Cardiovascular magnetic resonance stress T1 mapping for the detection of myocardial ischemia and significant coronary artery disease: validation against [15O]H2O PET and fractional flow reserve

Background Patients with prior coronary artery disease (CAD) and recurrent angina are guideline-recommended to undergo non-invasive functional imaging to evaluate for the presence of myocardial ischemia. [1] A novel non-contrast cardiovascular magnetic resonance (CMR) technique, stress T1 mapping reactivity (∆T1), has been proposed as a method for ischemia detection. ∆T1 is based on myocardial T1 relaxation times at rest and during vasodilator stress, and prior studies have suggested that ∆T1 can distinguish between normal, ischemic and infarcted myocardium. [2-4] However, validation studies assessing the diagnostic performance of ∆T1 against [15O]H2O PET and fractional flow reserve (FFR) in patients with prior CAD are lacking. Purpose The aims of this study are: (a) to assess the correlation between ∆T1 and [15O]H2O PET–defined myocardial blood flow (MBF); (b) to evaluate the diagnostic performance of ∆T1 for the detection of [15O]H2O PET-defined myocardial ischemia and FFR-defined significant CAD. Methods This PACIFIC-2 sub-study included patients with prior CAD (prior myocardial infarction and/or percutaneous coronary intervention) and recurrent angina, who underwent both rest and adenosine-induced-stress CMR T1 mapping (ShMOLLI) at midventricular level and [15O]H2O PET, preceding invasive coronary angiography with FFR measurements in all major coronary arteries. T1 mapping analysis was performed per vascular territory using cvi42. Pearson's correlation analysis was used to evaluate the association between ∆T1 and MBF per vascular territory. The diagnostic performance of ∆T1 on a per-patient basis was assessed by calculating the area under the receiver operating characteristic curve (AUC) using stress MBF and FFR as reference standards. Results Of the 189 patients enrolled in the PACIFIC-2 trial, 70 were excluded: due to missing T1 mapping (n= 61), T1 mapping artifacts (n= 7), and incomplete PET data (n= 2). Consequently, 119 patients (mean age 64 ± 10 years; 23 [19%] female) were included. A weak correlation was observed between rest T1 and rest MBF (r = 0.116; p = 0.031) and between ∆T1 and stress MBF (r = 0.109; p = 0.043), whereas no correlation was observed between stress T1 and stress MBF (r = 0.025; p = 0.638). ∆T1 demonstrated an AUC of 0.591 [95% confidence interval (CI): 0.489 – 0.693; p = 0.088] for the detection of PET-defined myocardial ischemia. For the detection of FFR-defined significant CAD, the AUC was 0.497 [95% CI: 0.381 – 0.613; p = 0.960] for ∆T1. Conclusion In patients with prior CAD the correlation between ∆T1 values and [15O]H2O PET-defined MBF is weak. Furthermore, ∆T1 demonstrated poor diagnostic performance for the detection of PET-defined myocardial ischemia and FFR-defined significant CAD in these patients. Consequently, CMR-derived ∆T1 does not appear to be a suitable diagnostic tool for patients with prior CAD and recurrent angina.