Quantification of Intramyocardial Lipid by Proton Magnetic Resonance Spectroscopy

Because routine biopsy of the myocardium is not feasible, MRS is the most promising technique for the quantification of myocardial triglycerides. MRS is routinely used to precisely characterize metabolite concentrations in muscle and liver. 14-16 Studies such as monitoring the levels of deoxymyoglobin and real-time tracking of the postprandial accumulation of cellular lipids are examples of its diversity and potential.15,17,18 Generally, these studies suggest that the reproducibility of MRS is between 2 and 6%.18,19 In vivo cardiac MRS provides unique challenges because of the requirement to compensate for concurrent heart and lung motion. Using cardiac and respiratory gating to minimize motional artifacts, an initial validation study found a variation of 17% for sequential measurements, attributing the major error to residual motional effects. 20 Moreover, measurements were limited to the inter-ventricular septum. Using navigator and cardiac gating appeared to give a slight, 4%, improvement, but this was a preliminary study and no validation was done.21 For a comprehensive clinical validation, other reproducibility factors must be addressed. Variations due to post-processing, coil placement and calibration, trigger reproducibility, internal versus external standard, shimming, and protocol sequence variables such as pulse quality, gradient strength, voxel size, relaxation time, echo time, and the number of scan repetitions are all known sources of reproducibility. 17,19,22-24 All of these variables must be characterized in order to achieve optimal inter- scanner and subject reproducibility along with accurate treatment tracking capability. Therefore, 10 normal healthy volunteers were imaged to determine the reliability of the MRS protocol with test-re-test measurements. The 8 heart transplant patients were imaged prior to their routine heart biopsies, and then the myocardial biopsy tissue was measure and compared to the pre-biopsy images.