Quantitation of Myocardial Infarct Size From Thallium-201 Images: Validation of a New Approach in an Experimental Model

A new computer-based method has been developed to quantitate myocardial infarct size from the size of the regional thallium-201 deficit. The operator outlines the left ventricular myocardial activity with an ellipse. The program then plots the background-corrected activities of the highest mean value in a 3 pixel myocardial band perpendicular to and within the ellipse. The approach uses a new interpolative background correction. To determine the accuracy of this approach in assessing regional thallium deficit size, acute myocardial infarction was produced in six dogs by 24 hour occlusion of the proximal left anterior descending coronary artery. Infarct size was assessed from planar thallium images of the dog heart in three views, each with the chest opened and closed and with the heart excised and placed in a cradle. Before removal of the heart, triphenyltetrazolium chloride was infused to delineate normal from infarct tissue. Transverse slices of left ventricle were made and thallium images of the slices acquired. Infarct size delineated by triphenyltetrazolium chloride staining was expressed as a percent of the total left ventricular slice surface area (planimetric infarct size). Infarct size from whole heart and left ventricular slice thallium images was expressed as a percent of the total length of the left ventricular perimeter (perimetric infarct size). This was determined from points below a certain percent of normalized peak thallium activity in the computer-generated thallium activity curve. Six additional dogs had infarct size determined from whole heart thallium images 24 hours after left circumflex coronary artery occlusion. With left ventricular slices, perimetric infarct size derived from activity below 45% of the peak correlated most closely with the planimetric size. Perimetric infarct size from thallium images of the whole heart was computed using the same percent cut-off on the normalized thallium activity curves. Absolute infarct size from thallium whale heart images, expressed as a percent of total left ventricular weight, was calculated and compared with that estimated by triphenyltetrazolium chloride staining. For left anterior descending coronary artery occlusions, there were good correlations between infarct sizes estimated by the staining method and thallium images of whole heart with the chest opened (r = 0.87), with the chest closed (r = 0.89) and with the heart excised (r = 0.83). For left circumflex coronary artery occlusions, these correlations were also good (r = 0.94, 0.93 and 0.92, respectively). It is concluded that this new computer approach to quantitation of infarct size from thallium deficit size has been validated in an experimental model