Quantitative Perfusion Cardiac Magnetic Resonance for the assessment of microvascular disease in angina with non-obstructive coronary artery disease (ANOCA) (RF_TH_288)

Coronary microvascular dysfunction (MVD) is increasingly recognised as a leading cause of angina with non-obstructive coronary artery disease (ANOCA), currently requiring invasive coronary assessment to diagnose. We sought to assess the most accurate quantitative perfusion (QP) cardiac magnetic resonance (CMR) method to diagnose MVD.



Methods:

We recruited 28 patients with ANOCA who underwent invasive microvascular testing using the bolus thermodilution method. Microvascular dysfunction was defined as an index of microvascular resistance (IMR) >25 and/or a coronary flow reserve (CFR) < 2.5. All patients then underwent stress perfusion CMR on a 3T Siemens Magnetom Prisma including quantitative perfusion using a previously reported, automated inline dual sequence method, acquiring three short axis slices of the LV at every heartbeat. Any segments deemed to have late gadolinium enhancement were excluded from final myocardial blood flow (MBF) values.



Results:

13 (46%) patients were confirmed to have MVD on invasive testing. 10/13 (77%) of those with MVD were women, 9/15 (60%) of those with normal invasive parameters were men. There were no significant differences between the groups in conventional cardiovascular risk factors.

Using receiver operator characteristic curve analysis and taking global QP values, we found global myocardial perfusion reserve (MPR) of >3.55 to have a good specificity but a poor sensitivity (78.6% and 50.0%) to detect invasively defined MVD. Whilst global stress MBF of >1.84 had both a reasonable sensitivity and specificity (69.2% and 71.4%).

Between the three acquired slices, the mid LV slice had the best diagnostic performance to detect invasively defined MVD. Mid-slice endocardial MPR < 3.04 showed the highest sensitivity of any marker but a weak specificity (84.6% and 53.3%). Whilst a mid-slice stress endocardial/epicardial MBF ratio < 0.84 displayed the highest specificity of the mid-slice values obtained, with a weaker sensitivity (78.6% and 61.5%).


Combining these criteria, logistic regression analysis revealed 75% of cases correctly identified with an AUC 0.756 (p0.011) for diagnosis of MVD.


Conclusion: QP CMR is able to diagnose MVD with high accuracy, however global values may be less accurate than using only mid slice values. This may be due to more partial volume effects in basal and apical slices compared with the mid-LV. The combination of mid-slice endocardial MPR and mid-slice stress MBF endo-epi ratio displayed the strongest overall accuracy.