Lactate and Other Metabolic Changes in Neuronal Migration Disorders

Abstract

To the Editor: In Fig. 4 of the paper “Metabolic Changes in Neuronal Migration Disorders: Evaluation by Combined MRI and Proton MR Spectroscopy” by I. L. Simone et al. (Epilepsia 1999;40:872–879), the authors assign a broad negative peak to lactate, and they state that this lactate signal is seen at 1.4 ppm. The presented proton MR spectrum is from one of the 4 out of 15 patients with neuronal migration disorders, in whom the authors report increased lactate. In the spectrum of Fig. 4 the following peaks can be assigned; N-acetylaspartate (NAA) at the chemical shift position 2.02 ppm; total creatine (Cr) at 3.03 ppm; and total choline (Cho) at 3.22 ppm. The secondary peak of Cr that should appear at 3.9 ppm as in Figs. 2 and 3, is absent in Fig. 4, probably due to artefacts from water suppression. There is no peak at 1.4 ppm, but there is a broad negative peak between 1.5 and 1.8 ppm, this peak cannot be assigned to lactate, primarily because the chemical shift is at least 0.2 ppm from the correct value. In addition, there is no splitting of the doublet, which would appear with the spectral resolution estimated from the NAA, Cr, and Cho peaks. The broad negative signal is most likely an artefact from lipid signals originating from locations outside the volume of interest. Outer volume lipid signals can give rise to spurious peaks occurring at wrong chemical shift positions due to B0 inhomogeneities, and they are often inverted as seen here. Signals from outside the volume of interest are supposed to be suppressed when pulse sequences like PRESS are used. Nevertheless, artefacts from lipids do occur because the metabolite concentrations in the volume of interest are much lower than the lipid concentrations dominating outside. Using long TE PRESS to avoid lipid signal does guarantee freedom from artifact because the spurious signals giving rise to outer volume broad peaks are not necessarily from the three pulse double spin echo; the signals could be from residual unspoiled free induction decays or single spin echoes that are not suppressed. Even if a broad peak is located at 1.33 ppm, and it is inverted using PRESS TE=135ms, one cannot conclude that the peak is lactate. The basis of assigning lactate and differentiating lactate from lipid can be summarized as follows 1: Using a known metabolite, such as Cr or NAA, or both, as chemical shift references, the lactate double peak should be centered at 1.33 ppm. If the resolution is sufficient to separate Cho and Cr, then the magnetic field homogeneity is sufficient to see the a splitting of the lactate doublet at least half-way to the base of the peak (using 1.5 Tesla MR systems). The split should be 7 Hz. In cases where the magnetic field homogeneity is suboptimal or unknown, one cannot necessarily expect to see splitting of the lactate doublet, making the differentiation between real lactate and overlapping broad lipid signals that much more difficult.