British Journal of Radiology (2003) 76, 483-486
© 2003 British Institute of Radiology
doi: 10.1259/bjr/30212020
Non-invasive study of human gall bladder bile in vivo using 1H-MR spectroscopy
A S K Dzik-Jurasz, PhD, FRCS, FRCR
A P Prescot, BSc, MChem
M O Leach, PhD, FInstP, FMedSci
and
D J Collins, BA, MInstP
The Cancer Research UK Clinical MR Research Group, Institute of Cancer Research, Royal Marsden NHS Trust, Downs Road, Sutton SM2 5PT, UK
Correspondence: Dr A S K Dzik-Jurasz, GlaxoSmithKline Pharmaceuticals, 891–995 Greenford Road, Building 5, Floor 3, Room 13, London UB6 0HE
 |
Abstract
|
|---|
The sampling of gall bladder bile for analytical studies remains an invasive procedure. We demonstrate the application of the non-invasive methodology of 1H-MR spectroscopy to the qualitative and quantitative assessment of human gall bladder bile in vivo. Spectral profiles in vivo are shown in relation to model and porcine gall bladder bile and the quantitation in man of the trimethylamine (choline) and lecithin concentrations were estimated to range from 25.9 mM to 48.4 mM (mean: 35.8 mM, standard deviation: 9.8). The composition of human gall bladder bile together with the quantitation of various constituents can be studied non-invasively in vivo.
|
Introduction
|
|---|
The techniques currently available for sampling choledochal or gall bladder (GB) bile in humans remain invasive [1]. In addition, biliary sampling via choledochal surgical access has diminished considerably since the widespread introduction of endoscopic techniques, whilst serial perioperative sampling of normal GB bile carries considerable ethical and practical restrictions. Nevertheless, the composition of GB bile remains an important determinant of biliary lithogenicity [2], while many classes of drug undergo significant biliary excretion [3]. The development of a non-invasive methodology capable of interrogating the composition of bile in vivo would therefore be of considerable clinical and scientific benefit. In a previous publication we demonstrated the in vivo localization of the bile-acid conjugate of 
-fluoro-ß-alanine in the GB of patients receiving the anti-cancer agent 5-fluorouracil using 19F-MR spectroscopy (MRS) [4]. 1H-MRS on the other hand remains to be developed and applied to the study of GB bile other than for in vitro analytic studies [5, 6]. The aim of the present study was to qualitatively and quantitatively assess the available 1H-MR spectra from in vivo human GB bile.
|
Materials and methods
|
|---|
10 healthy male volunteers (mean age: 31.2 years, range: 24–47 years), fasted overnight, were studied using a 1.5 T Siemens (Erlangen, Germany) Vision system using a 16 cm receive surface coil placed over the surface-markings of the GB. The GB was localized using a T2 weighted image. A PRESS (point-resolved spectroscopy) sequence with echo time (TE)=60 ms and repetition time (TR)=2500 ms was used after automatically shimming the volume of the right upper quadrant to between 6 Hz and 25 Hz. The shortest echo-time available to the sequence (TE=60 ms) was chosen to maintain the signal from short T2 metabolites. A total of 128 individual signal acquisitions were acquired in each study. In addition, an unsuppressed water signal was acquired with four averages from the target voxel. In half the volunteers the exact same sequence was run with the voxel placed within the liver. This was done in order to assess whether there might be spectral hepatic metabolite signals that might contribute to GB spectra as a result of contamination. In order to establish the expected appearance of in vivo GB spectra the identical sequence was run in vitro on model [7] and porcine bile. The composition of porcine bile is very similar to that in humans [8] and acknowledges current difficulties in acquiring normal human bile for study. The model bile contained taurocholic acid (130 mM), cholesterol (13 mM) and phosphotidylcholine (30 mM). On the basis of the appearance of the in vitro spectra, all spectra in humans with a lipid –CH2 (0.9–1.3 ppm) to trimethylamine (TMA; 3.2 ppm) peak height ratio greater than one were excluded. In order to account for phase effects in individual signals arising from patient motion, all spectra were analysed in power mode [9]. Our methodological approach therefore acknowledged and attempted to address the issue of spectral contamination as a result of physiological motion. Spectral analysis was performed from within FELIX (v98.0, Molecular Simulations, Inc, USA). Details of the quantification procedure included fitting a Lorentzian line shape to the biliary TMA and unsuppressed water resonance. As the data were fitted in the power mode, the square root of the model fit signal amplitude was taken prior to peak area calculation. The peak areas were corrected for T1 and T2 relaxation using values determined from porcine bile (TMA/water; T1=471/1230 ms and T2=178/372 ms). A further correction was made to account for the different receiver gain used to acquire the unsuppressed water spectra. Finally, the biliary water concentration was assumed to be 48.3 M (solid fraction=13% w/v) for all estimations of biliary TMA concentration. Approval for the study was granted under the local institutional ethical protocol with all individuals giving written consent.
|
Results
|
|---|
Eight out of the 10 data sets yielded spectra useful for further analysis. A useful data set was therefore one in which at least one free induction decay out of 128 had a lipid to TMA ratio of less than 1. Two data sets gave no signals with a lipid height less than that of the TMA. A typical data set is illustrated in Figures 1 and 2
and compared with the appearance of model and porcine bile. All spectra demonstrated peaks assigned to lipid methyl (0.7–0.9 ppm) and methylene (0.9–1.3 ppm) and TMA (3.2 ppm) in keeping with the published analytical data [5, 6]. Assuming a biliary water concentration of 48.3 M (87% w/v water content of bile) estimates of the concentration of biliary TMA ranged from 25.9 mM to 48.4 mM with a mean of 35.8 mM and standard deviation of 9.8. No TMA resonance was recorded from the voxel positioned in the liver of volunteers. A broad and very low intensity resonance around 1.2 ppm assumed to be due to methylene protons was visible in the hepatic voxel of all individuals scanned. The intensity of this resonance was considerably lower than that recorded in the GB voxel or model/porcine bile.
View larger version (102K):
[in this window]
[in a new window]
|
Figure 1. (a) Transverse, (b) coronal and (c) sagittal T2 weighted images through the upper quadrant of a volunteer. The gall bladder is seen as the bright signal intensity structure at the meeting point of the dashed lines (the dashed lines are orthogonal localizer lines created by default by the system during scanning). A typical spectroscopic volume of interest (2.20 cm3) is shown centrally within the organ. The 1H-MR spectrum derived from this volunteer is given in Figure 2c .
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Figure 2. PRESS 60 1H-MR spectra of (a) model (b) porcine in vitro gallbladder bile and (c) human in vivo gallbladder bile. The various resonances are labelled (TMA, trimethylamine). The peak at 4.7 ppm in (c) is a residual water resonance that has not been fully suppressed during the in vivo experiment.
|
|
|
Discussion
|
|---|
The current results demonstrate that it is possible to record 1H-MR spectra from the human GB that are similar to that expected from in vitro models of human GB bile. The study design anticipated and acknowledged the likely confounding influence of motion on the results. The adoption of a PRESS sequence, although losing some sensitivity to short T2 metabolites by a limitation in minimum echo time is not prone to the influence of phase effects as other single-voxel techniques such as stimulated echo acquisition [10]. Acquiring individual signals allowed the data to be edited according to our proposed methodology including summing data in power mode thereby avoiding signal loss from phase incoherences of individual measurements. Although lipid from surrounding structures might have contaminated the biliary signal, no TMA signal was observed from surrounding structures including the liver. Although chemical shift contamination due to off-resonance effects probably added to the recorded lipid resonance of bile, this could be minimized in future measurements by acquiring the peak of interest on-resonance. Although assumptions were made in the quantitation of the TMA peak (bile water concentration and metabolite relaxation times) the results fall within the expected range of TMA concentrations [11]. Since TMA is part of the polar-head of lecithin, our calculated TMA concentration also gives the lecithin concentration of GB bile. We acknowledge this methodology is currently of limited clinical value but could be a useful tool in the study of diseases or conditions associated with altered bile composition.
The ideal validation of the current approach would involve correlating in vivo 1H-MR spectra with conventional invasive biliary sampling. Nevertheless, the recognized variation in biliary composition and concentration of individual components indicates that future studies should address the normal variation in the 1H-MR spectrum as a function of factors such as prandial state, diet, age and sex. The current study indicates that an important physiological and pharmacological compartment can be qualitatively and quantitatively studied non-invasively.
|
Conflict of interest statement
|
|---|
The authors declare no conflict of interest.
|
Footnotes
|
|---|
Funding from Cancer Research UK grant SP1780/0103 is gratefully acknowledged. 
Received for publication July 9, 2002.
Revision received March 12, 2003.
Accepted for publication April 9, 2003.
|
References
|
|---|
- Shah SK, Mutignani M, Costamagna G. Therapeutic biliary endoscopy. Endoscopy 2002;34:43–53.[CrossRef][Medline]
- Dowling RH. Review: pathogenesis of gallstones. Aliment Pharmacol Ther 2000;14 Suppl 2:39–47.
- Siegers CP, Watkins III JB. Biliary excretion of drugs and other chemicals. Stuttgart, Germany: Gustav Fischer; 1991.
- Dzik-Jurasz AS, Collins DJ, Leach MO, Rowland IJ. Gallbladder localization of (19)F MRS catabolite signals in patients receiving bolus and protracted venous infusional 5-fluorouracil. Magn Reson Med 2000;44:516–20.[CrossRef][Medline]
- Sequeira SS, Parkes HG, Ellul JP, Murphy GM. In vitro determination by 1H-NMR studies that bile with shorter nucleation times contain cholesterol-enriched vesicles. Biochim Biophys Acta 1995;1256:360–6.[Medline]
- Melendez HV, Ahmadi D, Parkes HG, Rela M, Murphy G, Heaton N. Proton nuclear magnetic resonance analysis of hepatic bile from donors and recipients in human liver transplantation. Transplantation 2001;72:855–60.[Medline]
- Groen AK, Goldhoorn BG, Egbers PH, Chamuleau RA, Tytgat GN, Bovee WM. Use of 1H-NMR to determine the distribution of lecithin between the micellar and vesicular phases in model bile. J Lipid Res 1990;31:1315–21.[Abstract]
- Chodash HB, Tsang TK, Pollack JM, Eisenman RE, Rege RM, Ostrow JD. Effect of temperature on stability of eight components of porcine gallbladder bile. Dig Dis Sci 1997;42:273–8.[CrossRef][Medline]
- Ziegler A, Decorps M. Signal-to-noise improvement in in vivo spin-echo spectroscopy in the presence of motion. J Magn Reson 1993;Series B, 102:26–34.
- Burtscher IM, Johansson E, Holtas S, Stahlberg F. Quality assessment of localization technique performance in small volume in vivo 1H MR spectroscopy. Magn Reson Imaging 1999;17:1511–9.[Medline]
- Bovee WMMJ, Groen AK. Study of the distribution of biliary lipids between the vesicular and micellar phase in model bile. In: de Certaines JD, editor. Magnetic resonance spectroscopy of biofluids. Singapore: World Scientific Publishing Co Pte Ltd; 1989;199–216.
Top
Abstract
Introduction
