CT-derived fractional flow reserve (FFR) improves the diagnostic accuracy of coronary CT angiography (CCTA) but typically requires off-site processing. We evaluated the agreement and diagnostic accuracy of a fast (∼8-min), on-site CT-FFR algorithm against invasive FFR and its correlation with myocardial blood flow. Patients with suspected chronic coronary syndrome and at least one ≥50 ​% diameter stenosis on CCTA underwent [15O]H2O PET and invasive coronary angiography with three-vessel invasive FFR. Thirty-eight patients with 89 vessels were included for analysis. Median invasive and CT-FFR were 0.92 (IQR 0.80-0.97) and 0.84 (IQR 0.68-0.91), respectively (rho ​= ​0.75). Against invasive FFR ≤0.80 on a per-patient level, CT-FFR ≤0.80 showed 84 ​% accuracy (95 ​% CI 70-93 ​%), 100 ​% sensitivity (95 ​% CI 85-100 ​%), 62 ​% specificity (95 ​% CI 39-82 ​%), a positive predictive value (PPV) of 79 ​% (95 ​% CI 60-90 ​%), negative predictive value (NPV) of 100 ​% (95 ​% CI 72-100 ​%). Compared with ≥50 ​% diameter stenosis on CCTA on a per-vessel level, CT-FFR ≤0.80 had similar sensitivity and NPV but higher accuracy, specificity, and PPV. Both CT-FFR and invasive FFR correlated weakly with stress myocardial blood flow (rho 0.30 and 0.35, respectively) and myocardial flow reserve (rho 0.20 and 0.38, respectively). A novel on-site CT-FFR algorithm demonstrates fast and very high rule-out capability and incremental rule-in potential beyond diameter stenosis for functionally significant disease determined by invasive FFR in patients with suspected stenosis at CCTA. Both measures of FFR correlated weakly with myocardial blood flow and flow reserve by [15O]H2O PET.

Coronary computed tomography angiography (CCTA) offers excellent negative predictive value for coronary artery disease (CAD). In suspected single-vessel disease, invasive coronary angiography (ICA) could be simplified by deferring visualization of the normal coronary artery as identified on CCTA. This ultraselective ICA strategy might reduce procedure time, contrast use, and radiation exposure. This prospective study evaluated the potential benefits of this strategy. This was a multicentre prospective registry in three hospitals in the Netherlands. The study included patients over 18 years undergoing ICA for suspected CAD limited to a single coronary artery on CCTA. The primary endpoint was accuracy of CCTA in predicting a normal coronary artery on ICA. Secondary endpoints included potential reductions in procedure time, radiation exposure, and contrast volume with an ultraselective ICA approach. In total 102 patients were included. CCTA was correct in predicting a non-obstructive coronary artery in all patients (100 ​%). By deferring visualization of the normal coronary, a CCTA-guided ICA would have led to a reduction in procedure time (49 ​% [IQR 41-59]; 6.64 ​min [5.07-8.68] vs 3.18 ​min [2.42-4.41]), radiation exposure (35 ​% [25-44]; 5.09 ​Gy ​cm2 [3.11-7.88] vs 3.17 ​Gy ​cm2 [1.96-5.08]), and the amount of contrast agent used (33 ​% [28-41]; 71 ​ml [60-84] vs 46 ​ml [37-55]). This study demonstrated excellent performance of CCTA in predicting a normal coronary artery in patients with suspected single-vessel CAD. A CCTA-guided ultraselective ICA is feasible and offers substantial reductions in procedure time and materials, contrast agent volume and radiation exposure.