MR spectroscopy in stroke - British Medical Bulletin

MR spectroscopy in stroke

Dawn E SaundersDepartment of Radiology, King's College Hospital, London, UK

Magnetic resonance spectroscopy (MRS) is a non-invasive in vivo method thatallows the investigation of biochemical changes in both animals and humans.The application of MRS to the study of stroke has made possible dynamicstudies of intracellular metabolism of cerebral ischaemia. The majority of thestroke studies have been carried out using proton [1H]-MRS which allows thedetection of N-acetyl aspartate (NAA), a neuronal marker. ['H]-MRS changes inhumans demonstrate that after an infarct, lactate appears, while NAA and totalcreatine are reduced compared to the contralateral hemisphere. Longitudinalstudies demonstrate a further reduction of NAA suggesting that ischaemic injurycontinues for more than a week following infarction.

Major advances in the treatment of acute stroke require the accurateprediction of the mortality of stroke patients Patients with large infarcts areknown to do badly. In patients with small infarcts, less than 80 cm3, the additionof core NAA concentrations and cerebral blood flow have enabled the identifi-cation of some of the patients likely to benefit from new drug treatment.

Correspondence toDr Dawn E Saunders,

Department ofNeuroradiology, Atkinson

Morley^ Hospital, CopseHill, Wimbledon,

London SW20 ONE. UK

Conventional magnetic resonance imaging (MRI) is a well-developedand much utilised method for visualising the radiological extent ofdisorders involving cerebral ischaemia. Magnetic resonance spectro-scopy (MRS) is a non-invasive in vivo method that allows the investi-gation of biochemical changes in both animals and humans. Theapplication of MRS to the study of stroke has made possible dynamicstudies of intracellular metabolism of cerebral ischaemia. Theconcentrations of metabolites detected by MRS are relatively low (2-20mM) compared to the concentration of water (83.4 M) detected byMRI, and the sensitivity to local magnetic field inhomogeneities anddifficulties in quantitation of peak areas has until recently limited thetechnique to use as a research tool. Despite the relatively low signal-to-noise ratio of spectroscopy, improvements in magnet and gradient design,and the wider availability of magnets at high field strength (1.5 T) nowenable good quality brain []H]-MRS spectra to be recorded on mostmodern clinical instruments.

['HJ-MRS and phosphorus [31P]-MRS are most commonly carried outin humans. In the brain, ['H]-MRS has two great advantages: the protonis 15 more times sensitive than 31P, and almost every compound in living

Bntish Medical Bulletin 2000, 56 (No 2) 334-345 C The British Council 2000

Dow

nloaded from https://academ

ic.oup.com/bm

b/article-abstract/56/2/334/303282 by guest on 10 April 2019


tissue contains hydrogen. The study of cerebral ischaemia in humans ispredominantly confined to PHJ-MRS, the subject of this chapter.

Until recently, data acquisition required the skills of experiencedspectroscopists and physicists for manual adjustment of parameters foreach and every scan resulting in long examination times. The develop-ment of automated software programs, like the single voxel protonbrain exam (PROBE, General Electric, Milwaukee, USA)1, and theirimplementation on clinical systems has gone a long way to overcomingsome of the problems of data collection. The small concentrations ofvisible metabolites requires the collection of data from a single voxelmany times larger than that required for MR imaging.

Chemical shift imaging (CSI) or multivoxel MR spectroscopic imagingis a more advanced form of spectroscopy that uses phase encoding tosubdivide a large volume of interest into smaller acquisition voxels,thereby allowing the study of large and heterogeneous areas of brain. Themajor disadvantages of CSI include complicated acquisition techniques,longer scan times, and a large volume of generated data. CSI is ideallysuited to the study of stroke, particularly in the detection of the ischaemicpenumbra. However, the complexity of image acquisition and dataprocessing often necessitates the availability of a dedicated spectroscopistand has Limited the number of stroke studies carried out using CSI.

The normal [1H]-MRS spectrum

Histochemical and cell culture studies have shown that specific cell typesor structures have metabolites that give rise to ['HJ-MRS peaks. Achange in the resonance intensity of these marker compounds mayreflect loss or damage to a specific cell type or compound.

The acquisition of long echo data (TE = 270 ms, TR = 3 s) only allowsthe detection of NAA, Cr/PCr and Cho in normal brain, and lactate inregions of abnormality. T2 losses results in lower signal-to-noise andincreases the complexity of quantitation methods. The acquisition ofshort echo time data (TE = 30 ms, TR = 2s) reduces the effects of signalloss due to T2 relaxation and, therefore, provides spectra with increasedsignal-to-noise. In addition, short echo time spectroscopy detectsadditional resonances from metabolites with complex MR spectra suchas myo-inositol, glutamate and glutamine. Signals from these meta-bolites cancel at long echo times due to phase modulation ('J-couphng').Whilst providing us with more information, short echo time datainclude a broad background signal consisting of low concentrationmetabolites, and macromolecules and lipids with short T2 relaxationtimes, which increases the difficulty of accurate peak area estimation.

British Medical Bulletin 2000,56 (No 2) 335

Dow


ic.oup.com/bm


Stroke

Fig. 1 Proton spectra (TE = 30 ms, TR = 2020 ms) obtained in the acute phase from the (a)contralateral and (b) infarct centre Resonance peaks are- glutamate and glutamine (Glx)

at 3.8 and 2.1-2.45 parts per million (ppm), myo-inositol (ml) at 3.56 ppm, cholinecontaining compounds (Cho) at 3.22 ppm, creatine and phosphocreatine (Cr/PCr) methyl

singlet at 3.03 ppm, N-acetyl aspartate (NAA) methyl singlet at 2.01 ppm, and lactatedoublet (Lac) at 1.33 ppm The concentration of NAA and Cr/PCr are reduced and lactate is

elevated in the infarct centre compared with the contralateral hemisphere.

336 Bntish Medical Bulletin 2000;56 (No 2)

Dow


ic.oup.com/bm



Short echo ['ITJ-MRS spectra acquired from an infarct and thecontralateral hemisphere are shown in Figure 1. The chemical shift ofthe peaks are assigned with respect to water which has been removedfrom the spectra.

N-acetyl aspartate (2.01 ppm)

The methyl resonance of N-acetyl aspartate (NAA) produces a largesharp peak at 2.01 ppm and acts as a neuronal marker as it is almostexclusively confined to neurones in the human brain, where it is foundpredominantly in the axons and nerve processes2. It has also been foundin the oligodendrocyte type II astrocyte progenitor cells in rats3, butthese cells represent only 2-3 % of the glial population in man4. Since thediscovery of NAA in 19562, there has been a large body of literature onthe synthesis, distribution and function, which has been reviewed byBirken and Oldendorf5 and Williams6. However, despite a large body ofwork, the function of NAA remains unknown.

The interpretation of NAA signal in the brain of children is com-plicated by increases in the concentration during development7 when itis thought to have a role in supplying acetyl groups for myelin synthesis6.In adults, the concentration of NAA is known to vary in different areasof the brain8. This can be overcome in the study of stroke patients byusing the contralateral hemisphere for comparison9.

Creatine (3.94 & 3.03 ppm)

Both creatine and phosphocreatine have signals at 3.94 ppm (methylenesinglet) and 3.03 ppm (methyl singlet) which makes it impossible todistinguish between the two compounds and therefore total creatme(Cr/PCr) signal changes are considered by PHJ-MRS. Other resonancesseen at 3.03 ppm arise from y-amino butyric acid (GABA)10 and cytosolicmacromolecules become incorporated into the Cr/PCr peak11-12.

Cr/PCr is found in both neurones and glial cells4 and acts as aphosphate transport system and energy buffer within the cell. As thesignal comes from the sum of creatine and phosphocreatine, littleinformation can be gleaned about phosphocreatine metabolism. Thecomplete absence of creatine signal probably reflects necrotic tissue, butthe interpretation of reduced total creatine levels requires more work.

Choline (3.22 ppm)

The trimethylamine resonance of choline-containing compounds is pre-sent at 3.2 ppm and has been proposed as a marker of membrane damage.


Dow


ic.oup.com/bm


Stroke

In normal brain, the choline (Cho) peak is thought to consist predomin-antly of glycerolphosphocholine and phosphocholine; both compoundsare involved in membrane synthesis and degradation13.

Lipid/macromolecule resonances

Lipid peaks are detectable by short echo proton spectroscopy but not seenat long echo due to T2 losses. Peaks assigned to lipids/macromolecules havebeen detected at 0.9, 1.3 and 1.45 ppm in normal appearing brains14'15.Increases in these peaks have been reported in stroke15, and demyelin-ation14-16.

The 0.9 and 1.3 ppm resonances are assigned to the methylene andmethyl groups of lipid, respectively17. Recent work in animal models hasshown that contributions to these resonances also arise from mobileproteins12, and correlate with peaks found in humans18. The significanceof the appearances of these peaks is as yet unknown and may onlyrepresent increased MRS visibility of cell membrane lipids following cellbreakdown. The assignment of these signals will, however, influence theinterpretation of proton spectra obtained from the cerebral cortex.

[1H]-MRS changes in cerebral ischaemia

[]H]-MRS studies in humans demonstrate that after acute cerebralinfarction, lactate appears, while NAA and total Cr/PCr are reducedwithin the infarct compared to the contralateral hemisphere9. Largevariations in the initial concentrations of Cho have been observed in theregion of infarction9-19.

Lactate (doublet at 1.33 ppm)

Lactate is not normally detected within the brain and, as the end productof glycolysis, is a particularly useful measure of metabolism. The con-centration of lactate rises when the glycolytic rate exceeds the tissue'scapacity to catabolise it or remove it from the blood stream. The rise inbrain lactate that results from the mismatch between glycolysis andoxygen supply has been demonstrated by numerous [1H]-MRS experi-ments9'19'20 making it a hallmark for the detection of cerebral ischaemia.

The persistence of lactate weeks or months following stroke onset hasbeen observed9-20. Removal of lactate depends on tissue perfusion, thepermeability of the blood brain barrier (BBB) and diffusion of the

338 British Medical Bulletin 2000;56 (No 2)

Dow


ic.oup.com/bm



metabolite through the damaged tissue. The fall of lactate concentrationhas been shown to occur during a period of hyperaemia20. Using [1-13C]-labelled glucose, the metabolic activity of the lactate pool associatedwith a 32-day-old stroke has demonstrated that all the cerebral lactatearises from glycolysis of serum glucose. This supports the hypothesisthat elevated lactate, present weeks after a stroke, is the product ofongoing lactate synthesis in the brain21. It is not possible to establishwhether the lactate is being produced by on-going ischaemic tissue ormetabolically active macrophages. The presence of lactate beyond thetime in which active phagocytosis occurs supports the theory of per-sistently ischaemic tissue22.

Spectroscopic imaging studies have demonstrated that lactate levels inregions adjacent to the T2 hyperintensities were not significantly differentfrom those found in the infarcted brain and that lactate is not confined toareas of infarction determined by T2-weighted imaging and were evenfound in the contralateral hemisphere in patients with large infarcts23.

NAA (2.01 ppm)

Longitudinal studies in humans demonstrated a decline in NAAfollowing initial reductions in NAA detected in the first spectra after theonset of cerebral infarction9'19"20. The continuing fall in NAA concent-ration over the course of the first week after stroke onset cannot beexplained simply by an increase in oedema because changes in theconcentration of water within the infarct have been corrected for byusing the contralateral water signal as the internal standard9. It has beensuggested that NAA is actively degraded by enzymes within the injuredneurones in the first few days or hours following infarction16. Thisremains a possibility, but it would appear unlikely that enzymes wouldremain active for up to 7-10 days within an ischaemic neurone. Thegradual decline in NAA concentration and the persistence of lactatewithin the region of infarction over a period of a number of days issuggestive of ongoing ischaemia9 (Fig. 2) and has important implicationsin the timing of therapeutic intervention.

Cr/PCr (3.94 & 3.03 ppm)

Initial reductions in Cr/PCr are identified following infarction andfurther reductions have been demonstrated up to 10 days following thetime of onset9. The reduction in NAA in the infarct region is moremarked than the reduction in Cr/PCR and this is thought to reflect theincreased sensitivity of neurones to ischaemia.


Dow


ic.oup.com/bm


Stroke

Fig. 2 A longitudinalstudy of the infarct

core in a 32-year-oldlady. NAA and Cr/PCr

continue to fall over 6weeks following

infarction. Lactate isvisible within 16 h and

persists for 6 weeksThe lactate peaks can

be seen to increasefurther with somecontribution fromunderlying lipids.

Choline increases up to6 weeks which reflects

increased visibility ofcholine containing

compounds within themembrane due to cellnecrosis, (see Fig. 1 for

peak assignments).

stroke 6 weeks

stroke 3 weeks

stroke 1 week

stroke 3 days

stroke 16 hours

contralateral

ppm

Choline (3.22 ppm)

Changes in the choline peak have been shown to be increase, decreaseand stay the same following infarction9'19"20. The changes in the cholinepeak are thought to reflect changes in the MR visibility of the cholinecontaining compounds that make up the cell membrane.

Glutamate and other amino acids

Measurements of amino acids, such as glutamate and glutamine, aredifficult due to their complex peak structures. Other amino acids, such asGABA, aspartate and alanine are difficult to detect because of their lowconcentration and/or overlap with more intense resonances from other

340 British Medical Bulletin 2000,56 (No 2)

Dow


ic.oup.com/bm



Fig. 3 Spectraacquired from a 62-

year-old stroke patient4 weeks following

stroke onset The topspectrum is the

'standard' spectrumshowing a large lipidpeak underlying the

asymmetrical peaks ofthe lactate doublet

The centre spectrum isthe metabolite 'nulled'

spectrum acquiredfrom underlying

lipid/macromoleculesand showing a large

lipid peak at theposition of the lactate

peak (1.3 ppm). Thebottom spectrum is the

'pure' metabolite or'difference' spectrum

and shows the lactatesignal as a doublet

without theunderlying baseline.

NAA = N-acetylaspartate, Cr/PCr =

total creatine, Cho =choline, Lac = lactate,

L1 and L2 = lipids, MM= macromolecule peak.

standard"

c) " Ippm

compounds24. To date, there is very little work in cerebral ischaemia on thedetection of glutamate and glutamine, reflecting the difficulty of spectralanalysis.

Lipid/macromolecules

Signals from lipids pose particular problems in ['H]-MRS by obscuringthe resonance from the methyl group of lactate. It has been postulatedthat membrane lipids in the brain under normal conditions are notmobile enough to generate sharp peaks in in vivo spectra. Duringischaemia, degradation of the membrane leads to the release of free fattyacids that produce well defined but broad resonances in the region of the


Dow


ic.oup.com/bm


Stroke

lactate peak resulting in difficulties in the accurate quantitation oflactate. The metabolite nulling technique has been used to separate thelipid from the lactate signal (Fig. 3)15'18.

Reversible changes detected by [1H]-MRS

The identification of the ischaemic penumbra by [JH]-MRS depends onthe observation of the reversibility of metabolic changes within theregion of infarction. Spectroscopy studies have reported a reversiblereduction of NAA in acute multiple sclerosis lesions14'16-25, andmitochondrial encephalopathy with lactic acidosis and stroke-likeepisodes (MELAS)25. No convincing reports of reversible NAA incerebral infarction have been published. In MELAS, the primary defectis mitochondrial and, as NAA is synthesised in mitochondria, it has beensuggested that in demyelinating lesions, the function of the mito-chondria is reversibly impaired during inflammation resulting in thereversible reduction of the production of NAA within the damagedneurone14. CSI studies have observed high levels of lactate return toundetectable levels, in normal appearing regions of the brain separatefrom the area of infarction. Although, slight reductions in NAA werelater observed, the presence of high lactate concentrations did not resultin complete infarction23.

Quantitation methods in

To move MRS further into the clinical domain and allow the compari-son of metabolite concentrations within and between institutions,automated methods of quantitation need to be implemented to providedata analysis protocols for clinical studies. Previously, peak areametabolite ratios calculated from long echo time spectra were used tostudy metabolic changes in both animal and man. Ratios can bemisleading when metabolite concentrations in the numerator anddenominator change in disease states. Although the majority of strokestudies have been carried out using metabolite ratios, some studies haveused absolute quantitation of brain metabolites915.

The accuracy of the metabolite estimation depends on the fitting of theaccuracy of the fitting method. Various fitting routines have beenincorporated in the frequency domain26 (the conventional domain ofMRS data analysis) and the time domain to overcome the problem ofcontributions to peak area estimates from the underlying baseline andadjacent peaks to improve the accuracy of metabolite quantitation27'28.

342 Bnthh Medical Bulletin 2000,56 (No 2)

Dow


ic.oup.com/bm



'Metabolite nulling' has been employed to simplify the spectrum andovercome the problems of the incorporation of the broad backgroundsignal into the peak area estimates of short echo proton spectroscopy18.

[1H]-MRS as a predictor of outcome

Major advances in the treatment of acute stroke require the accurateprediction of the mortality of stroke patients and, therefore, theidentification of patients likely to benefit from drug treatment. Singlevoxel and CSI studies of stroke have attempted to provide informationabout the possible outcome of stroke.

Infarct volume determined by T2-weighted MR imaging has been shownto be a good predictor of outcome. All patients with small infarcts (<80cm3) survived 3 months29. The addition of the NAA concentration toinfarct volume allows a better prediction of patients' morbidity, thaneither the NAA concentration or infarct volume alone30. In a smallnumber of serially studied patients, the best recovery was seen in thosepatients with relatively preserved NAA, total creatine and choline peaks31.Large cerebral infarcts have been shown to be associated with reducedNAA concentration and raised lactate as well as reduced blood velocity.The best predictor of outcome was infarct volume and not metaboliteconcentration alone32.

The importance of lactate in the pathogenesis of cerebral infarction hasbeen studied extensively in animals. The degree and extent of tissuedamage has been correlated with the content of lactic acid in animals33,and an increase in morbidity and mortality m hyperglycaemic animals hasbeen linked to excessive lactate production and the resulting acidosis34.The correlation of lactate with outcome in humans, has been variable.Although a single study has shown lactate to correlate with acute strokeseverity and outcome19, other studies have failed to demonstrate arelationship between outcome and infarct lactate concentration alone30-32.

Application of [1H]-MRS to the clinical study of cerebral ischaemia

MR spectroscopy and multivoxel MR spectroscopic imaging haveproven to be valuable tools in the study of cerebral vascular disease.Their role in the detection of potentially salvageable ischaemic tissuesuitable for treatment with thrombolytic and neuroprotective agents hasnot yet been fully realised. The combined use of diffusion-weighted andperfusion imaging with ['H]-MRS may prove more valuable, especiallyin light of the growing observation of the negative diffusion-weighted


Dow


ic.oup.com/bm


Stroke

images in acute stroke35. The ability of multivoxel chemical shiftimaging to detect changes throughout the brain makes it the most suitedspectroscopy technique for the study of cerebral ischaemia. Furtherdevelopment of the CSI and data processing techniques are required tomove the use of spectroscopy further into the clinical domain.

Acknowledgement

References

I thank Prof. S R Williams, Department of Imaging Science and Engin-eering, University of Manchester, UK for reading the manuscript.

1 Webb PG, Sailasuta N, Kohler SJ et al Automated single-voxel proton MRS: technicaldevelopment and multisite verification. Magn Resort Med 1994; 31 365-73

2 Tallan HH, Moore S, Stein WH N-acetyl-L-aspartic acid in brain. / Biocbem 1956; 219257-64

3 Urenjak J, Williams SR, Gadian DG, Noble M. Proton nuclear magnetic resonancespectroscopy unambiguously identifies different neural cell types. / Neurosct 1993, 13: 981-9

4 Wolswi]k G, Munro PMG, Riddle PN, Noble M Origin, growth factor responses, and ultrastructure characteristics of an adult-specific ghal progenitor-cell Ann N Y Acad Set 1991; 633;502-4

5 Birken DL, Oldendorf WH N-acetyl-L-aspartic acid: a literature review of a compoundprominent in 'H-NMR spectroscopic studies of brain Neurosa Biobehav Rev 1989; 13: 23-31

6 Williams SR. In vivo proton spectroscopy. Experimental aspects and potentials. In: Rudin M,Seelig J. (eds) NMR - Basic Principles and Progress: In vivo magnetic resonance spectroscopyHeidelberg: Springer, 1992, 55-72

7 Kreis R, Ernst T, Ross BD. Development of the human brain: in vivo quantification ofmetabolite and water content with proton magnetic resonance spectroscopy. Magn Reson Med1993; 30: 424-37

8 Kreis R, Ernst T, Ross BD. Absolute quantitation of water and metabolites in the human brainn Metabolite concentrations. / Magn Reson 1993, 102: 9-19

9 Saunders DE, Howe FA, van den Boogaart A, McLean MA, Griffiths JG, Brown MM.Continuing ischaemic damage after acute middle cerebral artery infarction in humansdemonstrated by short-echo proton spectroscopy. Stroke 1995, 26. 1007—13

10 Rothman DL, Petroff OAC, Behar KL, Mattson RH Localised 'H-NMR measurements of y-amino butyric acid in human brain in vivo. Proc Natl Acad Sci USA 1993; 90; 5662-6

11 Behar KL, Ogino T Characterisation of macromolecule resonances in the 'H-NMR spectrumof rat brain. Magn Reson Med 1993, 30: 38^t4

12 Kauppinen RA, Kokko H, Williams SR Detection of mobile proteins by proton magneticresonance spectroscopy in the guinea pig brain ex vivo and their partial purification. /Neurochem 1992; 58: 967-74

13 Miller BL. A review of chemical issues in 'H-NMR spectroscopy, N-acetyl-L-aspartate, creatineand choline. NMR Biomed 1991, 4. 47-52

14 Davie CA, Hawkins CP, Barker GJ, Tofts PS, Miller DH, McDonald WI. Serial proton magneticresonance spectroscopy in acute multiple sclerosis lesions. Brain 1994; 117: 49-58

15 Saunders DE, Howe FA, van den Boogaart A, Griffiths JG, Brown MM Discrimination ofmetabolites from lipid and macromolecule resonances in cerebral infarction in humans usingshort echo proton spectroscopy. / Magn Reson Imaging 1997; 7: 1116-21

344 British Medical Bulletin 2000,56 (No 2)

Dow


ic.oup.com/bm



16 Arnold DL, Matthews PM, Francis GS, O'Connor J, Antel JP. Proton magnetic resonancespectroscopic imaging for metabolic characterization of demyehnating plaques Ann Neurol1992, 31 235-41

17 May GL, Wright LC, Holmes KT et al. Assignment of methylene proton magnetic resonancesin NMR spectra of embryonic and transformed cells to plasma triglycende. / Btol Cbem 1986;261 3048-53

18 Behar KL, Rothman DL, Spencer DD, Petroff OAC Analysis of macromolecule resonances in•H-NMR spectra in human brain. Magn Reson Med 1994; 32 294-302

19 Graham GD, Blamire AM, Rothman DL et al Early temporal variation of cerebral metabolitesafter human stroke- a proton magnetic resonance spectroscopy study. Stroke 1993; 24 1891-6

20 Gideon P, Hennksen O, Sperling B et al Early time course of N-acetyl aspartate, creatine andphosphocreatine, and compounds containing choline in the brain after acute stroke. Stroke1992; 11: 1566-72

21 Rothman DL, Howseman A, Graham GD et al Localised proton NMR observation of [3-13C]-lactate in stroke after [l-13C]-glucose infusion. Magn Reson Med 1991, 21: 302-7

22 Houkin K, Kwee EL, Nakada T Persistent high lactate level as a sensitive MR spectroscopyindicator of complete infarction. Brain Pathol 1994; 4- 23—36

23 Gillard JH, Barker PB, van Zijl PCM, Bryan RN, Oppenheimer SM. Proton MR spectroscopyin acute middle cerebral artery stroke. AJNR Am J Neuroradwl 1996; 17: 873-86

24 Williams SR. Cerebral amino acids studied by nuclear magnetic resonance spectroscopy in vivo.Prog Nuclear Magn Reson Sped 1999, 34 301-26

25 De Stafano N, Matthews PM, Arnold DL. Reversible decreases in N-acetyl-aspartate after acutebrain injury Magn Reson Med 1995, 34: 721-7

26 Provencher SW Estimation of metabolite concentrations from localised in vivo proton NMRspectra. Magn Reson Med 1993; 30: 672-9

27 de Beer R, van Ormondt D, Pijnappel WWF, van der Veen JWC Quantitative analysis ofmagnetic resonance signals in the time domain. Isr ] Chem 1988, 28 249-61

28 van der Veen JWC, de Beer R, Luyten PR, van Ormondt D. Accurate quantification of in vivo31P NMR signals using the variable projection method and prior knowledge. Magn Reson Med1988, 6: 92-8

29 Saunders DE, Clifton AG, Brown MB Measurement of infarct size using MRI predictsprognosis in middle cerebral artery infarction Stroke 1995; 26- 2272-6

30 Pereira AC, Saunders DE, Doyle VL et al Measurement of initial N-acetyl aspartate usingmagnetic resonance spectroscopy and initial infarct volume using MR] predicts outcome inpatients with middle cerebral artery territory infarction. Stroke 1999; 30: 1577-82

31 Ford CC, Griffey RH, Marwiywoff NA, Rosenberg GA. Multi-voxel 'H-MRS of stroke.Neurology 1992; 42: 1408-12

32 Wardlaw JM, Marshall I, Wild J, Dennis MS, Cannon J, Lewis SC. Studies of acute ischaemicstroke with proton magnetic resonance spectroscopy Relation between time from onset,neurological deficit, metabolite abnormalities in the infarct, blood flow and clinical outcome.Stroke 1998; 29. 1618-24

33 Pulsinelh WA, Waldman S, Rawhnson D, Plum F. Moderate hyperglycaemia augmentsischaemic brain damage: a neuropathologic study in the rat. Neurology 1982; 32: 1239-46

34 Siemkowicz E, Gjedde A. Post ishcaemic coma in rats: effect of different pre-ischaemic bloodglucose levels on cerebral metabolic recovery after ischaemia. Ada Physiol Scand 1980, 110-225-32

35 Wang PY, Beauchamp NJ, Ulag AM, Barker PB, van Zijl PCM Diffusion negative MR in acutestroke [abstract] In: Proceedings of the Annual Meeting of the American Society ofNeuroradiology. Philadelphia PA: American Society of Neuroradiology, 1997


Dow


ic.oup.com/bm


Documents

MR spectroscopy in stroke - British Medical Bulletin