Comparison of second-line on-site computed quantitative flow ratio from coronary computed tomography angiography to PET perfusion imaging for detecting obstructive coronary artery disease

Abstract

Abstract Introduction In patients with suspected obstructive coronary artery disease (CAD) on coronary computed tomography (CTA), guidelines endorse second-line selective testing for hemodynamic evaluation of suspected CAD. A variety of non-invasive modalities are available, and myocardial perfusion imaging with Rubidium-82 positron emission tomography (PET) is an established method with high diagnostic performance. Recently, an on-site method estimating computed tomography-derived quantitative flow ratio (CT-QFR) showed promising results for discriminating obstructive CAD. However, no study has compared the diagnostic performances of PET and CT-QFR. Purpose To assess a possible non-inferiority of CT-QFR compared to PET in patients with suspected obstructive CAD at CTA using invasive coronary angiography (ICA) with fractional flow reserve (FFR) as reference. Methods Patients (n=1732, 57% males, age 59±9.5) referred on a clinical indication with symptoms suggestive of obstructive CAD underwent routine CTA. Patients with ≥50% diameter stenosis (DS) on CTA were referred for PET and subsequent ICA with FFR. CT-QFR was analyzed post-hoc blinded to PET and ICA results. Abnormal CT-QFR was defined as CT-QFR ≤0.80 in any vessel with a diameter ≥1.5mm. An independent core-lab evaluated PET scans as abnormal/normal with optional analyst-dependent application of pre-specified criteria; summed stress score of ≥4 in ≥2 contiguous segments, vessel-specific myocardial blood flow (MBF) <2.00 ml/g/min, global myocardial blood flow reserve ≤1.8, and/or transient ischemic dilatation ratio >1.13. Obstructive CAD was defined as ICA with FFR ≤0.80 or high-grade stenosis (≥90% DS). Results In total, 445/1732 patients (25%) had suspected obstructive CAD on CTA of whom 400/445 patients (90%) underwent subsequent PET and ICA. CT-QFR was successfully analysed in 383/400 (96%) patients classifying 174/383 (45%) patients as having disease. In comparison, PET classified 130/383 (34%) patients as having disease. In total, obstructive CAD by ICA with FFR was identified in 162 (42%) patients. There was no significant difference in area under the receiver-operating characteristic curves for CT-QFR compared to the best performing PET metric (lowest vessel-specific MBF); 0.84 (95% CI 0.80–0.89) vs. 0.81 (0.77–0.85), p=0.19)) (Fig. 1). Overall diagnostic accuracy of CT-QFR versus PET was similar (78% (95% CI 74–82) vs. 77% (72–81), p=0.70. Sensitivities for CT-QFR and PET were 78% (71–84) and 63% (55–70), p<0.01, respectively, and specificities 78% (72–84) and 87% (82–91), p=0.01, respectively (Fig. 2). Three-vessel or left main disease on ICA was correctly identified in 30/31 patients by both CT-QFR and PET. Conclusion In patients with suspected obstructive CAD by CTA, second-line CT-QFR was non-inferior to PET for discriminating obstructive CAD by invasive FFR; Although diagnostic accuracy was similar, CT-QFR demonstrated higher sensitivity while PET showed higher specificity Funding Acknowledgement Type of funding sources: Public Institution(s). Main funding source(s): Aarhus University PhD fellowshipRegion Mid Health Research Foundation