Abstract 1917: Defining Appropriate Cardiac MRI Parameters to Detect Anatomic and Hemodynamic Significant Coronary Artery Disease
Background- Angiography and fractional flow reserve (FFR) using intracoronary pressure wires are well established to define anatomical and hemodynamically significant coronary artery disease (CAD), respectively. Methods to evaluate CAD using cardiac MRI perfusion imaging have not been established. Our objective was to define the most appropriate parameter using cardiac perfusion MRI to detect CAD.
Methods and Results- We prospectively enrolled 35 patients who underwent coronary angiography, FFR and cardiac MRI perfusion. Stress was induced using adenosine for both MRI and FFR tests. Maximum up-slope and peak-intensity derived from time-intensity curves. Indexes were calculated by dividing stress by rest. Quantitative myocardial perfusion reserve (MPR) was calculated using deconvolution technique. Qualitative assessment of MRI images was performed by 2 experts blind to quantitative data. Accuracy of quantitative, semi-quantitative, and qualitative data to detect > 50% diameter stenosis (DS) by QCA (n = 108 coronary segments) or FFR ≤ 0.75 (n = 43 coronary segments) was tested. Measurements were performed in a core laboratory. Sensitivity and specificity for detection of hemodynamically significant CAD (FFR ≤ 0.75) were 91.7 % and 68.0% for MPR (cutoff 2.06); 75.0 % and 56.0 % for up-slope index (cutoff 1.26) and 75.0 % and 60.0 % for peak-intensity index (cutoff 1.05). Area under the curve (AUC) was 0.84 for MPR (p < 0.01), 0.66 for up-slope (p = NS) and 0.69 (p = NS) for peak-intensity index to detect FFR ≤ 0.75. Sensitivity and specificity for detection of > 50%DS were: MPR (cutoff 2.06) = 84.6 % and 51.9%; up-slope index (cutoff 1.26) = 74.4 % and 63.5%; peak-intensity index (cutoff; 1.08) = 82.1 % and 61.5%, respectively. AUC was 0.75 for MPR, 0.73 for up-slope and 0.69 for peak-intensity indexes to detect > 50%DS (all p < 0.05). Visual assessment yielded a sensitivity of 78.6% and specificity of 65.5% to predict FFR ≤ 0.75, and sensitivity of 74.5% and a specificity of 67.2% to predict > 50%DS.
Conclusions- MPR appears to be the most accurate index to detect anatomical and hemodynamically significant CAD. Standardization of such quantitative methods, with minimal operator dependency, would be useful for clinical and research applications.