Repeatability of fully automated, inline quantitative assessment of myocardial perfusion in patients with suspected coronary artery disease.

Recent developments in the field of myocardial perfusion assessment with cardiovascular magnetic resonance (CMR) enable the automated inline quantification of myocardial blood flow (MBF). Previous studies have assessed its repeatability in healthy volunteers. This study assessed the repeatability of this technique in patients with suspected stable coronary artery disease (CAD). Patients with suspected CAD were studied twice on separate days. CMR perfusion imaging was performed at rest and during adenosine stress using a dual-sequence T1-weighted saturation recovery gradient echo sequence. Inline automatic reconstruction and image post-processing were implemented within the Gadgetron software framework, calculating MBF using a blood tissue exchange model. Repeatability of global stress and rest MBF, and myocardial perfusion reserve (MPR) were evaluated using Bland-Altman plots and intraclass correlation coefficients. Fifty-four patients (mean age 67 ± 9 years, 78% male) were studied. The median interval between the two scans was 2 days (IQR 3). There was no significant interstudy difference in global stress MBF (1.46 ± 0.51 mL/min/g vs. 1.51 ± 0.59mLmin/g, P = 0.44), global rest MBF (0.54 ± 0.14 mL/min/g vs. 0.56 ± 0.16 mL/min/g, P = 0.48), or global MPR (2.72 ± 0.80 vs. 2.84 ± 1.13, P = 0.76) between the two scans. Stress MBF, rest MBF, and MPR showed intraclass correlations of 0.60 (95% CI 0.39-0.75), 0.63 (95% CI 0.36-0.77), and 0.39 (95% CI 0.09-0.62), respectively. In patients with suspected CAD, quantitative assessment of myocardial perfusion by fully automated inline myocardial mapping shows moderate repeatability for stress and rest MBF, but poorer repeatability with MPR.