Differentiating cardiomyopathy of coronary artery disease from nonischemic dilated cardiomyopathy utilizing positron emission tomography.
 To determine if imaging of blood flow (using N-13 ammonia) and glucose metabolism (using F-18 2-deoxyglucose) with positron emission tomography can distinguish cardiomyopathy of coronary artery disease from nonischemic dilated cardiomyopathy, 21 patients with severe left ventricular dysfunction who were evaluated for cardiac transplantation were studied.
 The origin of left ventricular dysfunction had been previously determined by coronary angiography to be ischemic (11 patients) or nonischemic (10 patients).
 Images were visually analyzed by three observers on a graded scale in seven left ventricular segments and revealed fewer defects in dilated cardiomyopathy compared with ischemic cardiomyopathy for N-13 ammonia (2.7 +/- 1.6 versus 5 +/- 0.6; p less than 0.03) and F-18 deoxyglucose (2.8 +/- 2.1 versus 4.6 +/- 1.1; p less than 0.03).
 An index incorporating extent and severity of defects revealed more homogeneity with fewer and less severe defects in subjects with nonischemic than in those with ischemic cardiomyopathy as assessed by imaging of flow (2.8 +/- 1.8 versus 9.2 +/- 3; p less than 0.001) and metabolism (3.8 +/- 3.3 versus 8.5 +/- 3.6; p less than 0.005).
 Diagnostic accuracy for distinguishing the two subgroups by visual image analysis was 85%.
 Using previously published circumferential count profile criteria, patients with dilated cardiomyopathy had fewer ischemic segments (0.4 +/- 0.8 versus 2.5 +/- 2 per patient; p less than 0.01) and infarcted segments (0.1 +/- 0.3 versus 2.4 +/- 1.4 per patient; p less than 0.001) than did patients with cardiomyopathy of coronary artery disease.
 The sensitivity for differentiating the two clinical subgroups using circumferential profile analysis was 100% and the specificity 80%.
 An index incorporating both number and severity of defects derived from circumferential profile analysis was significantly lower in subjects with dilated cardiomyopathy than in ischemic cardiomyopathy (0.3 +/- 0.8 versus 2.7 +/- 2.4; p less than 0.005).
 Thus, noninvasive positron emission tomographic imaging with N-13 ammonia and F-18 deoxyglucose is helpful in distinguishing patients with severe left ventricular dysfunction secondary to coronary artery disease from those with nonischemic cardiomyopathy, and a semiquantitative index such as circumferential profile analysis is superior to that of visual analysis alone.
