oral and maxillofacial radiology Editor: ALLAN

G. FARMAN,

BDS, PhD (Odont),

MBA

Division of Radiology and Imaging Sciences Department of Biology and Physical Sciences School of Dentistry, [Jniversity of Louisville Louisville, Kentucky 40292

Magnetic resonance spectroscopy of inflammation associated with the temporomandibular joint M. E. Alder, DDS, MS,” S. B. Dove, DDS, MS,a V. A. Murrah, DMD, MS,b F. Salinas, BS,’ and R. F. Williams, PhD,C San Antonio, Tex. UNIVERSITY

OF TEXAS

HEALTH

SCIENCE

CENTER

AT SAN

ANTONIO

Noninvasive early recognition and treatment of temporomandibular joint dysfunction remains a diagnostic challenge. This pilot study evaluated the use of phosphorus 31 magnetic resonance spectroscopy with magnetic resonance imaging to measure alterations in pH and high-energy phosphate metabolite ratios of muscle that is adjacent to an inflamed temporomandibular joint. Ten New Zealand white rabbits were used in this study. Two animals were used to develop signal acquisition protocols and to ensure that stable baseline data could be measured. In each of the eight animals used in the experiment, one temporomandibular joint was injected with a suspension of silica particles and the contralateral joint served as a control. Data were collected from control and experimental joints on days 0, 7, 14, 21, and 28, after the injection. At the end of the study, temporomandibular joints were block resected and histologically examined to confirm the presence of an inflammatory response. Results indicated that pH and metabolite ratios could be obtained by 31P-magnetic resonance spectroscopy. Changes in pH and some metabolite ratios in experimental joints showed statistical significance (p < 0.001). Differences were seen on day 2 and day 7 (p = 0.040 and p = 0.008, respectively) in the phosphocreatine/a-adenosine triphosphate ratios. This contrasts with phosphocreatinelb adenosine triphosphate ratios that showed significance that began at day 7 (p = 0.022) and continued to day 14 (p = 0.025). Histologic examination indicated that the tissue response within the joint capsule was less than the granulomatous reaction expected. This pilot study clearly demonstrates that pH changes that occur in muscle that is adjacent to an inflamed joint are detectable by 31P-magnetic resonance spectroscopy. (ORAL

SURC

ORAL

MED

ORAL

PATHOL

1992;74:515-23)

I

t has been reported that approximately 32% of all adults will suffer from someform of temporomandib-

Supported by Institutional Research Grant No. 89-O-336 University of Texas Health Science Center. aDepartment of Dental Diagnostic Science. bDepartment of Pathology. CDepartment of Radiology. 7/16/40355

from the

ular joint (TMJ) dysfunction in their lifetime.lM3 These disorders may manifest symptoms that range from mild localized pain to severe debilitating syndromes. Despite years of research, the early diagnosis of these TMJ disorders remains a challenge to modern dentistry. Treatment of localized inflammatory changes in joints is currently being determined either empirically or invasively and after long-term observation. The 515

516

AEder et al.

ORAL SURG

ORAL MED

ORALPATHOL October 1992

Phosphocreatlne-+ii

Beta Inorganic

I

ATP

Phosphate

11/1(IJ,I,II//,IIII,IIII,IIII/

10

5

0

-i

-10

-15

-20

PPM

Fig. 1. Typical spectrum of normal rabbit skeletal muscle shows peaks of high-energy phosphate metabolites. early diagnosis and initiation of noninvasive therapies seems paramount to the successful treatment of these disorders. The majority of diagnostic information is usually obtained only after the condition has caused considerable damage with accompanying pain and distress to the patient. Therefore early detection of the cellular response to inflammatory changes that occur in and around the TMJ may make it possible to determine a definitive diagnosis, noninvasively, before degenerative changes have occurred in the joint. Changes in the TMJ include degenerative conditions such as osteoarthritis, traumatic injuries (acute and chronic), autoimmune conditions (rheumatoid arthritis), and septic conditions caused by trauma or other bacterial invasions.4 A common feature of TMJ dysfunction is a localized inflammatory process associated with the TMJ. The inflammatory response can result from the inability of the host to mount sufficient high-energy metabolites to continue a successful defense, or the inflammation can be a therapeutic response to an irritant of any type? Cell migration and the associated metabolic changes within and surrounding the synovium are the hallmarks of the initiation of the inflammatory process.6 Death, ischemia, and anoxia on a cellular level are indications of distress. Associated muscles of mastication may also exhibit conditions that range from mild inflammation (slightly tender to palpation) to complete spasm (tetany). The presence of these conditions alters the chemical composition of the affected areas. The hydrogen ion concentration and the high-energy metabolites of phosphorus are subject to change.7, * If these changes involve an increase in hydrogen ion concentration, they will cause an increase in proton signal intensity in magnetic resonance images (MN) of the affected region. Alterations in the high-energy phosphorus metabolites are also visible as changes in

31P chemical shifts or phosphorus metabolite concentrations (peak areas) by magnetic resonance spectroscopy (MRS). Typical magnetic resonance spectra of phosphorus (31P) in muscle tissue show peaks for adenosine triphosphate (a, 0, and y ATP), phosphocreatine (PCr), phosphodiesterase, phosphomonoesterase, and inorganic phosphate (Pi) (Fig. 1). The concentrations of these important phosphates can be estimated. Hydrolysis of PCr and ATP produces increasing concentrations of Pi. Anoxia, ischemia, or cell death produces a breakdown of the high-energy phosphates (PCr and ATP) and a rise in the inorganic phosphorus and lactic acid production from glycolysis, which lead to a fall in PH.~ Tissue pH can be measured by the chemical shift of the Pi peak (relative to the PCr peak). Chih-Hsiung et al. lo demonstrated that localized MRS quantitates metabolic changes in skeletal muscle during hemorrhage and resuscitation. Their results show that the reduction of intracellular Pi and increased levels of PCr are correlated with cell survival. The literature is replete with investigations of changes in 31P metabolites in many organs including muscle,” the brain, 12-t7 the kidney, t* the liver, and other organs.t9> 2o MRS and its potential role in the diagnosis and management of pediatric patients who suffer ischemic cerebral episodes have been shown2’ Chemical shift changes in the various phosphorus metabolites in a malignant growth have also been demonstrated.22, 23 However, there is an absence of literature that addresses the effects of inflammation on the phosphorus metabolites in and around a synovial joint over time. The aims of this study were to determine whether the techniques of MRS and magnetic resonance imaging (MRI) could be used to identify and longitudi-

31PNMR spectroscopy of injamed TMJ muscle 517

Volume 74 Number 4

Fig. 2. Stereotaxic on surface coil (B).

device with rabbit mandible

in position over surface coil. Condyle (A) is seen centered

nally monitor alterations in intracellular pH and 31P metabolite ratios of muscle adjacent to an area of induced inflamma,tion in the rabbit TMJ, and to determine the number of samples necessary to validate the statistical analysis of these variables for future study. MATERIAL AND METHODS Data acquisition

The image and spectroscopic data were acquired with a 2.0 Tesla 45 cm bore General Electric (GE) CSI-II Imaging spectrometer (General Electric Co., Fairfield, Corm.) equipped with an Oxford superconducting magnet (Oxford Research Systems, Oxford, England). Animal protocol

Ten New Zealand. white rabbits, which weighed 2 kg to 3 kg, were used in the study. The animals were weighed on acquisition and fed a standard diet throughout the experiment. For each experimental animal (n = 8), one ‘TMJ, which was arbitrarily chosen, served as the treatment effect (experimental) joint. The contralateral joint in the same animal was not treated and thus served as a control joint. In the remaining two rabbits, both joints served as additional pure controls, with no treatment intervention other than the anesthetic protocol and data collection described. The animals were monitored for distress and weighed at intervals during the data collection periods. The care and use of the New Zealand white rabbits of this study were in accord with the guidelines and protocols of the Animal Care and Use Committee of the University of Texas Health Science Center at San Antonio.

The rabbits were anesthetized with a combination of agents at a dose of 1.O ml/ 1.5 kg total body weight given by intramuscular injection. The anesthetic consisted of ketamine hydrochloride (28.74 mg/kg), xylazine hydrochloride (6.37 mg/kg), and acepromazine maleate (0.94 mg/kg). Immobility while data were collected was maintained by an additional 0.5 ml of the anesthetic administered intramuscularly every 25 to 35 minutes without removing the animal from the magnet or changing the animal’s position. To induce inflammation in the experimental joint, the rabbits were anesthetized and the joint was injected with a suspension of silica particles (100 to 500 mesh) in sterile water. The muscles adjacent to both TMJs were imaged (proton) and a 31P spectral analysis was done at one session. The pure control group (n = 2) was imaged bilaterally on the same schedule as the experimental group. The rabbits were killedonday2(n= l),day7(n= l),day14(n= l), and day 28 (n = 5). The TMJ and adjacent muscles were immediately block resected. Histologic preparations were presented in a blind random method for histologic evaluation. The response of the pathologist was limited to a multiple choice of four levels (none, mild, moderate, or severe) in four categories (inflammatory response, inflammatory changes, presence of necrosis, and bone remodeling). Imaging

Before any experimental manipulation of the animal, a baseline Tr weighted image (Repetition time/ Time to Echo (TR/TE) = 600/30 msec) and a T2 weighted image (TR/TE = 2500/60 msec) were acquired for both the left and right TMJ region with the

518

Alder et al.

ORALSURG

ORAL MEDORALPATHOL

October 1992

Fig. 3. Sagittai proton image of rabbit using 15cm volume coil with surface coil in place. Area of interference (diminished signal) corresponds to area that coil is able to excite when tuned. Importance of verification is demonstrated here. This coil (I) was not positioned directly over TMJ (2).

useof a 25mm surface coil. The field of view was 50 mm X 50 mm, with a slice thickness of 3 mm. Image data were collected as 256 phase-encodedstepswith a phase-encodetime of 4.0 msec. In the case of the pure control animals (n = 2), the same schedule of imaging was used except that only one T2 weighted image was acquired. After experimental manipulation (silica injection) and in accord with the protocol schedule, Ti imageswere taken to verify the location of the desired volume regions for subsequent spectroscopy and analysis. Spectroscapy Spectra were obtained with a two-turn 25-mm surface coil tunable to both proton and phosphorus (‘H and 31P). The radio frequency (RF) coil was placed in a cradle that was supported in the y-dimension and the TMJ of the rabbit was placed directly against the coil as shown (Fig. 2). This stereotaxic device reproducibly positions the rabbit within the magnet and on the surface coil. The stereotaxic device and RF surface coil were developed on site by the NMR Research Facility of the University of Texas Health Science Center at San Antonio. Magnetic field homogeneity wasmonitored by the proton line width and

ranged from 0.4 ppm to 0.9 ppm for all 31Pspectroscopy and proton-imaging experiments. The left and right TMJ area of each animal was spectroscopically analyzed at each data collection session.A nonlocalized spectrum was obtained for each sessionand followed by a phase-encoded sequenceto acquire localized spectra from 32 relatively discrete volumes. In phase-encoded spectroscopy, signal cutoff per volume is not sharp, but tends to bleed slightly into the next volume. However, phase encoding minimizes the influence of other tissuesthat are not of interest, such as skin and muscle. The same schedule was followed for all of the data collection sessionsfor control and experimental joints. Nonlocalized spectra were acquired with 128 acquisitions, an interpulse delay of 2 secondsand a pulse width of 4.5 degrees, which made the total data collection time about 7 minutes. Localized spectra were obtained in a one-dimensional experiment with 32 phase-encodedstepsover a 64-mm field of view to yield 2-mm contiguous slices. This phase-encoding sequenceusesonly one gradient to minimize homogeneity distortions.24 A 1200-msec interpulse delay with a pulse width of 40 psec gave optimum results with a total data collection time of about 20 minutes. Signal intensities and areas of the phosphorusmetabolites PCr, Pi, ATP (LY,,B,y), and total phosphate were calculated by the curve-fitting routine GEMCAP (GE.). The GEMCAP program permits an analysis of a spectrum of overlapping peaks and provides the positions, intensities, widths, and relative areasof up to 26 individual lines. The program allows a combination of Lorentzian and Gaussian shapes. Signal intensities (peak height and area) are proportional to the amount of each phosphorus compound present in the tissue being sampled; however, no attempt was made to obtain the absolute quantity of the high-energy phosphate compounds because an external reference was not used. The ratios of the peaksrelative to each other were determined and their values compared. All possible ratios were analyzed. Intracellular pH was computed from the chemical shift in ppm (a) of the Pi resonance peak relative to the PCr resonance peak25as shown in equation 1: pH = 6.72 i log (a - 3.27)/(5.69 - a)

(1)

epth and volume localization To verify the area excited by the surface coil and to correlate the MRI and the spectra, the dual-tuned (3iP and ‘H) surface coil was placed over the area of the TMJ. The rabbit was then placed in a 15cm volume imaging coil and a normal proton (‘H) image was taken with the ‘H-tuned coil in place. This pro-

Volume Number

31P NMR

74 4

spectroscopy

of injlamed TMJ muscle

519

Fig. 4. Tr-weighted image of rabbit TMJ (arrow] taken with 25-mm surface coil. Part of the lens of the eye is visible in lower portion of figure. Lines are superimposed as part of macro sequence as image is displayed; lines correspond to volumes acquired during 31P MR spectroscopy. duced an image with the area excited by the ‘H-tuned coil that appeared black. By detuning the proton portion of the surface coil but leaving it in place, the area of interest and the areas excited by the surface coil became visible, but with diminished brightness and with some ‘loss of overall resolution. Nevertheless, precise identification of the area of interest was possible (Fig. 3). Because the MR spectrometer used (GE. CSI 2T/45 G-eneral Electric Co., Fairfield, Conn.) is to measure and localize the area to be phase encoded directly from the image that appears on the monitor, this technique allows precise and reproducible positioning of the surface coil over the region of interest. To further localize the spectra from the total area excited by the surface coil, a program to generate and to define the 32 discrete volume elements 2 mm thick was used. These phase-encoded volumes are directly correlated with the MRI as superimposed lines on the image (Fig. 4). This same procedure was followed for all data collection sessions. Statistical

methods

A two-factor analysis of variance (ANOVA) and t test (SuperANOVA and StatView II respectively, Abacus Concepts, Inc., Berkley, Calif.) were u.sed to compare the calculated pH and metabolite ratios. All statistics were calcul.ated as two-tailed probabilities with significance set at p < 0.05. RESULTS Comparison of weight gains between pure control and experimental animals indicated no significant

Table I. A two-factor ANOVA comparison of joint (control and experimental), days (day 2, 7, 14, 28) and joint/days combined

Source

1 oj$EEn

Joint

1 4 3 67

Day Joint/day Residual

0.274 0.212 0.100 0.608

0.274 0.053 0.033 0.009

30.185 5.850 3.681 -

0.0001 0.0004 0.0162 -

Dependent variable: pH Note: One row has been excluded from calculations because of missing values.

Table II. Comparison of mean pH between control and experimental joints Changes

Day

Diflerence

2 7 14 28

-0.12 -0.25 -0.03 -0.11

in pH for Standard error 0.05 0.05 0.05 0.05

days and joints

t

test

-2.539 -5.404 -0.695 -2.400

P-value prob)

(btail

0.0127 0.0001 0.4F393* 0.0192

*Not significant.

differences. Histologic evaluation of all experimental joints indicated moderate to severe inflammation within the area of the TMJ. None of the control joints showed evidence of inflammation.

520

Alder et al.

ORALSURGORAL

WEDORAL

PATHOL

October 1992

7.200

6.800 6.600 6.400 6.200 6.000

Fig. 5. Comparison of mean pH control and experimental joint by day. Significant differences (JI < 0.05) are seen on day 2 and day 7.

Significant differences were seen on day 2 (p = 0.040) and day 7 (p = 0.008) in the PCr/ac ATP peak area ratios (Fig. 6). This contrasts with the phosphocreatine/a ATP ratios, which showed significant differences from day 7 (p = 0.022) to day 14 (p = 0.025) (Fig. 7). ANOVA components were used to determine sample sizes required to provide statistical power of 0.80 (a-tail probability, p < 0.05).26 The power analysis demonstrated that the number of joints required to verify the significance of these preliminary results was feasible for some but not all 31P ratios.

e 2 3.5 G g 3.0 CL 6 u) 2.5 ii P 2.0

DISCUSSION 2

7

28

Fig. 6. Comparison of control and experimentai joint mean and PCr/aATP ratios by day. Significant differences (p < 0.05) are seen on day 2 and day 7.

A two-factor ANOVA of calculated pH values for control and experimental joints indicated significant difference. (p < 0.001). Significance was also found between days (p = 0.0004) and joint/days combined (p = 0.0162) (Table I). In addition, the calculated pH values between control and experimental joints at day 2, day 7, and day 28 showed significance at the 0.05 level by paired t test (Table II and Fig. 5).

The rabbit TMJ demonstrates similarity to the human TMJ in both anatomy and function, which makes it an excellent experimental mode1.27 In addition, the use of the rabbit TMJ as a research model provides a method to monitor the degree of joint usage by periodically weighing the animal. The joint must be used to chew the normal diet provided; therefore guarding or decreased use of the joint will result in a weight loss or a failure to thrive. Because no weight loss or differences between the control and experimental animals were observed, there was no guarding or decreased use of the joint after injection of the irritant. Classic irritants frequently used in the arthritis model (Freund’s incomplete or complete adjuvant) may create alterations from the norm in the 31P me-

Volume 74 Number 4

31P NMR

spectroscopy

of inflamed

TMJ muscle

521

n Control q Experiment

2.5 -

0

2

7

14

28

Day 7. Comparisonof control and experimentaljoint meanandPCr/@ATP ratios by day. Significant differences(p < 0.05) are seenon day 7 and day 14.

Fig.

tabolites seen.28 For instance, Freund’s complete adjuvant contains whole organisms of Mycobacterium tuberculosis that have been heat-killed and freezedried.29 The exogenous cellular materials contain the residues of several phosphate compounds (phospholipids, phosphoproteins, Pi). These could alter the phosphate metabolite concentrations seen in the MRS spectrum. Therefore silica was chosen as an inflammatory-inducing agent because of its ability to sustain a physical inflamma.tory reaction and its lack of influence on the phosphate metabolite concentrations. The physical inflammatory reaction is due to the direct toxic actions of silica particles on the phagocyte. Inasmuch as silica cannot be degraded by the phagocyte after ingestion, the phagocyte dies. After the death of the phagocyte, the silica particles are released into the cellular milieu and the process is repeated, and thus inflammation is sustained or enhanced. Migration of the particles from the site of injection is also minimal. 3o Although the use of the silica-induced model of inflammation has been used extensively, correlation of the type of inflammation induced by this method and that type of clinical inflammation associa.ted with TMJ disorder has not been demonstrated. IBecause the intent of this study was to evaluate the use of MRS to assess inflammatory changes, no direct correlation can be made with clinical inflammation associated with TMJ disorders. For MRS to be a significant diagnostic tool, the region of interest must be localized. Surface coils have been successfully used to limit the volume of interest to an area immediately beneath the coil.‘” The observable tissue volume of a surface coil is approx-

imately equal to the diameter of the coil plus a depth equal to the radius.31 At shallow sample depths, surface coils provide dramatic advantages in signal-tonoise (S/N) ratio over volume coils because they are closer to the signal-generating nuclei and remote from a large fraction of the sample that contributes only noise. However, the high sensitivity to surface tissue and the contamination that results from magnetic resonance signals in surrounding tissues, deeper lying organs, or bone must be overcome when using surface coils.32 A large number of spatial localization schemes have been proposed and demonstrated.33-37 These localization sequences all strive to overcome the degradation of the spectroscopic resolution by eliminating as much of the extraneous excitation of nuclei as possible. MRS produces the highest resolution spectra in an extremely homogeneous magnetic field and, at the same time, provides a high S/N ratio. Gradients used to localize the tissue sample tend to spoil the homogeneity of the magnet and decrease the S/N ratio. All spatial localization methods must introduce gradients into the field to spatially encode spectra at depths within an in vivo system. These gradients reduce the homogeneity of the field and thus decrease the S/N ratios. Because some of the chemical shifts to be measured are exceedingly small (on the order of 1 ppm or smaller), the homogeneity of the magnetic field over the sample volume must be substantially better for spectroscopy than for imaging purposes.38 Therefore spatial localization techniques that introduce a minimum of gradients into the field give spectra of higher resolution and higher S/N ratios. This

study used a single gradient to localize a volume lim-

522

Alder et al.

ited in two dimensions by the size of the surface coil and in the third dimension by the phase-encoded slice, thus allowing a sample of tissue approximately equal to 500 mm3 to be analyzed. The volume excited by the surface coil is usually small in relation to the total volume of the tissue or organ system of the species to be studied and is normally at some depth relative to the surface. The use of a surface-coil generated MRI provides a noninvasive visual confirmation of the volume excited by the coil and eliminates the possibility that spectra would be obtained from areas other than those of interest to the diagnostician. The major metabolic pathways for energy production and biosynthesis in muscle are the high-energy phosphate reactions that occur within the mitochondria and cytoplasm of the cell. Anoxia or ischemia inhibits oxidative phosphorylation and diminishes ATP synthesis. If PCr is abundant, it will be used to maintain appropriate ATP levels. Once the use of ATP exceeds the rate of synthesis, ATP levels will fall. Furthermore, when oxidative production of ATP is slowed, glycolysis will be stimulated and result in an increased production of lactic acid and consequently tissue acidosis.39 In an acute localized inflammatory process, intracellular anoxia, ischemia, or both may be present as a result of the hemostasis that occurs with the vasodilatation. Injury to muscle fibers has been shown to change the pH of muscle, as well as some of the phosphate metabolite ratios.40 The results of these pilot data showed positive correlations between 3’P MRS metabolite changes and inflammation in muscle tissue adjacent to an induced, sustained inflammation in a TMJ. Because of the small sample size and the variability seen within each animal, these results must be viewed as preliminary data. A power analysis of the data indicated that validation of the study could be achieved with fewer than 30 joints for most of the metabolite ratios studied. Because no previous data exist on the use of MRS to assess inflammatory changes, all measurable metabolites were evaluated as potential markers. Instrumentation variables that may have influenced the results include variability in signal strength over the surface of the coil and variations in the degree of shimming accomplished that affects the homogeneity of the magnetic field and thus the line width of the spectral peaks. The lack of an external standard limits the ANOVA in ratio. No quantitation was possible, although the ratios could be compared. In fact, the significance of the earlier changes in PCr/a ATP ratio compared with the later changes in Per/@ ATP ratio requires additional analysis because of the possible contamination of the a! ATP resonance by other phosphate-containing species such as adenosine diphosphate and nicotinamide dinucleotide that

ORAL

SmG

ORAL

MEU ORAL PATK~L October 1992

may or may not be of significant concentration in the inflamed TMJ. In addition, the degree of inflammation varied, which may account for some of the trends in the PCr/ATP ratios observed. Currently, we are evaluating other irritants to determine their effectiveness as models for TMJ dysfunction. External standards are also being tested to provide for direct quantitation of metabolites and verification of the signal strength across the surface coil. In addition, larger sample sizes are being used to assure statistical validity. These are determined from the power analysis. Results of this initial work clearly demonstrate that high-energy metabolite changes that occur during inflammation are easily detected by MRS. The data indicate that pH has measurable, stable baseline values in muscle adjacent to a TMJ and significant measurable shifts in the pH of muscle tissue adjacent to an experimentally inflamed joint. No significance was demonstrated on day 14 in the mean pH; this may be due to the relatively small sample size and associated low statistical power of the test used. MRS is very technique sensitive, which may also have influenced the lack of significance. Changes in high-energy phosphate metabolites indicate trends that appear to correlate with the duration and degree of inflammation. We are in the process of investigating these trends further. Power analysis and sample size calculations indicate animal group sizes that are feasible and reasonable for further statistical validation. Consequently, the use of localized 31-P MRS has substantial potential as a noninvasive method of monitoring inflammation associated with TMJ dysfunction. The Support from the University of Texas Health Science Center Nuclear Magnetic Research Facility for instrument time and coil development is greatly appreciated. REFERENCES 1. Solberg WK, Woo MW, Houston JB. Preva!ence of mandibular dysfunction in youngadults.J Am Dent Assoc 1979;98:2534. 2. Kircos LT, Ortendahl DA, Mark AD, Arakawa M. Magnetic resonance imaging of the TMJ disk in asymptomatic volunteers. J Oral Maxillofac Surg 1987;45:852-4. 3. Guarlnick W, Daban LB, Merril RG. Temporomandibular joint afflictions. N Engl J Med 1978;229:123-9. 4. Goldenberg DL, Chisholm PL, Rice PA. Experimentai models of bacterial arthritis: a microbiologic and histopathologic characterization of the arthritis after the intra-articular injections of N. gonorrhoeae, S. aureus, group A streptococci, and E. coli. J Rheumatol 1983;10:5-11. 5. Hurley JV. The sequence of early events. In: inflammation. New York: Springer-Verlag 1978:26-7,48. 6. Di Rosa M. Inhibition of cell migration in vivo and granuloma formation. In: anti-inflammatory drugs. New York: SpringerVerlag 1979:225-6. 7. Wismer GL, Rosen BR, Buxton R, Stark DD, Brady TJ.

Volume Number

8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

18.

19.

20. 21.

22.

23.

24. 25.

74 4

Chemical shift imaging of bone marrow: preliminary experience. AJR 1985;145:1031-7. Bottomley PA, Hardy CJ. Rapid, reliable in vivo assays of human phosphate metabolites by nuclear magnetic resonance. Clin Chem 1989;35:392-5. Matson GB, Weiner M W. MR spectroscopy in vivo: principles, animal studies, and clinical applications. In: Stark DD, Bradley WG. eds. Magnetic resonance imaging. St. Louis: CV Mosby 19887%85,99. Chih-Hsiung WR, Bogusky T, Holcroft JW, Dramer GC. NMR monitoring of phosphate metabolism of rat skeletal muscle during hemorrhage and resuscitation. J Trauma 1988;28:757-64. Park JH, Vansant JP, Kumar NG, et al. Dermatom,yositis: correlative MR imaging and p-3 1 MR spectroscopy for quantitative characterization of inflammatory disease. Radiology 1990;177:473-9. Ackerman JJ, Grove TH, Wonh GG, Gadian DG, Radda GK. Mapping of metabolites in whole animals by 31P NMR using surface coils. Nature 1980;283:167-70. Brant-Zawadzki M, Weinstein P, Bartkowski H, Moseley M. MR imaging and spectroscopy in clinical and experimental cerebral ischemia: a review. AJR 1987;148:579-88. Prichard JW, Shulman RG. NMR spectroscopy of brain metabolism in vivo. Annu Rev Neurosci. 1986;9:61-85. Chang LH, Cohen Y, Weinstein PR, Chileuitt L, James TL. Interleaved ‘H and 31P spectroscopic imaging for studying regional brain injury. Magn Reson Imaging 1991;9:223.-7. Baker LL, Kucharczyk J, Sevick RJ, Mintorovitch J, Moseley ME. Recent advances in MR imaging/spectroscopy ad cerebral ischemia. AJR 1991:156:1133-43. Griffey RH, Brown MS, Bankhurst AD, Sibbitt RR, Sibbitt WL Jr. Depletion of high-energy phosphates in the central nervous system of patients with systemic lupus erythematosus, as determined by phosphorus-3 1 nuclear magnetic resonance spectroscopy. Arthritis Rheum 1990;33:827-33. Adam WR, Koretsky AP, Weiner MW. Measurement of renal intracellular pH in rats in vivo using 31P NMR: erects of acidosis and Kt depletion. Am J Physiol 1986;251:904-10. Koretsky AP, Wang S, Murphy-Boesch J, Klein MP, James TL, Weiner ML. 31P NMR spectroscopy of rat organs, in situ, using chronically implanted radio frequency coils. Proc Nat1 Acad Sci USA 1983;80:7491-5. Schaefer S. Clinical nuclear magnetic resonance spectroscopy: insight into metabolism. Am J Cardiol 1990;66:45F-50F. Leonard JC. Younkin DP. Chance B. et al. Nuclear magnetic resonance: an overview of its spectroscopic and imaging applications in pediatric patients. J Pediatr 1985;106:756-61. Griffiths JR, Stevens AN, Iles RA, Gordon RE, Shaw D. 3tP NMR investigation of solid tumors in the living rat. Biosci Rep 1981;1:319-25. Madden A, Glabolm J, Leach MO. An assessment of the sensitivity of in vivo 31P nuclear magnetic resonance spectroscopy as a means of detecting pH heterogeneity in tumours: a simulation study. Br J Radio1 1990;63:120-4. Freeman D. Spectroscopic imaging, CSI Application notes 1987:1-3. James TL. In vivo nuclear magnetic resonance spectroscopy.

31P NMR

26. 27.

28.

29.

30.

31.

32. 33.

34.

35. 36.

spectroscopy

of inflamed TMJ muscle

523

In: NMR, CT, and interventional radiology. San Francisco: Department of Radiology, University of California, 1984. Cohen J. Statistical power analysis for the behavioral sciences, rev. ed, New York: Academic Press, 1977:357-410. Holmlund A. Hellsina G. Bana G. Arthroscoov of the rabbit temporomanhibular joint. Int j Oral Maxill&c Surg 1986; 15:170-5. Di Rosa M, Giroud JP, Willoughby DA. Studies of the mediators of the acute inflammatory response induced in rats in different sites by carrageenan and turpentine. J Path01 1971; 104:15-29. Turk JL, Willoughby DA. Immunological and para-immunological aspects of inflammation. In: Vane JR, Ferreira SH, eds. Inflammation: handbook of pharmacology. Berlin: SpringerVerlag 1985;5O/l:P 231 Sedgwick AD, Willoughby DA. Initiation of the inflammatory response and its prevention. In: Bonta IL, Bray MA, Parnham MJ, eds. Pharmacology of inflammation. In: Glynn LE, Houck JC, Weissmann G, series eds. Handbookof inflammation. New York: Elsevier Science 1985;5:37-8. Bottomley PA, Edelstein WA, Hart HR, Schenck JF, Smith LS. Spatial localization in 31P and r3C NMR spectroscopy in vivo using surface coils. Magn Reson Med 1984;1:410-3. Bottomley PA. Spatial localization in NMR spectroscopy in vivo. Ann N Y Acad Sci 1987;508:333-48. Aue WP, Muller ST, Cross A, Seelig J. Volume-selective excitation: a novel approach to topical NMR. J Magn Res 1984;56:350-4. Bendall MR, Gordon RE. Depth and refocusing pulses designed for multipulse NMR with surface coils. J Magn Res 1983;53:365-85. Bottomley PA. A versatile field gradient control system for NMR imaging. J Phys E, Sci Instrum 1981;14:1801. den Hollander JA, Luyten PR. Image-guided localized ‘H and 31P NMR soectroscoav of humans. Ann N Y Acad Sci 1987:

508:386-98: -37. Ordidge RJ, Mansfield P, Lohman

JAB, Prime SB. Volume selection using gradients and selective pulses. Ann N Y Acad Sci 1987;508:376-85. 38. Kushmerick MJ. Spectroscopic applications of magnetic resonance to biomedical problems. Cardiovasc Intervent Radio1 1986;8:382-9. 39. Koretsky AP, Weiner MW. 31-Phosphorous nuclear magnetic resonance magnetization transfer measurement of exchange reaction in vivo. In: Biomedical magnetic resonance. San Francisco: Radiology Research and Education Foundation,

1984. 40. McCully

KK, Argov Chance B. Detection magnetic resonance (11):212-6.

Z, Boden BP, Brown RL, Bank WJ, of muscle injury in humans with 31-P spectroscopy. Muscle Nerve 1988;

Reprint requests. Marden E. Alder, DDS, MS Department of Dental Diagnostic Science University of Texas Health Science Center 7703 Floyd Curl Drive San Antonio, TX 78284-7919

at San Antonio