Author + information
- Konstantinos C. Siontis, MD,
- Bernard J. Gersh, MB, ChB, DPhil,
- Eric E. Williamson, MD,
- Thomas A. Foley, MD,
- J. Wells Askew, MD and
- Nandan S. Anavekar, MB, BCh∗ ()
- ↵∗Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905
Myocardial computed tomography perfusion (CTP) imaging is used as an alternative to established myocardial perfusion imaging modalities (1). Although coronary computed tomography angiography (CTA) has value in ruling out coronary artery disease (CAD), it is also characterized by high false positive rates and the inability to distinguish functionally significant from insignificant lesions (2). The addition of CTP to CTA may improve its diagnostic performance by providing functional information of coronary stenoses. Results of studies attempting to define the diagnostic characteristics of CTP with or without CTA have been variable. In this meta-analysis, we synthesized available evidence on the diagnostic performance of CTP with or without coronary CTA in reference to invasive coronary assessment.
Two investigators searched the MEDLINE, EMBASE, and CENTRAL databases using relevant key words for studies published before December 2014. References of eligible studies were perused for additional eligible studies. We included studies evaluating pharmacological stress CTP with or without CTA in reference to quantitative coronary angiography (QCA) or fractional flow reserve (FFR) for the diagnosis of CAD. We only included studies in which qualitative perfusion assessment was performed. We constructed 4 × 4 diagnostic performance tables adhering to QCA and FFR cutoffs adopted by individual studies for CAD definition. When results for different cutoffs were available, for consistency we used QCA stenosis >50% and FFR <0.80 thresholds. We calculated summary sensitivity, specificity, and areas under the receiver-operating characteristic curves (AUCROC) with 95% CI using bivariate random-effects meta-analysis on a per-vessel basis. Although traditional diagnostic meta-analysis unifies sensitivity and specificity into 1 measure (the summary ROC), bivariate meta-analysis maintains their distinct characteristics and takes into account their potentially negative correlation. Heterogeneity was quantified with the I2 statistic.
Twelve studies including 920 patients (median n = 39) and 1,563 coronary vessels (median n = 104) were eligible. Stenosis of 50% (n = 5 studies) or 70% (n = 2) per QCA and FFR <0.8 (n = 4) or <0.75 (n = 1) were the reference cutoffs. By QCA or FFR, a median 51.5% of patients (interquartile range: 35% to 71%) had significant CAD. The diagnostic performance of CTP alone and CTP/CTA was assessed in 8 and 9 studies, respectively. Per-vessel summary sensitivity, specificity, and AUCROC (95% CI) for CTP were 0.87 (0.74 to 0.94), 0.84 (0.74 to 0.91), and 0.92 (0.89 to 0.94), respectively. The respective summary values for CTP/CTA were 0.88 (0.78 to 0.94), 0.91 (0.88 to 0.93), and 0.91 (0.88 to 0.93) (Figure 1). No significant differences were observed in separate analyses for QCA and FFR as reference methods or when studies using QCA stenosis >50% as reference were excluded. Heterogeneity was significant (I2 > 50%) in the main and sensitivity meta-analyses of both CTP and CTP/CTA.
This bivariate meta-analysis indicated favorable diagnostic performance of CTP compared with “gold-standard” invasive methods for CAD assessment. Addition of CTA to CTP resulted in slightly improved specificity without significantly improved sensitivity or overall performance. We note the significant heterogeneity as a potential limitation of this analysis. Such degree of heterogeneity can be attributed to varying pre-test CAD probabilities among the included patient populations, variations in CTP imaging techniques, and slightly different diagnostic reference cutoffs. Despite its suggested diagnostic superiority compared with single-photon emission computed tomography (3), CTP has not yet gained widespread popularity in clinical practice. Although alternative well-established perfusion assessment methods remain available (single-photon emission computed tomography, positron emission tomography, stress echocardiography, and cardiac magnetic resonance), the possible superior diagnostic characteristics of CTP should be weighed against the cost, radiation exposure (4), and risk for contrast-induced nephropathy, even though the latter has been disputed and remains the subject of further investigation (5). In addition, the prognostic value of CTP-detected perfusion defects and the downstream effects on patient management remains to be determined. The ultimate clinical significance of the favorable diagnostic characteristics of CTP imaging will be established on the basis of patient outcome improvements in randomized trial comparisons in which different diagnostic modalities inform clinical decision making.
Please note: The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- American College of Cardiology Foundation
- Blankstein R.,
- Shturman L.D.,
- Rogers I.S.,
- et al.
- Budoff M.J.,
- Dowe D.,
- Jollis J.G.,
- et al.
- Einstein A.J.,
- Berman D.S.,
- Min J.K.,
- et al.