| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Technical Developments |
1 From the Department of Diagnostic Imaging, Hôpital Raymond Poincaré, 104 Blvd Raymond Poincaré, 92380 Garches, France. Received March 21, 2003; revision requested June 13; final revision received March 9, 2004; accepted May 12. Address correspondence to R.Y.C. (e-mail: robert.carlier@rpc.ap-hop-paris.fr).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
14°; mean, 6.43°) was obtained in cases with estimated reducibility greater than 5° (31 cases, 67%), which is a 34% (6.5° of 19.1°) mean reduction. A significantly greater level of kyphosis reduction was observed in cases with intravertebral clefts (20 cases, 43%) at hyperlordosis than in those without (7.2° vs 4.9°; P < .01). Vertebroplasty may reduce kyphosis due to localized collapsed vertebrae; intravertebral mobility and cleft suggest this possibility. © RSNA, 2004
Index terms: Osteoporosis, 30.56 • Spine, curvature, 30.862 • Spine, fractures, 30.411 • Spine, vertebroplasty, 30.126
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
It is known that, independent of pain, an osteoporotic vertebral compression fracture hampers physical function and quality of life (9–12) because of spinal deformity. Compression fractures lead to changes in spinal biomechanics and to kyphosis (8), which in the thoracic spine contribute to a decreased lung capacity (13,14) and in the lumbar spine contribute to a reduction in abdominal space, which causes loss of appetite (15) and secondary problems of poor nutrition. However, to our knowledge, no precise values of kyphosis angle as a threshold value of lung and bowel dysfunction are available in the literature. On the other hand, results of epidemiologic and statistical studies have shown that the presence of one osteoporotic vertebral compression fracture increases the risk of subsequent fractures by four to five times (16–18). Moreover, a prevalent vertebral deformity can enable prediction of an increased mortality (19).
Because of these problems, attempting to realign the spinal column is important for improving the quality of life of these patients. To our knowledge, this has been tried in the last few years with kyphoplasty (7,20) but not with vertebroplasty.
In patients with osteoporotic vertebral collapse, forced hyperlordosis may, in some cases, show a correction of the compression fracture with volume modification of the vertebral body and relative reduction of the localized kyphosis. Intravertebral vacuum cleft phenomenon can sometimes be observed on radiographs and has been described as a criterion of benign vertebral collapse (21,22). The purpose of our study was to evaluate the early radiographic outcome of PV with cement injection performed in a forced hyperextension position for localized kyphosis reduction.
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
Patient Population
Forty-six PV procedures were consecutively performed in 30 patients (24 women, six men). The mean age was 72.8 years (range, 57–83 years); it was 72.5 years (range, 68–77 years) in men and 72.8 years (range, 57–83 years) in women. There was no significant difference in terms of age (Student t test).
The vertebrae to be treated were located in the midthoracic section (T6 through T8) in eight cases, the lower thoracic section (T9 through T12) in 19 cases, the upper lumbar section (L1 through L2) in 14 cases, and the lower lumbar section (L3 through L4) in five cases (Table). The indications included painful primary osteoporotic vertebral compression fractures (24 patients) or secondary osteoporotic compression fractures (steroid-induced in four patients, radiation therapy–induced in two patients) that were unresponsive or minimally responsive to conventional treatments (analgesics, bisphosphonates, bed rest, bracing) for at least 3 months. No additional attention was given to the age of the fractures.
|
Symptomatic levels were identified by comparing clinical and preoperative imaging data. No decision was made before the team had considered the imaging results. Exclusion criteria were as follows: first, radiologic-clinical discordance with no reliability between the levels of collapsed fracture and pain; second, discovery at computed tomography (CT) and/or magnetic resonance (MR) imaging of a more plausible cause of pain than a collapse; and third, absence of signal abnormality in the collapsed vertebral fracture, which indicated an absence of persistent bone remodeling.
Imaging Assessment
For each patient, digital radiography, unenhanced CT, and MR imaging were performed in the days preceding PV (1–13 days preceding for radiography, 1–11 days for CT, and 1–13 for MR imaging).
Digital radiographs were always obtained with the same C-arm table (Polystar; Siemens Medical Solutions, Erlangen, Germany) so as to have comparable views. They were obtained with a 15-cm field of view, a 90-kV tension, and automatic adaptation of the milliamperage and time of exposure. The measurements were determined with computer assistance to reduce the variations in kyphosis angle measurement; this was in accordance with previous studies that compared measurements of angles of deformity in kyphosis and scoliosis that were performed, respectively, on conventional radiographs and on digital radiographs with computer assistance. Only the kyphosis of the vertebral bodies was taken into consideration; local kyphosis measured from the plates of the adjacent vertebrae was not considered valuable because of the considerable mobility of the disk spaces.
Radiography was performed with the C-arm table used for PV, which could be placed in an upright position. Lateral views in a standing position (neutral position) and in hyperextension (forced hyperlordosis) were obtained. If standing was impossible, a sitting position was used. The hyperextension view was obtained with the patient in a supine position and with a hard cushion placed under the patient’s back at the level of the compression fracture. The hard cushion was adapted to the morphology of each patient and to the level at which the hyperlordosis had to be imposed. The cushion was made of one or more sufficiently rigid rolled cotton wraps.
The digital radiographs were exported to a workstation (EasyVision; Philips Medical Systems, Eindhoven, the Netherlands). The localized kyphosis angle was then calculated for both neutral and hyperextended positions by using the endplates of the vertebral body on the lateral views. The possibility of localized kyphosis restoration was then evaluated as the angle in hyperextension subtracted from the angle in neutral position. Visualization of an intravertebral cleft in the neutral position and in the forced hyperextension position was also noted.
Measurement of kyphosis angles and evaluation for an intravertebral cleft were first performed separately by two experienced musculoskeletal radiologists (C.V. and R.Y.C., 23 and 15 years of experience in spinal radiology, respectively). The interobserver variations in kyphosis angle measurement were then evaluated by the same two senior radiologists. Three of 64 lateral views were rejected because the analysis of the vertebral endplates was confusing. Thus, interobserver variations were calculated from the measurements of 61 consecutive lateral views to assess the minimal angle that could be considered as a kyphosis reduction or intravertebral mobility. These lateral views were obtained in the first 37 patients referred for PV (age range, 56–88 years; mean, 71.4 years). Only 18 of these patients underwent PV (ie, 19 of 37 patients were excluded and were only included as control subjects). The 12 remaining patients in our study population (of 30 patients) were not considered in the calculation of interobserver variations. Absence of local kyphosis angle reduction was not a contraindication to PV, because pain relief constituted the main goal of PV.
Helical CT (CT Twin; Philips Medical Systems) images with frontal and sagittal reconstructed views were obtained to determine the extent of vertebral collapse and the continuity of the posterior wall and to exclude other causes of pain, such as disk herniation or facet joint arthrosis. Presence of an intravertebral cleft was also reported. Volumetric CT acquisition of 1.3-mm-thick transverse sections with 0.6-mm overlap was performed (120 kV, 200 mAs, 240-mm field of view, 512 x 512 matrix, pitch of 1, high spatial resolution). The results of CT examinations were interpreted independently by the same two radiologists (C.V. and R.Y.C., 20 and 15 years of experience in spinal CT, respectively). A consensus was obtained in case of disagreement.
MR imaging (MR Max; GE Medical Systems, Milwaukee, Wis) was performed at 0.5 T. Imaging consisted of sagittal (30-cm field of view, 192 x 192 matrix, 5-mm section thickness) spin-echo intermediate- and T2-weighted (repetition time msec/echo time msec of 2000/40 and 2000/100, respectively; one signal acquired, acquisition time of 6 minutes 24 seconds) and spin-echo T1-weighted (480/20, two signals acquired, acquisition time of 3 minutes 4 seconds) sequences, as well as transverse gradient-echo T2-weighted (1090/30, 25-cm field of view, 192 x 192 matrix, 5-mm section thickness, 25° flip angle, two signals acquired, and acquisition time of 6 minutes 55 seconds) sequences. Imaging was performed before and after contrast material enhancement with 0.1 mmol of gadoterate meglumine (Dotarem; Guerbet, Aulnay-sous-Bois, France) per kilogram of body weight.
First, MR images were evaluated for other causes of pain that would contraindicate PV, such as disk herniation, facet joint effusion (which could indicate acute arthrosis), or tumors located in the spinal canal.
Second, MR images were evaluated for changes in bone marrow signal intensity that would be consistent with persistence of bone remodeling at the levels of the compression fractures, either as simple bone marrow edema with high signal intensity on T2-weighted images, low signal intensity on nonenhanced T1-weighted images, and normal signal intensity on contrast-enhanced T1-weighted images or as an intravertebral cleft with well-demarcated linear or ellipsoid areas with signal intensity prolongation on T2-weighted images when fluid-filled or signal intensity void when gas-filled. The results of MR examinations were interpreted independently (C.V. and R.Y.C., 16 and 15 years of experience in spinal MR imaging, respectively). A consensus was obtained in case of disagreement. Absence of intravertebral cleft and normality of bone marrow were contraindications to PV.
PV Technique
Patients were always informed of potential complications of the procedure, and informed consent was obtained.
PV was performed in all patients by either one of two radiologists (C.V. and R.Y.C., 12 and 9 years of experience in PV, respectively). Conscious sedation with intramuscular injection of 1.0–2.5 mg of midazolam (Hypnovel, Hoffmann-La Roche, Basel, Switzerland) and local anesthesia with 20 mL of 0.5% lidocaine hydrochloride (from skin to periosteum; Xylocain, Astra, Sodertalje, Sweden) were used. No general anesthesia was induced, in order to keep the patient awake and reactive.
Patients were placed in a prone oblique position on the table, and C-arm fluoroscopy was used to guide a vertebroplasty needle (Escoffier Frères, Thonon-les-Bains, France) through a right parapedicular posterolateral approach into the vertebral body. This position imposes hyperlordosis because the patient lies on the anterolateral side of his or her shoulder, abdomen, and hip with elevated and flexed upper limbs and slight flexion of the lower limb ipsilateral to the punctured side; the knee is used for support. Direct manual pressure on the top of the kyphosis may also help to impose hyperlordosis, especially before injection. With a prone patient, elevation of the shoulders off the table may also produce hyperextension of the thoracic spine, but the posterolateral approach seems more difficult in this position.
The needle was advanced to the center or toward the anterior part of the vertebral body by using C-arm fluoroscopic guidance. The cement material was prepared by combining polymethylmethacrylate powder (Merck Biomaterial, Darmstadt, Germany) with 1 g of tungsten powder (Balt, Montmorency, France), which was used to render the polymethylmethacrylate more radiopaque, and the mixture was then injected into the vertebral body after positioning the patient in a maximum forced-hyperlordotic position.
The needle was positioned as near as possible to the intravertebral cleft when visible at forced hyperlordosis. Continuous fluoroscopic monitoring was used during injection to prevent overfilling and leakage into the spinal canal, neural foramen, or epidural venous system. Injection was immediately stopped if cement was observed within one of these spaces. In most cases, injection was terminated when no additional cement could be injected and/or when intravertebral cleft filling was observed. If a visible cleft was not filled, the needle was then repositioned so as to achieve this filling. Complete filling of vertebrae was not attempted in all cases, in order to reduce the probability of spinal canal or foraminal compromise or venous embolus.
The volume of injected cement, which was easily determined from the syringe graduation, ranged from 1.0 to 8.5 mL (mean, 5.16 mL). In 18 patients, only one vertebra was treated. Eight patients underwent treatment of two vertebrae, and four patients underwent treatment of three vertebrae.
All patients underwent radiography at the end of the procedure. The localized kyphosis angle was measured in the same way as before the procedure—on a lateral image obtained with the patient in a sitting or standing position—and the degree of kyphosis reduction was calculated as the posttreatment angle subtracted from the pretreatment angle (with both pre- and posttreatment angles determined in the same position, either standing or sitting).
Follow-up and Pain Relief
After the procedure, all patients were transferred to the medical unit for a 24-hour observation period and were then discharged. During this period, it is possible to manage immediate postoperative pain, to make sure there is no adverse effect, and to perform short-term follow-up. A visual analogic scale with 10 levels was used to evaluate the pain relief at the 24th hour. Patients were asked to indicate their improvement on a scale of 0 (no improvement or worsening situation) to 10 (total improvement). Pain relief was considered good if the score was 8–10, partial if the score was 4–7, and insufficient or null if the score was 0–3.
Statistical Analysis
For interobserver variations in localized kyphosis angle, the difference between measurements made separately by the two radiologists was calculated. Means and standard deviations of the localized kyphosis angle measurement were thus obtained. Statistical analysis was performed by using Mann-Whitney and Student t tests. Spearman correlation coefficient was calculated between the volume of cement used for filling the vertebral body and the degree of kyphosis reduction. For comparisons of groups, Mann-Whitney and Student t tests with a significant difference indicated at P < .05 were used. For the analysis of differences between groups, one vertebral body treated in each patient was randomly selected to account for lack of independence among cases of vertebral collapse.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
The threshold of valuable vertebral mobility was determined to be approximately equal to 2 standard deviations in the interobserver local kyphosis angle measurement. This is equivalent to the coefficient of repeatability proposed by Cheung et al (23) in a study of the reliability of quantitative analysis on digital images of the scoliotic spine.
The procedure was well tolerated by all patients. The Table gives the radiographic localized kyphosis measurements for 46 treated levels before and after PV. The treated levels ranged from T6 through L4, and the majority of levels were from T9 through L1. The localized kyphosis angle ranged from 3° to 40° (mean, 17.97°) in the neutral position (sitting or standing) before treatment, from 0° to 35° (mean, 12.32°) in hyperlordosis, and from 0° to 30° (mean, 13.0°) in the neutral position after PV.
The vertebral levels could be considered in two groups before treatment: those in which there was no or only a slight kyphosis reducibility (less than 5°; 15 of 46 levels, 32%) and those in which there was a reducibility of 5° or more at hyperlordosis (31 of 46 levels, 68%). After PV, poor kyphosis reduction (0°–6°; mean, 1.86°) was obtained in the group in which reducibility (determined before PV) was less than 5°. In the group in which reducibility before PV was 5° or more, the reduction ranged 2°–14°(mean, 6.4°); this represents a mean gain of 34% in kyphosis correction (difference between mean kyphosis angles before [19.1°] and after [12.6°] PV). A significant difference in local kyphosis correction was thus obtained between the group with poor (<5°) and that with good (
5°) kyphosis reducibility during the hyperlordosis test (Mann-Whitney test, P < .001, with only one vertebral body taken into account in each of the 30 patients).
In the group with good local kyphosis reducibility at hyperlordosis, intravertebral vacuum cleft phenomenon was visible in 20 of 31 cases (65%), either at the radiographic hyperlordosis test or at CT. Intravertebral vacuum cleft phenomenon was detected in only two cases on lateral radiographs obtained in the neutral position. Water- or gas-filled clefts (13 and two cases, respectively) were visible on MR images. The mean angle of kyphosis restoration after PV differed significantly between cases in which the cleft was visible and those in which it was not (7.02° vs 4.90°, respectively; P < .01 with Student t test). However, two vertebrae that manifested an intravertebral vacuum cleft phenomenon showed a poor level of reduction (Table).
In 10 procedures, the level of restoration was greater than that predicted before the procedure; in 15 procedures, the level of correction was the same as that predicted; and in 21 procedures, the level of correction was less than that predicted.
In 21 vertebrae that showed poor (<5°) local kyphosis reduction after PV, the amount of injected cement ranged from 1.0 to 8.0 mL (mean, 4.95 mL). In the 25 other vertebrae, in which local kyphosis reduction was 5° or more, the amount of injected cement ranged between 2.0 and 8.5 mL (mean, 5.16 mL). This indicated that there was no direct relationship between the volume of cement used to fill the vertebral body and the degree of kyphosis reduction (correlation coefficient, r = 0). An illustrative case is shown in Figures 1–7.
|
|
|
|
|
|
|
|
|
|
Complications
Cement leakage occurred in 17 (37%) of 46 procedures, as follows: A small amount of cement entered the disk space in nine cases (20%), the spinal canal in four cases (9%), the epidural venous plexus in two cases (4%), the paravertebral tissues in one case (2%), and the needle route in one case (2%). There were no immediate clinical complications related to these leakages. No other clinical complications, such as pulmonary embolism or rib fractures, were noted.
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
|
|
Intravertebral clefts have been radiographically recognized for decades. They traditionally appeared as vacuum or air-filled clefts inside a vertebral body and were usually associated with previous fractures. These clefts were initially presumed to represent a sign of avascular necrosis. The clefts were more easily detected in the supine position or with hyperlordosis (24).
Variable MR imaging appearances of the radiographically detectable clefts have been described depending on whether gas or fluid filled the cleft. These contents are variable over time in the supine position, because the gas is being progressively replaced by fluid (25). A cleft is detectable on T2-weighted MR images as an area of high signal intensity when it contains liquid and of low signal intensity with susceptibility artifacts when it contains gas (25). The link with avascular necrosis has since been revised, and the cleft has been reported as a good sign of benign collapse (22).
In a retrospective study, Lane et al (26) found that 75 (31%) of 236 treated vertebral collapses showed intravertebral clefts at the time of vertebroplasty (compared with 43% in our study). In that study, the radiographic preprocedural evaluation never included a lateral view with hyperlordosis, and the score of detection on radiographs was very low; there were only eight vertebrae with clefts. The score at MR imaging was better; 28 vertebrae with intervertebral cleft were detected. Those authors assessed the detection of cleft during vertebroplasty as the visualization of a cavity filled with cement, which is different from the trabecular filling that is classically visualized (26). In our study, things were not so predictable. Even in a case in which the needle is positioned into an easily detectable cleft before PV, the cement opacification can extend widely into the trabecular bone. The radiographic posttreatment appearance is then due to the filling of the intertrabecular spaces, and the cleft is retrospectively no more detectable. Lane et al also stated that the clefts most often detected at the thoracolumbar junction correspond to nonhealed collapses in an area of important spinal mobility (26).
In our study, the number of clefts was important (43%; 20 clefts detected among 46 treated vertebral bodies). Hyperlordosis testing and CT with multiplanar reconstructions explain this high number of detected clefts. MR sensitivity for depiction of water- or gas-filled clefts (13 and two cases, respectively) was less reliable than were lateral radiographs obtained at hyperlordosis and CT images with multiplanar reconstructions.
According to other authors, a 3-month period seems to be adequate for improvement with only medical treatment in patients with vertebral compression fractures (27). When the patients who require PV are chosen on the basis of this criterion, (ie, patients who have had a fracture for at least 3 months), the number of nonhealed fractures is important; intravertebral cleft is then very frequently found. The pathophysiology of persistent cleft is well established. However, the presence of a cleft explains the good results of PV even in cases in which the compression fracture is more than 3 months old. In such cases, the painful persistent intravertebral mobility disappears with simple cement filling of the cleft.
In our study, mobility in vertebral collapsed fracture is not always associated with a detectable intravertebral cleft. A cleft was detected in only 20 (43%) of 46 cases in our study, but in 31 (67%) of 46 cases there was a local kyphosis angle reduction of 5° or more. Good levels of kyphosis reduction after PV have been obtained in cases of intravertebral mobility without intravertebral cleft. This tends to suggest that part of the mobility is added during hyperlordosis into the trabecular network of the vertebral body.
PV is typically used only to treat pain, with no attempt to eliminate spinal deformity (7,8).
Belkoff et al (28), in an ex vivo study of osteoporotic cadaveric vertebral bodies, compared the biomechanical properties of isolated fractured osteoporotic vertebral bodies treated by means of kyphoplasty and those treated by means of PV. Kyphoplasty resulted in significant restoration of vertebral body height lost after compression (97%; P < .05), whereas PV treatment resulted in a significantly lower restoration of lost height (30%; P < .05). To our knowledge, there is no in vivo comparative study of kyphoplasty versus PV. Moreover, forced hyperlordosis may be difficult to apply to an isolated fractured vertebral body in a cadaver, and intravertebral cleft cannot be observed.
Results of our study show that vertebral body mobility in compression fractures may lead to reduction of the local kyphosis after PV. Despite the small number of cases in our study, the visualization of an intravertebral cleft reinforces this prediction of favorable results for kyphosis reduction and suggests repositioning of the needle if the cleft is not filled.
Substantial vertebral body mobility and visualization of an intravertebral cleft may be similar to the action of the inflatable bone tamps into the fractured vertebral body that are used in kyphoplasty.
Results of our in vivo study show that PV performed with the patient in hyperlordosis can be an alternative to kyphoplasty. Both techniques should now be prospectively compared, and cost and benefit-risk data should be evaluated (29). In our study, only the kyphosis angle has been taken in consideration, although the height of the vertebral body has been measured in some kyphoplasty studies (8,28). The evaluation of vertebral mobility should be limited to the vertebral body and avoid measurement of the motion present in the disk, contrary to the evaluation used in some studies (7,8,30). In one kyphoplasty study (30), a mean 8.8° of correction of the local kyphosis was reported, but the authors used the Cobb technique for calculating the angle (ie, "measurement taken from the superior endplate of the vertebra one level above the treated vertebra to the inferior endplate of the vertebral body one level below the treated vertebra" [30]). On the other hand, differences in the biomechanics of the treated vertebrae may appear; filling an artificial cavity created during kyphoplasty may not have the same consequences as filling the natural interconnecting spaces located within the trabecular network of the vertebral bodies’ cancellous bone.
A parapedicular posterolateral approach was intentionally chosen in our study: This is the usual practice of the authors, and hyperextension may be easier to obtain with a prone oblique position (Fig 9) than it is with the strict prone position often required for the transpedicular approach. Moreover, this position may be more comfortable for the patient. Cement leakage was not observed to have occurred more frequently in our study than is reported in literature (from 30% to 67%) (3,4,31–33).
|
The population of our study was relatively small, and follow-up of the patients concerned only the short-term with no information about mid- and long-term pain relief and eventual occurrence of new collapses. However, the efficiency of the procedure in regard to the primary pain relief seems to be the best predictor of midterm clinical outcome after PV (33).
In conclusion, in some patients vertebral compression fracture can be efficiently treated by means of PV for reduction of localized kyphosis. Vertebral body mobility in hyperlordosis may suggest the reduction of localized kyphosis after PV.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, C.V., R.Y.C.; study concepts, R.Y.C., C.V., H.G.; study design, R.Y.C., C.V., A.F.; literature research, H.G., R.Y.C., D.M.M., N.V.; clinical studies, C.V., R.Y.C., H.G.; data acquisition, C.V., R.Y.C.; data analysis/interpretation, R.Y.C., C.V., H.G.; statistical analysis, R.Y.C., C.V., A.F.; manuscript preparation, R.Y.C., D.M.M., H.G., C.V., N.V.; manuscript definition of intellectual content, C.V., R.Y.C.; manuscript editing, revision/review, and final version approval, R.Y.C., C.V., D.M.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and MethodsResults DiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. Hiwatashi, P.-L.A. Westesson, T. Yoshiura, T. Noguchi, O. Togao, K. Yamashita, H. Kamano, and H. Honda Kyphoplasty and Vertebroplasty Produce the Same Degree of Height Restoration AJNR Am. J. Neuroradiol., April 1, 2009; 30(4): 669 - 673. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C.C. Lin, I.H. Chen, T.C. Yu, A. Chen, and P.S. Yen New Symptomatic Compression Fracture after Percutaneous Vertebroplasty at the Thoracolumbar Junction AJNR Am. J. Neuroradiol., June 1, 2007; 28(6): 1042 - 1045. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Hiwatashi and P.-L.A. Westesson Vertebroplasty for Osteoporotic Fractures with Spinal Canal Compromise AJNR Am. J. Neuroradiol., April 1, 2007; 28(4): 690 - 692. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.J. Cloft and M.E. Jensen Kyphoplasty: An Assessment of a New Technology AJNR Am. J. Neuroradiol., February 1, 2007; 28(2): 200 - 203. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Robin Goodfellow (44-3) Rheumatology, March 1, 2005; 44(3): 418 - 418. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
