| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Musculoskeletal Imaging |
1 From the Departments of Diagnostic Radiology and Organ Imaging (J.F.G., G.E.A., J.K.T.W., A.T.A.), Orthopaedics and Traumatology (S.M.K., K.H.C., K.M.C., P.C.L.,), Medicine (D.S.C.H., A.K.L.W.), Anaesthesia and Intensive Care (G.M.J.), and Centre for Epidemiology and Biostatistics (A.Y.K.C.), Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, New Territories, Hong Kong. Received January 19, 2004; revision requested March 19; final revision received June 30; accepted July 26. Address correspondence to J.F.G. (e-mail: griffith@ruby.med.cuhk.edu.hk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The hospital ethics committee approved the study, and all patients provided written informed consent. A total of 254 patients with confirmed SARS treated with steroids underwent evaluation with magnetic resonance (MR) imaging for osteonecrosis. Clinical profiles, joint symptoms, relevant past medical and drug history, steroid dose, and radiographic and MR imaging evidence of osteonecrosis and other bone abnormalities were evaluated. Mann-Whitney, Kruskal-Wallis, and Pearson exact 
2 tests were performed, and univariate and multivariate logistic regression analyses were applied.
RESULTS: One hundred thirty-four (53%) of 254 patients had recent onset of large joint pain, but 211 (80%) of 264 painful joints were not associated with abnormality on MR images. MR images in 12 (5%) of 254 patients showed evidence of subchondral osteonecrosis in the proximal femur (n = 9), distal femur (n = 2), and proximal and distal femora and proximal tibiae (n = 1). Additional nonspecific subchondral and intramedullary bone marrow abnormalities were present in 77 (30%) of 254 patients. Results of multiple logistic regression analysis confirmed cumulative prednisolone-equivalent dose to be the most important risk factor for osteonecrosis. The risk of osteonecrosis was 0.6% for patients receiving less than 3 g and 13% for patients receiving more than 3 g prednisolone-equivalent dose. No relationship was found between additional nonspecific bone marrow abnormalities and steroid dose.
CONCLUSION: An appreciable dose-related risk was found for osteonecrosis in patients receiving steroid therapy for SARS. Additional nonspecific bone marrow abnormalities were frequent. Joint pain was common after SARS infection and was not a useful clinical indicator of osteonecrosis.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Many patients with SARS received high-dose steroid therapy, which was administered for a variable period. A serious side effect of steroid therapy is osteonecrosis. Unlike many other side effects (such as immunosuppression, myopathy, and reduced bone density), osteonecrosis, once established, does not regress after discontinuation of steroid therapy. High doses of steroids administered over a short period to patients who are not predisposed to osteonecrosis seem to confer little or no risk of osteonecrosis (2). On the other hand, high doses of steroids administered over a longer period to osteonecrosis-predisposed patients (such as patients with systemic lupus erythematosus, rheumatoid arthritis, malignancy, or organ transplant) are associated with a dose-related risk of osteonecrosis of between 4% and 52% (3,4).
The majority of patients treated for SARS in our hospital were not predisposed to osteonecrosis. They received variable doses of steroid therapy for a variable period.
Thus, the purpose of our study was to evaluate whether there is a relationship between steroid treatment and risk for osteonecrosis of the hip and knee in patients with SARS.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
The 50 patients who recovered from SARS infection and who were not evaluated with MR imaging were classified as nonparticipants. To ensure uniformity between participants (n = 254) and nonparticipants (n = 50), the variables of age, sex, maximum radiographic score, days in hospital, and intensive care unit admission were analyzed.
Questionnaire and Steroid Dose
Prior to MR imaging, the patient’s height and weight were measured and a four-page questionnaire was completed to determine whether there were any preexisting factors that predisposed the patient to osteonecrosis. The patient was questioned about the presence and duration of joint symptoms, past medical and trauma history, past and current medications (including topical steroids, Chinese herbal medicines, oral contraceptives, and nonsteroidal antiinflammatory agents), past and present smoking habits, and alcohol consumption.
All patients who showed no response to antibiotic therapy within 48 hours of initial administration received intravenous ribavirin at a dosage of 24 mg/kg/d and hydrocortisone at a dosage of 10 mg/kg/d.
Pulsed intravenous methylprednisolone (500–1000 g/d) was administered to patients who were affected by progressive deterioration (6). Hospital databases were accessed for steroid dosage information. Prednisolone-equivalent doses were calculated by adjusting the hydrocortisone or methylprednisolone dose to the prednisolone-equivalent dose on the basis of antiinflammatory potency. Conversion factors of 0.20 and 1.25 were used to calculate prednisolone-equivalent doses of hydrocortisone and methylprednisolone, respectively.
Osteonecrosis risk was compared with patient age, sex, weight, body mass index, smoking, alcohol intake, preexisting comorbidity, intake of other medications, maximum radiographic score (7), intensive care unit admission, days in the intensive care unit, days in hospital, alanine transaminase (ALT) level at admission, lactate dehydrogenase (LDH) level at admission, maximum LDH level, cumulative prednisolone-equivalent dose, cumulative prednisolone-equivalent dose adjusted for weight and body mass index, and average daily prednisolone-equivalent dose to determine whether there was a correlation. The relationship between hip or knee pain and the presence of osteonecrosis or other bone abnormalities was analyzed, as was the relationship between the severity of SARS infection (determined with reference to radiographic score, ALT level at admission, and LDH level at admission or maximum LDH level) and osteonecrosis (6).
MR Imaging Protocol
A modified MR imaging protocol was used that was designed specifically for osteonecrosis evaluation (8). Both hips were examined simultaneously, and then both knees were examined.
MR imaging was performed with a 1.5-T unit (Magnetom; Siemens, Erlangen, Germany). For imaging of the hips, a coronal T1-weighted spin-echo sequence (590/20 [repetition time msec/echo time msec]) was applied, with a section thickness of 3 mm, intersection gap of 0.3 mm, field of view of 350 mm, and matrix of 256 x 512. A coronal T2-weighted fat-suppressed short inversion time inversion-recovery sequence (5170/56/160 [repetition time msec/echo time msec/inversion time msec]), with a section thickness of 3 mm, intersection gap of 0.3 mm, field of view of 300 mm, and matrix of 384 x 512, also was applied in patients with reported hip pain.
For imaging of the knees, a coronal T1-weighted spin-echo sequence (520/20) was applied with a section thickness of 3 mm, intersection gap of 0.3 mm, field of view of 300 mm, and matrix of 256 x 512. A coronal T2-weighted fat-suppressed short inversion time inversion-recovery sequence (4050/56/160), with a section thickness of 3 mm, intersection gap of 0.3 mm, field of view of 280 mm, and matrix of 384 x 512, also was applied in patients with reported knee pain.
For both examinations, a phased-array spinal coil was used in conjunction with a phased-array body coil. Total examination time was 15 minutes for the T1-weighted sequences and 25 minutes if two additional T2-weighted fat-suppressed sequences were applied.
MR Image Interpretation
MR images were interpreted with consensus by two musculoskeletal radiologists (J.F.G. and G.E.A., with 9 and 5 years of experience in musculoskeletal imaging, respectively). Findings were classified as (a) no osteonecrosis, (b) nonspecific bone marrow abnormality, or (c) osteonecrosis. Nonspecific bone abnormalities were categorized as either subchondral or intramedullary in location (9–12). Subchondral abnormalities were defined as those located immediately beneath the articular surface. Intramedullary abnormalities were defined as those located within the medullary canal, removed from the articular surface. The location and size of intramedullary and subchondral abnormalities were recorded on a predesigned template. For calculating the size of abnormalities, the major and minor axis lengths were measured directly on the hard-copy images.
Osteonecrosis was defined as an area, either subchondral or intramedullary in location, demarcated by a distinct marginal rim with low signal intensity that encompassed medullary fat on T1-weighted images. The severity of osteonecrosis in the femoral head was graded by using the University of Pennsylvania system (13). According to this system, grades IA, IB, and IC reflect osteonecrotic involvement of less than 15%, 15% to 30%, and more than 30% of the volume of the femoral head, respectively, on MR images obtained in patients with normal plain radiographs (13).
Forty-four patients either with nonspecific subchondral bone marrow abnormalities not obviously related to degenerative change or with osteonecrosis were asked to return for a more detailed MR imaging examination, as well as a radiographic examination. At the detailed MR imaging examination, the affected hip or knee joint was evaluated individually by using standard coils and orthogonal sequences, depending on the site of the initial abnormality. For lesions of the femoral condyles, T1-weighted (518/14), T2-weighted (3970/74), and T2-weighted spectral presaturation and inversion recovery (SPIR) (3970/74/160) sagittal images and intermediate-weighted (3500/43) coronal images were obtained. For lesions of the patella, a small-field-of-view surface coil was used, and T1-weighted (518/14), intermediate-weighted fat-suppressed (2480/31), and T2-weighted SPIR (4080/67/160) transverse images were obtained. For lesions of the proximal femora (nearly all intramedullary lesions), T1-weighted (599/20) and T2-weighted SPIR (5170/66/160) oblique coronal images were obtained. The duration of each detailed imaging examination of a joint was approximately 30 minutes. Radiographic examination comprised anteroposterior and lateral views of the affected hip or knee joint. Radiographs were interpreted by two authors (J.F.G., G.E.A.) in consensus, who compared radiographs with both initial and detailed MR images specifically to identify early radiographic signs of osteonecrosis (subchondral osteopenia, osteosclerosis, or early subchondral collapse). Similar criteria were used for interpretation of both the detailed MR images and the initial MR images. Incidental unrelated pathologic entities depicted on radiographs and MR images also were noted. Irrespective of the imaging findings, all cases were reviewed at a special orthopedic clinic, and patients were informed of the results.
Statistical Analysis
Median and range were calculated to characterize continuous variables, and percentages were calculated for discrete variables. Mann-Whitney and Kruskal-Wallis tests were applied to analyze group distribution, and the Pearson exact 2 test was applied to compare proportions among groups. Univariate logistic regression models were applied to identify significant risk factors that were predictive of osteonecrosis or of a negative result at MR imaging. A stepwise multivariate logistic regression model was developed to identify important risk factors. The adjusted odds ratio for prediction of osteonecrosis was determined for each risk factor, while controlling for the effects of other covariates, by using the multivariate logistic regression model. Statistical software (SPSS version 11.5 for Windows; SPSS, Chicago, Ill) was used for all statistical analyses. A 5% significance level was applied for all tests (P < .05).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Predisposition to Osteonecrosis
One patient (with normal findings at MR imaging) had a history of systemic lupus erythematosus. None of the other patients evaluated had a recognizable predisposition to osteonecrosis (Table 1).
|
Nonspecific bone marrow abnormality.—Nonspecific subchondral and intramedullary bone marrow abnormalities were present on MR images in 77 (30%) of 254 patients (29 male, 48 female; median age, 42 years; range, 16–89 years). Patients with nonspecific bone marrow abnormalities were older, on average, than patients with no osteonecrosis and no bone marrow abnormalities (P < .001).
Subchondral bone marrow abnormality.—One hundred sixty-one subchondral bone marrow abnormalities (Figs 1–3) were present in 64 (25%) of the 254 patients evaluated. The locations of these abnormalities are diagrammed in Figure 4. Maximum length ranged from 3–26 mm (mean, 6.5 mm). Twelve (19%) of the 64 patients had appreciable associated degenerative change.
|
|
|
|
|
|
|
|
|
|
|
|
|
The 12 affected patients received cumulative prednisolone-equivalent doses of 8.74, 7.71, 7.29, 5.53, 5.15, 4.96, 4.15, 4.02, 3.52, 3.67, 3.01, and 0.76 g, respectively (mean, 4.57 g) (Fig 8). The patient with the most severe osteonecrosis (in both hips and knees) received a 5.53-g cumulative prednisolone-equivalent dose and, in addition, a nonsteroidal antiinflammatory drug (naproxen). The patient who received only a 0.76-g prednisolone-equivalent dose, at 18 years of age, was the youngest of the cohort with osteonecrosis. Other than steroid treatment, this patient had no past or current risk factor for osteonecrosis.
|
Fifteen potential risk factors were assessed, namely age, sex, weight, body mass index, maximum radiographic score, admission to the intensive care unit, days in the intensive care unit, days in hospital, ALT level at admission, LDH level at admission, maximum LDH level, cumulative prednisolone-equivalent dose, cumulative prednisolone-equivalent dose adjusted for weight and body mass index, and average daily prednisolone-equivalent dose. Intensive care unit admission, days in hospital, cumulative prednisolone-equivalent dose, cumulative prednisolone-equivalent dose adjusted for weight, cumulative prednisolone-equivalent dose adjusted for body mass index, and average daily prednisolone-equivalent dose were found to be significant predictors of osteonecrosis (Table 2). The results of analysis with a stepwise multivariate logistic regression model based on these six significant indicators, which were individually identified in univariate logistic regression analysis, revealed that cumulative prednisolone-equivalent dose was the only significant predictor of osteonecrosis (P = .001). In addition, it was the most important risk factor: The adjusted odds ratio and 95% confidence interval, with control for the effects of age, sex, and intensive care unit admission, was 1.589 (1.143–2.208) (P = .006) (Table 3).
|
|
No relationship between the severity of SARS infection (as judged from radiographic score, ALT level at admission, and LDH level at admission or maximum LDH level) and the development of osteonecrosis was observed (P > .05).
Joint Pain
One hundred thirty-four (53%) of 254 patients had new onset of joint pain after admission for SARS treatment. The pain involved only one joint in 18 (7%) of 254 patients and more than one joint in 116 (46%). The joints affected were the hip in 51 (20%), the knee in 104 (41%), the shoulder in 86 (34%), and the ankle in 46 (18%) of 254 patients. Details such as time of onset and site of hip and knee pain are shown in Table 4. Among 264 painful hip or knee joints in the 254 patients screened, findings on MR images in 210 (80%) joints were normal. Patients with osteonecrosis of the hip were more likely to have pain (P < .001), yet the majority of patients with hip pain did not have osteonecrosis. Patients with osteonecrosis of the knee were no more likely to have knee pain than were other patients evaluated (P = .93).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Steroids are postulated to induce osteonecrosis by causing a decrease in local blood flow by means of either (a) marrow fat hypertrophy, which leads to increased intraosseous pressure and impairment of venous flow, or (b) lipid emboli and lipid-loaded fibrin-platelet thrombi, which occlude subchondral arterioles and capillaries (17). Core decompression, the main surgical treatment for early osteonecrosis, is aimed at reversing this process by reducing intraosseous pressure and providing a conduit for angiogenesis to revascularize subchondral bone (18).
The clinical symptoms of osteonecrosis are nonspecific, and patients may be asymptomatic, as were four of the 12 patients with osteonecrosis in this study (19). Radiography does not depict osteonecrosis until a relatively late stage of disease (19,20). MR imaging is a sensitive test for detection of early osteonecrosis, especially in the femoral head (20,21). In this study, T1-weighted coronal imaging in the hips and knees was found useful in the evaluation of patients for osteonecrosis (8).
While steroids administered over a prolonged period increase osteonecrosis risk in otherwise predisposed patients (3,4,19), short-duration steroid therapy appears to be safe in nonpredisposed patients, even if administered in very high doses. Wing et al (2) reported the results of a study of 59 patients with acute spinal cord injury, most of whom received 11–15 g methylprednisolone (equivalent to 13.75–18.75 g prednisolone) over a period of 24 hours. No osteonecrosis was apparent on MR images obtained 6 months after initiation of steroid therapy (2).
Little is known regarding the osteonecrosis risk with high-dose steroids administered to nonpredisposed patients over a more prolonged period. Patients in this study were generally healthy prior to infection with SARS; thus, we had the opportunity to study the relationship between steroid dosage and osteonecrosis in a population without known predisposition. We analyzed as many variables as possible in an effort to establish factors predictive of osteonecrosis.
Twelve (5%) of 254 patients evaluated had osteonecrosis. Cumulative prednisolone-equivalent dose was the most important risk factor predictive of osteonecrosis. The risk of osteonecrosis was 0.6% for patients receiving less than 3 g and 13% for patients receiving more than 3 g. No additional unrelated variable was found to be predictive of osteonecrosis risk. This finding does emphasize the individual susceptibility to osteonecrosis even in patients not known to be predisposed to the condition.
Subchondral bone marrow abnormalities were present in a greater number of patients than expected. Two recent MR imaging–based studies, performed in 100 (mean subject age, 42.7 years) and 115 asymptomatic knees (mean subject age, 47.5 years), reported 3% and 11.3% prevalence of subchondral bone marrow abnormalities, respectively (18,19). All abnormalities occurred on the medial side of the joint and in patients older than 40 years (18,19). Mean subject age in the current study was only 35.5 years, yet the prevalence of subchondral bone marrow abnormalities (24 [25%] of the 254 patients examined) was more than twice that previously reported. Patients with nonspecific bone marrow abnormalities were analyzed as a separate group, in case the abnormality in some patients eventually would be found to represent a mild form of osteonecrosis.
Subchondral bone marrow abnormalities mainly comprised areas of low signal intensity on T1-weighted images and high signal intensity on T2-weighted images, findings that are compatible with a "bone marrow edema" pattern (9,10). The MR imaging appearance of these subchondral marrow abnormalities was comparable with that observed in severe degenerative disease (10), and the abnormalities did tend to occur in patients who were older than those with normal findings or osteonecrosis at MR imaging. However, cartilage thinning (as a manifestation of degenerative disease) was not a feature in many cases, and this fact makes it less likely that degenerative disease or focal chondral injury was the sole cause of the observed abnormalities.
One probable cause of subchondral marrow abnormality may be a combination of steroid-induced subchondral bone demineralization and increased physical activity after recovery from SARS, which could lead to subchondral trabecular impaction and edema similar to those described in patients after renal transplantation (11,12,22–25). Another probable cause of subchondral marrow abnormality is early subchondral osteonecrosis (9). The distribution of the marrow abnormalities around the knee was similar to that observed in osteonecrosis (26–29). Possibly because of a difference in blood supply, osteonecrosis around the knee tends to affect discrete areas of bone marrow, unlike osteonecrosis in the proximal femur, which nearly always affects the anterosuperior aspect of the femoral head. On detailed MR images obtained in our patients, the deep thin subchondral line, described by Lecouvet et al (9) as one useful discriminatory criterion to distinguish early irreversible subchondral osteonecrosis from transient subchondral bone marrow edema, was not a feature of these lesions. Transient subchondral bone marrow abnormality may represent "reversible" subchondral osteonecrosis (9). A further study is currently under way specifically to investigate the significance of these subchondral bone abnormalities.
Small intramedullary abnormalities, occasionally associated with fluid or edema, were present in 15% of patients. The incidence of intramedullary abnormalities in the normal population is not known. Many may simply represent venous lakes or bone islands that were too small to be classified. Although they may potentially represent very small areas of osteonecrosis, they are by virtue of their small size and location not likely to cause symptoms, structural weakening, or long-term problems. Although no relationship was found between nonspecific bone abnormality and steroid dose, this may be a reflection of our small sample size.
New onset of joint pain occurred in 53% of patients after infection with SARS. The majority of painful joints showed no abnormality on MR images. Joint pain in the aftermath of viral infections is not uncommon, and there is a recognized association between joint pain and viruses such as hepatitis C, rubella, and human T-cell lymphotrophic virus type 1 (30). Joint pain, therefore, is not a reliable clinical indicator when evaluating for osteonecrosis after initiation of steroid treatment for SARS.
There are three main limitations of this study. First, the time range from initiation of steroid treatment to MR imaging was 3.3–9.7 months, and a small number of patients were examined early (ie, before 6 months). Osteonecrosis, however, is considered to develop soon after the initiation of steroid treatment. In a longitudinal study in which serial MR imaging was performed in 72 patients with systemic lupus erythematosus, all 32 patients who developed osteonecrosis did so within 5 months (mean, 3.1 months) of initiation of steroid treatment. Thereafter, no new case of osteonecrosis was detected with MR imaging up to a period of 12 months (19). The cohort in the current study received high-dose steroids for a short period of about 3 weeks, at the start of the illness, after which the doses were gradually decreased. Second, as all patients were treated with steroids for SARS, no control group was available for study. The possibility that SARS infection itself may have contributed to the development of osteonecrosis has not been refuted. Autopsy in patients with SARS has revealed fibrin thrombi in and intimal swelling of pulmonary vessels (31). The results of the current study, however, revealed no relationship between markers of the severity of SARS infection (radiographic score, ALT level at admission, LDH level at admission, and maximum LDH level) and osteonecrosis. Third, it was assumed that no preexisting osteonecrosis was present, as the cohort studied was healthy prior to infection with SARS.
In conclusion, the results of this MR imaging–based study show a dose-related risk of osteonecrosis in patients who receive steroid therapy for SARS, with cumulative prednisolone-equivalent dose being the most important predictor. Nonspecific subchondral and intramedullary bone marrow abnormalities were a frequent observation. Joint pain was common after SARS infection and was not a useful clinical indicator of osteonecrosis.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantor of integrity of entire study, J.F.G.; study concepts and design, J.F.G., G.E.A., S.M.K., D.S.C.H., G.M.J., A.K.L.W., K.H.C., K.M.C., P.C.L., A.T.A.; literature research, J.F.G., G.E.A.; clinical studies, S.M.K., D.S.C.H., G.M.J., A.K.L.W., K.H.C., K.M.C., P.C.L.; data acquisition, J.F.G., G.E.A., J.K.T.W., D.S.C.H., G.M.J., A.K.L.W., K.H.C.; data analysis/interpretation, J.F.G., G.E.A., A.Y.K.C.; statistical analysis, all authors; manuscript preparation and revision/review, A.Y.K.C., J.F.G.; manuscript editing, J.F.G., G.E.A.; manuscript definition of intellectual content and final version approval, all authors
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N.-F. Zhang, Z. R. Li, H.-Y. Wei, Z.-H. Liu, and P. Hernigou Steroid-induced osteonecrosis: THE NUMBER OF LESIONS IS RELATED TO THE DOSAGE J Bone Joint Surg Br, September 1, 2008; 90-B(9): 1239 - 1243. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. K. C. Wong and W. S. Poon Risk of High Dose Hydrocortisone in Patients With Aneurysmal Subarachnoid Hemorrhage Stroke, January 1, 2008; 39(1): e12 - e12. [Full Text] [PDF]
|
|
|
|
 |
|
 V. C. C. Cheng, S. K. P. Lau, P. C. Y. Woo, and K. Y. Yuen Severe Acute Respiratory Syndrome Coronavirus as an Agent of Emerging and Reemerging Infection Clin. Microbiol. Rev., October 1, 2007; 20(4): 660 - 694. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R.-c. Chen, X.-p. Tang, S.-y. Tan, B.-l. Liang, Z.-y. Wan, J.-q. Fang, and N. Zhong Treatment of Severe Acute Respiratory Syndrome With Glucosteroids: The Guangzhou Experience Chest, June 1, 2006; 129(6): 1441 - 1452. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Mont, L. C. Jones, and D. S. Hungerford Nontraumatic Osteonecrosis of the Femoral Head: Ten Years Later J. Bone Joint Surg. Am., May 1, 2006; 88(5): 1117 - 1132. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
