Usefulness of electrocardiographicgated stress technetiumm-99m sestamibi tomography to differentiate ischemic from nonischemic cardiomyopathy single-photon emission computed

The noninvasive differentiation between ischemic and nonischemic cardiomyopathy is frequently difficult. We examined the clinical value of stress electrocardiographic gated (ECG-gated) single-photon emission computed tomography (SPECT) to identify ischemic cardiomyopathy and detect coronary artery disease (CAD) in 164 patients without known CAD, ejection fraction less than or equal to40% by ECG-gated SPECT, and subsequent coronary angiography. Summed stress, rest, and difference scores were measured from the SPECT studies, and regional wall motion variance was calculated from the ECG-gated images. Sensitivity and 95% confidence intervals for the diagnosis of ischemic cardiomyopathy and for detection of any CAD (>50% diameter stenosis) were estimated using previously defined cutoffs for summed stress score and regional wall motion variance. For the diagnosis of ischemic cardiomyopathy, sensitivity of stress SPECT (summed stress score >8) was 87% (95% confidence interval [CI] 78 to 95), with a specificity of 63% (95% Cl 60 to 82). The addition of wall motion information (summed stress score >8 or regional wall motion variance >0.114) increased sensitivity to 88% (95% Cl 80 to 96) and decreased specificity to 45% (95% Cl 35 to 55). If reversibility was also taken into account (summed stress score >8, regional wall motion variance >0.114, or summed difference score >0), sensitivity further increased to 94% (95% CI 88 to 100) and specificity decreased to 32% (95% Cl 23 to 41). For detection of any CAD, the combined approach using stress perfusion, reversibility, and region of wall motion had a sensitivity of 94% (95% CI 89 to 99) and a specificity of 45% (95% Cl 35 to 57). Therefore, ECG-gated SPECT is very sensitive for detection of ischemic cardiomyopathy and CAD among patients with moderate to severe systolic dysfunction. (C) 2004 by Excerpta Medica, Inc.