Myocardial fractional flow reserve (FFR) has uncertain validity in patients with recent myocardial infarction and the use of FFR in this setting is controversial. We performed a prospective study to assess the diagnostic accuracy of FFR in patients with a recent non-ST segment myocardial infarction (NSTEMI).Methods
NSTEMI patients who had been referred for early invasive management were included. FFR was measured in all major patent epicardial coronary arteries with a visual stenosis estimated at ≥30% severity. Where clinically appropriate, an FFR assessment following PCI was also performed. Patients were scheduled for a stress perfusion 3T MRI following discharge from hospital.In a subsetof patients MRI was performed prior to coronary angiography/PCI. Baseline stress (I.V. adenosine 140 µg/kg/min) and rest perfusion MRI images were analysed side-by-side using dedicated software (Argus Dynamic Signal, Siemens, Erlangen, Germany). The stress and rest perfusion scans were viewed simultaneously, areas of hypoperfusion were assigned to coronary territories using the American Heart Association coronary arterial segment model. The analyses were performed independently by two observers who were blinded to the FFR results.Methods
In each patient, the coronary artery territories with abnormal perfusion were recorded. In cases of disagreement between observers, a third blinded observer adjudicated.Results
106 NSTEMI patients (mean age 56.7 ± 9.8 years, 82.6% male) were included. The mean time between the FFR evaluation and MRI was 5.8 ± 3.1 days. The mean ± SD left ventricular ejection fraction was 58.2 ± 9.1%. Mean infarct size was 5.4 ± 7.1% and mean troponin was 5.2 ± 9.2 g/L. A total of 1696 segments were available for analysis. 34 segments were excluded from the analysis due to problematic image quality so 1664 segments were finally included.Of these, 824 segments were available for comparison with FFR. 156 coronary arteries were assessed 92 in the infarct-related arteries and 64 in the non-infarct-related arteries. Of these, 28(17.1%) and 33(21.1%) arteries had an FFR ≤ 0.75 and ≤0.80, respectively. There was a negative correlation between the number of ischaemic segments and FFR (r = -0.79, 0 < 0.0001).The sensitivity, specificity, PPV and NPV for FFR ≤ 0.8 was 91.17%, 95.7%, 91.2% and 95.7% respectively.Results
ROC analysis defined the optimal FFR cut off value for identification of reversible ischaemia on MRI to be ≤0.8 (AUC 0.94 (0.89–0.99), p < 0.0001). An FFR ≤ 0.8 was associated with a sensitivity of 88.6% and a specificity of 94%.Conclusion
FFR measured in patients with recent NSTEMI has a high level of accuracy for inducible perfusion abnormalities revealed by 3T stress MRI, a non-invasive reference method.