Aims: The longitudinal myocardial blood flow (MBF) gradient derived from positron emission tomography (PET) has been proposed as an emerging non-invasive index of haemodynamically significant coronary artery disease (CAD). This study aimed to investigate the diagnostic value of longitudinal MBF gradient for the presence of haemodynamically significant CAD.
Methods and results: A total of 204 patients (603 vessels) with suspected CAD underwent [15O]H2O PET followed by invasive coronary angiography with fractional flow reserve (FFR) of all major coronary arteries. Longitudinal base-to-apex MBF gradients were assessed by two methods, using MBF in apical and mid (Method 1) or in apical and basal (Method 2) myocardial segments to calculate the gradient. The hyperaemic longitudinal MBF gradient was only weakly correlated with FFR (Method 1: r = 0.12, P = 0.02; Method 2: r = 0.22, P < 0.001). The hyperaemic longitudinal MBF gradient (by both methods), had lower diagnostic value when compared with hyperaemic MBF for the presence of haemodynamically significant CAD, defined as an FFR ≤ 0.80. No significant correlations between longitudinal MBF gradients and FFR were noted in proximal lesions, whereas longitudinal MBF gradients and FFR were significantly correlated in non-proximal lesions (r = 0.57, P < 0.001).
Conclusion: PET measured longitudinal flow parameters had lower diagnostic value when compared with hyperaemic MBF for the presence of haemodynamically significant CAD. Since lesion location was found to affect the correlation of PET measured longitudinal flow parameters and FFR, presence of a longitudinal flow gradient may be partly caused by normalization to a relatively normal perfused areas.