In hepatic encephalopathy (HE) – a frequent complication of chronic liver dysfunction – the neuro-psychological impairments range from mild deficits in psychomotor and visiopractic abilities to confusion and finally stupor in higher grades due to decreased hepatocellular detoxification and synthesis functions. In HE, specific changes of brain tissue concentrations of myo-Inositol (mI) (decrease), glutamine/glutamate (Glx) (increase) and Choline (Cho) (decrease) have been reported [1]. For spectroscopic detection of Glx, short echo times and highest possible field strength are preferred. In this study, therefore, we used a short echo time PRESS sequence (TE = 30 ms) to achieve high signal of the J-coupling resonances mI and Glx. Of principal interest is the differentiation of glutamine (Gln) and glutamate (Glu) which is generally very difficult to achieve at the standard field strength of 1.5T. This report shows that this differentiation is possible at a whole body unit of 3T (Siemens Trio), as demonstrated in a patient with chronic liver disease and accompanying HE. For comparison, measurements were acquired at 3T and 1.5T (Siemens Sonata) and the spectra were compared with those of a healthy volunteer at 3T and 1.5T.

We examined a 56-year-old man with ethyltoxic liver cirrhosis (Child-Pugh C). The referring hepatologist performed a clinical examination for detection of the HE including a number connection test (NCT) [2] immediately before the MR examinations. Pathological NCT was found with 90 s while the NCT score can be classified as normal provided the time for finishing the test is less than ten times the subjects age in decades [3]. Additionally, the patient suffered from drowsiness as well as elevated ammonia blood levels during the last days in hospital. This made a reliable clinical diagnosis of Hepatic encephalopathy possible.

Two subsequent MRS and MRI examinations of the same patient were performed within one day on a 1.5T whole-body MR System (MAGNETOM Sonata, Siemens AG, Erlangen, Germany) using a conventional circularly polarized head coil and on a 3T whole-body MR System. (MAGNETOM Trio, Siemens AG, Erlangen, Germany) using a standard head coil. For the follow-up examination, exact reproducibility of the spectral localization was of special importance. This was achieved by acquiring 3 orthogonal (sagittal, axial, coronal) gapless heavily T2-weighted (turbo factor 15) turbo spin echo (TSE) data sets within 2 minutes for 3-dimensional positioning of the volume of interest (VOI).

Since it has been shown that occipital gray and white and parietal white matter locations exhibit essentially identical MRS metabolite changes in HE [1], we decided to place the VOI in the medial part of the occipital lobe. This area contains grey and white matter and allows the measurement of a VOI size of 2x2x2 cm3 with good homogeneity of the magnetic field (Fig. 1).

MR imaging consisted of nonenhanced axial T1 weighted (T1w) spin echo (SE) and T2 weighted (T2w) TSE images. Measurement parameters (TR/TE/acquisition time) were 580ms / 12ms / 2min for the T1w and 3000ms / 90ms / 3min for the T2w imaging. Slice thickness SL was 5 mm with a gap of 1 mm.

In chronic liver disease, signal hyperintensities in the basal ganglia on T1w images are found most pronounced in the globus pallidus [4]. They probably represent manganese deposition due to liver dysfunction which is potentially reversible if liver function improves or where there has been a liver transplant [5].

For the detection and quantification of mI and Glx, it is essential to use spectroscopic sequences with a short echo time (TE) to minimize the influence of J-coupling and T2 relaxation (Glx, mI), as both factors lead to a signal decrease.

Therefore 1H-MRS was performed using a single voxel double spin echo sequence (PRESS, TR 3000, TE 30, NA = 64) with a measurement time of 3minutes per voxel. For each voxel a short reference measurement without water suppression was performed for an eddy current correction during postprocessing.

Spectral postprocessing was performed using an optimized LCmodel [6] for absolute quantification. In our LC-model, the basis data set was completed by lipid and macromolecule resonances which are especially important in demyelinating and neoplastic lesions [7]. The results of the LC-model evaluation are depicted in Figs. 2 and 3 for both measurements at 1.5 Tesla and 3 Tesla. The left/right column in Fig. 2 shows the results for 3T/1.5T of the healthy volunteer. The upper row depicts the LC-model fit of the spectra, the middle row the glutamate (Glu) and the lower row the glutamine (Gln) signal. It is clearly shown that normals cerebral concentration of Gln is relatively low [1.9 mMol/l] in comparison to the excitatory neurotransmitter (Glu) [6.5 mMol/l].

Of major importance is the separation of the β- (3.77 ppm) and γ- (2.05-2.45 ppm) resonances of Gln and Glu at 3T due to increased chemicalshift while they are slightly together at 1.5T. This increased chemical-shift allows better separation of Glu and Gln in the LC-model fit at 3T compared to 1.5T.

More prominently, this is depicted in Fig. 3 with the corresponding spectra of the patient with hepatic encephalopathy. The upper row shows again the fit of the LC-model with an HE-typical decrease of choline (Cho) and myo-inositol (mI). Additionally, in comparison with Fig. 2 an elevated glutamine signal can be detected, best seen in the range of the b- and g-resonances. The elevated glutamine concentration [5.3 mMol/l] in the patient leads to a similar signal strength of Glu and Gln. Together with the higher chemical-shift at 3T, proper separation and reliable quantification of glutamine and glutamate is possible with the LC-model evaluation (Fig. 4). Usually this is not possible at 1.5T, leading to frequently used combined evaluation of both metabo-lites as a resonance Glx=Glu+Gln at lower field-strengths.

[ 1 ] Ross B D, Jacobson S, Villamil F, Korula J, Kreis R, Ernst T, et al.
Subclinical hepatic encephalopathy: proton MR spectroscopic abnormalities. Radiology 1994;193:457-463.

[ 2 ] Parsons-Smith B, Summerskill W, Dawson A, Sherlock S. The electroencephalograph in liver disease. Lancet 1957;2:867-871.

[ 3 ] Van Gorp W, Satz P, Mitrushina M.
Neuropsychological processes associated with normal aging. Dev Neurpsychol 1990;6:279-290.

[ 4 ] Brunberg JA, Kanal E, Hirsch W, Van Thiel DH.
Chronic acquired hepatic failure: MR imaging of the brain at 1.5. AJNR Am J Neuroradiol 1991;12:909-914.

[ 5 ] Naegele T, Grodd W, Viebahn R, Seeger U, Klose U, Seitz D, Kaiser S, Mader I, Mayer J, Lauchart W, Gregor M, Voigt K:
MR imaging and 1H spectroscopy of brain metabolites in hepatic encephalopathy: time-course of renormalization after liver transplantation. Radiology 2000; 216:683 – 691

[ 6 ] Provencher S W:
Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magn-Reson-Med. 1993; 30: 672-679

[ 7 ] Seeger U, Klose U, Mader I, Grodd W, Nägele T:
Parameterized evaluation of macromolecules and lipids in proton MR spectroscopy of brain diseases. Magn Reson Med 2003; 49: 19-28

---

Thomas Nägele¹, Uwe Seeger¹, Holger Hass², Wilhelm Küker¹, Stefan Heckl¹, Uwe Klose<sup>1

1</sup> Department of Neuroradiology,
² Department of Gastroenterology and Hepatology Medical Clinic, University of Tübingen