| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Musculoskeletal Imaging |
1 From the Departments of Radiology and Nuclear Medicine (G.S., A.W., M.J., O.K.) and Orthopaedic Surgery (R.W.), St Josef Hospital, and the Institute of Pathology, Ruhr-Universität Bochum, Gudrunstrasse 56, 44791 Bochum, Germany (C.K.). From the 1999 RSNA scientific assembly. Received October 31, 2002; revision requested March 6, 2003; final revision received October 1; accepted October 21. Address correspondence to G.S. (e-mail: gebhard.schmid@ruhr-uni-bochum.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Fifty-one patients underwent MR imaging before microsurgical removal of extruded lumbar disk herniation material. Marrow signal intensity changes along the cartilaginous endplates were classified according to Modic types 1–3. Severity of changes was evaluated with respect to extension along the endplate in the anteroposterior diameter (0%, <33%, 33%–66%, >66%). The existence of a dorsal vertebral corner defect was evaluated in relation to the existence of hyaline cartilage in the disk extrusion material.
RESULTS: Mean tissue composition of herniated material in all patients was 63% anulus fibrosus, 30% nucleus pulposus, and 8% cartilaginous endplate. Twenty-five of the 51 patients had hyaline cartilaginous material in the extrusion (range, 5%–50%). Patients without marrow signal intensity changes along the cartilaginous endplate showed significantly less cartilaginous material in the extruded disk (P = .023, Fisher exact test). Mean percentage hyaline cartilage in patients without changes was 2% ± 4 (SD) (Modic type 1, 16% ± 15; type 2, 10% ± 12). When the changes extended 33% of the vertebral endplate, there was cartilaginous endplate material in the extruded disk (P = .006). Cartilage from the endplate was present in the extruded disk material in 40% (16 of 40) of patients without a vertebral corner defect and in 82% (nine of 11) of patients with a vertebral corner defect (P = .019).
CONCLUSION: Avulsion-type disk herniation seems to be common, and vertebral endplate marrow signal intensity changes on MR images are indicative of cartilaginous material in the extruded disk herniation material.
© RSNA, 2004
Index terms: Cartilage, MR, • Magnetic resonance (MR), tissue characterization, 31.121411, 32.121411, 33.121411 • Spine, intervertebral disks, 31.316, 32.326, 33.336 • Spine, MR, 31.121411, 32.121411, 33.121411
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Although disk herniation is a common disease, there are few data in the literature regarding histologic findings in herniated material and its composition of nucleus pulposus, anulus fibrosus, and cartilage from the vertebral cartilaginous endplate (6–13). Knowledge of tissue composition is of interest because some authors (6,10,14–16) suggest that different tissue compositions may lead to different inflammatory responses. Of special interest is that although the intervertebral disk is generally without a vascular supply, extruded disk material often shows neovascularization, and this is thought to be part of the natural healing mechanism. Findings in histopathologic studies (17–20) show that hyaline cartilage material in the extruded disk material can suppress neovascularization and thus suppress subsequent size reduction of the herniated material. This could be one of the reasons why some disk extrusions resolve with time and others do not. Thus, patients with suppressed neovascularization may be candidates for surgical removal of the disk herniation, while patients with adequate neovascularization may be treated conservatively or with epidural corticosteroid injection.
To be of use for patient treatment, histologic findings in the herniated fragment, especially the amount of hyaline cartilage, should be predictable on the basis of clinical or imaging data. Modic et al (21) showed that vertebral endplate marrow signal intensity changes are associated with fissures in the vertebral endplate. If hyaline cartilage from the vertebral endplate is found in the herniated material, endplate avulsion must have preceded the process of disk herniation. Therefore, vertebral endplate signal intensity changes may be regarded as osteocartilaginous fracture signs similar to other skeletal manifestations. The presence of this avulsion-type disk herniation has been confirmed in histopathologic studies and seems to be the predominant type of disk herniation in the elderly (7,22,23).
The purpose of this study was to compare preoperative MR imaging findings with data for tissue composition of herniated disks in patients after microsurgical removal of herniated material.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Nine of the 51 patients had previously undergone spinal surgery in a lumbar segment different from the one in the current study.
Criteria for inclusion in the current study were a clinical history with evidence of disk herniation, direct preoperative MR imaging findings in the lumbar spine of extruded disk herniation, surgical removal of herniated disk material, and histopathologic assessment of the surgical specimen. Consecutive patients who met the inclusion criteria were included during a 2-year period (from September 1996 through September 1998). Patients who had undergone previous surgery or percutaneous nucleotomy (laser or chemonucleolysis) in the same lumbar segment and those with a history of malignant diseases were excluded. All patients gave written consent for the MR imaging examination that was performed as a routine preoperative procedure on the basis of accepted clinical indications. Institutional review board approval and informed consent from each patient were obtained for the study.
MR Imaging
All MR imaging examinations were performed with a 1.5-T superconducting system (Magnetom Vision; Siemens Medical Systems, Erlangen, Germany) with a spinal surface coil. The following pulse sequences were performed. (a) In all patients: sagittal T2-weighted turbo spin echo (repetition time msec/echo time msec of 4,700/120, echo train length of 15, section thickness of 4 mm, field of view of 280 x 280 mm, matrix of 240 x 256, and one signal acquired). (b) In all patients: sagittal T1-weighted spin echo (600/14, section thickness of 4 mm, and other parameters as in sequence a). (c) In all patients: transverse T2-weighted turbo spin echo in one level or multiple levels depending on the pathologic condition seen with sequences a and b (4,500/120, echo train length of 15, section thickness of 4 mm, field of view of 200 x 200, and matrix of 240 x 256 mm). (d) In patients after previous spinal surgery: transverse T1-weighted spin echo (600/14) before and after administration of 0.1 mmol per kilogram of body weight gadopentetate dimeglumine (Magnevist; Berlex Laboratories, Wayne, NJ). Repetition times varied in individual patients according to changes required to adhere to the specific absorption rate.
Surgical and Histopathologic Findings
All but two patients underwent microdiscectomy. In the two patients, free fragments of the herniated disk could not be extracted with this approach, and microdiscectomy was changed intraoperatively to hemilaminectomy.
With microscopic guidance, all epidural or submembraneous parts of the disk herniation were removed. The intradiscal space was examined, and loose parts of the disk were also removed to prevent reherniation. Immediately after removal, the intradiscal and extruded parts of the removed disk material were placed separately into 5% formaldehyde solution, labeled, and sent for histopathologic examination.
The surgical specimens were brought immediately to the Institute of Pathology, where they were embedded in paraffin and completely evaluated by one pathologist (C.K.). Specimens were first inspected macroscopically and weighed. Fragments with a diameter larger than 1 cm were cut with a knife to obtain a representative cross section. The material was cut with a microtome into 4-µm-thick slices, which were stained with hematoxylin-eosin to differentiate anulus fibrosus, nucleus pulposus, and hyaline cartilaginous endplate. The proportion of each finding was determined semiquantitatively, and results are given as percentages. Precision error with this technique was approximately 5%–10%.
While the hyaline cartilaginous endplate is easily identifiable, nucleus pulposus and anulus fibrosus may not always be readily distinguished, especially in severely degenerated disk material. The term cartilaginous material is used in the current study exclusively for hyaline cartilage from the vertebral endplate (not for fibrocartilage from the anulus fibrosus or the degenerated nucleus pulposus). The pathologist used the following criteria to discriminate between nucleus pulposus, anulus fibrosus, and hyaline cartilaginous endplate (24).
Anulus fibrosus.—This tissue is characterized as fibrocartilage that includes fibrous matrix material with regularly intermingled bland looking small cartilage cells. Typically, the fibrous structure resembles interdigitating fascicles. No apparent mucoid-rich areas are included.
Nucleus pulposus.—Histologically, nucleus pulposus shows loosely arranged bland looking cartilage cells that are arranged in an abundant (ie, predominantly mucoid or myxoid) matrix. This matrix is composed of nearly granular material that does not show features of fascicular collagen fibers, as in anulus fibrosus.
Cartilage endplate.—This structure is composed of hyaline cartilage. The cartilage cells exhibit a clearly visible perinuclear halo, which is a result of fixation and preparation of the tissue. The matrix is homogeneously structured without fascicular (as in anulus fibrosus) or smoothly granular (as in nucleus pulposus) features. Some fibers of anulus fibrosus may merge in the hyaline cartilage of the endplate.
Granulation tissue and neovascularization were not examined separately. Therefore, as a result of its peripheral localization, most granulation tissue will be included in anulus fibrosus.
Evaluation
Two MR image readers, with 10 (G.S.) and 2 (A.W.) years of experience in MR imaging of the lumbar spine, interpreted the MR images independently, and the final result was obtained with consensus. The readers were blinded to the histopathologic results. Signal intensity changes in bone marrow adjacent to the cartilaginous endplate were evaluated only in the segment with the herniated disk. Because the changes along the vertebral endplates on the preoperative images were evaluated postoperatively, the MR image readers knew which lumbar segment was treated at surgery.
Marrow signal intensity changes were classified according to Modic et al (21). Type 1 changes had increased marrow signal intensity in T2-weighted MR images and decreased signal intensity in T1-weighted MR images. Type 2 changes had increased signal intensity in T1- and T2-weighted MR images. Type 3 changes had decreased signal intensity in T1 -and T2-weighted MR images. For simplification, no signal intensity changes and only tiny spots of signal intensity changes in the bone marrow adjacent to the cartilaginous endplates were assigned type 0.
We quantified the extension of marrow signal intensity changes along the cartilaginous endplate as no change or tiny spots of change, less than 33% of the anteroposterior diameter of the cartilaginous endplate, 33%–66%, and more than 66%. Extension of the marrow signal intensity changes was estimated subjectively with consensus by the two MR image readers.
Findings in a previous study (22) showed that avulsion fractures of the cartilaginous endplate take place in the middle third of the sagittal diameter of the endplate. Thus, we compared patients with and those without marrow signal intensity changes in the middle third of the cartilaginous endplate. The current study was focused on avulsion-type disk herniation; therefore, we examined the dorsal aspect of the vertebral body adjacent to the cartilaginous endplate to see if there was an obvious vertebral corner defect, as described previously (25,26). One experienced musculoskeletal radiologist (G.S.) subjectively compared the vertebral corners at the level of the disk herniation with those at other levels in T1- and T2-weighted sagittal MR images. Then, he divided the patients into two groups: group A, no definite vertebral corner defect or indefinite unsharp vertebral corner; group B, definite vertebral corner defect.
Statistical Analysis
To determine the strength of dependency of two metric variables, we calculated correlation coefficients. To compare distributions of dichotomous variables in patient groups, we used the Fisher exact test. To test for differences in median values of metric variables in two independent patient groups, we used the Mann-Whitney U test. To compare more than two patient groups, we used the Kruskal-Wallis test. These calculations were performed with software (SPSS for Windows 10.0; SPSS, Chicago, Ill). Differences with a P value less than .05 were considered to be statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
On average, the composition of herniated material in all patients was 63% anulus fibrosus, 30% nucleus pulposus, and 8% cartilaginous endplate. In five patients, we also found small bone fragments. The relative tissue composition changed slightly with age in all patients, with a moderate reduction in nucleus pulposus and cartilaginous endplate and a moderate increase in anulus fibrosus (difference not significant, Kruskal-Wallis test for nucleus pulposus and anulus fibrosus). Tissue composition was not significantly different in the lumbar vertebral segments from L4 through L5 and from L5 through S1 (proportion of anulus fibrosus and nucleus pulposus, Mann-Whitney U test; existence of cartilaginous endplate, Fisher exact test). There was only one disk herniation (in the L3–4 disk) in the current study. Twenty-five of the 51 patients had cartilaginous material in their extrusion, which accounted for 5%–50% of its total mass.
The proportion of extrusions with cartilaginous material and its dependency on age and sex are shown in the Table. Nineteen of 37 (51%) disk herniations in men contained cartilaginous material as did seven of 14 (50%) in women. The average percentage of cartilaginous material was larger in women (9% ± 13) than in men (7% ± 10), even though the women were slightly younger (mean age of men, 41.1 years ± 11.3; women, 38.4 years ± 13.3). Because of the small number of women in the current study, the difference in the percentage of cartilaginous material was not significant (Mann-Whitney U test). In addition, the age-dependant percentage of cartilaginous fragments in the herniated material differed, with higher percentages in the middle age group compared with those in the youngest and oldest age groups. The proportion of extrusions with cartilaginous material (Fig 1) was also highest in the middle age group (difference not significant, Fisher exact test). In a comparison of extrusions weighing up to 2 g, greater than 2 g and less than 3.9 g, and more than 4 g (Fig 2), the larger extrusions showed a significantly higher proportion that contained cartilaginous fragments (P = .008, Fisher exact test).
|
|
|
Marrow Signal Intensity Changes and Cartilaginous Material
In six of 21 (29%) patients without marrow signal intensity changes (type 0) and in 19 of 30 (63%) patients with types 1 and 2 changes, the extrusion contained cartilaginous material (Fig 3). In patients with types 1 and 2 changes, the average proportion of cartilaginous material was greater than that in patients without changes (type 0). The mean percentage of cartilage in patients without changes was 2% ± 4 compared with 16% ± 15 in patients with type 1 changes and 10% ± 12 in those with type 2 changes. As types 1 and 2 changes increased in size along the vertebral endplates, there was a higher proportion of extrusions that contained cartilaginous material (P = .025, Fisher exact test; Fig 4).
|
|
We found that in 70% (14 of 20) of patients with marrow signal intensity changes in the middle third of the endplate had cartilaginous material in the extrusion compared with 35% (11 of 31) of patients without such changes (P = .023). The mean percentage of cartilaginous material was 5% ± 9 (median, 0%) in patients without changes in the middle third of the endplate compared with 12% ± 13 (median, 10%) in patients with such changes.
Vertebral Corner Defects
When evaluating vertebral corner defects (Fig 5), we found that cartilaginous material was present in 16 of 40 (40%) patients in group A (no definite vertebral corner defect) compared with nine of 11 (82%) patients in group B (definite vertebral corner defect) (P = .019, Fisher exact test) (Fig 6). Average percentage of cartilaginous material in the disk material increased from 4% ± 7 (median, 0%) in group A to 19% ± 16 (median, 15%) in group B. Size of the extrusion was not significantly different between groups A and B (Mann-Whitney U test).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
Tanaka et al (22) report that avulsion-type disk herniation predominates in the elderly. In this type of herniation, cartilaginous endplate rupture was found in severely degenerated disks. The authors show that there is only a loose connection between the cartilaginous endplate and the subchondral bone, but there is a strong connection between the inner fibers of the anulus fibrosus and the cartilaginous endplate. Avulsion-type disk herniation may also occur in a substantial percentage of young patients.
This might be of interest because it has been shown that mechanical pressure cannot be the only reason for sciatica: Biochemical, vascular, and immunologic factors can produce sciatica without severe mass effect or nerve root compression (4,5,27). Few investigators have evaluated the presence and effect of various amounts of different histologic tissue types (nucleus, anulus, cartilaginous endplate) on disk herniation. In another study (7), cervical spine fragments of hyaline cartilage and fibrocartilage were present in every specimen, and cartilaginous-endplate herniation was the predominant type of herniation at this site. Similar results were found at the lumbar spine, where one research group (6) in a detailed study of 100 cases found that free fragments consisted mainly of nucleus pulposus in 54% of patients and of endplate material in 44% of patients. Surprisingly, those authors found only one patient with disk herniation consisting predominantly of anulus fibrosus. In another report (11), the herniation model of nuclear fragmentation, migration, and extrusion was favored because nucleus pulposus was present in 89% of their study material, while they found some endplate material in 49% of their cases. Yasuma et al (15) state that where complete extrusion of sequestrated material has occurred, this tissue almost exclusively consists of anulus fibrosus. Thus, there are conflicting theories about the origin of the extruded disk.
Findings in the current study show that avulsion-type disk herniation with hyaline cartilage material occurs frequently (in nearly 50% of patients). The amount of cartilage may be as much as 50% of the extruded material, and bone fragments were observed in five patients. Although internal derangement of the intervertebral space increased with age, larger cartilage fragments were found in the middle age group, and the mean percentage of cartilage decreased with age in a comparison of the middle age group with the oldest age group. Women had more avulsion-type disk herniations, and the largest cartilage avulsions were found in the group aged 30 years and younger. This might be explained in part by the larger herniation mass in younger patients.
There is evidence from experimental studies that blood vessels from the surrounding fibrovascular tissue infiltrate into the herniated anulus fibrosus but not into the herniated cartilaginous endplate (17). The same authors confirmed in an animal study that anulus fibrosus induces vascular sprouting and inflammation with subsequent decrease in size of the anulus material. They showed that when cartilaginous endplate was implanted, vascular sprouting and inflammation were inhibited and the implanted cartilage material did not decrease in size. Even more, when they implanted both anulus and endplate material, neovascularization and inflammation reaction were depressed, possibly as a result of inhibitors of neovascularization found in cartilage (18). Indirect confirmation of their findings may be those of other authors (6), who showed that single free fragments, which contain cartilage endplate less frequently than do multiple free fragments, are associated more often with inflammatory granulation tissue around the fragments.
Bone erosion caused by lumbar intervertebral disk herniation is well known from CT studies. It has been studied in detail with MR imaging to assess the prevalence of this finding in patients with disk herniations (25). In their semiquantitative evaluation of 46 disk levels, those authors found no severe, three moderate, and 23 mild defects of the vertebral corner. The defects were seen more often in young patients and in patients with large herniations or substantial cranial or caudal migration. We evaluated this sign because a vertebral corner defect was obvious in some of the larger disk herniations, and we suspected that a substantial amount of cartilaginous endplate could be derived from this site. As our results show, the likelihood of the presence of cartilage material in the herniated disk increases when a vertebral corner defect is suspected. Thus, the vertebral corner is another potential source of endplate material.
Signal intensity changes in the bone marrow adjacent to the vertebral endplates in degenerative spine disease are well known. These changes were classified by Modic et al (21) into three types, and this classification system is widely used. Type 1 changes were seen in 4% of the patients referred for lumbar spine imaging, and type 2 changes were seen in approximately 16% of these patients. While type 2 changes seem to be stable, without substantial alteration during many years, it is known that type 1 changes may convert within 1–3 years to type 2 changes. Type 3 changes show signal intensity reduction and represent sclerosis of the endplates and the adjacent bone marrow. Type 1 changes seem to reflect the more acute stage, type 2 the more intermediate to chronic stage, and type 3 the end stage of the same degenerative process.
This process of osteochondrosis starts with internal derangement of the disk space, fissuring of the cartilaginous endplate, and sprouting of granulation tissue into the bone marrow. Previously, some studies dealt with the clinical importance of these changes. In one study (28), these changes had a low prevalence in asymptomatic young people and may be a marker of painful disk degeneration. Toyone and co-workers (29) showed that type 1 changes are highly correlated with low back pain and segmental hypermobility, but type 2 changes are not. In a recent study (30), types 1 and 2 changes along the endplate were shown to be a good predictor of painful disk derangement with concordant pain at discography.
Our results confirm that there is cartilaginous material in a high proportion of extruded disk herniations. The amount of cartilage in the herniation material is usually less than 10%, but it can be as much as 50%. The association of the amount of cartilaginous material with endplate abnormalities supports the theory that avulsion of the vertebral endplate is one source of disk herniation. The good correlation of marrow signal intensity changes in the middle third of the endplate with cartilaginous material in the disk herniation further supports the histopathologic findings by Tanaka et al (22) that most avulsions occur in the inner or transitional zone of the anulus-endplate interface. Although this avulsion-type disk herniation is probably the predominant mechanism of disk herniation in elderly patients, it may also happen in young people, especially when large disk herniations occur. Signal intensity changes along the endplates may be caused not only by edema, fissuring of the endplates, and formation of granulation tissue but also in part by endplate avulsion. The clinical relevance of cartilage material in the disk herniation is not completely clear. However, experimental data indicate a suppression of inflammatory reaction and neovascularization. Therefore, more knowledge about the role of neovascularization in the natural healing mechanism is necessary, as well as more knowledge about the modulation of this process by cartilage and other tissues. A high degree of signal intensity changes along the endplate or a substantial vertebral corner defect may be in vivo indicators of cartilage material in extruded disk herniation.
A limitation of the current study is that granulation tissue around the herniated material was not evaluated. This should be done in future studies to directly correlate the amount of hyaline cartilage in the extrusion to the extent of granulation tissue and neovascularization. Although a potential memory bias regarding evaluation of the vertebral corner defects cannot be excluded, it is unlikely because the MR image reader of the vertebral corner defects and the marrow signal intensity changes along the endplate was blinded to the results at histopathologic examination.
In conclusion, avulsion-type disk herniation seems to be common, and vertebral endplate abnormalities on MR images are associated with cartilage material in the extruded disk herniation.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Rahme and R. Moussa The Modic Vertebral Endplate and Marrow Changes: Pathologic Significance and Relation to Low Back Pain and Segmental Instability of the Lumbar Spine AJNR Am. J. Neuroradiol., May 1, 2008; 29(5): 838 - 842. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
