Purpose

Myocardial blood flow can be assessed by positron emission tomography myocardial perfusion imaging (PET MPI) with high accuracy. Additionally, electrocardiogram (ECG)-gated PET acquisition allows for the evaluation of left ventricular (LV) volumes and function. In the current study we performed a head-to-head comparison of functional parameters obtained from 13N-ammonia PET versus the gold standard cardiac magnetic resonance (CMR) with both acquisitions performed simultaneously on a hybrid PET/MR device.

Materials and Methods

Fifty-one patients underwent ECG-gated 13N-ammonia rest PET MPI and CMR on a PET/MR device. Left-ventricular end-diastolic volumes (LVEDV [ml]), end-systolic volumes (LVESV [ml]), stroke volumes (LVSV [ml]), and LV ejection fractions (LVEF [%]) were obtained from both modalities and compared using Pearson correlation and Bland-Altman (BA) analysis with CMR serving as the standard of reference.

Results

We found a very high correlation between LVEDV (r=0.9), LVESV (r=0.9), LVSV (r=0.8), and LVEF (r=0.9) obtained from PET and CMR (all p<0.0001). BA analysis revealed narrow limits of agreement of -0.5% to 1.6% for LVEF and only minimal bias of 0.6%. By contrast, PET underestimated LVEDV, LVESV, and LVSV with a mean bias of -28.9 ml, -11.2 ml, and -17.5 ml, and limits of agreement of -24.6% to 33.1%, -7.9% to 14.5%, and -14.3 % to 20.7 %, respectively.

Conclusion

LVEF obtained from 13N-ammonia PET are interchangeable with those obtained from the gold standard CMR. By contrast, 13N-ammonia PET substantially underestimates LV volumes as compared to CMR.