Diagnostic performance of noninvasive myocardial perfusion imaging using single-photon emission computed tomography, cardiac magnetic resonance, and positron emission tomography imaging for the detection of obstructive coronary artery disease: a meta-analysis

Abstract

Objectives: This study aimed to determine the diagnostic accuracy of the 3 most commonly used noninvasive myocardial perfusion imaging modalities, single-photon emission computed tomography (SPECT), cardiac magnetic resonance (CMR), and positron emission tomography (PET) perfusion imaging for the diagnosis of obstructive coronary artery disease (CAD). Additionally, the effect of test and study characteristics was explored.

Background: Accurate detection of obstructive CAD is important for effective therapy. Noninvasive myocardial perfusion imaging is increasingly being applied to gauge the severity of CAD.

Methods: Studies published between 1990 and 2010 identified by PubMed search and citation tracking were examined. A study was included if a perfusion imaging modality was used as a diagnostic test for the detection of obstructive CAD and coronary angiography as the reference standard (≥50% diameter stenosis).

Results: Of the 3,635 citations, 166 articles (n = 17,901) met the inclusion criteria: 114 SPECT, 37 CMR, and 15 PET articles. There were not enough publications on other perfusion techniques such as perfusion echocardiography and computed tomography to include these modalities into the study. The patient-based analysis per imaging modality demonstrated a pooled sensitivity of 88% (95% confidence interval [CI]: 88% to 89%), 89% (95% CI: 88% to 91%), and 84% (95% CI: 81% to 87%) for SPECT, CMR, and PET, respectively; with a pooled specificity of 61% (95% CI: 59% to 62%), 76% (95% CI: 73% to 78%), and 81% (95% CI: 74% to 87%). This resulted in a pooled diagnostic odds ratio (DOR) of 15.31 (95% CI: 12.66 to 18.52; I(2) 63.6%), 26.42 (95% CI: 17.69 to 39.47; I(2) 58.3%), and 36.47 (95% CI: 21.48 to 61.92; I(2) 0%). Most of the evaluated test and study characteristics did not affect the ranking of diagnostic performances.

Conclusions: SPECT, CMR, and PET all yielded a high sensitivity, while a broad range of specificity was observed. SPECT is widely available and most extensively validated; PET achieved the highest diagnostic performance; CMR may provide an alternative without ionizing radiation and a similar diagnostic accuracy as PET. We suggest that referring physicians consider these findings in the context of local expertise and infrastructure.