INTRODUCTION

Measurement of lactate (Lac), a by-product of glycolysis, has been hindered by the low concentration in the healthy brain (~1mM) and overlap of the methyl doublet at 1.31 ppm in ¹H spectra by lipid and macromolecule signals¹. Inversion of the doublet at longer echo times can reduce signal overlap, but better precision can be achieved by spectral editing. MEGA-PRESS (MEshcher-GArwood Point RESolved Spectroscopy)² (TE ~ 144 ms) offers robust editing of lactate signal but necessarily includes a time and signal penalty. Asymmetric PRESS (A-PRESS, TE1/TE2 25/85 ms) also inverts the lactate doublet and is a single shot method but does not resolve overlapping signals (Figure 1). Here, we investigated whether selective editing combined with A-PRESS (MEGA-APRESS) could improve the precision of lactate detection and repeatability compared to MEGA-PRESS and A-PRESS.

METHODS

¹H MRS data were acquired at 3T (Philips Ingenia CX) using a dStream-32 head coil in 8 healthy adults (3 women, 31 ± 7 years) from a 3 × 3 × 3 cm³ volume-of-interest in the left parietal lobe and a 2L phantom (GE Braino base with 10 mM sodium lactate). All spectra were acquired with TR = 3000 ms, 1024 data points using VAPOR water suppression, B1 = 22 μT, using high bandwidth π/2 (FREMEX05) and π (FREMREF04) pulses. A-PRESS scan duration 4 min 54 s, 80 averages, Ed BW = 42 Hz), MEGA-PRESS and MEGA-APRESS (scan duration 8 min 15 s, 80 ON/OFF pairs, Ed BW = 27 Hz) were collected twice in randomised order to test intra-session repeatability along with a water reference. Data were analysed using QUEST (jMRUI, v4) fitting basis sets simulated using NMRSCOPE or, in the case of edited spectra, acquired from the phantom. For each participant, lactate estimates from edited spectra were referenced to the averaged water integral from A-PRESS (Lac/H₂0). Repeatability was analysed by computing the coefficient of variation. Fitting accuracy was assessed using the standard deviation of the fit generated by jMRUI. Repeatability and precision between the three methods were compared using repeated measures one-way ANOVA, followed by Tukey’s multiple comparisons test (GraphPad Prism, v9.2), while statistical differences in Lac/H₂0 estimates between MEGA-PRESS and MEGA-APRESS was determined with a paired t-test. Spearman correlation was used to investigate the relationship between Lac/H₂0 estimates.

RESULTS

Spectra from the phantom edited by either editing approach are shown in Figure 2. In vivo, the coefficient of variation was the lowest for the MEGA-APRESS (4.4 ± 4.3 %) when compared to MEGA-PRESS (8.6 ± 5.8 %) and A-PRESS (13.8 ± 13.1%) (Figure 3A). Intrasession repeatability of Lac/H₂O did not differ between all three methods (P = 0.20, n = 8), although MEGA-PRESS, and MEGA-APRESS showed good quality fits (18.0 ± 2.1 % and 18.5 ± 2.0 %, respectively), with both significantly lower than A-PRESS (42.2 ± 21.6%, P = 0.0005, n = 16) (Figure 3B). Lac/H₂0 estimates obtained from edited spectra did not differ between MEGA-APRESS and MEGA-PRESS (0.44 ± 0.03, and 0.42 ± 0.05, paired t-test, P = 0.08, n = 16) (Figure 3C). However, the confidence interval of Lac/H₂0 from the MEGA-APRESS scan was about half that obtained using MEGA-PRESS (95% confidence interval of the mean, [0.42-0.46] vs [0.39-0.45]). There was a weak correlation between different editing estimates (Spearman, r = 0.52, P = 0.04, n = 16) (Figure 3D).

DISCUSSION

Using MEGA-APRESS with TE 25/85 for lactate detection allowed more robust quantification in healthy adult brains. It yielded an excellent intrasession reproducibility not achieved by the standard MEGA-PRESS and A-PRESS, without loss in fitting accuracy and with improved consistency.

CONCLUSION

Edited spectroscopy with MEGA-APRESS improved in vivo brain lactate quantification in healthy adults. It will provide more certainty for measurement of lactate in pathology.

Acknowledgements

The authors acknowledge the facilities and scientific and technical assistance of the National Imaging Facility, a National Collaborative Research Infrastructure Strategy (NCRIS) capability, at Neuroscience Research Australia and the University of New South Wales (Australia).