Some works have raised during last years with the aim of provide optimized quantitation of brain tumors metabolites detected by MR spectroscopy. However, it should be noted that comparable data from different imaging centers and then, from different series, is difficult to obtain because of the heterogeneity of scanners, which is particularly important in spectroscopy, as the signal is varying in huge amounts (until two folds) from 1.5T to 3T magnets for a metabolite considered [21]. This may lead to important errors for (1) differential diagnosis from DLGG to other tumors; (2) follow-up characterization, as the successive exams for a same patient may be performed in a same center alternatively on magnets with different magnetic field strengths. Moreover, the data post-processing is also heterogeneous as numerous softwares are available in many MR centers, whatever commercial, freeware or in-house, with different solutions for quantitation. Some interactive quantification methods for the single voxel spectroscopy (SVS) such as AMARES or VARPRO, bring user lots of repetitive and unbearable work. What’s more, the quantitative result is hard to be reproduced because of the difference in user’s prior knowledge and state. Therefore, the automatic quantification is required to replace the interaction quantification for the sake of more efficiency and robustness. In their paper, Dou et al. [22] present a new automatic quantitative approach based on the convex envelope (AQoCE), compared with LC Model used on Siemens magnets. This method leads to increase both specificity and sensitivity of MR spectroscopy, thus suggesting that automated processing of MR data may avoid intra individual variability [23, 24]. This kind of software is based on mathematical algorithms, which supposes to have local specific competencies. The results, while repeatedly provided with a rigorous method in a same center, may be difficult to compare with those arising from another center. The reproducibility of some interesting results may be then random. Moreover, spectroscopic parameters are integrated in a global multiparametric analysis including contrast enhancement, related cerebral blood volume, ADC values extracted from diffusion sequence. Thus, the UCSF team [24] recently published a study on 120 patients with recurrent DLGG. They found several multivariate models with similar accuracy for predicting the grade II and the malignant transformation to grade III or IV. Hence, this kind of tools within the “kinetic” approach they allow are used in routine in some centers with expertise in the field, for the therapeutic follow-up and when the surgical resection is not possible [24]. Moreover, new proton high resolution multivoxel sequences, so-called “Laser or Mega Laser”, allow 3D metabolic mapping, including 2 hydroxyglutarate and glutamate, thus marking the tumoral spatial heterogeneity [25].

Numerous publications have emphasized a pro malignant role of metabolic enzymes as isocitrate deshydrogenase, succinate deshydrogenase and fumarate deshydrogenase [25]. The most investigated one until today, as the cytosolic isoform, named IDH1, is of interest because (1) its key-role in tumorigenesis; (2) the ability of proton MRS to detect and to quantify some of their metabolic counterparts (Fig. 14.1). Once mutated, the IDH1 enzyme leads to convert α-ketoglutarate to 2-hydroxyglutarate, which is identified to play a crucial role in the initiation of tumorigenesis in mutant IDH1 cells (please refer to the dedicated chapter in this book). Using point resolved spectroscopy within spectral difference editing from proton magnetic resonance spectroscopy, it is possible to detect and quantify 2-hydroxyglutarate resonance, a so-called oncometabolite, which accumulate [30]. Moreover, some authors have demonstrated the interest of longitudinal 2-hydroxyglutarate quantitation for monitoring treatment response in IDH1 mutant patients [31–33]. The authors developed a specific 3D sequence for over sampling the tumor and avoid difficulties due to heterogeneity. It is potentially available in centers with routinely use of proton magnetic resonance spectroscopy.

In their recent experimental works, Viswanath et al. [34] demonstrated that the IDH1 mutation may lead to a decrease of lactate production from pyruvate. They used multinuclear magnetic resonance spectroscopy, and especially ¹³C-MRS with dynamic nuclear polarization for pyruvate-lactate fluxes assessment, and ³¹P-MRS for steady-state assessment of intracellular pH and PCr. Furthermore, they showed that reduced flux of hyperpolarized [1-¹³C] pyruvate to hyperpolarized [1-¹³C] lactate could be due to reduced MCT1 and MCT4 expression in IDH1 muted cells. Increased intracellular pH in both type of cells was not modified by the drop of intracellular lactate.

Interestingly, the vascular counterpart of this “biochemical re-programming” is also accessible by MRI, during the same time of examination. Kickingereder et al. [30] provided a study in which they demonstrated that a one-unit increase in rCBV corresponded to a two-third decrease in the odds for an IDH mutation and correctly predicted IDH mutation status in 88% of patients. It has been established that increasing levels of 2 hydroxyglutarate leads to indirectly decrease of hypoxia-inducible factor 1-alpha, thus limiting angiogenic growing.

Whereas some of those results are obtained in experimental conditions with limited extrapolation to human- in vivo conditions, we must notice that they are partly consistent with theoretical conjectures expressed in previous papers [35, 36]. Yet, at this stage, before the recent knowledge assessing the links between genomic and metabolic modifications, alterations of MCT properties have been suggested by the previously published model. In this concept, glioma is considered as a general parametrical system evaluating in a viability domain, with specific ranges a value of each parameter [17, 35, 37]. Especially, high concentrations of Lactate (resulting from transport rate by MCT and pHi, are known to be not compatible with metabolic function of the parametrical system, e.g. the glioma (Fig. 14.2). This field of investigation seems to be at its beginning as the technical possibilities are increasing from year to year, especially in multinuclear magnetic resonance imaging, with the ability of studying different molecules according the nucleus studied, e.g. proton, phosphorus, carbon, sodium. Here is arising a new dimension of magnetic resonance, as metabolic one, which cross the link with the isotopic approach.

Moreover, this approach leads to interrogate potential therapeutic ways, based on MCT targeting, with potential therapeutic monitoring of a putative response using multinuclear spectroscopy.