Parameter optimization for reproducible cardiac 1H-MR spectroscopy at 3 Tesla

• Published in: Journal of magnetic resonance imaging Vol. 44; no. 5; pp. 1151 - 1158
• Main Authors: de Heer, Paul, Bizino, Maurice B., Lamb, Hildo J., Webb, Andrew G.
• Format: Journal Article
• Language: English
• Published: Nashville Blackwell Publishing Ltd 01.11.2016; Wiley Subscription Services, Inc
• Subjects: Breathing; cardiac proton magnetic resonance spectroscopy; Chess; Coefficient of variation; Compensation; Data acquisition; Excitation; Heart; Heart diseases; Magnetic fields; Magnetic resonance imaging; Magnetic resonance spectroscopy; motion compensation; myocardial triglyceride content; Optimization; Reproducibility; Respiration; Spectroscopic analysis; Spectroscopy; Spectrum analysis; triglycerides; water suppression
• ISSN: 1053-1807; 1522-2586
• DOI: 10.1002/jmri.25254

Purpose To optimize data acquisition parameters in cardiac proton MR spectroscopy, and to evaluate the intra‐ and intersession variability in myocardial triglyceride content. Materials and Methods Data acquisition parameters at 3 Tesla (T) were optimized and reproducibility measured using, in total, 49 healthy subjects. The signal‐to‐noise‐ratio (SNR) and the variance in metabolite amplitude between averages were measured for: (i) global versus local power optimization; (ii) static magnetic field (B0) shimming performed during free‐breathing or within breathholds; (iii) post R‐wave peak measurement times between 50 and 900 ms; (iv) without respiratory compensation, with breathholds and with navigator triggering; and (v) frequency selective excitation, Chemical Shift Selective (CHESS) and Multiply Optimized Insensitive Suppression Train (MOIST) water suppression techniques. Using the optimized parameters intra‐ and intersession myocardial triglyceride content reproducibility was measured. Two cardiac proton spectra were acquired with the same parameters and compared (intrasession reproducibility) after which the subject was removed from the scanner and placed back in the scanner and a third spectrum was acquired which was compared with the first measurement (intersession reproducibility). Results Local power optimization increased SNR on average by 22% compared with global power optimization (P = 0.0002). The average linewidth was not significantly different for pencil beam B0 shimming using free‐breathing or breathholds (19.1 Hz versus 17.5 Hz; P = 0.15). The highest signal stability occurred at a cardiac trigger delay around 240 ms. The mean amplitude variation was significantly lower for breathholds versus free‐breathing (P = 0.03) and for navigator triggering versus free‐breathing (P = 0.03) as well as for navigator triggering versus breathhold (P = 0.02). The mean residual water signal using CHESS (1.1%, P = 0.01) or MOIST (0.7%, P = 0.01) water suppression was significantly lower than using frequency selective excitation water suppression (7.0%). Using the optimized parameters an intrasession limits of agreement of the myocardial triglyceride content of ‐0.11% to +0.04%, and an intersession of ‐0.15% to +0.9%, were achieved. The coefficient of variation was 5% for the intrasession reproducibility and 6.5% for the intersession reproducibility. Conclusion Using approaches designed to optimize SNR and minimize the variation in inter‐average signal intensities and frequencies/phases, a protocol was developed to perform cardiac MR spectroscopy on a clinical 3T system with high reproducibility. J. Magn. Reson. Imaging 2016;44:1151–1158.