Diagnostic Accuracy of Noninvasive 64-row Computed Tomographic Coronary Angiography (CCTA) Compared with Myocardial Perfusion Imaging (MPI): The PICTURE Study, A Prospective Multicenter Trial

Abstract

Although multiple studies have shown excellent accuracy statistics for noninvasive angiography by coronary computed tomographic angiography (CCTA), most studies comparing nuclear imaging to CCTA were performed on patients already referred for cardiac catheterization, introducing referral and selection bias. This prospective trial evaluated the diagnostic accuracy of 64-row CCTA to detect obstructive coronary stenosis compared to myocardial perfusion imaging (MPI), using quantitative coronary angiography (QCA) as a reference standard. Twelve sites prospectively enrolled 230 patients (49% male, 57.8 years) with chest pain. All patients underwent MPI and CCTA (Lightspeed VCT/Visipaque 320, GE Healthcare, Milwaukee, WI, USA) prior to invasive coronary angiography (ICA). All patients were evaluated, and those found to have either an abnormal MPI or CCTA were clinically referred for ICA. CCTAs were graded on a 15-segment American Heart Association model by three blinded readers for presence of obstructive stenosis (>50% or >70%); MPI was graded by two blinded readers using a 17-segment model for estimation of the % myocardium ischemic or with stress defects. ICAs were independently graded for % stenosis by QCA. The efficacies of MPI and CCTA were assessed including all vessel segments for per-patient and per-vessel analyses. The prevalence of stenosis ≥50% by ICA was 52.1% (25 of 48). The sensitivity of CCTA was significantly higher than nuclear imaging (92.0% vs 54.5%, P < 0.001), with similar specificity (87.0% vs 78.3%) when obstructive disease was defined as ≥50%. CCTA provided superior sensitivity (92.6% vs 59.3%, P < 0.001) and similar specificity (88.9% vs 81.5%) using QCA stenosis ≥70%. For ≥50% stenosis, the computed tomographic angiography odds ratio for ICA disease was 51.75 (95% CI = 8.50-314.94, P < 0.001). For summed stress score ≥5%, the odds ratio for ICA CAD was 12.73 (95% CI = 2.43-66.55, P < 0.001). Using receiver operating characteristic curve analysis, CCTA was better at classifying obstructive coronary artery disease when compared to MPI (area = 0.85 vs 0.71, P < 0.0001). This study represents one of the first prospective multicenter, controlled clinical trials comparing 64-row CCTA to MPI in the same patients, demonstrating superior diagnostic accuracy of CCTA over myocardial perfusion single photon emission computed tomography (MPS) to reliably detect >50% and >70% stenosis in stable chest pain patients.