Abstract 2120: Myocardial Perfusion Reserve Determined by Model Based Quantitative Analysis of Stress-Rest Perfusion MRI Permits Improved Accuracy in Predicting Coronary Artery Diseases When Compared with Semiquantitative Analysis
PURPOSE: The purpose of this study was to determine if myocardial perfusion reserve (MPR) determined by model based, fully quantitative analysis of rest and stress myocardial blood flow (MBF) can provide higher diagnostic accuracy in detecting coronary artery disease (CAD) as compared with MPR index (MPRI) by semi-quantitative up-slope analysis.
METHOD AND MATERIALS: Twenty-nine patients (18 men, 72.1±7.7 years) underwent X-ray coronary angiography and stress-rest myocardial perfusion MRI within 4 weeks. Luminal diameter narrowing of >=50% was considered to be significant on coronary angiography. Myocardial perfusion MRI was obtained with a saturation recovery balanced TFE sequence, by injecting 0.05mmol/kg of Gd-DTPA during vasodilator stress and in the resting state. Dual bolus method was used to correct for blood saturation. Absolute MBF and MPR were quantified by using a model based Patlak analysis, with correction for flow-dependent alteration in extraction fraction of Gd-DTPA. Semi-quantitative MPRI was determined from up-slopes of blood and myocardial signal curves.
RESULTS: Significant CAD was detected in 15 (52%) of 29 patients on coronary angiography. In patients without CAD, semi-quantitative MPRI (1.2±0.3) was significantly lower than absolute MPR (2.8±0.8) (p<0.01). ROC analysis demonstrated that absolute MPR, compared with MPRI, had better diagnostic accuracy for detecting CAD (Az 0.88 vs 0.75, p=0.04). The optimal threshold in predicting CAD was 2.0 for absolute MPR and 1.1 for MPRI. On vessel based analysis, the sensitivity and specificity in predicting significant CAD were 75.8% and 75.9% for absolute MPR, and 69.7% and 68.5% for MPRI. The sensitivity and specificity in detecting patients having significant CAD were 86.7% and 78.6% for absolute MPR, 86.7% and 64.3% for MPRI.
CONCLUSION: The results in this study demonstrate that semi-quantitative up-slope analysis of first-pass MRI substantially underestimates absolute value of MPR. Model based quantification of MPR from rest-stress myocardial perfusion MRI can provide improved prediction of flow-limiting CAD with reduced manual interaction and post-processing time.