmentation of the left ventricular wall boundaries and oedema regions were 0.85±0.02 and 0.74±0.05, respectively. Compared to standard manual approaches, the new semi-automated method for estimating myocardial oedema is straightforward and accurate.