Psychosomatic Medicine Faster Service from Outside North America
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Geiser, F.
Right arrow Articles by Liedtke, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Geiser, F.
Right arrow Articles by Liedtke, R.
Related Collections
Right arrow Hematology
Right arrow Eating Disorder
Right arrow Radiology and Brain Imaging
Psychosomatic Medicine 63:631-637 (2001)
© 2001 American Psychosomatic Society

ORIGINAL ARTICLES

Magnetic Resonance Spectroscopic and Relaxometric Determination of Bone Marrow Changes in Anorexia Nervosa

Franziska Geiser, MD, Dipl-Psych, Petra Mürtz, PhD, Dipl-Phys, Götz Lutterbey, MD, Frank Träber, PhD, Dipl-Phys, Wolfgang Block, Dipl-Biol, Katrin Imbierowicz, MD, Guntram Schilling, MD, Dipl-Psych, Hans Schild, MD and Reinhard Liedtke, MD

From the Psychosomatic Clinic (F.G., K.I., G.S., R.L.) and the Clinic for Radiology (P.M., G.L., F.T., W.B., H.S.), University of Bonn, Bonn, Germany.

Address reprint requests to: Dr. Franziska Geiser, Klinik und Poliklinik für Psychosomatische Medizin und Psychotherapie der Universität Bonn, Sigmund-Freud-Str. 25, D-53105 Bonn, Germany. Email: Reinhard.Liedtke{at}ukb.uni-bonn.de


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
OBJECTIVE: The objective of this study was to assess and quantify bone marrow changes in patients with anorexia nervosa using 1H magnetic resonance spectroscopy and relaxometry.

METHODS: The bone marrow fat fraction and the longitudinal and transverse relaxation times (T1 and T2, respectively) of water were measured in the lumbar and femoral marrow of 20 patients with anorexia nervosa and 19 healthy control subjects.

RESULTS: Patients with anorexia nervosa showed significant hyperhydration and reduction of the fat fraction in their bone marrow, predominantly in the proximal femur. These changes were associated with hematological abnormalities. In a retest of seven patients after psychotherapy and gain of weight, the pathological changes in marrow proved to be largely reversible in correlation with the increase in body mass index.

CONCLUSIONS: Fat depletion and excess of tissue water in the bone marrow in anorexia nervosa can be quantified by 1H magnetic resonance spectroscopy and relaxometry. The distribution of the pathological changes in the lumbar and femoral marrow follows the pattern of normal bone marrow conversion from hematopoietic to cellular during childhood.

Key Words: anorexia nervosa • bone marrow • magnetic resonance • 1H spectroscopy • relaxometry

Abbreviations: AN = anorexia nervosa; BMI = body mass index; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, fourth edition; MR = magnetic resonance; SCL-90-R = revised Symptom Checklist-90; SE = spin echo; T1 = longitudinal (spin-lattice) relaxation time; T2 = transverse (spin-spin) relaxation time; TE = echo delay time; TI = inversion delay time; TR = repetition time; TSE = turbo spin echo.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
Hematological abnormalities associated with AN (1) are believed to be the result of bone marrow changes characterized as "serous marrow." These changes were first described some 30 years ago (2–4). They include hypoplasia of the bone marrow and serous atrophy (fat and hematopoietic cell atrophy accompanied by an accumulation of an extracellular matrix rich in hyaluronic acid). Serous marrow was also found in the bones of elderly or cachectic patients and in patients suffering from renal insufficiency, cancer, or HIV infection (4–6). Modern imaging techniques now allow systematic, noninvasive assessment of bone marrow changes.

Mayo-Smith et al. (7) found reduced intravertebral fat content, measured with computed tomography, in AN patients. Vande Berg et al. (8) used T1- and T2-weighted MR images to assess bone marrow changes in AN. They found a waterlike signal intensity pattern in the lumbar, pelvic, and femoral marrow in 6 of 14 patients. Results of a biopsy performed on one patient showed that the marrow had characteristics typical of serous atrophy. This marrow transformation correlated with a low BMI and low blood cell counts. In a second study (9), a predominance of marrow changes was found in the distal lower limbs of anorectic patients. The authors also reported more changes in the diaphyses than in the epiphyses. This distribution pattern is the reverse of that seen in most bone marrow disorders. Lambert et al. (10) reported high correlations between serous atrophy of bone marrow (as assessed by MR imaging) and hematological abnormalities in AN patients. Serous marrow was also associated with a depletion in total body fat mass in this study. Furthermore, MR imaging has been used to ascertain serous atrophy of the clival marrow in patients with AN (11). Several studies have shown that reversal of bone marrow hypoplasia and reduction of gelatinous material occurs during clinical recovery (2, 3, 11).

Until now the determination of MR signals suggestive of serous atrophy of the bone marrow was made on the basis of a qualitative judgment by experienced radiologists. Higher severity of changes was defined by a prolongation of both the T1 and T2 times and by the involvement of a higher number of bone marrow spaces (9).

The aim of our study was to perform a quantitative analysis of the severity of bone marrow changes in patients with AN. Using localized 1H MR spectroscopy and relaxometry, we measured T1 and T2 times and the fat/water relation in the bone marrow of patients with AN and a control group of healthy subjects. Correlations with BMI and hematological findings were examined. Through MR measurements in three different locations, we investigated the distribution of marrow changes. A retest of a subgroup of patients after therapy allowed us to study the reversibility of bone marrow changes.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
Sample Description and Study Design
Subjects in the patient group were recruited from all patients admitted for inpatient or outpatient treatment in our psychosomatic clinic within a time range of 6 months. Inclusion criteria were as follows: 1) a diagnosis of AN following the criteria specified in DSM-IV (maintenance of body weight of <85% of that expected (BMI <17.5 m/kg2) by restrictive and/or purging eating behavior, intense fear of gaining weight, disturbance in the way one’s body weight is experienced, and amenorrhea) (12); 2) duration of illness >6 months; 3) female aged between 15 and 60 years; and 4) absence of consuming illness or psychotic disorder. Written informed consent in accordance with the requirements of the clinic’s ethics committee was obtained for all measurements. For the control group, 19 female university employees agreed to undergo MR testing. Control subjects had to have a BMI >18 m/kg2. They underwent a short clinical interview to exclude those with a history or presence of eating disorders or other severe somatic or psychological disorders.

The patient sample comprised 18 inpatients and 2 outpatients. Median patient age was 23 years (range, 15–56 years). Ten patients had had AN from 6 months to 2 years; 7 patients, from 3 to 6 years; and 3 patients, from 11 to 31 years (chronic anorexia). Twelve patients had the restrictive type of AN, and eight had the purging type. Five patients had been obese at an earlier point in life. The patients’ lowest BMI since the onset of AN ranged from 9.7 to 16.4 m/kg2. One patient had a muscle disease (Kearns-Sayre syndrome) that was nonconsuming; the other patients had no somatic complaints unrelated to AN. All patients were assessed with a modified version of the Structured Clinical Interview for DSM-IV (13) before treatment. Eight patients were given a diagnosis of personality disorder (following DSM-IV criteria) in addition to a diagnosis of AN. Three were given a diagnosis of borderline personality disorder; three, obsessive-compulsive disorder; and two, mixed personality disorder. After 3 weeks of treatment, a second interview was conducted by the clinical director of the department, who had not been directly involved in the treatment, and the diagnoses were reconsidered in a diagnostic conference with staff members. The original AN diagnosis was confirmed for all patients; of the comorbid personality disorder diagnoses, one (mixed personality disorder) was dropped and one was converted from obsessive-compulsive disorder to schizoid personality disorder. For the study, these revised diagnoses were used. Although cognitive and emotional rigidity could be observed in most of the other 13 patients, these characteristics were not sufficiently extended to justify a personality disorder diagnosis, and they were attributed to the psychological and somatic impairment brought on by the AN. The SCL-90-R (14) was administered before treatment to assess symptom distress. On the General Symptom Index, all patients reached a percentage rank (in relation to a normal population) of 80% or more; 10 patients scored higher than 95% of the normal population. Table 1 shows the mean scores on the SCL-90-R subscales. In regard to duration of illness, weight, and psychological distress, our patient sample comprised predominantly severe cases of AN. This reflects the nature of our clinic, where most patients are sent for inpatient treatment because of the severity of their psychosomatic disorder.


View this table:
[in this window]
[in a new window]
  		Table 1. Means and Standard Deviations of the AN Group on the SCL-90-R Subscales
 
Patients underwent MR imaging within the first 14 days of therapy. Routine laboratory studies were reviewed. Patients’ weight at time of investigation ranged from 32.7 to 51.0 kg (median, 40 kg; mean, 38.1 kg). The range of actual BMI was 11.5 to 17.4 kg/m2 (median, 14.4 kg/m2; mean, 14.5 kg/m2; no significant deviation from normal distribution was detected by the Kolmogorov-Smirnov test).

In seven patients a second MR study was performed after significant weight gain was achieved (after 4–16 weeks of psychosomatic therapy). Treatment comprised individual and group psychoanalytic sessions, nonverbal therapy groups, a behavioral program for the improvement of eating behavior, and addition of high-calorie food to the diet (500–1500 kcal/d). The mean age and mean BMI before therapy of the retest subsample did not differ from those of the patient sample (22.9 years and 14.5 kg/m2); five patients had the restrictive type of AN, and two had the purging type. At the time of MR retesting, these patients had gained between 3 and 12 kg (median, 4.6 kg) with a mean increase in BMI of 2.3 m/kg2 (range, 0.8–4.5 m/kg2).

The control group underwent MR testing in the same way as the AN group. Their BMI ranged from 18.2 to 27.5 kg/m2 (median and mean, 22 kg/m2). Their mean weight was 60 kg (range, 50–75 kg). The age range was similar to that of the AN group (21–56 years), although the mean was slightly higher (27 years). Because the MR investigation was very time consuming (see next paragraph), it was not possible to perform complete measurements at all three locations on each volunteer. Each location was measured on a minimum of 8 volunteers.

MR Imaging and 1H MR Spectroscopy
The MR investigations were performed on 1.5 Tesla whole-body systems (Gyroscan S15/ACS II and ACS-NT, Philips Medical Systems, Best, The Netherlands) equipped for imaging and spectroscopy. For MR imaging and 1H MR spectroscopy of the femur, the standard body coil of the MR system was used; MR images and spectra of the lumbar vertebrae were acquired using the body coil as transmitter and a 17-cm-diameter surface coil as receiver. The MR imaging protocols were as follows: for the femur, multislice coronal T1-weighted SE (TR/TE, 600/14 ms) and T2-weighted TSE acquisitions (TR/TE, 2700/120 ms with and 2500/120 ms without fat suppression) with a slice thickness of 5 mm; for the spine, sagittal T1-weighted SE (460/15 ms) and T2-weighted TSE (2000/120 ms with and without fat suppression) sequences with a slice thickness of 4 mm.

We chose as locations the marrow spaces of the femur epiphysis and diaphysis and of the lumbar spine. Localized 1H MR spectra were obtained from cubic voxels of 8 cm3 in the femoral head and in the selected vertebral body of the lumbar spine. In the femoral shaft a bar-shaped volume of 1 x 1 x 3 cm was selected, guided by the preceding MR imaging acquisitions. Chemical shift–selective T1 times for the water component at 4.7 ppm and for the lipid resonance of -(CH2)n- and terminal -CH3 molecular groups at 1.2 ppm were obtained from an inversion recovery series (TR/TE 3000/40 ms) of seven spectra with inversion delays TI increasing from 10 to 2000 ms. Using eight signal averages, each T1 measurement was completed within 3 minutes. T1 values were calculated by separate nonlinear least-squares fits to the respective spectral line integrals S(TI) using the following equation: S(TI) = S0 [1 - 2 x exp(-TI/T1) + exp(-TR/T1)] x exp(-TE/T2). Transverse relaxation times (T2) for water and fat were measured by variation of the echo delay (TE) from 40 to 150 ms in a series of seven SE spectra (Figure 1,a and b). Acquisition time was again 3 minutes with a TR of 3000 ms and eight signal averages. T2 and S0 were determined from a linear regression of ln S(TE) according to the equation S(TE) = S0 x exp(-TE/T2). From the calculated ratio S0f/S0w of the equilibrium magnetizations, the relative fat/water content of bone marrow could be obtained. Detailed descriptions of the MR spectroscopy acquisition parameters used and the procedures applied for processing and fitting of the MR spectra are given elsewhere (15). Because of multiexponential relaxation of the lipid component and neglect of minor abundant components of the triacylglyceride chains (eg, CH2-CO- and -CH2-CH=H at 2.1 to 2.3 ppm, and -HC=CH- at 5.4 ppm), the true fat/water ratio of bone marrow is slightly underestimated by our procedure. However, this systematic effect concerns patient and control group measurements in equal magnitude and thus does not influence the interpretation of our results.



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 1. T2 measurement by SE series of localized 1H spectra in a 29-year-old patient with AN. a) Display of a 10 x 10 x 30-mm voxel of interest in the left femur diaphysis on T2-weighted TSE (2500/120 ms) image before treatment. b) Stacked plot of 1H spectra with increasing echo delays (TE), obtained before treatment (BMI = 13.9 kg/m2). c) Stacked plot of 1H spectra after 3 months of psychosomatic treatment (BMI = 15.4 kg/m2).

 
Although spectroscopic T1 and T2 measurements lasted only 6 minutes for each voxel, the preparation phase, including localized shimming of the magnetic field, had to be repeated for each selected anatomic region. Therefore, the duration of the whole MR imaging and MR spectroscopy protocol was about 1 hour, and the complete set of T1 and T2 measurements in all three desired locations could not be performed in all cases.

Because of the chemical shift of about 3.5 ppm between the water and the lipid resonance frequencies, T1 and T2 values for the water and fat components, respectively, can be measured separately by MR spectroscopic relaxometry. In addition, this technique allows quantitative determination of the fat/water ratio in the investigated tissue volume. In conventional nonselective MR imaging sequences with monoexponential calculation of relaxation times, however, only "mean" tissue T1 and T2 values can be calculated from the measured data. Alterations in the relative fat/water ratio may then lead to apparent variations of these mean T1 and T2 values even if the "true" relaxation times of water and lipids did not change at all.

In the following text, "T1" and "T2" refer to the T1 and T2 relaxation times of water.

Statistics
Statistical data were analyzed with SPSS for Windows, version 9.0. We used Student’s t test for unpaired or paired samples and one-factor analysis of variance for comparisons with more than two samples. Because the aim of the study was descriptive and not confirmatory, levels of significance were not set a priori. Statistical probability levels for each individual test are given.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 
Quantitative Analysis of MR Marrow Changes
Figure 2 shows box plots of the MR outcome variables in the AN group and the control group. Means and standard deviations are given in Table 2. Patients with AN showed a higher variation of marrow fat and free water than control subjects. Because norms for these measurements do not yet exist, we classified as severely abnormal any results that exceeded the mean of the control group by more than 2 SDs (equal to a 5% error probability criterion). Table 2 shows that the most frequent severe marrow change in AN was a prolongation of T1 in the femur epiphysis, followed by a prolonged T1 in the femur diaphysis and reduced fat content in the femur epiphysis.



View larger version (11K):
[in this window]
[in a new window]
  		Fig. 2. Box plots (median, 25th to 75th percentiles, and range) of MR marrow measurements in the AN and control groups.

 

View this table:
[in this window]
[in a new window]
  		Table 2. Means and Standard Deviations of MR Findings in the AN and Control Groups
 
We performed Student’s t test to examine the differences between the AN and control groups (Table 2). Marked differences (significant at a 5% level) between the AN and the control groups were found in the relative fat content of the femur epiphysis bone marrow (AN group showed less fat). Also, T1 (representing the water component) was significantly prolonged in the AN group in all three locations (femur epiphysis, femur diaphysis, and lumbar spine). There were no relevant differences in T2.

Intercorrelation of MR Findings
Within the AN group we found high correlations between fat in the femur epiphysis and diaphysis (r = 0.81, p < .001) and between T1 times at these locations (r = 0.81, p = .003), but not between T2 times at these two locations (r = 0.25, p = .52). Fat and T1 measurements were more highly correlated in the femur epiphysis (r = -0.88, p < .001) than in the femur diaphysis (r = -0.64, p = .03) or the lumbar spine (r = -0.64, p = .005). Results of measurements less likely to show pathological changes (Table 1), like T2 times or measurements in the lumbar spine, were less associated with other findings than results of measurements with a high rate of pathological findings (and a higher standard deviation, like fat and T1 relaxation time in the femur epiphysis).

Correlations of Marrow Changes With Clinical Data
Within the AN group, correlations between MR marrow measures and BMI were moderate. Only the T1 relaxation times in the femur epiphysis and diaphysis showed correlations with the BMI that were significant or close to significant (femur epiphysis: Pearson‘s r = -0.49, p = .04; femur diaphysis: r = -0.53, p = .07) (Figure 3). The duration of illness, type of anorexia (purging or restrictive), personality disorder diagnosis, and psychological symptom distress as measured by the SCL-90-R (General Symptom Index and subscales) were not associated with MR variables.



View larger version (15K):
[in this window]
[in a new window]
  		Fig. 3. BMI and T1 relaxation time in the femur epiphysis in AN patients (r = -0.49, R2 (regression analysis) = 0.24).

 
A reduction of fat in the femur epiphysis was associated with low leukocytes (r = 0.45, p = .04) and low basophile (r = 0.59, p = .01) and eosinophile counts (r = 0.77, p < .001). These parameters were also correlated with an elevated T1 time in the femur epiphysis (leukocytes: r = -0.52, p = .03; basophiles: r = -0.53, p = .04; eosinophiles: r = -0.60, p = .02). Anemia was associated more closely with an elevated T2 time in the femur epiphysis (erythrocytes: r = -0.39, p = .09; hematocrit: r = -0.37, p = .12; mean corpuscular hemoglobin: r = 0.46, p = .07; mean corpuscular hemoglobin concentration: r = 0.44, p = .09) than with a reduced fat content (erythrocytes: r = 0.32, p = .16; mean corpuscular volume: r = -0.38, p = .08). Patients with high T1 times in the lumbar spine showed reduced serum albumin (r = -0.51, p = .07) and elevated serum {gamma}-globulin (r = 0.65, p = .02) levels.

Reversibility of Bone Marrow Abnormalities
To assess the reversibility of bone marrow changes, we performed a second MR imaging study in seven AN patients after psychosomatic treatment and substantial gain of weight (see Materials and Methods). The increase in BMI (mean = +2.3 kg/m2, p = .004) was accompanied by a marked increase of fat content in the femur epiphysis and diaphysis (p = .04 and p = .003, respectively; see Figure 1,b and c, and Table 3). All other measured variables exhibited less important changes in the direction of normality.


View this table:
[in this window]
[in a new window]
  		Table 3. Means and Standard Deviations of MR Findings Before and After Therapy in 7 AN Patients
 
We found a high correlation between the increase in BMI and the fat ratio after therapy (femur epiphysis: Pearson’s r = 0.81; diaphysis: r = 0.95). The duration of time between the two measurements was not associated with the extent of marrow improvement, indicating that the speed of weight gain seems to be less relevant for the remission of bone marrow abnormalities.


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ACKNOWLEDGMENTS
 REFERENCES
 
In this study a new technique of 1H MR spectroscopic relaxometry was used. This technique allows quantification of the proportion of fat (as a percentage) and the content of tissue water in bone marrow, measured by T1 and T2 relaxation times. Ranges of normal results for this technique were calculated from the findings in a control group of healthy subjects. The MR measurements performed on AN patients confirmed prior findings of "serous" bone marrow changes (8, 9). We found a substantial reduction of fat, predominantly in the marrow of the femur epiphysis but also in the marrow of the femur diaphysis, accompanied by a prolongation of T1 relaxation times (in order of importance, in the marrow of the femur epiphysis, femur diaphysis, and lumbar spine). T2 relaxation times were within the limits of our defined norms in most AN patients and were prolonged only in cases of extensive marrow abnormalities. This result supports earlier findings that the T2 time is less sensitive than the T1 time in the detection of altered fat/nonfat marrow balance (16–18). Marrow changes were associated with low leukocyte counts and anemia. Correlations of marrow changes with the BMI were moderate (around r = 0.50). This finding indicates that BMI does not give reliable evidence of the state of bone marrow in AN.

In contrast to the findings of Vande Berg et al. (9), who described more frequent involvement of the diaphysis marrow of the long bones than the adjacent epiphysis, we found the most frequent and most marked alterations in the femur epiphysis, followed by the femur diaphysis. Similar to Vande Berg et al., we observed few changes in the lumbar spine. Vande Berg et al. concluded that marrow changes in AN occur with decreasing frequency from the distal to the proximal end of the bones of the lower limbs. This pattern contradicts a fundamental feature of the distribution of other bone marrow disorders. In contrast, our results indicate that the marrow changes in AN follow the pattern of normal bone marrow conversion from hematopoietic to cellular during childhood, beginning in the epiphyseal ossification centers, followed by the phalanges, diaphysis, flat bones, and metaphysis (19, 20).

Because changes in the femur epiphysis are the most frequent and severe, and correlate with changes in the other locations evaluated in this study, further research on correlations between marrow changes and clinical data should concentrate on measurements in this location. Focusing on the femur epiphysis reduces the MR imaging and MR spectroscopy protocol from 1 hour (for three locations) to about 20 minutes.

An important issue for treatment of AN is the question of reversibility of somatic complications under treatment. Confirming the results of Pearson (2), Lampert and Lau (3) and Kuwashima et al.(11), we found that bone marrow changes were reversible during psychosomatic treatment (with additional high-calorie feeding) in conjunction with an increase of BMI.

One limitation of our study is the relatively small sample, which led to small numbers of patients being analyzed in some statistical comparisons. A small sample was used because of the complexity and high cost of the method of measurement. These factors led to equal or smaller sample sizes in previous studies as well (7–11). To compensate for the low numbers of subjects, unequal sample sizes and sample variances were taken into account in statistical calculations (Levene test; t test with unequal variances, Ref. 21), a sufficiently large control group was included, and only results with both statistical and clinical relevance were considered.

Although MR imaging is too costly to be used for routine clinical testing in patients with AN, it has contributed significantly to our understanding of the nature of processes underlying somatic symptoms associated with AN. The quantitative assessment of marrow changes performed in this study is another step in this direction.


   ACKNOWLEDGMENTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
REFERENCES
 
We thank Carolyn Watkins and Thressa Newell for their English editorial assistance.

Received for publication November 22, 1999.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ACKNOWLEDGMENTS
 REFERENCES
 

  1. Treasure J, Szmukler G. Medical complications of chronic anorexia nervosa. In: Szmukler G, Dare C, Treasure J, editors. Handbook of eating disorders. Chichester: John Wiley & Sons; 1995. p. 197–220.
  2. Pearson HA. Marrow hypoplasia in anorexia nervosa. J Pediatr 1967; 71: 211–5.[Medline]
  3. Lampert F, Lau B. Bone marrow hypoplasia in anorexia nervosa. Eur J Pediatr 1976; 124: 65–71.[Medline]
  4. Seaman JP, Kjeldsberg CR, Linker A. Gelatinous transformation of the bone marrow. Hum Pathol 1978; 9: 685–92.[Medline]
  5. Smith RR, Spivak JL. Marrow cell necrosis in anorexia nervosa and involuntary starvation. Br J Haematol 1985; 60: 525–30.[Medline]
  6. Mehta K, Gascon P, Robbay S. The gelatinous bone marrow (serous atrophy) in patients with acquired immunodeficiency syndrome. Arch Pathol Lab Med 1992; 116: 504–8.[Medline]
  7. Mayo-Smith W, Rosenthal DI, Goodsitt MM, Klibanski A. Intravertebral fat measurement with quantitative CT in patients with Cushing disease and anorexia nervosa. Radiology 1989; 170: 835–8.[Abstract/Free Full Text]
  8. Vande Berg B, Malghem J, Devuyst O, Maldague BE, Lambert MJ. Anorexia nervosa: correlation between MR appearance of bone marrow and severity of disease. Radiology 1994; 193: 859–64.[Abstract/Free Full Text]
  9. Vande Berg B, Malghem J, Lecouvet FE, Lambert M, Maldague BE. Distribution of serouslike bone marrow changes in the lower limbs of patients with anorexia nervosa: predominant involvement of the distal extremities. AJR Am J Roentgenol 1996; 166: 621–5.[Abstract/Free Full Text]
  10. Lambert M, Hubert C, Depresseux G Vande Berg B, Thissen J-P, Nagant de Deuxchaisnes C, Devogelaer J-P. Hematological changes in anorexia nervosa are correlated with total body fat mass depletion. Int J Eat Disord 1997; 21: 329–34.[Medline]
  11. Kuwashima S, Nishimura G, Yamato M, Fujioka M. Magnetic resonance imaging of clival marrow in patients with anorexia nervosa. Acta Paediatr Jpn 1996; 38: 114–7.[Medline]
  12. DSM-IV. Diagnostic and statistical manual of mental disorders. 4th ed. Washington, DC: American Psychiatric Association; 1994.
  13. Wittchen HU, Zaudig M, Fydrich T. Strukturiertes Klinisches Interview für DSM-IV (SKID-I und SKID-II). Göttingen: Hogrefe; 1997.
  14. Derogatis LR. Administration, scoring, and procedures manual-I for the R (revised) version of the psychopathology rating scale series. Baltimore: John Hopkins University School of Medicine; 1977.
  15. Träber F, Block W, Layer G, Bräucker G, Gieseke J, Kretzer S, Hasan I, Schild HH. Determination of 1H relaxation times of water in human bone marrow by fat-suppressed turbo spin echo (TSE) in comparison to MR spectroscopic methods. J Magn Reson Imaging 1996; 6: 541–8.[Medline]
  16. Steiner RM, Mitchell DG, Rao VM, Murphy S, Rifkin MD, Burk DL, Ballas SK, Vinitski S. Magnetic resonance imaging of bone marrow: diagnostic value in diffuse hematologic disorders. Magn Reson Q 1990; 6: 17–34.[Medline]
  17. Smith SR, Williams CE, Davies JM, Edwards RH. Bone marrow disorders: characterization with quantitative MR imaging. Radiology 1989; 172: 805–10.[Abstract/Free Full Text]
  18. Vande Berg BC, Malghem J, Lecouvet FE, Maldague B. Classification and detection of bone marrow lesions with magnetic resonance imaging. Skeletal Radiol 1998; 27: 529–45.[Medline]
  19. Taccone A, Oddone M, Dell’Acqua AD, Occhi M, Ciccone MA. MRI "road map" of normal age-related bone marrow. II. Thorax, pelvis and extremities. Pediatr Radiol 1995; 25: 596–606.[Medline]
  20. Waitches G, Zawin JK, Poznanski AK. Sequence and rate of bone marrow conversion in the femora of children as seen on MR imaging: are accepted standards accurate? AJR Am J Roentgenol 1994; 162: 1399–406.[Abstract/Free Full Text]
  21. Ramsey PH. Exact type I error rates for robustness of Student’s t-test with unequal variances. J Educ Stat 1980; 5: 337–49.



This article has been cited by other articles:


Home page
 ImagingHome page
A J P Goddard and A Gholkar
Diagnostic and therapeutic radiology of the spine: an overview
Imaging, October 1, 2002; 14(5): 355 - 373.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Geiser, F.
Right arrow Articles by Liedtke, R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Geiser, F.
Right arrow Articles by Liedtke, R.
Related Collections
Right arrow Hematology
Right arrow Eating Disorder
Right arrow Radiology and Brain Imaging

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS