This study aimed to investigate the correlation and assess the interchangeability between myocardial flow reserve (MFR) quantified by positron emission tomography (PET) and coronary sinus flow (CSF) reserve measured by cardiac magnetic resonance (CMR) in patients with ischemic heart disease, utilizing a hybrid PET/MR system.

Consecutive 119 patients (mean 68.5 ± 13.1ys, male 95) who underwent rest-pharmacological stress 13N ammonia PETMR were enrolled. Myocardial blood flow (MBF mL/g/min) at rest and stress, along with MFR, were quantified from PET data. Myocardial perfusion defect was assessed using the summed stress score (SSS). CSF (mL/min/g) was obtained by phase-contrast cine CMR for rest and stress, and CSF reserve was calculated as the stress-to-rest ratio.

Rest and stress CSF showed a significant but weak correlation to MBF (r = 0.40, p < 0.0001; r = 0.44, p < 0.0001 for rest and stress, respectively). CSF reserve showed a significant but moderate correlation with MFR (r = 0.43, p < 0.0001) and a weak inverse correlation with SSS (r = − 0.27, p = 0.003). Brand-Altman plots revealed a poor agreement with a proportional error (bias − 0.60 ± 1.34, 95%limit of agreement − 2.02 to 3.21).

Despite a moderate correlation, CSF reserve and PET-derived MFR demonstrate poor agreement and are not interchangeable. The divergence between these metrics necessitates caution when interpreting global coronary flow reserve derived from CMR in patients with established ischemic heart disease.