Quantitation of size of relative myocardial perfusion defect by single-photon emission computed tomography.

The purpose of this study was to validate a semiautomatic method for quantitating the size of relative myocardial perfusion defects from single-photon emission computed tomographic (SPECT) images. We compared the size of the image defect in vivo, expressed as percent of involved left ventricle, as determined by this method with the anatomic size of the defect in vitro in 19 dogs. To test the method under optimal conditions, we first labeled the left ventricular myocardium in nine dogs by left atrial injection of 99mTc-labeled macroaggregated albumin particles after acute occlusion of one coronary artery. The "defect volume" was defined as the volume of the left ventricular myocardium for which counts fell 2 or more SDs below the distribution of counts in the myocardium supplied by a normal coronary artery in a series of animals. The relative in vivo defect volume by SPECT occupied 26.46 +/- 12.7% of the left ventricular volume (mean +/- SD), compared with a relative defect size in vitro of 33.3 +/- 13.7% (p = NS) of left ventricular volume as determined by well counting of myocardial samples. There was a close correlation between the two measurements (r = .92). However, myocardial relative defect volumes involving less than 5% of myocardium were not identified by SPECT. The defect volume weighted for the relative reduction in flow within the defect zone or the relative "reduced perfusion volume" was also determined. The correlation between the estimates by SPECT and those made in vitro for relative reduced perfusion volume was also high (r = .94).(ABSTRACT TRUNCATED AT 250 WORDS)