Improved correction of spatial inhomogeneities of surface coils in quantitative analysis of first‐pass myocardial perfusion imaging

PURPOSE: To test whether image normalization using either a separate 3D proton-density (PD)-weighted prescan, or 2D PD-weighted images prior to the perfusion series, improves correction of differences in spatial sensitivity induced by radiofrequency (RF) surface receiver coils. Originally, this correction was applied using the baseline signal in the myocardium before arrival of the contrast agent. This is of importance, as quantitative analysis of magnetic resonance (MR) myocardial perfusion using deconvolution with the arterial input assumes equal signal sensitivity over the heart. MATERIALS AND METHODS: First-pass myocardial perfusion measurements were obtained in 13 patients without known coronary artery disease. Absolute perfusion values were assessed for 18 myocardial segments without any normalization and using the three different normalization methods. RESULTS: Using 2D or 3D PD-weighted normalization, similar mean perfusion values were found, but with reduced spatial variance over the 18 segments. The relative dispersion of perfusion at rest was 23% and 35% for the 3D prescan normalization and the baseline normalization, respectively. With 2D and 3D PD-weighted prescan normalization the relative dispersion was closer to the expected physiological heterogeneity. CONCLUSION: PD-weighted prescan normalization proved to be a valuable addition to quantitative analysis of myocardial perfusion, and better than baseline-based normalization.