Metabolic Characterization of Human Soft Tissue Sarcomas in … · olism of sarcomas in vivo is to use 31P MRS3. The 3IP MR spectrum has signals from phospholipid metabolites, NTPs

[CANCER RESEARCH 56. 2964-2972. July I. 1996]

Metabolic Characterization of Human Soft Tissue Sarcomas in Vivo and in VitroUsing Proton-decoupled Phosphorus Magnetic Resonance Spectroscopy1

Chun-Wei Li, Annette C. Kuesel, Kristin A. Padavic-Shaller, Joseph Murphy-Boesch, Burton L. Eisenberg,

Richard G. Schmidt, Reinhard W. von Roemeling, Arthur S. Patchefsky, Truman R. Brown, andWilliam G. Negendank2

Departments of Nuclear Magnetic RÃ©sonanceand Medical Spectroscopy Â¡C-W.L. A. C. K.. K. A. P-S.. J. M-B., T. R. B.. W. G. N.J. Surgical Oncology Â¡B.L E.. R. G. SJ. Medical

Oncology Â¡R.W. V. R.Â¡.ami Pathology Â¡A.S. P.], Fox Chase Cancer Center. Philadelphia. Pennsylvania Will

ABSTRACT

We applied 'H-decoupling and nuclear Overhauser enhancement toobtain well-resolved "I' magnetic resonance spectra accurately localized

to 20 soft tissue sarcomas in vivo, using three-dimensional chemical shift

imaging. Fifteen spectra had large phosphomonoester signals (21% oftotal phosphorus) that contained high amounts of phosphoethanolamine(compared to those of phosphocholine) but no signals from glycerophos-

phoethanolamine, and glycerophosphocholine was detected in only fourcases. Prominent nucleoside triphosphates (52% of phosphorus) and lowinorganic phosphate (10% of phosphorus) indicated that a large fractionof these 15 sarcomas contained viable cells, and this impression wasconfirmed histologically in 13 of the sarcomas. High-resolution in vitro "I'

spectra of extracts of surgical specimens of four of the sarcomas studied invivo and six additional sarcomas confirmed the in vivo assignments ofmetabolites and revealed considerable inter- and intratumoral variations

of metabolite concentrations associated with histolÃ³gica) variations in therelative amounts of cells and of matrix materials or spontaneous necrosis.Seven sarcomas, all high grade with pleomorphic or round cells ratherthan spindle cells, contained an unidentified phosphodiester signal in vivn;its absence in the extract spectra indicates that it may be from an abnormally mobile membrane component. We have documented a means toobtain new information about in vivo metabolism in human sarcomas andto provide a basis on which to examine the uses ol "I1 magnetic resonance

spectroscopy in the clinical management of sarcomas.

INTRODUCTION

Observations that indicate the importance of aspects of energymetabolism and phospholipid metabolism in malignant behavior,treatment sensitivity, and resistance both in experimental models (1.2) and in patients (3-7) have stimulated interest in the biochemistry of

sarcomas. An attractive way to obtain information about the metabolism of sarcomas in vivo is to use 31P MRS3. The 3IP MR spectrum

has signals from phospholipid metabolites, NTPs. and other energy-

related metabolites, and it provides a means to measure intracellularpH. The "P MR spectra of approximately 100 human soft tissue

sarcomas in vivo have been reported (1, 3-6, 8-16). In general, these

spectra had relatively strong signal intensities in the PME and PDEregions and an intracellular pH (determined from the position of the PÂ¡signal on the frequency axis) of approximately 7.25. However, mostof the reported spectra were incompletely localized to the sarcomas:(a) many were heavily contaminated with signals from muscle; (b)

Received 4/17/96: aecepled 4/24/96.The cosls of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by NIH Grants CA56960. CA54339. and CA41078 and by Siemens

Medical Systems (Iselin, NJ).2 To whom requests for reprints should be addressed, at Department of Nuclear

Magnetic Resonance and Medical Spectroscopy, Fox Chase Cancer Center, 7701Burholme Avenue. Philadelphia. PA 19111

' The abbreviations used are: MRS. magnetic resonance spectroscopy: NMR, nuclear

magnetic resonance; MRI. magnetic resonance imaging: PME. phosphomonoester: PDE.phosphodiester; NTP. nucleoside triphosphate: PCr. phosphocreatine; PEth. phosphoethanolamine: PChol. phosphocholine: GPEth. glycerophosphoethanolamine; GPChol. glycerophosphocholine; NOE. nuclear Overhauser effect: CSI. chemical shift imaging: NDP.nucleoside diphosphate; MR. magnetic resonance: RIF-1. radiation-induced fibrosarcoma.

many had insufficient resolution to clearly distinguish overlappingPME. PÂ¡,and PDE signals: and (c) none had sufficient resolution todistinguish the major components within the PME and PDE regions.

Two factors contribute to the poor resolution of metabolites in the3'P MR spectra: (a) the inhomogeneity of the magnetic field within

the region of interest. The adequacy of magnetic field homogeneitydepends upon the extent of the efforts devoted to shimming procedures: and (b) the broadening of the 31P signal peaks by couplingbetween magnetic fields of 3'P nuclei and those of neighboring 'H.This effect may be reduced by radio frequency irradiation of 'Hduring the acquisition of 3'P signals, a technique referred to as'H-decoupling. In addition, the irradiation of 'H between acquisitionscan increase some of the 3IP signal intensities by a process called

NOE enhancement.We have recently applied the combination of 'H-decoupling and

full NOE enhancement in vivo to 3IP MRS studies of brain, calfmuscle, liver, and non-Hodgkin's lymphomas (17-20). We report here

the use of this technique, in conjunction with the means to optimizethe magnetic field homogeneity automatically within the region ofinterest (autoshimming; Ref. 21) to improve the resolution within thePME and PDE regions of the spectrum in patients with sarcomas. Weused MRI-directed, three-dimensional CSI to accurately localize 31P

MR spectra to the regions of interest (22). To permit application ofthese techniques in various anatomic sites, we constructed dual-tuned(3IP and 'H) surface coil arrangements. This approach enabled us to

obtain more information about the in vivo metabolic characteristics ofsoft tissue sarcomas than has heretofore been available.

Soft tissue sarcomas are heterogeneous histologically. Many contain large amounts of matrix materials of a fibroid, lipoid. or myxoidcharacter, and many undergo spontaneous focal necrosis. These factors can reduce the fraction of viable cells within a region studied byMRS and therefore reduce the metabolite signals relative to the noisein the spectrum. This could account in part for the considerablevariations in quality reported among in vivo 3IP spectra in human

sarcomas (1, 3-6, 8-16). To examine this issue, we determined the

histopathological characteristics of the sarcomas that were studied invivo before surgery, and we obtained high-field "P MR spectra of the

water-soluble extracts of surgical specimens of soft tissue sarcomas,

some of which were also studied in vivo. This approach enabled us toconfirm the in vivo assignments of metabolites to observed signals, toobtain molar concentrations of the metabolites, and to examine theaspects of inter- and intratumoral heterogeneity.

PATIENTS AND METHODS

Patient Population. Eligibility for the in vivo 3IP MRS study required a

biopsy-proven diagnosis of soft-tissue sarcoma, a tumor mass ot approximately3-cm diameter or larger located within 10 cm of the surface of the body, an

absence of the standard contraindications to MRI. and signed informed consentas approved by the Institutional Review Board. Of 20 patients. 16 were newlydiagnosed. 3 were recurrent after prior treatment, and 1 was resistant tochemotherapy. From four of these patients and six additional patients, portionsof surgical specimens were obtained and extracted for in vitro "P MRS study.

2964

on June 30, 2021. © 1996 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

"P MRS OF SARCOMAS IN VIVO AND IN VITRO

Characteristics of the patients and their sarcomas are summarized in Table 1,where they are lettered a-z and referred to as such throughout the manuscript.

Histopathological Analysis. Surgical specimens were obtained in 23 of the26 cases; most of these were from resections performed after the "P MRS

study. In one additional case (case k), only a needle aspiration biopsy wasobtained. In another case (case g), the recurrent tumor was not biopsied, andthe original resection was several years old. In case m, the resection (whichshowed >95% necrosis) was performed 3 months after the patient receivedneoadjuvant chemotherapy, and the original biopsy was not available from anoutside institution. The categorization of each sarcoma followed current WHOtyping (23). The grade was determined based on the tumor differentiation,mitosis count, and extent of necrosis (24). The predominant cell type wasdesignated as spindle, pleomorphic, round, or giant cell. The relative amountsof cells and noncellular materials were noted, as was the variability of thisfeature within the tumor. The extent and nature of matrix (myxoid, fibroid-

collagen, chondroid, and so forth) and the nature and extent of necrosis (none,focal, or extensive) were noted.

Procedures for in Vivo 3IP MRS. The procedures for in vivo 3IP MRS

were described recently in detail (20), so only a brief description will be givenhere, in vivo studies were performed at 1.5 Tesla in a Siemens Magnetomclinical imager/spectrometer (Siemens AG, Erlangen, Germany). The surfacecoils constructed to receive NMR signals from tumors in various anatomicsites in vivo included a 12- or 15-cm diameter, single-turn, 3IP coil of coppertubing underlying a 22 X 14-cm butterfly 'H coil of 1-cm-wide copper foil.

The patient was positioned and made comfortable on the instrument table, thecoil assembly was placed over the region of interest, and the coils were tunedand isolated at the "P and 'H frequencies. 'H MR images of 10-mm thickness

were obtained in nine slices in each of three dimensions and were used todefine (and, if necessary, to adjust) the coil position relative to the tumor massand were later used to guide the registration of CSI voxels with the regions ofinterest. The magnetic field was then shimmed on the 'H signal to optimize

local static magnetic field homogeneity, using an autoshimming algorithmbased on three-dimensional 'H CSI (21, 25). Broadband 'H-decoupling wasaccomplished using Waltz-4 modulation during acquisition of 3IP, and NOE

enhancement was accomplished by delivering continuous-wave, low-level

power between acquisitions. A triphenylphosphate reference signal from asmall sample placed near the center of the "P coil was used to calibrate the

radiofrequency field plot of the coil so that a 45-degree pulse angle could beapplied within the center of the sarcoma to be studied. Nonlocalized "P NMRspectra were obtained with and without 'H-decoupling and NOE enhancement.

Three-dimensional CSI was then performed in an 8 x 8 x 8 array of cubicvoxels of 16-64 cc (usually 27 cc), depending on the size of the mass. The45-degree, rectangular, 250-/J.S excitation pulse was followed by the 0.8-ms.triangular, phase-encoding gradients applied simultaneously in three dimen

sions, and then by the acquisition of 512 data points in 256 ms. With arepetition time between excitations of 1 s, a full dataset of one acquisition/phase-encoding step required 8.5 min. Four full dataseis were obtained in 34

min.In Vivo MRS Data Processing. In vivo MRS data were processed on a

Spare Station 2 (Sun Microsystems) computer, using programs created in ourlaboratory (17, 26). Sagittal, axial, and coronal MR images were created andoverlaid with 8x8 grids indicating the "P MRS voxel positions. The CSI

spectra were plotted in 8 X 8 voxel arrays in each of three dimensions. Thespectra from one or more voxels within the regions of interest were extractedand processed using NMR1 software (New Methods Research, Syracuse, NY).Peak areas were estimated by fitting to Gaussian lineshapes. Metabolite peaksignal intensities were expressed as fractions of the total phosphorus signal.The pH was determined from the position of P( relative to a-NTP and applyingthe Henderson-Hasselbalch equation (27). Peak assignments were based on thepositions of known metabolites in high-resolution NMR spectra (28) and wereconfirmed by high-field 3IP NMR studies of the extracts of sarcoma surgical

specimens. The frequency scale was expressed in parts/million (ppm) and wasset by placing the center of the a-NTP peak at -10.0 ppm; this peak was

selected because its position is not affected by pH within the range ofphysiological values (27), and because it is present in all cases. PCr occurs asa contaminant from muscle either directly because of muscle within the voxelor indirectly as a "bleed" artifact associated with the phase-encoding used to

localize signals in CSI (29). The NTPs (adenosine-. cytosine-, guanosine-,thymidine-, and UTPs) are nonspecifically designated NTP because they occur

at the same frequency and cannot be distinguished in vivo.In Vitro 3IP MRS of Surgical Specimens. Between one and three biopsies

(0.3-3.5 g) were obtained during surgery and were freeze-clamped within 10

min of removal of the tumor from the patient. Locations of the biopsies wereselected to represent solid tumor, and evident necrotic regions were avoided.

Table 1 Characteristics of patients and histopathologicai review

Caseabcdefghijk1mnOpqr8tUVWXyzAge(yr)4448678045733421756354735464468072604469657767523454SexFMFFFFMFFFMFMFMMFFMMFFFFMMSiteThighThighCalfCalfThighAnkleParaspinalShoulderPelvisAnkleSacrumThighArmThighThighShoulderThighFlankArmThighAbdomenAbdomenRetroperThighThighRetroperClinicalstatus"NewReeNewNewNewNewResNewNewNewNewNewNewNewNewReeNewReeNewNewReeNewNewNewReeReeMRS

studies

in vivo invitro++

+++

+++++++++++++

+++

+++++++++Type*MFHLiposarcLeiomyoLeiomyoLeiomyoMFHUndiffEwing'sLeiomyoMFHChordomaFibrosarcMFHRhabdoMFHLiposarcLiposarcLiposarcRhabdoLiposarcLeiomyoLeiomyoLeiomyoMFHNFLiposarcGradeHighHigh"HighHighHighHighHighHighHighIntermHighLowHighHighHighHighLowIntermHighHighHighIntermHighHighLowLowHistologyPrimary

features0Cellular

(PI), focalnecrosisCellular(Ro), myxoidmatrixCellular

(Sp)Cellular(Sp)Cellular(Sp)Cellular

(Gn,Sp)NotavailableCellular(Ro)Cellular

(PI), focalfibrosisCellular(Sp)Variable:

cells (PI) andmatrix'Fibrous

matrix, cells (Sp). focalhyalineNotavailableCellular

(PI, Sp), muscle, focalnecrosisExtensivenecrosis, cells(PI)Variable:

cells (PI) andmatrixMaturefat, focalnecrosisCells

(PI), fibrosis, focalhyalineCysticwith rim of sarcoma cells(myoblast)Variable:

cells (Gn. Sp), myxoid. focalnecrosisExtensivenecrosisCellular(Sp)Cellular(Sp)ExtensivenecrosisCellular(Sp)Mature

fat" Ree, recurrent. Res, resistant to chemotherapy.'' MFH. malignant fibrosis histiocytoma; NF. neurofibrosarcoma; Liposarc. liposarcoma; Leiomyo. leiomyosarcoma; undiff. undiffereniiated; Fibrosarc. fibrosarcoma; Rhabdo.

rhabdosarcoma; Ewing's, Ewing's sarcoma.' Gn. giant cells; PI, pleomorphic cells; Ro. round cells; Sp, spindle cells; variable, variations from region to region.11Myxoid liposarcoma with transformation to high grade round cell.' Needle aspiration biopsy.

2965




The tissue was stored at -70Â°C until extraction. The tissue was extracted with

the dual-phase extraction method of Tyagi et al. (30), modified for use withfrozen tissue specimens. The frozen tissue was powdered with a Bio-Pulverizerâ„¢ (BioSpec Products, Bartlesville, OK) cooled in liquid nitrogen. Thepowder was homogenized with a Tissue-tearerâ„¢ (BioSpec Products) in 10

v/w of ice-cold methanol containing 0.4 mM phenylphosphonic acid as thequantitation standard. An equal amount of ice-cold chloroform was added,followed by homogenization, the addition of an equal amount of ice-colddistilled water, and re-homogenization. The phases were separated by centrif-ugation at 1000 X g for 10-60 min at 4Â°C,and the tissue residue was

re-extracted as described above (except that the methanol did not contain the

standard). The water phases were combined and dried on a rotating evaporator.The dried residue was redissolved in 6-9 ml of water, the pH was adjusted to6.3, and the solution was freeze-dried. The freeze-dried residue was taken upin 400 /il D2O and 1.3 ml of buffer solution (88 mM Â¿V-[2-hydroxyethol]-piperazine-/v"-3-propane-sulfonic acid; Sigma Chemical Co.) containing 44mM rran.s-l,2-diaminocyclohexane-/v',/V,/v",yV'-tetraacetic acid (Aldrich Chem

ical Co.). The pH was adjusted to 8.0. The buffer prevented the pH changes (ofup to 0.5 pH units) that would otherwise have occurred during the dataacquisition.

'H-decoupled "P NMR spectra were obtained at 162 MHz on a 9.4-TeslaBruker AM 400 spectrometer (Bruker, Karlsruhe, Germany) at 25Â°C.A 16 K

free-induction decay was acquired in 1.02 s using a 45 degree pulse (8.6 /is),broadband composite-pulse 'H-decoupling, and a total repetition time of 2.7 s.

Several acquisitions with 1000 or 2500 transients were accumulated. The totalnumber of transients varied between 1000 and 25000, depending on the size ofthe sample and the signal intensity. The free-induction decays were multipliedwith a Lorentz-Gaussian function (Bruker parameters: LB, â€”¿�1; GB, 0.01),

were zero filled to 32 k, and were Fourier-transformed. After baseline correc

tion, the resonances of the metabolites of interest were quantified by fittingGaussian lines, using WIN-NMR software (Bruker). The resonance areas

obtained were recalculated to yield nmol metabolite/g tissue. The chemicalshifts were referenced to that of GPChol at 0.49 ppm.

RESULTS

The influence of 'H-decoupling and NOE enhancement on the invivo 3IP MR spectrum is observed by comparing nonlocalized cou

pled and decoupled spectra from a thigh containing a malignantfibrous histiocytoma (case a; Fig. 1). Two separate components ineach of the PME and PDE regions are resolved, and the signalintensities of many of the peaks (especially PME, NTP, and PCr) areincreased, resulting in an increased signahnoise ratio. The method by

PCr

PME ' a-NTP ÃŸ-NTP

Decoupled

Coupled

10 -5 -10 -15 -20 -25

Chemical Shift (ppm)Fig. I. Effects of 'H-decoupling and NOE enhancement on the "P NMR spectrum

obtained without localization from case a (a malignant fibrous histiocytoma in the thigh).Bottom, the spectrum acquired without 'H-coupling or NOE enhancement; top, thespectrum acquired with 'H-decoupling and NOE enhancement. PCr is from muscle

surrounding the sarcoma.

(A)

(B)

,*Art*JWÂ«HiA'vwv

PChol

o -s -10Chemical Shift ( ppm )

â€¢¿�25

Fig. 2. MRl-direcled "P MRS study in case a. A. an axial MR image with pan of the8X8 array of 3 X 3 X 3 cm3 CSI voxels overlying the image. Highlighted region, theCSI voxels that contain the spectra shown in B. B, 'H-decoupled, NOE-enhanced, "P

NMR spectra from the voxels highlighted in A. The two spectra within the dashed borderswere accurately localized to the tumor by voxel-shifting the dataseis. Signals in the othervoxels, which are in muscle, are dominated by the high PCr peak. C. a spectrumsummmed from the two highlighed spectra in B.

which we obtained MRI-directed 31P MR spectra localized in three

dimensions using CSI is illustrated in Fig. 2. The benefit obtained bylocalizing with three-dimensional CSI is evident when one compares

the localized spectrum in Fig. 1C to the nonlocalized spectra from thesame tumor (Fig. 1). In the localized spectrum, there is a nearlycomplete elimination of the intensity of the PCr signal from surrounding muscle.

The spectrum localized to the sarcoma in Fig. 2C has strong signalsfrom the PMEs (PEth and PChol) and PDE signals at 0.5 and -0.3

ppm but no detectable signal from GPEth at 1.0 ppm. The nonlocalized 'H-decoupled spectrum in Fig. 1 contains the same two PDE

signals. The stronger signal near 0.5 ppm is at the position of GPChol,but the weaker signal at â€”¿�0.3ppm is unidentified. In the spectrum

localized to the sarcoma in Fig. 2C, the unidentified signal is stronger(hence from the sarcoma), whereas the GPChol signal is weaker, as

2966



T MRS OF SARCOMAS IN VIVO AND IN VITRO

Fig. 3. "P MR spectrum from the water-soluble extract of

a portion of the surgical specimen of a leiomyosarcoma (cased). /, a-glycerol phosphate: 2. unassigned PME; 3, IMP/GMP; 4. AMP; 5, unassigned PME; 6. a-phosphate of NDP-hexoses and NDPn (e.g.. NAD); 7. CDP-choline; 8. ÃŸ-phos-phate of NDP-hexoses. Insel, the in vivo "P MR spectrum

obtained before surgery. The signal cutoff (top) is PCr frommuscle contaminating the voxel. U, an unidentified PDEsignal at -0.3 ppm, which did not appear in the extract

spectrum.

PEth

3Cho

< 5

IL

GPChol

GPEth

7-NTP o-NTP

Y-NTP

PCr

ÃŸ-NDP

PP

Chemical ShiÂ«(ppm)

a-NTP

a-NDP

ÃŸ-NTP

-5 -10

Chemical Shift (ppm)

-15 -20

expected (18), from muscle. The strong signals from the NTPs lead usto infer that a significant part of this sarcoma is composed of meta-

bolically viable cells and is therefore relatively well perfused.The in vitro 31P spectrum of the water-soluble extract of a leiomy

osarcoma (case d) is shown in Fig. 3, and the in vivo 31P spectrum

obtained from this sarcoma before surgery is shown in the inset. Thein vivo spectrum is quite similar to the spectrum from a malignantfibrous histiocytoma shown in Fig. 1C: the relative signal intensitiesof the PMEs, the PÃŒ,and the NTPs are essentially the same. However,the PDE regions differ in that the spectrum in the inset in Fig. 3 hasa smaller signal from the unknown metabolite at -0.3 ppm and a

higher signal at the position expected for GPChol. The in vitrospectrum confirms the in vivo metabolite assignments. It shows thatthe highest signals within the PME region are from PEth and PChol,that GPChol is present as observed in vivo, and that the unknown PDE

metabolite observed in vivo does not appear in the extract spectrum.The relative signal intensities in the extract spectrum differ from thein vivo ones in part because of the different spin-lattice relaxation

times at different magnetic fields, and because of the different degreesof signal saturation caused by the shorter repetition time used in vivo(1 s) than in vitro (2.7 s).

We have obtained 'H-decoupled, NOE-enhanced 3IP MR spectra

localized to each of 20 soft tissue sarcomas in vivo. Sixteen spectracontained 31P signals and were Ã©valuable.These are displayed in Fig.

4. Eleven spectra (a-k) were of high quality, with high signal:noise

ratios and good or excellent resolution of adjacent peaks. Nine of thesarcomas from which these spectra were obtained were examinedhistologically, and all nine were predominantly or entirely cellular(Table 1). Five spectra (/-/>) were of poorer quality but were adequate

to quantitate the PME, PÂ¡,and, in some cases, the unidentified mobile

Fig. 4. 'H-decoupled, NOE-enhanced 3IP NMR spectra from

each of 16 sarcomas. Spectra are lettered according to the casedesignations in Table 1 and were plotted so that the height of thedominant PME peak is the same in all spectra (except those withhigh PCr). In processing, the spectra were line-broadened by a3-6-Hz Gaussian filter.

-5 -10 -15 -20

ppm

2967

0 -5 -10 -15 -20

ppm

-5 -10 -15 -20

ppm



T MRS OF SARCOMAS IN VIVO AND IN VITRO

Table 2 Metabolic characteristics of sarcomas derived from the in vivo ' P spectra in Fig. 4.

Metabolite signal intensities are expressed as fractions of total phosphorus signal in each spectrum. PEth:PChol ratio is derived from peak heights; it is designated as NR when thePME signal was too broad to resolve PChol from PEth. pH is predominantly intracellular pH, determined from the position of Pt relative to a-NTP.

CasePMEa

0.14b0.16C0.23d0.14e0.24f0.22g0.27h0.26i0.18j0.20k0.1210.32m0.22n0.22o

0.17Average0.21SD0.06CV

(%)" 27PÂ¡0.080.180.150.080.160.130.060.010.100.110.080.100.070.110.110.100.0442PDE0.480.290.070.100.060.150.050.020.310.130.130.140.210.100.160.170.1167NTP0.300.370.550.680.540.500.620.530.420.570.670.440.490.570.560.520.1020PEthiPChol1.71.51.21.42.72.82.32.83.62.41.62.02.0NRNR2.10.733pH7.177.497.147.087.507.357.247.107.277.177.147.517.057.577.157.260.172

1CV. coefficient of variation.

PDE (*) peak intensities. These spectra were from sarcomas that

contained large amounts of matrix or necrosis relative to cells (Table1). Four spectra (data not shown in Fig. 4) contained little or nosignals above the noise level. In three of these spectra, the explanationfor the lack of signals was evident when the tumors were surgicallyresected (Table 1). One (case q) was a low grade liposarcoma composed of mature fat and fibrous tissue with only scattered cells. Thesecond (case r) contained a large fraction of fibrotic matrix andhyalinized areas relative to the fraction of cells. The third (case s)contained predominantly a large cystic region that in the MRI was notdistinguished from a rim of viable tumor found at surgical resection.In retrospect, this tumor had undergone extensive partial resection atthe time it was biopsied. In the fourth case (t), subsequent surgicalresection showed large regions of myxoid matrix and focal necrosis.

With the exception of p. the 16 spectra displayed in Fig. 4 have twomajor features in common: (a) prominent signal intensities in thePME region, with the signal intensity of PEth (4.2 ppm) greater thanthat of PChol (3.8 ppm); and (b) prominent NTP signals. The PDEregion is more variable among these spectra. Seven spectra ( a, b, i, m,n, o, and p) contain a prominent signal (*) at approximately -0.3

ppm. Four spectra (d, e.f, and i) have signals at the position expectedfor GPChol (0.49 ppm).

The metabolic characteristics of soft tissue sarcomas derived fromthe 15 spectra (from cases a-o) in Fig. 4 are summarized in Table 2.The results are expressed as fractions of total 3IP signal after exclud

ing the contaminating signal from muscle PCr in those cases in whichit was present. The mean PME signal intensity is 21% of the totalphosphorus signal, and in cases in which PChol as well as PEth isresolved, the PEth:PChol peak height ratio is 2.1. The mean signal inthe PDE region is 17% of the total but has the greatest variability inaccord with the impression gained in the presentation of the cases inFigs. 2 and 3. The mean PÂ¡signal is 10% of the total and occurs atpositions corresponding to a mean pH of 7.26. The NTPs account forthe remaining 52% of the total signal.

Â¡nvitro 3IP spectra from the extracts of surgical specimens of four

leiomyosarcomas are shown in Fig. 5. These include the case (d) inwhich an in vivo spectrum (d) was also obtained (Fig. 3). From aqualitative standpoint, these spectra, as well as those (data not shown)from a malignant fibrous histiocytoma (case x) and a neurofibrosar-

coma (case y), have features in common with one another and with thein vivo spectra. These features include prominent PME signals with ahigher concentration of PEth than PChol, very low concentrations ofGPEth, variable concentrations of GPChol (but in all cases lower than

that of PChol), and the presence of NTPs. The spectra from extractsof four liposarcomas, on the other hand, had much more variablefeatures (Fig. 6). Two spectra (b and r) did have greater concentrations of PEth than PChol as well as high concentrations of NTPs, likethe in vivo spectra (Fig. 4) and the extract spectra from leiomyosarcomas (Fig. 5). These two liposarcomas, however, differed by havingvery high concentrations of GPChol. In case r, the in vivo spectrum (r)

J

xl

xl

u

LidxO.lII

.Ill50-5

-10 -15 -20

Chemical Shift ( ppm )Fig. 5. 3IP MR spectra of extracts of surgical specimens of four leiomyosarcomas.

Spectra are lettered according to the case designations in Table 1. See Fig. 3 for peakassignments.

2968




x JO

-5 -10 -15

Chemical Shift ( ppm )

-20

Fig. 6. "P MR spectra of extracts of surgical specimens of four liposarcomas. Spectra

are lettered according to the case designations in Table 1. See Fig. 3 for peak assignments.

contained only noise, which was explained by the histolÃ³gica! examination, which showed considerable fibrosis and hyalinized matrix(Table 1). The spectra of the third and fourth liposarcomas (p and Â¿)in Fig. 6 show a virtual absence of PME and NTP signals, but only PÂ¡,some PDEs (including PChol), and, in case p, NDP-hexoses. The in

vivo spectrum from case p (p, Fig. 4) also showed an absence of PMEand NTP signals as well as a high PÂ¡signal.

Several limitations must be taken into account when comparing thein vivo and the in vitro 31P spectra: (a) the different degrees of

saturation of the signals in vivo and in vitro, greater in those with thelonger spin-lattice relaxation times (e.g., PDEs) than in those with

short ones (e.g., NTPs; Ref. 19); (b) the greater likelihood of musclecontamination in vivo; (c) the intratumoral tissue heterogeneity; the invitro spectra are from small regions selected for sampling because

they appeared solid and viable, whereas the in vivo spectra obtainedfrom 27-cc voxels are from signals averaged over heterogeneous

regions within the tumors; and (d) the ability to quantitate the in vivometabolite signals only by reference to the total phosphorus signal orto one another; actual concentrations of the metabolites could varyconsiderably among the tumors. The extract spectra give the opportunity to examine the extent of inter- and intratumoral heterogeneity

both qualitatively and quantitatively because molar concentrations areobtained. Concentrations of the major metabolites, summarized inTable 3, reveal a marked heterogeneity among tumors. Even amongleiomyosarcomas, which qualitatively have similar spectra (Fig. 5),the concentration ranges span an order of magnitude. The total metabolite concentrations varied from 0.3-6.1 /nmol/g tissue and theNTP + NDP from 0-1.5 fj.mol/g tissue. The tissue metabolite con

centrations are primarily a function of cellularity; all five surgicalspecimens with above-median total metabolite concentrations were

entirely or predominantly cellular (case d, r, v, w, and y in Table 1);whereas all five with below-median total metabolite concentrations

were not. Cases u and x had extensive necrosis, and case z waspredominantly mature fat. Case p varied from region to region andalso had a very low signal:noise ratio in vivo (Fig. 4). In case b, thelow metabolite concentrations seem to contradict the in vivo study thatproduced a high-quality spectrum (Fig. 4). The histolÃ³gica! examina

tion provided an explanation for this discrepancy (Table 1). Case bwas a recurrent myxoid liposarcoma with transformation to a high-

grade, round cell liposarcoma; some regions were predominantlymyxoid matrix, some were entirely cellular, and some were mixed.The specimen used to obtain the extract was from a myxoid region,whereas the in vivo spectrum contains signals averaged over a largepart of the tumor.

Data indicating the extent of intratumoral heterogeneity are summarized in Table 4. Each of two samples from cases u and v werequite similar. In the other cases, metabolite concentrations varied bya factor of two or more. The heterogeneity of metabolite concentrations evident in the data in Tables 3 and 4 was not simply due tovariable cellularity because different metabolites were affected todifferent degrees. For example, there was no correlation at all betweenthe concentrations of the NDP-hexoses and those of NTP + NDP

among the 10 cases (Table 3).An unidentified PDE signal at â€”¿�0.3ppm was evident in seven in

vivo spectra (a, b, i, m, n, o, and p in Fig. 4). All of the sarcomascontaining this signal were high-grade. Histological examination was

available in six cases: five had pleomorphic cells, and one was intransformation to a round cell liposarcoma (Table 1). This PDE signalwas not observed in any of the spectra obtained from extracts ofsurgical specimens (e.g., Fig. 3).

The signals from NDP-hexoses were obtained in the extract spectrain most (8 of 10) cases. As indicated in Fig. 3, their a-phosphateresonances appear around -10.5 to â€”¿�11ppm, and their ÃŸ-phosphate

Table 3 Concentrations of major metabolites tData in cases u-y are means of two or three samples shown separately in Table 4.

i in soft tissue sarcomas

CaseduVWbPrZXyDx"LeioLeioLeioLeioLipoLipoLipoLipoMFHNFPEth1.140.172.221.910.14bd1.010.060.181.20PChol0.470.030.620.370.04bd0.090.020.060.09GPEth0.060.020.020.080.040.240.270.020.060.07GPChol0.360.060.050.130.330.820.620.150.220.07CDP-chol0.080.010.070.07bd0.040.10bdbd0.06IMP/GMP0.270.070.770.470.19bd0.18bd0.020.32AMP0.040.100.070.040.01bd0.08bd0.020.10NTP+NDP1.050.251.450.470.54bd1.14bd0.430.69NAD0.110.050.290.19bd0.110.27bd0.100.10NDP-hex40.03bd*0.080.07bd0.160.16bd0.030.08NDP-hex50.040.100.140.08bd0.310.28bd0.040.13

' Dx. diagnosis: Leio. leiomyosarcoma; Lipo, liposarcoma; MFH. malignant fibrosis histiocytoma; NF. neurofibrosarcoma.*bd, below detection limit.

2969




Table 4 Concentrations of selected metabolites (pmol/g) for multiple samples offÃ¬vesarcomas

CasePEthu

0.160.18v

2.172.27w

2.451.36x

0.270.08y

1.890.880.82PChol0.07bd0.650.590.490.240.080.030.120.090.06GPEth0.020.020.020.010.060.090.070.040.100.060.04GPChol0.050.070.040.050.090.170.390.050.090.080.05CDP-chol0.01bd0.050.080.100.05bdbd0.090.050.03NTP

+ NDP0.200.291.461.450.310.620.87bd0.910.600.55NDP-hex4bd"bd0.080.080.070.07bd0.060.110.060.05NDP-hex50.02bd0.140.130.080.08bd0.080.230.130.03

' bd. below detection limit.

resonances appear around â€”¿�12 to â€”¿�13ppm. Although five different quality typical of those shown in Fig. 4. The variations in spectral

NDP-hexoses were found, the two labeled NDP-hexose 4 and NDP-

hexose 5 were detected most frequently, were present in the highestconcentrations, and (from their positions) are tentatively assigned toUDP-yV-acetyl-galactosamine and UDP-A'-acetylglucosamine, respec

tively (31). Their concentrations, which were included in Tables 3 and4, did not correlate with any particular tissue type or histologicalcharacteristic.

DISCUSSION

By implementing 'H-decoupling and NOE enhancement in con

junction with dual-tuned surface coils, adequate shimming to obtain

good homogeneity of the magnetic field within regions of interest, andaccurate localization of 3IP MR spectra to regions of interest using

three-dimensional CSI, we have been able to obtain more information

about the in vivo metabolic characteristics of soft tissue sarcomas thanhas heretofore been available. We confirm the high PME signalintensity and the slightly alkaline pH reported previously in sarcomas(1, 3-6, 8-15). In addition, we show that the PME region is domi

nated by PEth over PChol, and that the PDE region contains nodetectable GPEth and only occasionally detectable GPChol. ThatGPEth would be detected in 'H-decoupled 31P MR spectra if it were

present in significant quantities has been demonstrated in the brain(17) and liver (19). The absence of detectable GPChol and GPEth wasobserved by Bachert et al. (15) in a malignant fibrous histiocytoma inthe only previously reported study of a human sarcoma in vivo using'H-decoupled 3IP MRS. Redmond et al. (4) found the same charac

teristics in an extract of surgical tissue from one patient with amalignant fibrous histiocytoma, and, interestingly, Evanochko et al.(2) observed almost exactly the same spectrum in an extract of amurine RIF-1 sarcoma. We have confirmed the essence of theseanecdotal observations in our high-resolution 3IP NMR studies of

tissue extracts from 10 sarcomas, 4 of which were studied in vivo.Several sarcomas had in their in vivo spectrums a prominent mobilePDE signal from an unidentified metabolite. Because it did not appearin the water-soluble extracts of sarcoma surgical specimens, it may be

associated with either an abnormally mobile phospholipid moiety or,conceivably, the fragments of nucleic acids. The in vitro studiesrevealed marked variations among sarcomas in the actual concentrations of metabolites. To a great extent, this correlated with the relativeamounts of cells and of noncellular regions (matrix or necrosis) withinthe tumor.

By using appropriate surface coil designs, we have managed tostudy sarcomas in several different anatomic sites. Nevertheless, theapplication of 31P MRS remains limited by the inability to access

sarcomas deeper than 10-12 cm from the surface and the difficulty instudying sarcomas smaller than approximately 3-cm diameter. In most

cases, it is necessary to acquire four averages of the CSI dataset toensure high enough signal relative to noise to obtain spectra of the

quality among Ã©valuablespectra are likely to be determined primarilyby variations in the relative amounts of malignant cells and matrix ornecrosis within them. Some sarcomas (4 of 20 in this series) have somuch matrix or so much spontaneous necrosis that they do not containenough viable cells to produce an Ã©valuablespectrum. The histological heterogeneity that characterizes many sarcomas is associated withmetabolic heterogeneity, as we have documented in cases in whichtwo or three specimens could be obtained from the same surgicalspecimen.

The prominent NTP signals in vivo in 14 of our cases suggest thatpart of the solid fraction of many sarcomas in vivo contains viablecells with well-preserved energy metabolism. The alkaline intracellu-

lar pH (mean 7.26) in sarcomas might also be taken as an indicationof well-preserved energy metabolism. However, a similar pH measured by 31P MRS seems to be typical of many other types of human

cancers, some of which are expected to be uniformly well perfused(e.g., lymphomas; Ref. 20), and some of which are not (32). Althoughthe pH is slightly acidic relative to the plasma pH (7.4), it is morealkaline than the pH measured using 31P MRS in muscle (7.10) and in

the brain (7.05). This observation in cancers contradicted expectationsthat cancer cells might be more acidic because of the production oflÃ¡clate from excessive glycolysis, and its possible mechanism andsignificance have been discussed and debated by a number of authors(11, 33-36).

Our observations are relevant to the development of experimentalsarcoma models because our results indicate that some experimentalmodels have phospholipid metabolite patterns that differ from those ofsarcomas in vivo in patients. For example, although RIF-1 tumorsgrowing in the mouse had 3IP MR spectra remarkably similar to those

we observed in patients (including a predominance of PEth overPChol; Ref. 2), RIF-1 cells in culture contained mostly PChol (2).

This situation arises partly because the levels of PEth and PChol incultured cells can be markedly affected by the concentrations ofcholine and ethanolamine in the medium as well as by its pH (31,37-40). We believe that our observations should help guide the

development of experimental models that have metabolic characteristics similar to those that occur in vivo in patients.

The mechanisms underlying the pattern of phospholipid metabolites we observed in sarcomas are unknown. The general pattern,however, is very similar to that observed using 'H-decoupled 3IPMRS in non-Hodgkin's lymphomas (20) and in a variety of cancersurgical specimens studied by high-field 31P NMR spectroscopy (4,

41-46). Based on reports of similar patterns in experimental cancer

models (47, 48), we have hypothesized that the high concentration ofPEth relative to PChol, occurring in the presence of a low concentration of GPEth, is in part a manifestation of a sustained activation of aphosphatidylethanolamine-specific phospholipase D (20).

Shinkwin et al. (14) suggested that the amounts of PME and PDE

2970




relative to NTP in 31P MRS might distinguish low and high grade

soft tissue sarcomas. This result was not statistically significant,however, and the study was limited by poor spectral resolution.Improvements in acquisition of 31PMR spectra, as outlined in this

report, should permit a rigorous test of the ability of MRS to gradesarcomas according to their metabolic characteristics. Because thehistolÃ³gica!grade remains one of the most reliable (49, 50) potential markers of biological or clinical aggressiveness of a soft tissuesarcoma, it would be interesting to explore the ability of MRS toprovide metabolic markers that correlate with histolÃ³gica! grade.Our results to date do not indicate different metabolic features in3IP MRS among different grades of sarcomas or among non-Hodgkin's lymphomas (20), except that the unidentified PDE

signal occurred only in high-grade sarcomas.The significance of the presence of NDP-hexoses in sarcomas is

uncertain. UDP-hexoses play a role in the synthesis of glycoproteins,glycolipids, and proteoglycans. UDP-/V-acetylgalactosamine andUDP-/V-acetylglucosamine, in particular, are involved in the anabo-lism of proteoglycans that are constituents of the extracellular matrix.It would not be surprising, therefore, that NDP-hexoses are present insarcoma cells that have a mesenchymal origin, although we found noevident correlation with histolÃ³gica!features.

Characteristics in 31PMR spectra may correlate with the sensitivityor resistance of an individual patient's sarcoma to treatment. Koutcher

et al. (3) found a higher PME:PDE ratio in the baseline spectra ofthree sarcomas that responded to chemotherapy than in three sarcomasthat did not. Sostman et al. (5) found that the mean pH (7.30) of 10soft tissue sarcomas that responded to treatment with neoadjuvantradiotherapy and hyperthermia was higher than the mean pH (7.16) ofthe 10 sarcomas that did not. Sijens et al. (6) found that an earlydecrease in PME correlated with response (as measured by the fraction of necrosis at surgery performed 2 months later) in 11 extremitysarcomas treated with TNF-a + melphalan via isolated limb perfusion, and that an early decrease in PÂ¡correlated with the extent ofmeasurable shrinkage of the sarcoma. Other investigators (3,4, 9, 51)obtained baseline and follow-up 3IP MR spectra in 11 sarcomas

treated with chemotherapy: all 7 sarcomas that eventually respondedhad a decrease in the PME signal intensity in follow-up spectraobtained within 2 weeks after initiation of treatment, whereas none ofthe 4 that failed to respond to treatment did so. On the other hand,an increased PMErNTP ratio after the first hyperthermia treatmentcorrelated with the eventual response of soft tissue sarcomas toradiation and hyperthermia as measured by the fraction of necrosisin the surgical specimen (1). These reports suggest that earlymetabolic changes occur consistently in sarcomas destined to respond to treatment, but that the particular metabolic event maydepend upon the nature of the treatment. This phenomenon couldbe useful in the clinical management of patients. Because over 50%of patients with advanced, recurrent, or metastatic soft tissuesarcomas fail to respond to chemotherapy, and only 10% havecomplete responses (52), the ability to identify patients whosesarcomas are going to be resistant to treatment is likely to becost-effective. The technique we report here should permit a more

rigorous test of the hypothesis that metabolic changes predictsensitivity of a sarcoma to a particular treatment. Using thistechnique, we have recently shown that early treatment-induceddecrease in PEth specifically correlates with the eventual responseof non-Hodgkin's lymphomas to treatment (53). We believe that

the study reported here provides a good technical basis on which toexamine potential clinical uses of 31P MRS in the management of

soft tissue sarcomas as well.

ACKNOWLEDGMENTS

We thank Peter Z. Liu and Tariq Javaid for assistance in automatic shimming; Libo He and Chris Elsasser for assistance in coil construction; CalvinShaller for assistance in specimen processing; Drs. Margaret von Mehren andArthur Staddon for referring patients; and Drs. Suzanne Franks and FrankKappler for helpful discussions.

REFERENCES

1. Prescott. D. M.. Charles. H. C., Sostman, H. D.. Dodge, R. K., Thrall, D. E., Page,R. L., Tucker, J. A., Harrelson, J. M., Leopold, K. A., Oleson. J. R., and Dewhirst.M. W. Therapy monitoring in human and canine soft tissue sarcomas using magneticresonance imaging and spectroscopy. Int. J. RadiÃ¢t.Oncol. Biol. Phys., 28: 415-423,1993.

2. Evanochko, W. T., Sakai, T. T., Ng. T. C.. Krishna, N. R., Kim, D. H., Zeidler. R. B.,Ghanta. V. K., Brockman. R. W., Schiffer, L. M., Braunschweiger, P. G., andGlickson, J. D. NMR study of in vivo RIF-1 tumors: analysis of perchloric acidextracts and indentification of 'H. "P, and I3C resonances. Biochim. Biophys. Acta.

805: 104-116. 1984.

3. Koutcher, J. A., Ballon, D., Graham, M.. Healey, J. H., Casper. E. S., Heelan, R., andGerweck, L. E. "P NMR spectra of extremity sarcomas: diversity of metabolic

profiles and changes in response to chemotherapy. Magn. Reson. Med., 16: 19-34,1990.

4. Redmond, 0. M.. Bell, E., Stack, J. P., Dervan, P. A., Carney, D. N., Hurson. B., andEnnis, J. T. Tissue characterization and assessment of preoperative chemotherapeuticresponse in musculoskeletal tumors by in vivo 3IP magnetic resonance spectroscopy.

Magn. Reson. Med.. 27: 226-237, 1992.5. Sostman, H. D.. Prescott, D. M., Dewhirst, M. W., Dodge. R. K.. Thrall. D. E.. Page,

R. L., Tucker, J. A., Harrelson. J. M.. Reece. G., Leopold. K. A., Oleson, J. R., andCharles, H. C. MR imaging and spectroscopy for prognostic evaluation in soft tissuesarcomas. Radiology, 190: 269-275, 1994.

6. Sijens, P. E., Eggermont, A. M. M., van Dijk. P.. and Oudkerk, M. "P magnetic

resonance spectroscopy as predictor of clinical response in human extremity sarcomastreated by single dose TNF-a + melphalan isolated limb perfusion. NMR Biomed..8: 215-224. 1995.

7. Adler, L. P., Blair. H. F.. Makley, J. T.. William, R. P., Joyce, M. J., Leisure, G.,Alkaisi, N.. and Miraldi. F. Noninvasive grading of musculoskeletal tumors usingpositon emission tomography. J. NucÃ-.Med., 32: 1508-1512. 1991.

8. Griffiths, J. R., Cady, E., Edwards, R. H., McCready, V. R., Wilkie. D. R., andWiltshaw, E. "P NMR studies of human tumor in situ. Lancet, 25: 1435-1436, 1983.

9. Ross, B., Helsper, J. T.. Cox, I. J., Young, I. R., Kempf. R., Makepeace, A., andPennock, J. Osteosarcoma and other neoplasms of bone: magnetic resonance spectroscopy to monitor therapy. Arch. Surg., /22: 1464-1469, 1987.

10. Semmler, W., Gademann, G., Bachert-Baumann, P.. Zabel. H. J.. Lorenz. W. J.. andVan Kaick, G. Monitoring human tumor response to therapy by means of P-31 MRspectroscopy. Radiology, 166: 533-539, 1988.

11. Negendank. W. G., Crowley, M. G., Ryan. J. R.. Keller, N. A., and Evelhoch. J. L.Bone and soft tissue lesions: diagnosis with combined H-l MR imaging and P-31 MRspectroscopy. Radiology. 173: 181-188, 1989.

12. Dewhirst, M. W., Sostman, H. D., Leopold, K. A., Charles, H. C., Moore, D., Bum.R. A., Tucker, J. A., Harrelson, J. M., and Oleson, J. R. Soft tissue sarcomas: MRimaging and MR spectroscopy for prognosis and therapy monitoring. Radiology, 174:847-853. 1990.

13. Sostman, H. D., Charles, H. C.. Rockwell. S., Leopold, K.. Beam. C., Madwed, D.,Dewhirst, M., Cofer. G., Moore. D., Burn, R.. and Oleson, J. Soft tissue sarcomas:detection of metabolic heterogeneity with P-31 MR spectroscopy. Radiology, 776.837-843, 1990.

14. Shinkwin, M. A., Lenkinski, R. E., Daly, J. M., Zlatkin. M. B., Frank, T. S.. Holland,G. A., and Kressel, H. Y. Integrated magnetic resonance imaging and phosphorusspectroscopy of soft tissue tumors. Cancer (Phila.), 67: 1849-1858, 1991.

15. Bachert, P., Bellemann, M. E.. Layer, G., Koch, T.. Semmler, W., and Lorenz, W. Invivo 'H, 31P-['H], and "C-['H] magnetic resonance spectroscopy of malignant

histiocytoma and skeletal muscle tissue in man. NMR Biomed., 5: 161-170. 1992.16. Hoekstra, H. J.. Boeve. W. J., Kamman, R. L., and Mooyaart, E. L. Clinical

applicability of human in vivo localized phosphorus-31 magnetic resonance spectroscopy of bone and soft tissue tumors. Ann. Surg. Oncol., /: 504-511. 1994.

17. Murphy-Boesch, J., Stoyanova, R., Srinivasan. R., Willard, T., Vigneron, D., Nelson.S., Taylor, J. S., and Brown, T. R. Proton-decoupled 3IP chemical shift imaging of the

human brain in normal volunteers. NMR Biomed.. 6: 173-180, 1993.18. Brown, T. R., Stoyanova, R.. Greenberg, T., Srinivasan, R., and Murphy-Boesch, J.

NOE enhancements and T, relaxation times of phosphorylated metabolites in humancalf muscle at 1.5 Tesla. Magn. Reson. Med., 33: 417-421, 1995.

19. Li, C. W., Negendank, W. G., Padavic-Shaller. K., Murphy-Boesch. J.. and Brown.T. R. Molar quantitation of hepatic metabolites in vivo in proton-decoupled, NOE-enhanced 3IP NMR spectra localized by three- dimensional CSI. NMR Biomed.. in

press, 1996.20. Negendank, W. G., Padavic-Shaller, K. A., Li, C. W.. Murphy-Boesch, J.. Stoyanova,

R., Krigel, R. L., Schilder, R. J.. Smith, M. R., and Brown, T. R. Metaboliccharacterization of human non-Hodgkin's lymphomas in vivo using proton-decoupled

phosphorus magnetic resonance spectroscopy. Cancer Res., 55: 3286-3294. 1995.21. Hu, J., Javaid, T., Arias-Mendoza, F., Liu, Z., McNamara, R., and Brown, T. R. A

fast, reliable automatic shimming procedure using 'H chemical shift imaging. J.Magn. Reson. (Ser. B), 108: 213-219, 1995.

2971




22. Brown. T. R., Kincaid. B. M.. and Ugurbil, K. NMR chemical shift imaging in threedimensions. Proc. Nati. Acad. Sci. USA, 79: 3523-3526, 1982.

23. Weiss, S. W. Histological Typing of Soft Tissue Tumours, 2nd Edition. Berlin:Springer-Verlag. 1994.

24. Coindre, J. M., Trojani, M., Contesso, G., David. M.. Rouessse, J., Bui, N. B.,Bodaert. A.. De Mascarel, I.. De Mascarel, A., and Goussot. J-F. Reproducibility of

a histopathologic grading system for adult soft tissue sarcoma. Cancer (Phila.), 58:306-309, 1986.

25. Liu. Z.. Javaid. T.. Hu. J.. and Brown. T. R. CSI-Autoshim procedures for body andsurface coil. Proc. Soc. Magn. Reson., 3: 1174, 1994.

26. Vigneron. D. B.. Nelson. S. J.. Murphy-Boesch. J.. Kelley, D. A. C. Kessler, H. B.,Brown. T. R.. and Taylor. J. S. Chemical shift imaging of human brain: axial, sagittal,and coronal P-31 metabolite Â¡mage.Radiology. 777: 643-649. 1990.

27. Moon. R. B.. and Richards. J. H. Determination of intracellular pH by "P magnetic

resonance. J. Biol. Chem.. 248: 7276-7278. 1973.28. Robitaille. P-M. L.. Robitaille, P. A., and Brown. G. G. An analysis of the pH-

dependent chemical-shift behavior of phosphorus-containing metabolites. J. Magn.Reson., 92:73-84, 1991.

29. Murphy-Boesch, J., Jiang. H.. Carvajal, L., Stoyanova, R.. Kappler. F.. Negendank,

W., and Brown. T. R. Numerical correction for B, field inhomogeneity and pointspread function for quantitation of metabolites from CSI spectra. Proc. Soc. Magn.Reson., 2: 969, 1993.

30. Tyagi, R., Azrad. A.. Degani. H.. and Salomon. Y. Simultaneous extraction of cellularlipids and water-soluble metabolites: evaluation by NMR spectroscopy. Magn. Reson.Med.. 35: 194-200. 1996.

31. Shedd, S.. Lutz. N. W.. and Hull. W. E. The influence of medium formulation onphosphomonoester and UDP-hexose levels in cultured human colon tumor cells asobserved by "P NMR spectroscopy. NMR Biomed., 6: 254-263. 1993.

32. Negendank, W. Studies of human tumors by MRS: a review. NMR Biomed., 5:303-324, 1992.

33. Evelhoch, J. L. The pH of human tumors: facts and problems. In: W. C. Dewey,M. Eddington, R. J. M. Fry, and G. F. Whitmore (eds.). Radiation Research: ATwentieth-Century Perspective, pp. 778-783. San Diego, CA: Academic Press, Inc.,

1992.34. Griffiths, J. R. Are cancer cells acidic? Br. J. Cancer, 64: 425-427. 1991.

35. Tannock. I. F.. and Rotin, D. Acid pH in tumors and its potential for therapeuticexploitation. Cancer Res., 49: 4373-4384, 1989.

36. Vaupel. P.. Kallinowski. F.. and Okunieff. P. Blood flow, oxygen and nutrient supply,and metabolic microenvironment of human tumors: a review. Cancer Res., 49:6449-6465, 1989.

37. Horwitz, A. F. Manipulation of the lipid composition of cultured animal cells. In: G.Poste and G. L. Nicolson (eds.). Dynamic Aspects of Cell Surface Organization, pp.295-305. New York: North Holland. 1977.

38. Navon, G., Navon, R., Shulman. R. G., and Yamane, T. Phosphate metabolites inlymphoid, friend erythroleukemia. and HeLa cells observed by high-resolution "P

nuclear magnetic resonance. Proc. Nati. Acad. Sci. USA, 75: 891-895, 1978.39. Daly, P. F., Lyon, R. C., Faustino, P. J., and Cohen, J. S. Phospholipid metabolism in

cancer cells monitored by "P NMR spectroscopy. J. Biol. Chem., 262: 14875-14878,

1987.40. Kuesel. A. C., Graschew. G., Hull, W. E., Lorenz. W., and Thielmann, H. W. 3IP

NMR studies of cultured human tumor cells: influence of pH on phospholipidmetabolite levels and the detection of cytidine 5'-diphosphate choline. NMR

Biomed., 3: 78-89. 1990.41. Merchant, T. E.. Gierke. L. W.. Meneses, P.. and Glonek. T. "P magnetic resonance

spectroscopic profiles of neoplastic human breast tissues. Cancer Res., 48: 5112-5118, 1988.

42. Barzilai. A.. Horowitz. A.. Geier, A., and Degani. H. Phosphate metabolites andsteroid hormone receptors of benign and malignant breast tumors. Cancer (Phila.). 67:2919-2925, 1991.

43. Smith, T. A. D.. Glaholm. J.. Leach, M. O.. Machin. L., and McCready, V. R. Theeffect of intra-tumour heterogeneity on the distribution of phosphorus-contaningmetabolites within human breast tumours: an in vitro study using "P NMR spec

troscopy. NMR Biomed.. 4: 262-267, 1991.

44. Kalra, R., Wade. K. E.. Hands, L.. Styles. P.. Camplejohn. R.. Greenall. M.. Adams.G. E., Harris, A. L., and Radda. G. K Phosphomonoester is associated with proliferation in human breast cancer: a "P MRS study. Br. J. Cancer. 67: 1145-1153. 1993.

45. Kasimos. J. N., Merchant. T. E.. Gierke. L. W.. and Glonek, T. "P magnetic

resonance spectroscopy of human colon cancer. Cancer Res.. 50: 527-532, 1990.46. Cox. I. J., Bell. J. D.. Peden. C. J., Iles. R. A., Foster. C. S., Watanapa. P., and

Williamson. R. C. N. In vivo and in vitro 31P magnetic resonance spectroscopy of

focal hepatic malignancies. NMR Biomed., 5: 114-120. 1992.47. Gillham. H.. and Brindle, K. M. -"P NMR studies of phospholipid metabolism in

lipid-loaded cells. Proc. Soc. Magn. Reson., 3: 1364. 1994.

48. Kiss, Z.. Crilly. K. S., and Anderson, W. H. Carcinogens stimulate phosphorylationof ethanolamine derived from increased hydrolysis of phosphatidylelhanolamine inC3H/10T/2 fibroblasts. FEBS Lett., 336: 115-118, 1993.

49. Elias, A. D. Advances in the diagnosis and management of sarcomas. Curr. Opin.Oncol., 4: 681-688. 1992.

50. Coindre. J-M., Terrier, P.. Bui, N. B., Bonichon. F.. Collin, F., Le Doussal, V.,Mandard, A-M.. Vilain. M-O.. Jacquemier, J.. Duplay. H.. Sastre, X., Earlier. C..Henry-Amar. M., Mace-Lesech, J.. and Contesso. G. Prognostic factors in adult

patients with locally controlled soft tissue sarcoma: a study of 546 patients from theFrench Federation of Cancer Centers sarcoma group. J. Clin. Oncol.. 14: 869-877.1996.

51. Redmond, O. M.. Stack. J. P.. Dervan. P. A., Hurson, B. J.. Carney, D. N., and Ennis.J. T. Osteosarcoma: use of MR imaging and MR spectroscopy in clinical decisionmaking. Radiology./72: 811-815, 1989.

52. Palei, S. R., and Benjamin, R. S. The role of chemotherapy in soft tissue sarcomas.Cancer Control. Â¡994:599-605, 1994.

53. Negendank, W. G., Padavic-Shaller, K.. Li. C. W.. Merphy-Boesch. J., Schilder.

R. S., Smith. M. R.. and Brown. T. R. Phosphomonoester changes in vivo andtreatment response in human non-Hodgkin's lymphomas. Proc. Am. Assoc. Cancer

Res., 37: 196-197. 1996.

2972



1996;56:2964-2972. Cancer Res Chun-Wei Li, Annette C. Kuesel, Kristin A. Padavic-Shaller, et al. Magnetic Resonance Spectroscopy

Using Proton-decoupled Phosphorusin Vitro and VivoinMetabolic Characterization of Human Soft Tissue Sarcomas

Updated version

http://cancerres.aacrjournals.org/content/56/13/2964

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/56/13/2964To request permission to re-use all or part of this article, use this link


http://cancerres.aacrjournals.org/content/56/13/2964http://cancerres.aacrjournals.org/cgi/alertsmailto:[email protected]://cancerres.aacrjournals.org/content/56/13/2964http://cancerres.aacrjournals.org/

Documents

Metabolic Characterization of Human Soft Tissue Sarcomas in … · olism of sarcomas in vivo is to use 31P MRS3. The 3IP MR spectrum has signals from phospholipid metabolites, NTPs