Published online before print October 2, 2003, 10.1148/radiol.2292021156
(Radiology 2003;229:587-592.)
© RSNA, 2003
Human Gallbladder Bile: Noninvasive Investigation in Vivo with Single-Voxel 1H MR Spectroscopy1
Andrew P. Prescot, MChem,
David J. Collins, BA,
Martin O. Leach, PhD and
Andrzej S. K. Dzik-Jurasz, PhD, FRCS, FRCR
1 From the Cancer Research UK Clinical Magnetic Resonance Research Group, Institute of Cancer Research, Surrey, England (A.P.P., M.O.L., A.S.K.D.J.); and Royal Marsden National Health Service Trust, Sutton, England (D.J.C.). Received September 9, 2002; revision requested November 18; revision received January 6, 2003; accepted February 25. Supported by Cancer Research UK grant SP1780/0103 and Medical Research Council grant GU 78/6400. Address correspondence to A.S.K.D.J., GlaxoSmithKline Pharmaceuticals, Building 5, Floor 1, Room 13, 890–995 Greenford Rd, London UB6 0HE, England (e-mail: andrzej.s.dzik-jurasz@gsk.com).
 |
ABSTRACT
|
|---|
Proton (hydrogen 1) magnetic resonance (MR) spectroscopy was used to study model and porcine bile in vitro. The method was subsequently developed to facilitate the acquisition of in vivo 1H MR spectra from the gallbladder bile of 10 human volunteers. Signals attributable to phosphotidylcholine and conjugated bile acid protons were observed in eight of the 10 volunteers. Phosphotidylcholine concentrations were estimated, and five values (mean = 35.8 mmol/L, SD = 9.8) were within the expected range of levels in human bile. Findings in this preliminary investigation indicate that human gallbladder bile can be qualitatively and quantitatively studied noninvasively with 1H MR spectroscopy.
© RSNA, 2003
Index terms: Bile ducts, MR, 768.12145 • Gallbladder, MR, 762.12145
|
INTRODUCTION
|
|---|
The chemical composition of gallbladder bile is an important determinant of gallstone formation (1). Also, many classes of drugs are excreted in the bile (2). The sampling of bile for chemical analysis can be useful in the determination of causes and treatment of biliary stones and in the evaluation of biliary excretion of drugs. Bile samples can be obtained directly from the biliary tree during surgery, from biliary drainage tubes, or from direct needle aspiration, but an invasive procedure is required (3). Development of a noninvasive technique that allows determination of the chemical composition of gallbladder bile in vivo would be of clinical and scientific interest. In a previous publication, Dzik-Jurasz et al (4) used fluorine 19 MR spectroscopy to demonstrate in vivo localization of bile acid conjugate of 
-fluoro-ß-alanine in the gallbladder of patients receiving the cytotoxic agent 5-fluorouracil. To our knowledge, hydrogen 1 magnetic resonance (MR) spectroscopy has been used only for detection and identification of gallbladder bile metabolites in vitro (5–8). However, the application of localized 1H MR spectroscopy to the study of human gallbladder bile in vivo is not straightforward, primarily because gallbladder motion throughout data acquisition results in substantial peak attenuation and potential spectral contamination from surrounding lipid. Thus, the purpose of our study was to develop a 1H MR spectroscopic method to noninvasively study human gallbladder bile in vivo.
|
Materials and Methods
|
|---|
Data Acquisition
All in vitro and in vivo 1H MR spectroscopic examinations were performed by the investigators with a 1.5-T whole-body MR imager (Vision; Siemens Medical Systems, Erlangen, Germany). The body coil was used for radiofrequency transmission, and a commercially available circularly polarized 16-cm surface coil was used for radiofrequency reception. A point-resolved spectroscopic (PRESS) sequence was used to acquire the 1H spectra (9). A conventional section-select sinc 90° radiofrequency pulse of 5.12-msec duration was used for signal excitation. The section-select 180° refocusing radiofrequency pulses were optimized with the algorithm described by Mao et al (10). The duration of each refocusing radiofrequency pulse was 10.24 msec, and both were executed in the presence of a linear magnetic field gradient (1.6 mT/m). The refocusing section-select gradient amplitude was chosen to maximize the voxel dimensions without exceeding the transmitter voltage capability of the body coil. To minimize the chemical shift artifact, the excitation section-select gradient amplitude was set to 4.0 mT/m. The spatial displacements of lipid calculated for the excitation and refocusing pulses were 1.3 and 3.2 mm, respectively. The static magnetic field (B0) crusher gradient pulses were placed symmetrically around each 180° radiofrequency pulse to destroy unwanted coherences produced by any radiofrequency pulse imperfections. These crusher pulses were placed on an axis orthogonal to the section-select direction to suppress the stimulation of zero-quantum coherences in J-coupled species. Water suppression was achieved with three identical frequency-select Gaussian radiofrequency pulses, each followed by strong B0 gradient pulses (11). The duration of each Gaussian pulse was 25.6 msec, which provided one excitation bandwidth of 60 Hz.
An automated localized shimming routine was used to optimize B0 homogeneity over the region of interest, and the full width at half maximum of the water signal was between 6 and 25 Hz in all cases. The carrier frequency was positioned at the water frequency (63.6 MHz) for all examinations. Optimization of the applied radiofrequency field (B1) was achieved by varying the radiofrequency transmitter reference voltage until a maximum fully relaxed unsuppressed localized water signal was obtained from the spectroscopic voxel.
All PRESS spectra were recorded with a repetition time of 1,500 msec, and the shortest echo time (60 msec) available to the sequence was selected to minimize signal losses due to T2. The raw data were acquired with 512 data points sampling a spectral bandwidth of 1,000 Hz. One hundred twenty-eight individual free-induction decays were acquired at each study, which gives a measurement time of 3.2 minutes.
In Vitro Measurements
Six 1H MR spectroscopic measurements were performed in vitro. A nominal voxel size of 3.4 mL was selected for all in vitro 1H MR spectroscopic measurements. Spectra were recorded initially from three phantoms (each 50 mL), which contained model bile, porcine bile, and an aqueous solution of taurocholic acid (130 mmol/L, pH 8.0). The model bile solution was prepared according to the technique of Groen et al (8), with sodium phosphate buffer (pH 8.0) in preference to tris-buffer to avoid additional proton signals in the MR spectrum. The composition of porcine bile is similar to that of human bile (6). The quantification procedure was validated by acquiring 1H MR spectra from the three phantoms, which contained known concentrations of phosphorylcholine (9.6, 25.1, and 49.4 mmol/L, respectively; pH 8.2). Unsuppressed water spectra were acquired from the target voxel (four signals acquired) positioned in the phosphorylcholine phantoms.
In Vivo Measurements
Ten healthy male volunteers (mean age, 31 years; age range, 24–47 years) fasted overnight and then were examined. None had a medical history of hepatobiliary disease, and physical examination was unremarkable. Written informed consent was obtained from all individuals before MR examination. The local institutional ethics committee approved the protocol used in this study. The gallbladder was localized with three orthogonal MR images acquired with a T2-weighted true fast imaging with steady-state precession sequence (repetition time msec/echo time msec of 6.32/3.0, flip angle of 70°, and four signals acquired). The images were subsequently used to position the spectroscopic voxel (mean, 3.3 mL; range, 2.2–4.1 mL) within the boundaries of the gallbladder. Free (quiet) breathing was continuous throughout the in vivo 1H MR spectroscopic measurements. In addition, an unsuppressed water spectrum was acquired from the target voxel (four signals acquired). Total measurement time for each examination was between 30 and 45 minutes.
Data Postprocessing
Spectral analyses were performed off line (FELIX 98; Accelrys, San Diego, Calif). In general, each free-induction decay was zero filled once and multiplied by an exponential apodization function (line broadening of 2 Hz). After Fourier transformation, the individual spectra were referenced to the large phosphotidylcholine singlet resonance at 3.2 ppm and presented in the power mode. The in vitro 1H MR spectra were constructed by averaging the total selected individual power spectra.
On the basis of the appearance of the in vitro data, all individual power spectra in humans with a lipid saturated methylene proton signal (-CH2-, 1.2 ppm) to phosphotidylcholine trimethylamine proton peak (-N+[CH3])3, 3.2 ppm) height ratio greater than 1 were excluded. All other power spectra were retained unscaled for evaluation. The signal-to-noise ratio of the individual spectra was sufficient to perform this analysis in most individuals and provided justification for rejection of individual spectra. The resulting mean power spectrum (comprising n summed power spectra following rejection of spectra) was used to quantify biliary phosphotidylcholine.
Quantification of Biliary Phosphotidylcholine
A Lorentzian line shape was fitted to the biliary phosphotidylcholine trimethylamine signal (3.2 ppm) and to the unsuppressed water signal (4.7 ppm) with the simulated annealing algorithm (FELIX 98; Accelrys). As the data were fitted in the power mode, the square root of the model fit signal amplitude was obtained prior to peak area calculation. Peak areas were corrected for T1 and T2, with values determined from porcine bile (T1 of phosphotidylcholine, 471 msec; T1 of water, 1,230 msec; T2 of phosphotidylcholine, 178 msec; T2 of water, 372 msec). A further correction was made to account for the different receiver gains used to acquire the unsuppressed water spectra. The biliary water concentration was assumed to be 48.3 mol/L (solid fraction of 13% wt/vol) for all estimations of biliary phosphotidylcholine concentration. Phosphorylcholine was used to validate the quantification method owing to the improved water solubility and its inherent trimethylamine proton resonance.
|
Results
|
|---|
The 1H MR spectrum recorded from the model bile preparation is presented in Figure 1, A. The peaks were assigned according to chemical shift values on the basis of previously reported 1H MR spectroscopic studies performed on bile (5–8). The phosphotidylcholine proton resonances were assigned as follows: The choline head group trimethylamine singlet resonance was at 3.2 ppm, and the relatively broad signal at 1.2 ppm was assigned to the saturated -CH2- protons of the phosphotidylcholine fatty acid moieties. An extra peak was seen as a shoulder on the phosphotidylcholine trimethylamine resonance at 3.3 ppm. The 1H MR spectrum recorded from porcine bile is shown in Figure 1, B. The similarity between this spectrum and the model bile spectrum is striking. Compared with the model bile spectrum, however, an extra peak is observed at 3.8 ppm.
View larger version (14K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 1. A, PRESS 1H MR spectrum recorded from model bile. B, PRESS 1H MR spectrum recorded from porcine bile. C, PRESS 1H MR spectrum recorded from taurocholic acid phantoms. All spectra were constructed with 128 individually processed power spectra. Spectra have arbitrary scaling. 1 = phosphotidylcholine trimethylamine (-N+[CH3]3), 2 = phosphotidylcholine (-CH2-), 3 = conjugated bile acid -CH3 protons.
|
|
The 1H MR spectrum recorded from the taurocholic acid phantom is presented in Figure 1, C. The methyl (-CH3) protons of the cholic acid moiety are clearly seen at 0.6 and 0.8 ppm. These assignments are readily translated to the model bile and porcine bile spectral data. However, the cyclic methylene protons of taurocholic acid (1.0–2.5 ppm) are not observed. The peak at 3.1 ppm corresponds to the upfield -CH2- protons of the taurine moiety.
A typical spectrum recorded from the gallbladder of a volunteer is illustrated in Figure 2, together with the MR images used to determine the voxel coordinates. The gallbladder appears as a hyperintense structure in the true fast imaging with steady-state precession MR images as a result of the relatively long T2 of biliary water. Use of the postprocessing method described earlier resulted in the exclusion of 61 of the 128 power spectra. The metabolite peaks in the spectrum are similar to those in the model bile and porcine bile 1H MR spectra. The importance of acquiring nonaveraged free-induction decays is demonstrated in Figure 3, which shows two of the 128 individually processed spectra from the same individual. While the first spectrum is within the limits of our postprocessing criteria, the second spectrum is clearly contaminated by lipid and was excluded.
View larger version (67K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 2. T2-weighted true fast imaging with steady-state precession MR images (6.32/3.0, flip angle of 70°, and four signals acquired) in a 27-year-old volunteer show typical placement of PRESS voxel in gallbladder. A, Transverse MR image. B, Coronal MR image. C, Sagittal MR image. D, Corresponding 1H MR spectrum constructed with 67 individually processed power spectra. Spectrum has arbitrary scaling. 1 = phosphotidylcholine trimethylamine (-N+[CH3]3), 2 = phosphotidylcholine (-CH2-), 3 = conjugated bile acid -CH3 protons, 4 = residual water.
|
|
View larger version (19K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 3. Two of 128 individually processed 1H MR spectra for volunteer in Figure 2. A, Individual spectrum fulfills postprocessing criteria and contributes to final spectrum. B, Individual spectrum shows lipid contamination, which results in its exclusion from spectral averaging. Spectra have arbitrary scaling. 1 = phosphotidylcholine trimethylamine (-N+[CH3]3), 2 = phosphotidylcholine (-CH2-), 3 = conjugated bile acid -CH3 protons, 4 = residual water.
|
|
A more extreme example is illustrated in Figure 4, which shows 1H MR spectra recorded from the gallbladder of a second volunteer. The first of the two spectra was generated with all 128 individual power spectra and shows the biliary phosphotidylcholine trimethylamine proton resonance to be clearly dominated by contaminant lipid components. The second spectrum, however, was constructed from only the 11 power spectra that satisfied the postprocessing criteria. The resulting spectrum shows a substantial reduction of contaminant lipid signal and is qualitatively similar to the model and porcine bile 1H MR spectra.
View larger version (19K):
[in this window]
[in a new window]
[Download PPT slide]
|
Figure 4. PRESS 1H MR spectra from the gallbladder of a 24-year-old volunteer. A, Power spectrum shows all 128 free-induction decays used to construct it. B, Power spectrum illustrates that 11 of the individual power spectra showed a phosphotidylcholine trimethylamine resonance greater than the lipid signal and were averaged to produce it. Spectra have arbitrary scaling. 1 = phosphotidylcholine trimethylamine (-N+[CH3]3), 2 = phosphotidylcholine (-CH2-), 3 = conjugated bile acid -CH3 protons, 4 = residual water.
|
|
Eight of the 10 in vivo data sets yielded spectra suitable for quantitative analysis. Two of the 10 data sets gave no free-induction decays with a detectable phosphotidylcholine trimethylamine proton peak. Estimates of biliary phosphotidylcholine concentrations are presented in the Table, with the number of power spectra included for signal averaging for each volunteer. Concentration estimates calculated for phosphorylcholine samples are also shown to demonstrate the validity of the quantification procedure.
|
Discussion
|
|---|
Findings in this study demonstrate that the single-voxel 1H MR spectroscopic method has the potential to noninvasively interrogate the metabolic composition of human gallbladder bile in vivo. In agreement with previous 1H MR spectra of human gallbladder bile, spectral peaks that arise from phosphotidylcholine and conjugated bile acid protons are observed in the model bile and in vivo 1H MR spectra. For a variety of reasons, the number of peaks detected from these biliary components is substantially reduced, primarily owing to the low sensitivity and spectral resolution associated with the performance of in vivo 1H MR measurements at 1.5 T. Second, in contrast to the pulse-and-acquire experiments used routinely in analytic 1H MR spectroscopic studies, conventional single-voxel 1H MR spectroscopic pulse sequences incorporate one or more echo time periods. Hence, further signal losses due to T2 attenuation are expected.
Therefore, the PRESS sequence was adopted to maximize the signal obtained from the sampled volume and because of its insensitivity to phase effects compared with other single-voxel techniques, such as stimulated echo acquisition mode (12). Furthermore, numerically optimized 180° radiofrequency pulses were implemented to minimize additional signal loss due to incomplete spin refocusing. The minimum echo time (60 msec) permitted with the MR imager was used to reduce further losses of short T2 metabolite signals. This choice of echo time, however, leads to substantial peak suppression of J-coupled spins (where J equals approximately 7 Hz) where their coherence states are approaching antiphase at the start of signal acquisition. This explains why the cyclic methylene protons of taurocholic acid are suppressed in the corresponding 1H MR spectrum. For the same reason, the -CH2- protons of the taurine moiety are substantially attenuated compared with the -CH3 signals. Moreover, the downfield taurine -CH2- peak (3.5 ppm) is completely eliminated from the 1H MR spectrum. For this reason, the resulting 1H MR spectra are further simplified and consist predominantly of singlet resonances. The additional resonance seen at 3.3 ppm in the model bile 1H MR spectrum is probably due to phosphotidylcholine occurring in a different chemical microenvironment. An extra peak was also observed in the porcine bile spectrum (3.8 ppm), which has previously been assigned to the glycine -CH2- moiety of glycocholic acid (7). Future studies include the development and similar application of the PRESS method with a lower echo time (<30 msec) to investigate short T2 metabolites and J-coupled species in human gallbladder bile.
Detection and identification of biliary micellar cholesterol have been demonstrated previously in analytic 1H MR spectroscopic studies with olefinic and -CH3 proton resonances (13). In a number of studies with conflicting results, however, these cholesterol signals have not been detected (7,14). In the present study, a resolved signal peak could not be assigned to cholesterol. If the cholesterol -CH3 protons were detectable, their identification would be extremely difficult due to coincident chemical shifts with the bile acid -CH3 resonances. Furthermore, the frequency-select water-suppression module together with the effect of J modulation may lead to attenuation of the cholesterol olefinic proton peak downfield of the water signal.
We recognized that the macroscopic motion associated with the human gallbladder in vivo produces motion-induced signal phase shifts. These phase variations are a major source of peak attenuation and contamination where conventional averaging of free-induction decays is performed. Therefore, the acquisition of individual free-induction decays followed by the rejection of contaminated data and the concomitant addition of individual power spectra were used to eliminate this motion artifact (15). It should be noted that this method retains the natural signal line widths. A second advantage of acquiring separate free-induction decays is that it enables the inspection of individual spectra prior to signal averaging. Therefore, with apriori assumptions based on well-defined in vitro models, the data were edited to minimize lipid contamination due to motion. The application and effectiveness of a suitable criterion for minimizing contaminant lipid signal has been proposed and illustrated in two in vivo measurements. In the first example, approximately 50% of the power spectra satisfied the postprocessing criteria, whereas in the second example only 10% of the acquired data were used to construct the final 1H MR spectrum. In the latter example, the data were also presented with all 128 power spectra to provide an indication of how online averaging would have given rise to severely contaminated data. Individual 1H MR spectra with severe lipid contamination were observed in all individuals. The number of spectra used to construct the in vivo data sets showed a degree of variability indicative of significant interpatient differences (eg, extent of respiration-induced gallbladder motion).
The data processing methods yielded eight in vivo 1H MR spectra that showed close similarity with the model bile data and that were suitable for further quantitative analysis. Two examinations gave 128 individual 1H MR power spectra that showed dominating lipid signal. However, accurate voxel positioning in the gallbladder was difficult in both volunteers due to poorly distended gallbladders (which remained so in both men at a subsequent visit). Therefore, severe lipid contamination due to gallbladder motion was unavoidable. The phosphotidylcholine trimethylamine singlet resonance was chosen for quantification because of its simple line-shape structure, discrete chemical shift, and known concentration range in human gallbladder bile. Although assumptions were made for the concentration estimates of phosphotidylcholine (water concentration and bile metabolite relaxation times), five of the values are within the expected range of phosphotidylcholine levels in human bile (mean, 35.8 mmol/L; SD, 9.8) (16). However, the remaining three concentrations were found to be high. In these three cases, contaminant lipid signals were observed in the water reference data. As a consequence, the gallbladder water concentration was represented by a substantially lower water signal integral, which in turn gave rise to overestimation of the phosphotidylcholine concentrations. To circumvent this problem, we strongly recommend the acquisition of at least 32 individual water reference spectra and the inspection of each spectrum for lipid contamination. We acknowledge that the quantification of the bile acid components with similar methods would be difficult as a result of coincident chemical shifts (primarily with phosphotidylcholine fatty acid methylene protons) and signal that may arise as a result of the postprocessing criteria. Furthermore, the exact species of conjugated bile acid (taurine or glycine derivative) would need to be determined before any attempt at quantification. This may be facilitated by using a shorter echo time (ie, <60 msec) and acquiring the data at a higher B0 field strength.
In summary, an acquisition strategy and data processing method has been developed for acquiring 1H MR spectra from the human gallbladder in vivo. The method accounts for gallbladder motion and allows spectral editing to alleviate the problem of lipid contamination. The ideal validation of the current approach would involve the correlation of in vivo 1H MR spectra with results at conventional invasive biliary sampling. Future studies will also address the advantages with low echo times with an improved water suppression module. Normal variation in the 1H MR spectrum as a function of diet, prandial state, age, and sex should also be determined. Finally, findings in this study indicate that the composition of human gallbladder bile can be qualitatively and quantitatively studied noninvasively with 1H MR spectroscopy.
|
ACKNOWLEDGMENTS
|
|---|
The authors thank Geoffrey Payne, PhD (Cancer Research UK Clinical Magnetic Resonance Research Group) for his assistance with implementation of optimized radiofrequency pulses and all the volunteers who participated in this study.
|
FOOTNOTES
|
|---|
Abbreviation: PRESS = point-resolved spectroscopy
Author contributions: Guarantor of integrity of entire study, A.S.K.D.J.; study concepts, A.S.K.D.J.; study design, D.J.C., A.P.P.; literature research, D.J.C., A.P.P., A.S.K.D.J.; clinical studies, A.P.P., A.S.K.D.J.; experimental studies, A.P.P.; data acquisition, A.P.P., A.S.K.D.J.; data analysis/interpretation, all authors; manuscript preparation, A.P.P.; manuscript definition of intellectual content and editing, A.S.K.D.J., D.J.C., M.O.L.; manuscript revision/review, all authors; manuscript final version approval, A.S.K.D.J., D.J.C., M.O.L.
|
REFERENCES
|
|---|
- Dowling RH. Review: pathogenesis of gallstones. Aliment Pharmacol Ther 2000; 14(suppl 2):39-47.
- Siegers CP, Watkins JB, III. Biliary excretion of drugs and other chemicals Stuttgart, Germany: Gustav-Fischer, 1991.
- Shah SK, Mutignani M, Costamagna G. Therapeutic biliary endoscopy. Endoscopy 2002; 34:43-48.[CrossRef][Medline]
- Dzik-Jurasz ASK, Collins DJ, Leach MO, Rowland IJ. Localisation of 19F MR spectroscopy catabolite signals in patients receiving bolus and protracted venous infusional 5-fluorouracil. Magn Reson Med 2000; 44:516-520.[CrossRef][Medline]
- Halvorsen RA, Ribeiro A, Blinder R, Waters C, Thompson W. Magnetic resonance spectroscopy of bile: in vitro stability over time and component identity. Invest Radiol 1989; 24:903-908.[CrossRef][Medline]
- Chodash HB, Tsang TK, Pollack JM, Eisenman RE, Rege RM, Ostrow JD. Effect of temperature on stability of eight components of porcine bile. Dig Dis Sci 1997; 42:273-278.[CrossRef][Medline]
- Ishikawa H, Nakashima T, Inaba K, et al. Proton magnetic resonance assay of total and taurine-conjugated bile acids in bile. J Lipid Res 1999; 40:1920-1924.[Abstract/Free Full Text]
- 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; 31:1315-1321.[Abstract]
- Bottomley PA. Spatial localization in NMR spectroscopy in-vivo. Ann N Y Acad Sci 1987; 508:333-348.[Medline]
- Mao J, Mareci TH, Andrew ER. Experimental study of optimal selective 180-degree radiofrequency pulses. J Magn Reson 1988; 79:1-10.
- Haase A, Frahm J, Hanicke W, Matthaei D. 1H NMR chemical shift selective (CHESS) imaging. Phys Med Biol 1985; 30:341-344.[CrossRef][Medline]
- Frahm J, Bruhn H, Gyngell ML, Merboldt KD, Hanicke W, Sauter R. Localised high-resolution proton NMR spectroscopy using stimulated echoes: initial application to human brain in-vivo. Magn Reson Med 1989; 9:79-93.[Medline]
- Ellul JPM, Murphy GM, Parkes HG, Slapa RZ, Dowling RH. Nuclear magnetic resonance spectroscopy to determine micellar cholesterol in human bile. FEBS Lett 1992; 300:30-32.[CrossRef][Medline]
- Graaf MP, Groen AK, Bovee WMMJ. Analysis of micellar and vesicular lecithin and cholesterol in model bile using 1H- and 31P-NMR. MAGMA 1995; 3:67-75.
- Ziegler A, Decorps M. Signal-to-noise improvement in in-vivo spin-echo spectroscopy in the presence of motion. J Magn Reson Series B 1993; 102:26-34.[CrossRef]
- Bovee WMMJ, Groen AK. Study of the distribution of biliary lipids between the vesicular and micellar phase in model bile In: de Certaines JD, ed. Magnetic resonance spectroscopy of biofluids. Singapore: World Scientific, 1989; 199-216.
This article has been cited by other articles:
|
|
|
|
 
A Dzik-Jurasz
The development and application of functional nuclear magnetic resonance to in vivo therapeutic anticancer research: 2002 Sir Godfrey Hounsfield lecture delivered at the President's Day, Manchester
Br. J. Radiol.,
April 1, 2004;
77(916):
296 - 307.
[Abstract]
[Full Text]
[PDF]
|
|
|
TOP
ABSTRACT
INTRODUCTION
