Richard Edden

  MRS of Lactate

  Spinal Cord MRSI


  Contact me



Lactate is often elevated in brain ischemia, mitochondrial diseases and other neurological disorders. In in vivo proton spectroscopy, the methyl signal of lactate (at 1.3 ppm) may sometimes be difficult to distinguish from overlapping lipid signals. One way to differentiate lactate from lipid is to use echo times (TE) of 136-144 msec (1/J, J~7 Hz), where, under ideal conditions, the lactate methyl doublet will be inverted due to evolution of the scalar coupling. However, at high magnetic field strengths, the chemical shift difference between the lactate methyl and methine (CH) resonances may become similar to the bandwidth of the slice selective pulses used for spatial localization. This alters the modulation of the lactate methyl group, often resulting in inefficient inversion and detection. This abstract presents theoretical and experimental results that demonstrate that lactate detection can be optimized by trimming the PRESS-excited volume with outer-volume suppression (OVS) pulses.

Single-voxel PRESS spectroscopy (3x3x3 cm) and MRSI (28x28 resolution) of a 5mM lactate and 12.5 mM NAA phantom were performed on a Philips Intera 3T system with a transmit-receive head coil. RF pulses were transmitted with a maximum field of 20 mT. Spectra were acquired at an echo time of 144 ms (TE1 = 30, TE2 = 114 ms) and a recycle time of 2000 ms. Refocusing was achieved with 10 ms sinc-Gaussian pulses of bandwidth 900 Hz.

The evolution of the lactate methyl spins can be modeled by assuming that the lactate methine resonance (4.1 ppm) experiences both (Figure 1a), either or neither 180° refocusing pulse. The relative size of the four regions is determined by the bandwidth of the RF pulses and the chemical shift difference between the lactate resonances (360 Hz at 3T) and a single voxel spectrum can be modeled as the weighted sum of the four individual multiplets, shown Figure 1b. A PRESS MRSI experiment (Figure 1c) demonstrates the heterogeneity of the lactate signal excited and spectra (Figure 1d), extracted from the voxels marked in 1c, agree well with the model spectra. By enlarging the PRESS volume and suppressing regions B, C and D with high-bandwidth, slice-selective saturation pulses, a single-voxel spectrum of region A can be acquired. Because the coupling evolution is homogenous throughout that volume, losses of signal intensity due to the chemical-shift artifact are avoided.

The spectrum shown above (far left) was acquired using traditional PRESS localization. The predicted signal attenuation is 80%, but the actual loss is 87%. The poor phase of the lactate peak and the additional signal loss are due to ‘edge’ effects, caused by the non-rectangular profile of the slice-selective pulses. The corresponding spectrum (centre left) was acquired using an enlarged PRESS volume with saturation pulses applied to suppress signal from regions B, C and D. There is a sevenfold improvement in signal intensity, with loss of only 4%, as measured against the integral of the NAA peak.

The PRESS-IVS method was applied to a stroke patient, with significant improvement seen in the intensity of the inverted lactate signal (far right).

Web site and all contents © Richard Edden 2006, All rights reserved. Last updated: 2nd June 2006

Web Template From:

Valid CSS!