| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Department of Radiology, Dartmouth Medical School, Dartmouth-Hitchcock Medical
Center, One Medical Center Dr., Lebanon, NH 03756.
2 Skirball Institute–NYU Medical Center, 540 First Ave., New York, NY
10016.
3 Millennium Technology Inc., 465-5600 Parkwood Way, Richmond, BC, Canada.
4 Department of Orthopaedic Surgery and Physical Rehabilitation, UMass Memorial
Medical Center, 55 Lake Ave. N, Worchester, MA 01655.
5 National Orthopedic Imaging Associates, 1260 S Eliseo Dr., Greenbrae, CA
94904.
Received May 23, 2003;
accepted after revision August 5, 2003.
Address correspondence to D. W. Goodwin.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Spin-echo images of 11 resected tibial plateaus acquired at 7 T were compared with the structure of the extracellular matrix as shown by fracture sectioning the samples in the plane of imaging. Four samples were scanned at two different orientations relative to the main magnetic field (B0). T2 maps were acquired in two orientations on three of these four samples.
RESULTS. On the basis of the presence of reproducible regional variations in the shape of the matrix, a characteristic matrix architecture was described. The location of peak signal intensity and T2 on MRI correlated with the level at which the matrix was estimated to be aligned at approximately 55° to B0 (r = 0.91). This correlation of matrix orientation relative to B0 with T2 and signal intensity on MRI was not altered by regional variations in the shape of the matrix or by imaging samples at two different orientations.
CONCLUSION. The structure of the extracellular matrix, through its orientation-dependent influence on T2 decay, exerts a strong influence on the MRI appearance of cartilage. At the tibial plateau, a characteristic matrix architecture is associated with an equally characteristic MRI appearance.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
On MRI of articular cartilage, variation in T2 decay across the depth of the tissue produces a typical layered or laminar appearance [3, 4]. A longer T2 transitional layer separates the shorter T2 surface and deep layers (Fig. 1). Earlier investigators have suggested that variations in water or proteoglycan concentration or perhaps collagen fibril orientation could explain the presence of these layers on MRI of articular cartilage [5–7]. However, no study, to our knowledge, has established such a relationship. We believe that it is the organization of the extracellular matrix at the macroscopic level and not the orientation of individual collagen fibrils that influences T2 decay and produces layers.
|
The morphology of the matrix is evident only when cartilage is sectioned by fracturing rather than by cutting the tissue. Sectioning by fracturing is particularly useful in displaying large structures that would be obscured with routine sectioning that cuts through the structure of interest. When it is fractured, cartilage cleaves along lines of least resistance, producing a curved fracture plane that reflects the underlying tissue organization [8, 9]. The highly structured and anisotropic organization of proteoglycans and collagen restricts water mobility and enhances dipole interaction. Consequently, the T2 of cartilage is short. However, in highly structured tissues such as cartilage, T2 is also dependent on tissue orientation relative to the main magnetic field (B0). T2 decay lengthens as the orientation of the internuclear vector is tilted away from B0, reaching maximal prolongation at approximately 55° [3, 4, 7]. This orientation dependence of T2, termed the "magic angle effect," is most evident on shorter TE images because of the generally rapid T2 decay of the tissue. The curved shape of the extracellular matrix and its orientation relative to B0, therefore, provide an explanation for the variation in T2 and the presence of layers on MRI of cartilage. A study showing correlation between the location of layers and the shape of the matrix supports this assertion [10].
The laminar appearance of cartilage on MRI, however, is not uniform. It varies depending on the joint and the location in the joint [11–13]. For example, on MRI of the knee, we have observed predictable regional variations in the laminar appearance of the tibial plateau (Fig. 1). Reproducible variations in the MRI findings imply the presence of predictable regional variations in the shape of the extracellular matrix. We believe that the extracellular matrix of the tibial plateau has a characteristic global architecture that is associated with an MRI appearance that is equally characteristic. We compared T2 and signal intensity on spin-echo images and T2 maps of the tibial plateau with the structure of the extracellular matrix as displayed through fracture sectioning in the plane of imaging. On the basis of these observations, we describe the typical matrix architecture of the tibial plateau. An awareness of the matrix architecture is essential for understanding the normal variations in T2 decay and the laminar appearance displayed on MRI of the tibial plateau and other joint surfaces. The interpretation of MRIs of articular cartilage must account for these variations in underlying structure.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Terminology
The term "fibril" describes microscopic collagenous structures
apparent only at high magnification. Fibrils are arranged with proteoglycans
and other extracellular matrix constituents into a larger macroscopic
structure that is evident after fracture sectioning. This macroscopic
structure has been variously described as "fibers" in some
scanning electron microscopy studies and as "leaves" in others. We
have instead chosen to use the term "matrix structure" or
"macroscopic structure" to describe this tissue organization and
have consciously avoided use of the term "fiber." The term
"matrix architecture" describes the global matrix organization of
an entire joint surface and predictable regional variations in the shape of
the matrix. The use of this term reflects our belief that there is a
characteristic architecture for each joint surface, much as trabecular
patterns are characteristic for bones. Layers refer to the laminar variations
in signal intensity and T2 seen on MRI of cartilage obtained with the surface
oriented perpendicular to B0. The deep layer refers to the low
signal intensity and short T2 layer abutting the subchondral bone. A
high-signal-intensity, long T2 transitional layer separates this layer from
the low-signal-intensity, short T2 surface. The term "zone" is
reserved for descriptions of regional variations in cartilage histology.
Sample Preparation
Eleven tibial plateaus were studied. Six lateral and four medial tibial
plateaus were obtained after above-knee amputations from four women and two
men (age range, 27–74 years; mean age, 50 years). One lateral tibial
plateau was harvested from the knee of a 79-year-old man during arthroplasty
for degenerative disease of the medial compartment. Two medial plateaus
obtained after amputation were too large to fit into the coil used for
imaging. Indications for amputation included acute vascular occlusion (two
knees), chronic vascular insufficiency (two knees), acute traumatic amputation
(one knee), and synovial cell sarcoma (one knee). By inspection of gross
specimens, we selected for imaging only articular surfaces free of obvious
cartilage injury.
The samples were harvested by dissecting away soft tissues and cutting through bone parallel to the articular surface. Scores cut into the subchondral bone defined the plane of imaging and subsequent fracture sectioning. Samples were refrigerated in saline solution until the time of imaging, which occurred no later than 1 week after amputation or surgery.
Imaging
All samples were placed in saline-filled containers and scanned with an
SMIS spectrometer scanner (Surrey Medical Imaging Systems, Guildford, England)
and a 7-T horizontal bore magnet (Magnex Scientific, Abingdon, England).
Single-slice spinecho images (TR/TE range, 1,000/14–20; field of view,
45–55 mm; matrix size, 512 x 512; slice thickness, 1 mm; and
number of signals averaged, 16–20) of nine tibial plateaus were
obtained. The read gradient was oriented parallel to B0 in a
direction appropriate to shift the marrow fat signal away from the cartilage.
Five samples were imaged in both the midsagittal and the midcoronal planes.
The influence of orientation was evaluated by scanning the remaining four
samples in a single plane, first with the articular surface perpendicular to
B0 and subsequently with the surface oriented at approximately a
55° angle to B0. Two of these samples were imaged in the
midcoronal plane, one in the midsagittal plane, and one in the sagittal plane
through the submeniscal region of the plateau—the peripheral region of
the plateau normally covered by the meniscus.
To better investigate the entire tibial plateau, we obtained multislice acquisitions in both the sagittal and coronal planes in two additional samples. The matrix size was decreased to 256 x 512 and other imaging parameters were unchanged, producing seven images (1 mm thick, with a 4-mm slice separation) in each plane in both samples.
T2 maps of three samples were obtained with the articular surface oriented perpendicular to B0 and again at approximately 55° to B0. T2 values were measured using a multiecho sequence with a selective 90° pulse (90x) and a nonselective composite refocusing pulse (90x – 180y – 90x), as previously described [14]. Acquisition parameters were TR, 9 sec; sweep width, 50 kHz; TE, 12 msec; 8.8-msec echo spacing, 16 echoes; bandwidth, 50 kHz; field of view, 45–55 mm; matrix size, 128 x 128; slice thickness, 1 mm; signals averaged, 8–10. To facilitate correlation of the T2 measurements with the spin-echo images, these maps were acquired in the same plane as higher resolution spin-echo images without moving the sample.
Matrix Structure
After scanning, we fixed samples in formalin, demineralized the subchondral
bone in 5% nitric acid for 48 hr, and rinsed the samples in saline. After
freezing, the samples were fractured in the plane of scores cut in the
subchondral bone before scanning. The first two samples were fractured after
submersion in liquid nitrogen. The remainder were simply placed in a freezer
at –4°C overnight and then fractured by hand. The latter technique
was preferred because it provided greater control of the fracture plane. One
sample was first cut into a 2-mm-thick slab in the plane of imaging and
subsequently frozen and fractured multiple times perpendicular to the imaging
plane.
Samples were then dehydrated through a graded series of ethanol, then 50% ethanol and 50% hexamethyldisilazane, and finally 100% hexamethyldisilazane [15]. After fixation and drying the samples, we obtained photographs of all samples digitally with a macrolens, using oblique lighting to optimally display the matrix structure.
Scanning electron micrographs of four samples were obtained after coating with a gold and palladium target sputter-coater (Technics Hummer V, Anatech, Springfield, VA) using a scanning electron microscope (DSM-962, Zeiss USA, Thronwood, NY). On these four samples, a partial proteoglycan digest was performed before dehydration by placing samples in a 37°C bath of 40 mg of hyaluronidase type IV and 20 mg of trypsin type I in 20 mL of phosphate-buffered saline for 1 hr. We obtained low-magnification, large-field-of-view images of the fractured surface of the sample and higher-magnification images of selected regions of interest. Because matrix structure was more easily and completely displayed on inspection of the unmagnified specimens, electron microscopy was limited to only the first four samples. Before photography, the subsequent samples were fixed and dried like those for scanning electron microscopy, but no proteoglycan digest was performed because this step is necessary only when visualizing microscopic structures.
Split Lines
Surface split lines can be created by puncturing the surfaces of joints
with a round awl. These splits in the articular surface are oriented
perpendicular to the orientation of the matrix curvature and therefore reflect
the underlying structure of the cartilage matrix
[16]. We created split lines
in the tibial plateau removed at above-knee amputation for acute vascular
occlusion in a 27-year-old woman. A round awl punctured the surface of the
medial plateau to a depth of 2 mm across the entire joint surface. The sample
was photographed after it had been coated with india ink, rendering the split
lines more easily visible. We then fixed and decalcified the sample, as
described previously, and fractured it in multiple planes.
Image Analysis and Correlation
A single observer reviewed all MRIs, electron micrographs, and photographs
using ScionImage software (Scion, Frederick, MD). T2 measurements and signal
intensities on spin-echo images were compared pixel by pixel after correcting
for the differences in resolution. Signal intensity was plotted across the
depth of cartilage in a 10-pixel-wide plot, in two different locations on each
of 10 samples. A central measurement was obtained at the center of the image,
and a submeniscal measurement was obtained at an arbitrarily chosen location
at the edge of the joint. Only a submeniscal region measurement was obtained
on the one sample that did not include the center of the joint. The location
of peak signal intensity was measured and recorded as a percentage of tissue
thickness from subchondral bone to the articular surface. On the photographs
of the corresponding fractured specimen, the point at which tissue orientation
was approximately 55° was estimated and recorded as a percentage of tissue
thickness. These estimations were made at locations corresponding to the level
of MRI measurements using the imaging software to measure from the edge of the
joint. We calculated the Pearson product-moment correlation coefficient to
describe the correlation between location of peak signal intensity, recorded
as a percentage of tissue thickness, and the estimated location of 55°
matrix orientation, also recorded as percentage of tissue thickness.
On the basis of these observations, the same observer proposed a model of the typical matrix architecture of the tibial plateau. The model was intended to be consistent with our hypothesis that the orientation of the matrix relative to B0 determines the T2 and signal intensity of cartilage. The model was also designed to be consistent with the pattern of surface split lines, which are known to be oriented perpendicular to the orientation of underlying matrix curvature [16].
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
Regional variations in the MRI appearance of the tibial plateau were apparent (Figs. 2A,2B,3A,3B,3C,3D,3E,4A,4B,4C,5A,5B,5C). In the central region of the tibial plateau, a thick deep layer extended across almost the entire depth of the tissue. Only thin transitional and surface layers were seen, and frequently, it was difficult to clearly visualize the surface. Striations within the deep layer were oriented perpendicular to the subchondral bone.
At the joint periphery, or submeniscal region, transitional and surface layers were much larger. The deep layer gradually decreased in the more peripheral regions of the joint and was often absent as the transitional and surface layers became more prominent.
At the tibial eminence, the orientation of the striations within the deep layer was no longer perpendicular to the underlying bone. Instead, striations were aligned at an angle to the subchondral bone, approaching an orientation parallel to the weight-bearing axis of the knee. At the margin of the tibial eminence, a prominent transitional layer similar to that at the submeniscal region was present. The regional variability of layers evident on spin-echo images paralleled changes in the T2 displayed on the T2 maps (Figs. 2A,2B and 3A,3B,3C,3D,3E).
When the articular surface of the tibial plateau was imaged at a 55° orientation to B0, the appearance differed in a reproducible and characteristic fashion (Fig. 3A,3B,3C,3D,3E). In comparison with images obtained with the surface perpendicular to B0, signal intensity within the deep layer of cartilage increased. In the transitional layer, striations of lower signal intensity were frequently evident. These changes in the transitional layer, however, were not observed uniformly throughout the entire image of most samples. Anisotropy in the transitional layer was most apparent in the submeniscal region. As on images obtained with the surface perpendicular to B0, the T2 maps acquired at an angle to B0 indicated a direct relationship between T2 and signal intensity (Fig. 3A,3B,3C,3D,3E).
When samples were sectioned, the fractured surface was uneven and corrugated rather than smooth (Figs. 4A,4B,4C and 5A,5B,5C). Large columnar structures radiated from the subchondral bone in parallel arrays curving into the plane of the articular surface at more superficial levels. The appearance suggested the presence of large fibers. These structures, however, were not easily separable from adjacent fibers and sometimes appeared more like sheets of tissue. The surface of the two samples fractured after freezing with liquid nitrogen was considerably smoother than that of samples that had been fractured after freezing at –4°C (Fig. 3A,3B,3C,3D,3E).
Regional variations in cartilage structure were present and similar in all samples, indicating the presence of a characteristic global architecture of the extracellular matrix of the tibial plateau. This architecture is summarized in Figure 6. In the central portion of the tibial plateau, a columnlike arrangement of tissue oriented perpendicular to the subchondral bone predominated (Fig. 4A,4B,4C). This structure curved into the plane of the joint surface only at a superficial level. In the submeniscal region, a more complex structure was present. The matrix extended from the bone in an oblique direction, curving both away from the center of the joint and in a second plane so that the plane of curvature rotated by approximately 90° (Fig. 6). This complex curve produced a large overhanging edge at the periphery of the joint after fracture sectioning (Fig. 5C). At the tibial eminence, parallel arrays again extended from the subchondral bone. Alignment, however, was more in the direction of the weight-bearing axis than strictly perpendicular to the subchondral bone (Fig. 4C). At the edge of the tibial eminence, the tissue curved obliquely away from the center of the joint. The appearance was similar to the curvature in the submeniscal region, although a large overhanging edge was not observed.
|
The surface split lines created by puncturing the cartilage with a round awl were most obvious in the submeniscal region of the tibial plateau and were not as clearly identified in the central portion of the articular surface. In part, this difference was due to the presence of surface fibrillation centrally. Split lines were oriented in a wheel-spoke configuration, perpendicular to the annular direction of cartilage curvature we observed at the joint surface in the submeniscal region.
The location of peak signal intensity correlated with the estimated location of a 55° orientation of the matrix with an r value of 0.91. Regional variations in matrix structure paralleled regional variations in signal intensity and T2 (Figs. 4A,4B,4C and 5A,5B,5C). This correlation persisted when images and T2 maps were acquired at two different orientations (Fig. 3A,3B,3C,3D,3E). The striations seen on MRI appeared to correlate well with the columnlike structure apparent in the surface of the fracture planes, not only in the radial layer but also in the transitional layer when samples were scanned at 55° (Figs. 3A,3B,3C,3D,3E,4A,4B,4C,5A,5B,5C).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The study of cartilage morphology at the macroscopic level requires the use of fracture sectioning techniques. If, instead, routine sectioning is used, the plane of sectioning cuts through the curved extracellular matrix, concealing the presence of a complex 3D structure [8, 9, 17, 18]. Although there is considerable variability in previously published descriptions of the cartilage matrix, most reports suggest that a collagenous structure or structures radiate from the subchondral bone, arcing into the plane of the surface at more superficial levels. Differences in published descriptions likely reflect interspecies variability in cartilage morphology, the use of different techniques, and artifact related to sample preparation [19].
Surface split line patterns provide additional information regarding the shape of the underlying matrix. These splits in the cartilage surface, produced by puncturing the surface with a round pin, are orientated perpendicular to the direction of matrix curvature [16]. These lines form patterns that are typical for each joint surface [20]. This finding supports our contention that different joints have matrix architectures that are characteristic for that joint surface.
Our description of cartilage matrix structure, including our description of regional variations in form, is similar to that in previous studies. Although we are unaware of earlier attempts to describe the global organization of the entire tibial plateau cartilage, our findings are consistent with those of prior studies of tibial cartilage. Clark [21], in a scanning electron microscopy study of tibial plateau fragments, also found a strong vertical organization in the center of the joint, which became more oblique and complex in the submeniscal region. Detailed scanning electron microscopy studies of smaller cartilage samples from the tibial plateau have emphasized the strong vertical organization of the matrix, presumably reflecting the selection of samples from the central region of the joint surface [19, 21, 22]. An earlier description of the surface split line pattern of the tibial plateau is also similar to our findings [23].
The shape of the extracellular matrix provides an explanation for the T2 heterogeneity and anisotropy that characterize MRI of cartilage. Restriction of water mobility by the matrix enhances dipole interactions and shortens T2 decay. Because T2 is dependent not only on the distances between dipoles but also on their orientation relative to B0, T2 decay is anisotropic [24]. This T2 anisotropy explains not only the influence of orientation on MRI of cartilage but also the presence of layers. As the matrix curves into the plane of the articular surface, the magic angle effect produces gradual changes in the T2 and signal intensity.
Previous studies of articular cartilage support our description of the characteristic MRI appearance of the tibial plateau. T2 decay varies with the depth of cartilage, producing a layered appearance that is dependent on the orientation of cartilage relative to B0 [3, 4, 14, 25]. Earlier studies have also reported marked differences in the MRI appearance of cartilage when the margins of joints are compared with central regions of the same joint surface [11–13, 26]. Several authors have also noted the presence of striations [13, 26–28].
Establishing a quantitative correlation between matrix structure and signal intensity or T2 on MRI is difficult, and our comparison was descriptive for several reasons. First, the complex structure of the matrix usually curves out of the plane of imaging, rendering a direct comparison impossible. Second, measurements of matrix orientation made from photographs are influenced by the angle from which photographs were taken and by the unavoidable distortion of tissue during drying. Finally, fracture sectioning is imprecise even when closely controlled.
Although the correlation between signal intensity and tissue orientation in our study was qualitative, all our observations reaffirm our conclusion that matrix orientation determines T2. Regional variations in the structure of the matrix matched similar variations in the MRI and T2 measurements (Figs. 3A,3B,3C,3D,3E,4A,4B,4C,5A,5B,5C). Moreover, the correlation between matrix orientation and T2 persisted when images and T2 maps of cartilage were acquired in two different orientations (Fig. 3A,3B,3C,3D,3E). The inconsistent finding of lower signal intensity and T2 in the transitional layer on the 55° images should be expected. When cartilage is tilted, only portions of the curved region of the matrix align with B0. The region that aligns will vary depending on the direction that the sample is tilted 55°. The results of an earlier study comparing signal intensity with matrix structure also support our conclusion that matrix orientation relative to B0 determines T2 [10].
The apparent correlation of striations within low-signal-intensity layers with the corrugated fracture surface of cartilage is a further indication of the strong influence of matrix structure on MRI. Although these striations were not discernible on our low-resolution T2 maps, earlier studies have noted that the striations reflect additional T2 heterogeneity, different from the well-known layers [28]. We believe that the striations are, in fact, a reflection of matrix organization into the fibrous macroscopic structure seen on fracture samples. The striations are evident in the transitional layer of images obtained at 55°, consistent with the presence of a continuous structure.
Our results have specific implications for the interpretation of MRI of the knee. A prominent longer T2 transitional layer should be expected in the submeniscal region of the tibial plateau and should not be mistaken for cartilage injury. Conversely, injuries to cartilage in this region may be difficult to identify unless a sequence with a TE long enough to minimize the influence of the magic angle effect is used.
Our results also have implications for the use of MRI after the surgery for cartilage injury. The relationship between cartilage structure and MRI suggests that the findings of images obtained after the surgical repair of cartilage injury will not appear normal unless the overall structure of the extracellular matrix is reestablished. The value of MRI in this setting is unclear. A successful chondrocyte transplantation may result in a viable graft that remains distinct in appearance from surrounding cartilage. Conversely, a normal MRI appearance, indicating the presence of a normal macroscopic structure, may signal successful graft incorporation. Similar considerations must be taken into account in the evaluation of cartilage after mosaicplasty. On postsurgical MR images, the appearance of the implant will be similar to the appearance of the harvest site and potentially unlike the cartilage adjacent to the implant.
The structure of the extracellular matrix determines the form and physical properties of articular cartilage. Presumably, the matrix architecture of the tibial plateau is ideally suited to meet the physical demands faced by that joint surface. Any joint surface will also have a characteristic matrix architecture optimally suited to the function of that joint. An equally characteristic MRI appearance should therefore be expected at any joint. Accordingly, the choice of MRI sequences and even positioning relative to B0 would benefit from consideration of the unique matrix architecture of each joint surface.
Limitations in our study should be recognized. We studied human adult cartilage obtained after amputation. Consequently, the influence of degenerative changes associated with aging was unavoidable. Minimal surface fibrillation was apparent in the central region of almost all samples. Unfortunately, finding adult cartilage that is free of any wear is extremely difficult [29]. The study of a younger population is limited by the substantial differences between immature and mature articular cartilage [30]. Because the findings in our study were consistent in all samples, we believe our conclusions are reasonable despite this potential limitation.
The high field strength and long imaging times we used produced images of far higher resolution than that of clinical images. This may limit the relevance of our results to clinical imaging. However, in our experience, the layers and striations seen in this study are evident on clinical images of cartilage obtained on a 1.5-T scanner (Fig. 1). These techniques facilitated correlation of MRI with the fractured specimen; therefore, our findings provide an explanation of the normal variations in signal intensity seen on MRI of the tibial plateau. T2 decay is more rapid at higher field strengths [31]. Although the actual values we have reported are therefore shorter than would be measured at lower field strengths, this result does not influence the pattern of T2 variation we have described.
In conclusion, we have described both the 3D architecture of human articular cartilage of the tibial plateau and the typical MRI appearance of this joint surface. The characteristic appearance of the tibial plateau on MRI reflects the influence of an equally typical joint-specific macroscopic architecture of the extracellular matrix on T2 relaxation. MRI is, therefore, an ideal technique for the study of articular cartilage structure. It also follows that the interpretation of MRI of articular cartilage requires consideration of the joint-specific structure of the articular surface and orientation relative to B0.
Acknowledgments
We thank Charles Daghlian and Louisa Howard for assistance with sample
preparation and for performing the electron microscopy and Thomas Osborne for
drawing Figure 6.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  L. M. White, M. S. Sussman, M. Hurtig, L. Probyn, G. Tomlinson, and R. Kandel Cartilage T2 Assessment: Differentiation of Normal Hyaline Cartilage and Reparative Tissue after Arthroscopic Cartilage Repair in Equine Subjects. Radiology, November 1, 2006; 241(2): 407 - 414. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. G. Potter and L. F. Foo Magnetic Resonance Imaging of Articular Cartilage: Trauma, Degeneration, and Repair Am. J. Sports Med., April 1, 2006; 34(4): 661 - 677. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. P. Recht, D. W. Goodwin, C. S. Winalski, and L. M. White MRI of Articular Cartilage: Revisiting Current Status and Future Directions Am. J. Roentgenol., October 1, 2005; 185(4): 899 - 914. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
