Relationship between CT coronary angiography and stress perfusion imaging in patients with suspected ischemic heart disease assessed by integrated PET-CT imaging

Abstract

Although computed tomography (CT) coronary angiography (CTA) provides detailed assessments of the anatomic extent of coronary artery disease (CAD), its value for predicting myocardial ischemia is unclear. We examined the value of CTA to identify the presence of ischemia, as determined by stress perfusion imaging, using integrated positron emission tomography (PET)-CT imaging. We studied 110 consecutive patients (median age, 57 years; 55% male) with suspected CAD undergoing stress rubidium 82 myocardial perfusion PET imaging and CTA in the same setting. Increasing degrees of CTA-detected luminal narrowing (<50%, 50%-70%, and >70%) were associated with reduced sensitivity with commensurate improvements in specificity for identifying myocardial ischemia both on a per-vessel basis and on a per-patient basis. Consequently, with increasing degrees of CTA-detected stenosis severity, the positive predictive value increased (14%, 26%, and 53%, respectively, on a per-vessel basis [P < .001] and 29%, 44%, and 77%, respectively, on a per-patient basis [P = .005]), whereas the negative predictive value was unchanged (97%, 97%, and 96%, respectively, on a per-vessel basis [P = not significant (NS)] and 92%, 91%, and 88%, respectively, on a per-patient basis [P = NS]). Receiver operating characteristic analysis revealed no differences between these 3 anatomic criteria (receiver operating characteristic areas of 0.66 +/- 0.07, 0.73 +/- 0.06, and 0.71 +/- 0.07, respectively [P = NS]) for identifying ischemia. Nearly half of significant angiographic stenoses (47%) occurred without evidence of myocardial ischemia, whereas 50% of normal PET studies were associated with some CTA abnormality. Despite an excellent negative predictive value, CTA is a poor discriminator of patients with myocardial ischemia. Conversely, a normal stress PET study is a poor discriminator of patients without evidence of non-flow-limiting (subclinical) coronary atherosclerosis. These results suggest potentially complementary roles of CT and perfusion imaging in the evaluation of patients with suspected CAD.