15-O-water myocardial flow reserve PET and CT angiography by full hybrid PET/CT as a potential alternative to invasive angiography

Combined myocardial flow reserve (MFR) by PET and CT coronary angiography (CTA) is a promising tool for assessment of coronary artery disease. Prior analyses of MFR/CTA has been performed as side-by-side interpretation, not as volume rendered, full hybrid analysis, with fused MFR/CTA. We aimed to: (i) establish a method for full hybrid analysis of MFR/CTA, (ii) validate the inter- and intra-observer reproducibility of MFR values, and (iii) determine the diagnostic value of side-by-side versus full hybrid MFR/CTA with 15-O-water PET. Forty-four outpatients scheduled for invasive coronary angiography (ICA) were enrolled prospectively. All underwent rest/stress 15-O-water PET/CTA with ICA as reference. Within two observers of different experience, the Pearson r at global and territorial level exceeded 0.953 for rest, stress, and MFR values, as determined by Carimas software. Within and between observers, the mean differences between rest, stress, and MFR values were close to zero and the confidence intervals for 95% limits of agreement were narrow. The diagnostic performance of full hybrid PET/CTA did not outperform the side-by-side approach, but performed better than MFR without CTA at vessel level: specificity 93% (95% confidence limits: 89-97%) versus 76% (64-88%), p = 0.0004; positive predictive value 71% (55-86%) versus 51% (37-65%), p = 0.0001; accuracy 90% (84-95%) versus 77% (69-84%), p = 0.0009. MFR showed high reproducibility within and between observers of different experience. The full hybrid model was not superior to side-by-side interpretation of MFR/CTA, but proved better than MFR alone at vessel level with regard to specificity, positive predictive value, and accuracy.