Detection of the following metabolites:

- Neurotransmitters (GABA/Glu)

- Onco-markers (2HG)

- Antioxidants (GSH/Asc)

- Anaerobic metabolism (Lac)

In principle proton MR spectroscopy (1H-MRS) is able to detect a comprehensive metabolic profile of the human brain and other organs noninvasively. However, in practice the full range of 1H-MRS detectable metabolites cannot be reliably quantified in vivo due to a spectral overlap of resonances from different compounds, which is routinely observed in particular at lower magnetic field strength. In case of partial spectral overlap, the compounds can often still be separated well via elaborate spectral fitting routines at higher magnetic field strength. For metabolites with complete spectral overlap “Spectral editing” or “2D-NMR” is mandatory (1). In stable conditions such as in the brain or muscle, the use of J-difference editing methods is straightforward and achieves optimal sensitivity and highly reproducible results. In this educational talk, the application of spectral editing 1H-MRS techniques and their parameter as well as sequence design optimization to the following highly important metabolites will be discussed in detail, but also spectral editing of other metabolites will be briefly introduced:The in vivo detection of cerebral GABA is the most popular applications of J-difference editing (2). GABA is the major inhibitory neurotransmitter in the mammalian central nervous system with ~1-2 mM concentration and alterations in GABA concentration have been found in many neurological and psychiatric disorders (e.g., epilepsy, depression, and schizophrenia (2)), during drug treatment (3), as well as in the healthy brain (4). GABA has six 1H-MRS visible protons in three methylene groups, which form an A2M2X2 spin system. Its proton multiplet signals correspond to the three methylene CH2 groups in the molecule resonating at: 1.88, 2.28, and 3.02 ppm (3). Unfortunately, all three resonances of GABA are overlapping with other more intense resonances (i.e., creatine and phosphocreatine (tCr), with Glutamate and Glutamine (Glx), with N-acetyl aspartate and N-acetyl aspartyl glutamate (tNAA), and with macromolecules (MM) (4, 5). This makes direct observation impossible. For spectral editing the GABA resonance and all other resonances at 1.9 ppm are suppressed (including also Glx at 2.1 ppm and tNAA at 2.0 ppm) via editing pulses. This suppression also affects the signal modulation of the J-coupling partners (i.e., Glx at 3.75 ppm and GABA at 3.02 ppm) (2, 6). However, J-difference 1H-MRS does not separate the GABA signal entirely from contamination by co-edited signals at 3.0 ppm that are also coupled to nearby spins (i.e., MMs at 1.7 ppm, Glu at 3.75 ppm (7), and homocarnosine). As contributions of up to 50% from MMs can be expected, the derived GABA signal is usually labeled as GABA+ (4), unless special techniques are employed that remove most of the MM contamination (2). The co-editing of Glu is not of concern, since Glu-H2 and GABA-H4 do not overlap. Rather it can be even turned into an advantage, since Glu can be directly detected via the edited spectrum.A mutation of the IDH1 and the IDH2 gene results in a gain of enzymatic function, leading to excess metabolization of α-ketoglutarate (αKG) and accumulation of the onco-metabolite 2HG (8). The increased intracellular concentration of 2HG of up to 5-35mM(8), is an excellent target for in vivo detection via 1H-MRS allowing for non-invasive detection of this metabolite(9, 10). IDH-mutant cells produce this metabolite, but only trace levels are found in IDH-wildtype cells. 2HG is, therefore, an ideal biomarker for IDH-mutant gliomas(11). IDH-mutant glioma patients exhibit unique clinical characteristics: a generally younger age at initial presentation, improved 5-year survival and reduced rates of malignant progression than patients with IDH1 wild-type gliomas(12). As IDH mutations cause a depletion of NADPH, which is lowering the reductive capabilities of tumor cells(13), it has been speculated(14, 15) that IDH-mutant gliomas initial results indicate that they are susceptible to treatments such as radiotherapy that create free radicals. This can be visualized via special 1H-MRS sequences such as spectral edited 1H-MRS. For 2HG the detection challenge arises from the fact that the 2HG spectrum is largely overlapping with Glx, which are abundant and have a similar five-spin system. Peaks in the region of 2.6 to 2.4 ppm, while the 2HG multiplet at 4.02 ppm is overlapped by the large myo-inositol resonance at 4.05 ppm, lactate at 4.09 ppm and tCr at 3.91 ppm. However, this 2HG resonance at 4.02 ppm can be well edited via its J-coupled resonance at 1.9 ppm. The co-editing of Glu at 3.75 ppm and GABA at 3.01 ppm, which are also coupled to 1.9 ppm, is not a problem (9).Glutathione is the major antioxidant in the human body, particularly in major organs like brain, liver, and kidney. In the human brain, glutathione levels are in the range of 1–3 mM (16). Glutathione exists in both reduced (GSH) and oxidized (GSSG) states. GSH plays a major role in our body’s defense system against reactive oxygen species and oxidative stress (17, 18). GSH protects our cells by acting as non-enzymatic scavenger of free radicals (19). In addition it is an electron donor for glutathione peroxidase in the enzymatic detoxification of hydrogen peroxide. An increased GSSG-to-GSH ratio is considered an excellent biomarker of oxidative stress (20). Oxidative stress plays a pathological role in numerous diseases (e.g., multiple sclerosis (21, 22), Alzheimer’s (23, 24), Parkinson’s(24, 25), Huntington’s disease(26) amyotrophic lateral sclerosis (27), bipolar disorders (28), autism (29)), is important in cancer development and progression (30, 31), and is generally associated with aging (18, 32). All resonances of GSH are overlapping with more intense signals from other metabolites, most noticeably Glx, tCr and tNAA. In vivo detection of GSH has, therefore, been mainly achieved through spectral editing methods, although it has been shown that GSH can be detected directly when combing short-TE spectra with spectral fitting at ultra-high magnetic field. At lower field strength the triplet cysteine β-CH2 resonance of GSH at 2.95 ppm can be well detected in vivo via selective editing pulses applied on the J-coupled cysteine α-CH resonance of GSH at 4.56 ppm without major problems of co-edited metabolites other than tNAA, which does not negatively affect quantification.In normal brain tissue, Lac only overlaps with macromolecular resonances. However, in tumors, stroke or with inadequate localization, Lac can be overlapping with large lipid resonances. In these cases, Lac is best observed with spectral editing techniques. Single-shot techniques such as multiple quantum coherence filtering have been found particularly attractive as they remove the uncoupled lipid signals very successfully (33). With spectral editing techniques, it is possible to obtain full yield of both the coupled target (via the difference spectra) and the non-edited spectra (1). In this case editing pulses are applied at the 4.1 ppm lactate peak, which is giving a J-coupled doublet resonance at 1.33 ppm. Tumors can contain large amounts of lactate, and an early drop in lactate may be correlated with therapy response (34)). Conversely, increases in lactate in normal-appearing peritumoral regions may predict progression. Measurement of lactate in tumors is particularly challenging due to overlapping lipid peaks. These lipid peaks feature unusual mobility (i.e., longer T2), which makes them visible even at long TE of 144 msec. Also T1 relaxation time of lipids in tumors is longer than in normal tissues. This renders T1-relaxation based suppression less effective.These are only the most common applications for spectral editing. There are far more J-coupled metabolites that can be edited. In particular, there is increased interest in editing several metabolites at once using double-editing (35) or HADAMARD-encoded spectral editing approaches. In general, spectral editing has long remained a method applied only via single voxel 1H-MRS, but recent developments show promising results that combine 2D- and 3D-mapping with spectral editing (36, 37). This will improve the understanding of spatially heterogeneous pathologic alterations of the brain biochemistry.