Real-World Comparison of Cardiac MRI and Nuclear Perfusion Imaging for Myocardial Scar Detection

Background: Cardiovascular magnetic resonance late gadolinium enhancement (CMR-LGE) is the reference standard for scar detection, but data comparing LGE with fixed perfusion defects (FPD) on nuclear myocardial perfusion imaging (MPI) are limited. This study aimed to evaluate concordance between myocardial scar detection by CMR-LGE and nuclear MPI. Methods: In this retrospective single-center study, patients undergoing contrast CMR and nuclear MPI within 1 year were included: single-photon emission computed tomography (SPECT) with CT attenuation correction (SPECT AC , n = 328), SPECT without attenuation correction (SPECT NoAC , n = 246), or 82-rubidium-positron emission tomography (PET) (n = 240). Scar was assessed by CMR-LGE and FPD on MPI. Minimal resting myocardial blood flow (MBF Min ) was determined for PET. Patients with myocardial infarction or revascularization between studies were excluded. Associations of LGE and FPD with major adverse cardiovascular events (MACE) were analyzed. Results: LGE was present in 354 patients (43%), predominantly nonischemic, whereas nuclear MPI detected scar less frequently (30%, P < 0.01). Nuclear scar was associated with ischemic LGE (OR: 5.86; 95% CI: 3.64-9.43; P < 0.001), but not nonischemic LGE. PET demonstrated higher specificity than SPECT for detecting ischemic and ≥50% transmural LGE ( P < 0.01). Patients with ≥50% transmural LGE had lower MBF Min (0.61 [0.53-0.68] vs 0.80 [0.74-0.85] mL/min/g; P < 0.001). Patients with both LGE and FPD had higher MACE risk than patients without these features (HR: 4.95 [95% CI: 3.01-8.11]; P < 0.001). Conclusions: MPI detected myocardial scar less frequently than CMR-LGE, with PET providing greater specificity than SPECT for ischemic and transmural scar. Combined LGE and FPD identified patients at highest risk of adverse outcomes, supporting PET MPI for myocardial scar characterization.