Performance of computed tomography-derived fractional flow reserve using reduced-order modelling and static computed tomography stress myocardial perfusion imaging for detection of haemodynamically significant coronary stenosis.

Abstract

To compare the diagnostic performance of a reduced-order computed tomography-derived fractional flow reserve (CT-FFR) technique derived from luminal deformation and static CT stress myocardial perfusion (CTP). Forty-six patients (84 vessels) with suspected coronary artery disease from a single institution planned for elective coronary angiography prospectively underwent research indicated invasive fractional flow reserve (FFR) and 320-detector CT coronary angiography (CTA) and static CTP. Analyses were performed in separate blinded core laboratories for CT-FFR and CTP. CT-FFR was derived using a reduced-order model with dedicated software on a standard desktop computer. CTP was assessed visually and quantitatively by transmural perfusion ratio (TPR). Invasive FFR was significant in 33% (28/84) of vessels. Overall per-vessel sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy for CT-FFR were 81%, 84%, 71%, 90%, and 83%, respectively, those of visual CTP were 54%, 92%, 79%, 77%, and 78%, respectively, and TPR were 64%, 48%, 42%, 70%, and 54%, respectively. Per-vessel receiver operator curve analysis demonstrated a significantly larger area under the curve (AUC) for CT-FFR (0.89) with that for visual CTP (0.72; P = 0.016), TPR (0.55; P < 0.0001), and CTA (0.76; P = 0.04). The addition of CT-FFR to CTA provided superior improvement in performance (AUC 0.93; P < 0.0001) compared with CTA alone, a combination of CTA with visual CTP (AUC 0.82; P = 0.007) and CTA with TPR (AUC 0.78; P = 0.0006). Based on this selected cohort of patients, a reduced-order CT-FFR technique is superior to visual and quantitatively assessed static CTP in detecting haemodynamically significant coronary stenosis as assessed by invasive FFR.