T2 relaxation time and delayed gadolinium ...

T2 Relaxation Time and Delayed Gadolinium-Enhanced MRI of Cartilage (dGEMRIC) of Human Patellar Cartilage at 1.5 T and 9.4 T: Relationships with Tissue Mechanical Properties E. Lammentausta,1 P. Kiviranta,2 M.J. Nissi,3 M.S. Laasanen,3 I. Kiviranta,4 M.T. Nieminen,5 J.S. Jurvelin1,3 1

Department of Applied Physics, University of Kuopio, POB 1627, FI-70211 Kuopio, Finland

2

Department of Anatomy, University of Kuopio, Kuopio, Finland

3

Department of Clinical Physiology and Nuclear Medicine, Kuopio University Hospital, Kuopio, Finland

4

Department of Orthopaedics and Traumatology, Jyva¨skyla¨ Central Hospital, Jyva¨skyla¨, Finland

5

Department of Diagnostic Radiology, Oulu University Hospital, Oulu, Finland

Received 5 April 2005; accepted 4 August 2005 Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/jor.20041

ABSTRACT: Quantitative magnetic resonance imaging (MRI) techniques have been developed for noninvasive assessment of the structure of articular cartilage. T2 relaxation time is sensitive to the integrity and orientation of the collagen network, while T1 relaxation time in presence of Gd-DTPA2 (dGEMRIC) reflects the proteoglycan content of cartilage. In the present study, human patellar cartilage samples were investigated in vitro to determine the ability of MRI parameters to reveal topographical variations and to predict mechanical properties of cartilage at two different field strengths. T2 and dGEMRIC measurements at 1.5 T and 9.4 T were correlated with the static and dynamic compressive moduli at six anatomical locations of the patellar surface. Statistically significant linear correlations were observed between MRI and mechanical parameters at both field strengths, especially between T2 and Young’s modulus. No significant difference was found between the T2 measurements at different field strengths in predicting mechanical properties of the tissue. Topographical variation of T2 values at both field strengths was similar to that of Young’s moduli. The current results demonstrate the feasibility of quantitative MRI, particularly T2 mapping, to reflect the mechanical properties of human patellar cartilage at both field strengths. ß 2005 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res

Keywords:

articular cartilage; human; T2; dGEMRIC; biomechanics

INTRODUCTION Articular cartilage constituents, that is, proteoglycans, the three-dimensional collagen network, and interstitial fluid, and their interactions control the loading response of articular cartilage. The anisotropic collagen network restricts the Correspondence to: Eveliina Lammentausta (Telephone: þ358-17-162341; Fax: þ358-17-162585; E-mail: [email protected]) ß 2005 Orthopaedic Research Society. Published by Wiley Periodicals, Inc.

swelling pressure of proteoglycans (PGs) induced by electrostatic repulsion forces and hydration. This produces a prestrain of collagen fibers without external loading. The composition of articular cartilage provides a poroelastic tissue with a nearly frictionless surface and the ability to absorb energy and distribute loads uniformly to the underlying bone.1 In normal articular cartilage, the PG content gradually increases towards the deep cartilage. The collagen content of normal cartilage is, inversely to PG content, highest in the superficial zone.2 The collagen fibrils are JOURNAL OF ORTHOPAEDIC RESEARCH 2006

1

2

LAMMENTAUSTA ET AL.

organized parallel to the articular surface in the superficial zone (5–15% of the total tissue thickness), more randomly in the transitional zone (1–15% of the total thickness), and perpendicular to the articular surface in the deep zone (70–90% of the total thickness).3 Mechanically, the collagen network is primarily responsible for the dynamic properties of cartilage by constraining transversal expansion, whereas PGs contribute predominantly to the interstitial fluid flow and to the equilibrium response of cartilage.4,5 The structural and mechanical properties vary within and between different articular surfaces.6,7 Several quantitative MRI techniques have recently been introduced for the noninvasive assessment of structural and mechanical properties of articular cartilage.8 T2 mapping is sensitive to the integrity of the collagen network, collagen content, and fibril orientation.9–11 T1 mapping in the presence of Gd-DTPA2 contrast agent, namely delayed gadolinium-enhanced MRI of Cartilage (dGEMRIC), reflects the PG distribution in cartilage via the inverse distribution of the ionic contrast agent.12,13 Both T2 mapping14–16 and dGEMRIC14,15,17,18 have been related to mechanical properties of animal and human cartilage at high magnetic fields; however, these relationships have not been studied in human cartilage using a clinical MRI system with clinically feasible imaging parameters. The aims of the present in vitro study were to determine the ability of T2 mapping and dGEMRIC to reveal the topographical variation and tissue characteristics in human patellar cartilage and to assess the ability of quantitative MRI techniques to predict the site-dependent mechanical properties at a typical clinical field strength. To accomplish these aims, T2 and dGEMRIC measurements at 1.5 T and 9.4 T were compared with measurements of static and dynamic compressive moduli at six different anatomical locations at the patellar cartilage surface. All MRI measurements were performed after equilibration at low concentration of GdDTPA2 enabling the concurrent measurement of T2 and dGEMRIC, because the T2 relaxation of cartilage is nearly unaffected at low Gd-DTPA2 concentrations.19

MATERIALS AND METHODS Human cadaver knee joints (N ¼ 14, 12 male, 2 female, mean age ¼ 55 18 years) were obtained within 48 h JOURNAL OF ORTHOPAEDIC RESEARCH 2006

postmortem from the Jyva¨skyla¨ Central Hospital, Jyva¨skyla¨, Finland, with permission from the national authority (National Authority of Medicolegal Affairs, Helsinki, Finland, permission 1781/32/200/01). Right knee joints were opened, and patellar, femoral, and tibial bones were dissected free and frozen separately for later use. The patellae were thawed overnight, and six topographical locations were assessed: superolateral (SL), superomedial (SM), central lateral (CL), central medial (CM), inferolateral (IL), and inferomedial (IM) (Fig. 1). The sites of interest showed mostly cartilage with smooth and intact surfaces; however, several samples also included early degenerative superficial changes. For the reproducible localization of sites of interest a grid was drawn on the cartilage surface using a fine felt-tipped pen, and marker cuts were made on the each side of patellae with a biopsy punch. Patellae were equilibrated overnight in 0.5 mM GdDTPA2 solution (Magnevist, Schering AG, Germany) and frozen. Prior to measurements, patellae were wrapped in Parafilm (Pechiney Plastic Packaging, Chicago, IL) to prevent dehydration, and MRI-visible capsules were attached at the measurement sites to serve as localization markers. A General Electric Signa TwinSpeed 1.5 T clinical scanner (GE Healthcare, Milwaukee, WI) was used together with a 300 receiving surface coil and the body coil as the transmitting coil. The articular surface of intact patellae were oriented parallel to the B0 field to emulate clinical patient positioning. The frequency encoding direction was oriented perpendicular to the articular surface to minimize chemical shift artefact. For T2 mapping, a series of images were acquired using a multislice

Figure 1. Test sites for MRI and mechanical measurements on the patellar articular surface. SM ¼ superomedial, SL ¼ superolateral, CM ¼ central medial, CL ¼ central lateral, IM ¼ inferomedial, IL ¼ inferolateral. DOI 10.1002/jor

T2 AND dGEMRIC OF HUMAN CARTILAGE

3

multiecho spin echo sequence (TR ¼ 1000 ms, TE ¼ 10.3, 20.6, 30.9, 41.2, 51.5, 61.8, 72.1, and 82.4 ms, ETL ¼ 8, 3-mm slice thickness, a field of view of 8 cm, and 256 256 matrix size to yield 0.313 mm in-plane resolution, measured at room temperature). In this prototype sequence, the slice profile was modified to decrease the contribution of the stimulated echoes to the signal. This was followed by a T1 relaxation time (dGEMRIC) measurement using a single-slice inversion recovery fast spin echo sequence (TR ¼ 1700 ms, TE ¼ 11 ms, TI ¼ 50, 100, 200, 400, 800, and 1600 ms, ETL ¼ 6). After the 1.5 T MRI measurements, full-thickness 4-mm cartilage disks without subchondral bone were prepared from each measurement site using a biopsy punch and a razor blade and were frozen. For 9.4 T measurements, an Oxford NMR vertical magnet (Oxford Instruments PLC, Witney, UK), a SMIS console (SMIS Ltd, Surrey, UK), and a 5-mm high-resolution volume spectroscopy probe (Varian Associates Inc., Palo Alto, CA) were used. Prior to measurements, the samples were thawed, sealed in a test tube (dia. ¼ 5 mm), and immersed in Gd-DTPA2 solution. The samples were located axially in the center of the coil, and the sample surface was oriented perpendicular to the B0 field, as limited by the coil construction. T2 was determined from

six single spin echo measurements (TR ¼ 1500 ms, TE ¼ 14, 20, 28, 40, 56, and 80 ms, 1-mm slice thickness, 0.039 mm resolution across cartilage depth, at 25 18C). Subsequently, T1 measurements were conducted using a saturation recovery sequence (TE ¼ 14 ms, TR ¼ 100, 180, 330, 600, 1100, and 2000 ms). The MRI parameter maps were fitted into exponential relaxation equations assuming a mono-exponential decay20 (MatLab, Mathworks Inc., Natick, MA). For 1.5 T data obtained from the multiecho sequence, the T2 maps were calculated using all eight echoes. For 1.5 T measurements, regions of interest were manually segmented at the measurement sites as marked by the vitamin capsules (Fig. 2a). To avoid a partial volume effect, the most superficial pixel was omitted from the analysis. Spatial depth-wise profiles were calculated by averaging 10 pixels for 1.5 T and 7 pixels for 9.4 T measurements along the cartilage surface to match the slice thickness between different field strengths. Further, the relaxation time profiles of the 9.4 T measurements were downsampled to match the depthwise resolution of the 1.5 T measurements. Profiles at 1.5 T were truncated to match the thickness of the 9.4 T measurements, because some of the deepest tissue typically remained on the bone when the cartilage disks

Figure 2. Localization of the sample sites from (a) 1.5 T MRI slices with MRI-visible capsules and (b) 9.4 T slices. Capsules were attached on a plastic film wrapped firmly around the patella. The regions of interest were chosen to be rectangular and approximately 10 pixels (3.125 mm) wide at 1.5 T and 7 pixels (1.094 mm) wide at 9.4 T.

Figure 3. Relaxation time maps for a representative sample obtained both at 1.5 T and 9.4 T field strengths. The 1.5 T data was truncated to match the sample thickness at 9.4 T. In all maps the cartilage surface locates in the upper edge. The 9.4 T and 1.5 T maps have widths of approximately 1 mm and 3 mm, respectively.

DOI 10.1002/jor

JOURNAL OF ORTHOPAEDIC RESEARCH 2006

4

LAMMENTAUSTA ET AL.

were isolated. From the profiles, relaxation times for the approximately 1 mm of the most superficial tissue (three pixels at 1.5 T and 24 pixels at original 9.4 T resolution) and bulk values covering the full thickness uncalcified cartilage were determined. Because the thickness of the superficial cartilage layer was beyond the resolution of 1.5 T measurements, the superficial values of MR parameters refer to the values calculated for the most superficial 1 mm of tissue. This is thicker than the histological superficial cartilage layer as determined by the collagen fibril orientation. Prior to biomechanical testing, samples were reequilibrated in phosphate-buffered saline solution including enzyme inhibitors (5 mM ethylenediaminetetraacetic acid (EDTA) and 5 mM benzamide HCl) for at least 2 h to wash out the ionic contrast agent that could possibly affect the swelling pressure of cartilage. Biomechanical testing was performed with a custommade high-resolution material testing device including a load cell with resolution of 5 mN (Honeywell Sensotec, Columbus, OH) and a precision motion controller with 0.1 mm resolution (Newport, Irvine, CA).21 A stressrelaxation test was performed in unconfined compression geometry. After establishing a proper surface contact, a 10% prestrain step was applied, followed by a 1-h relaxation. Subsequently, this was followed by three 2% steps at 1 mm/s ramp velocity and 40 min relaxation after each step. Young’s modulus (static compression) was determined from the equilibrium response. A dynamic test with 1-Hz frequency and 1% strain amplitude was conducted after the stress-relaxation test, and the dynamic modulus was determined as the stress/strain ratio from peak-to-peak values.14,22 For comparison of mechanical and MRI parameters, linear Pearson correlation coefficients were determined separately for different locations, for medial and lateral facets, and for all samples. The mixed linear model tests were applied to test the statistical significance of topographical variation of MRI and mechanical parameters. The linear mixed model provides an accurate comparison of data with missing values by modeling also variances and covariances of data. Advantageously, the possible sample interdependencies are acceptable,23 for example, the dependency between the samples at different locations obtained from the same patella. Statistical analyses were conducted by using the SPSS software (version 11.5, SPSS Inc, Chicago, IL).

the cartilage disks with nonuniform thickness could not be mechanically tested. Both T2 and dGEMRIC showed similar relaxation time patterns at different field strengths (Fig. 3). Depth-wise T2 profiles revealed a similar shape at 1.5 T and 9.4 T, with long T2 values in the more superficial tissue and short T2 values in the deeper cartilage (Fig. 4), despite the 90-degree orientation difference in the B0 field. Superficial T2 values varied between 29 and 105 ms and between 22 and 62 ms at 1.5 T and 9.4 T, respectively, while bulk T2 values varied between 32 and 116 ms and between 17 and 50 ms at 1.5 T and 9.4 T, respectively. dGEMRIC profiles at both field strengths showed increasing T1 values towards deep cartilage, but occasionally the spatial profiles showed a somewhat different shape (Fig. 4). Superficial dGEMRIC values varied between 188 and 392 ms and between 392 and 1082 ms at 1.5 T and 9.4 T, respectively. Bulk dGEMRIC values varied between 187 and 398 ms and between 407 and 1083 ms at 1.5 T and 9.4 T, respectively. The mean values SD of T2 and dGEMRIC at different locations at both field strengths are shown in Tables 1 and 2. For 1.5 T and 9.4 T field strengths, T2 relaxation times were linearly correlated by r ¼ 0.41 ( p < 0.01) and r ¼ 0.47 ( p < 0.01), and dGEMRIC

RESULTS The total number of the samples measured successfully by MRI at 1.5 T and 9.4 T and by biomechanical testing was 75 (for different locations, SL: n ¼ 14, CL: n ¼ 12, IL: n ¼ 12, SM: n ¼ 12, CM: n ¼ 12, IM: n ¼ 13). Some samples were unsuccessfully isolated from the bone, and JOURNAL OF ORTHOPAEDIC RESEARCH 2006

Figure 4. Depth-wise profiles of relaxation times calculated from the maps shown in Figure 3. One pixel in 1.5 T equals eight pixels in 9.4 T. The 9.4 T data has been downsampled to match the resolution of the 1.5 T profile. The orientation of sample surface was parallel to B0 field for 1.5 T measurements and perpendicular to B0 field for 9.4 T measurements. DOI 10.1002/jor


5

Table 1. Mean Values SD of the Measured MRI Parameters at 1.5 T for Different Topographical Locations of Patellar Cartilage Surface (Fig. 1), Medial (MED), and Lateral (LAT) Facets, and All Samples (ALL) Superficial ROIs Location SL CL IL SM CM IM LAT MED ALL

Bulk Tissue

dGEMRIC (ms)

T2 (ms)

dGEMRIC (ms)

T2 (ms)

276 55 296 47 310 52 298 51 298 31 305 50 293 52 300 44 297 48

57 11 68 21 70 20 58 18 65 15 60 14 65 18 61 15 63 17

277 56 317 48 315 49 295 53 304 37 312 51 302 53 304 47 303 50

52 15 51 12 55 16 53 23 53 12 47 10 53 14 51 16 52 15

values by r ¼ 0.23 ( p < 0.05) and r ¼ 0.31 ( p < 0.01) for superficial and bulk tissue, respectively. Young’s modulus at equilibrium (Es) and dynamic modulus (Ed) ranged from 0.02 to 2.56 MPa and from 0.18 to 12.29 MPa, respectively (Table 3). Es and Ed showed a strong linear correlation with each other (r ¼ 0.92, p < 0.01, N ¼ 75). The topographical variation of superficial T2 (or 1/T2, i.e., relaxation rate R2, Fig. 5) values and mechanical parameters showed similar trends at both field strengths, whereas dGEMRIC showed only a weak association with mechanical properties (Fig. 5 and Tables 1–3). The topographical variation of Young’s modulus and dynamic modulus showed statistically significant ( p < 0.05) differences between the locations SL–CL, SL– SM, CL–SM, CL–IM, SM–CM, and CM–IM. Additionally, variations of dynamic moduli were significant between sites SL–IL and IL–SM. For

superficial T2 at 9.4 T, the variations were significant between sites SL–CL, CL–IL, CL– CM, and CL–IM. Significant linear correlations between the MRI parameters and the mechanical moduli were observed at both field strengths; at 1.5 T, the highest linear correlations were observed between T2 and mechanical parameters (up to r ¼ 0.86), whereas at 9.4 T dGEMRIC showed higher linear correlations with mechanical moduli (up to r ¼ 0.83) (Tables 4 and 5). At the medial facet, T2 values at 1.5 T showed the strongest linear correlations with mechanical parameters (r ¼ 0.75, p < 0.01), while at the lateral facet, dGEMRIC values at 9.4 T showed the highest correlations (r ¼ 0.62, p < 0.01). For pooled data from all samples and locations, the highest linear correlation was that between the superficial T2 at 1.5 T and dynamic modulus (r ¼ 0.66, p < 0.01).

Table 2. Mean Values SD of the Measured MRI Parameters at 9.4 T for Different Topographical Location at Cartilage Surface (Fig. 1), Medial (MED) and Lateral (LAT) Facets, and All Samples (ALL) Superficial ROIs Location SL CL IL SM CM IM LAT MED ALL

DOI 10.1002/jor

Bulk Tissue

dGEMRIC (ms)

T2 (ms)

dGEMRIC (ms)

T2 (ms)

600 79 535 90 583 131 579 98 606 140 647 181 574 102 612 144 593 125

39 10 47 5 37 8 37 10 41 10 36 10 41 9 38 10 40 10

684 78 639 106 681 124 690 100 691 150 719 176 669 102 701 143 645 124

28 7 32 6 27 6 26 9 29 8 25 6 29 7 27 7 28 7


6

LAMMENTAUSTA ET AL.

Table 3. Mean Values SD of the Measured Biomechanical Parameters for Different Topographical Location at Cartilage Surface (Fig. 1), Medial (MED) and Lateral (LAT) Facets, and All Samples (ALL) Location SL CL IL SM CM IM LAT MED ALL

Es (MPa)

Ed (MPa)

0.67 0.26 0.24 0.13 0.52 0.69 0.73 0.34 0.31 0.30 0.66 0.39 0.49 0.45 0.57 0.39 0.53 0.42

5.68 1.88 1.95 0.97 3.23 3.50 5.34 2.66 2.37 2.31 4.29 2.27 3.73 2.78 4.01 2.65 3.87 2.70

Es ¼ Young’s modulus, Ed ¼ dynamic modulus.

At the medial facet, the differences in the strength of correlation between superficial T2 and both mechanical moduli were significant. At the lateral facet, linear correlations were significant between both superficial and bulk dGEMRIC and both mechanical moduli. For all samples, the correlation coefficients showed significant differences between the superficial T2 and dynamic modulus, superficial dGEMRIC and dynamic modulus and between bulk dGEMRIC and both mechanical moduli. No significant differences

were found in correlation coefficients between the MRI parameters obtained at single field strength and the mechanical moduli within individual locations.

DISCUSSION Human patellar cartilage was investigated to determine the topographical variation of quantitative MRI parameters at both a typical field strength used clinically (1.5 T) and a high field strength (9.4 T), which were further correlated with the site-matched mechanical reference parameters. Previously, strong linear correlations have been established between the dGEMRIC technique and Young’s modulus for bovine cartilage,14,15 and previous in vitro studies with human tissue have shown the ability of both dGEMRIC and T2 relaxation time mapping to predict the mechanical properties of cartilage.17,18 In the present study, the most significant linear correlation was obtained between T2 at 1.5 T and the dynamic modulus. The prediction of the tissue compressive stiffness with MRI varied from poor to excellent among the test locations and parameters, dGEMRIC showing a more accurate prediction at 9.4 T, whereas T2 yielded a better agreement at 1.5 T. Our MRI results suggest that

Figure 5. Topographical variation of the relaxation rate R2 (1/T2, for display purposes) and dGEMRIC obtained at 1.5 T (a,c) and 9.4 T (b,d) along with that of Young’s modulus Es (mean SD) at different locations of the patellar cartilage (first 0.938 mm in depth). JOURNAL OF ORTHOPAEDIC RESEARCH 2006

DOI 10.1002/jor


7

Table 4. Linear Correlation Coefficients between the Mechanical Parameters and T2 Relaxation Time at 1.5 T and 9.4 T for Different Topographical Locations of Patellar Cartilage (Fig. 1), Medial Facet (MED), Lateral Facet (LAT), and All Samples (ALL)

T2 at 1.5 T Superficial versus Es Superficial versus Ed Bulk versus Es Bulk versus Ed T2 at 9.4 T Superficial versus Es Superficial versus Ed Bulk versus Es Bulk versus Ed

SL

CL

IL

SM

CM

0.01 0.17 0.33 0.51

0.41 0.42 0.22 0.23

0.64* 0.70* 0.41 0.52

0.55 0.75** 0.44 0.63*

0.75** 0.73** 0.78** 0.67*

0.37 0.54* 0.52 0.60*

0.11 0.18 0.06 0.02

0.24 0.20 0.34 0.40

0.36 0.40 0.46 0.52

0.39 0.41 0.65* 0.64*

IM

MED

LAT

ALL

0.86** 0.81** 0.78** 0.67*

0.69** 0.75** 0.58** 0.59**

0.44** 0.52** 0.25 0.33*

0.55** 0.62** 0.40** 0.45**

0.29 0.23 0.27 0.21

0.39* 0.39* 0.48** 0.50**

0.34* 0.36* 0.38* 0.41*

0.37** 0.38** 0.43** 0.45**

Es ¼ Young’s modulus, Ed ¼ dynamic modulus. *Significant at p < 0.05 level; **significant at p < 0.01 level.

the dynamic mechanical properties of cartilage are predictable also at clinically relevant field strengths, whereas the static compressive strength is more closely reflected via dGEMRIC at higher field strengths. The scatter of superficial T2 at 1.5 T was quite wide. Early stages of the articular cartilage degeneration, possibly present in some of the samples, include fibrillation of the superficial zone of the cartilage extending into the transitional zone,24 which would influence the T2 values of tissue close to the articular surface. A significant difference in T2 between the two field strengths was observed. This probably relates to the different sample orientation (i.e., collagen arrangement), known to affect significantly the T2 relaxation time through dipolar interaction.25 In

addition, multiecho sequence has been reported to increase T2 relaxation time.26,27 To validate current results, T2 values were calculated for some regions of interest omitting the first echo. The results were similar. Despite the different sample orientation and field strength, a significant relationship between T2 and mechanical properties was observed at both field strengths. Formerly, the dependence of T2 on collagen orientation has been confirmed by polarized light microscopy, and the collagen-related role of T2 has been shown by enzymatic digestions.9–11 Further, the effect of collagen matrix degradation on the T2 values and cartilage compressive stiffness has been revealed by investigating the influences of enzymatic collagen depletion in animal samples.9,16

Table 5. Linear Correlation Coefficients between the Mechanical Parameters and dGEMRIC at 1.5 T and 9.4 T for Different Topographical Locations of Patellar Cartilage (Fig. 1), Medial Facet (MED), Lateral Facet (LAT), and All Samples (ALL)

dGEMRIC at 1.5 T Superficial versus Es Superficial versus Ed Bulk versus Es Bulk versus Ed dGEMRIC at 9.4 T Superficial versus Es Superficial versus Ed Bulk versus Es Bulk versus Ed

SL

CL

IL

SM

CM

IM

MED

0.12 0.18 0.28 0.42

0.33 0.25 0.38 0.30

0.37 0.33 0.39 0.35

0.44 0.35 0.42 0.30

0.39 0.31 0.37 0.29

0.51 0.45 0.40 0.38

0.41* 0.33* 0.31 0.23

0.36 0.26 0.22 0.26

0.14 0.10 0.35 0.27

0.61* 0.59* 0.53 0.57

0.66* 0.48 0.35 0.50

0.35* 0.35* 0.31 0.32

0.82** 0.80** 0.79** 0.83**

0.13 0.08 0.06 0.10

LAT 0.15 0.04 0.03 0.13 0.60** 0.62** 0.58** 0.55**

ALL 0.26* 0.17 0.14 0.03 0.45** 0.47** 0.43** 0.42**

Es ¼ Young’s modulus, Ed ¼ dynamic modulus. *Significant at p < 0.05 level; **significant at p < 0.01 level. DOI 10.1002/jor


8

LAMMENTAUSTA ET AL.

A recent in vivo study reported that T2 measurements may prove feasible in evaluating the severity of osteoarthritis.28 The current results demonstrate the feasibility of T2 mapping to reflect the mechanical properties of articular cartilage, both at 1.5 T and 9.4 T. The topographical variation of T2 at both field strengths was similar to that of Young’s modulus, and the linear correlations between T2 and Young’s modulus were significant both at 1.5 T and 9.4 T. In the present study, T2 relaxation time was measured in the presence of Gd-DTPA2, because the contrast agent has a minimal effect on T2 relaxation time at low equilibrating concentrations.19 Consequently, the present results also show that T2 in the presence of Gd-DTPA2 is indicative of the mechanical integrity of cartilage. Generally, dGEMRIC showed a very weak topographical variation that did not follow the variation in the mechanical moduli. The linear correlation coefficients between MRI and mechanical parameters were not as high as reported for animal tissue.14,15 This may be due in part to the heterogeneous sample population, that is, large variation in the age and tissue integrity, but also suggests that the mechanical properties of cartilage cannot satisfactorily be explained by a single MRI parameter. However, linear correlations within some of the individual test sites were high and significant, while at some sites no significant correlations existed. Similar results on human knee cartilage have been reported.18 In the light of our and previous results, we believe that a single structural component can significantly modulate the mechanical properties of normal cartilage, but in degenerated tissue the interaction between different constituents and the prediction of mechanical properties become more complex. Further, variations in loading conditions between different joint surfaces or topographical locations may be related to a differently adapted macromolecular composition that can further complicate the use of a single MRI parameter as a surrogate marker for mechanical properties.17 The 1.5 T MRI measurements were conducted using intact patellae with articular cartilage attached to subchondral bone, whereas for 9.4 T experiments and biomechanical testing cartilage disks without subchondral bone were prepared. Both the MRI and mechanical parameters may have been altered when isolating cartilage tissue from subchondral bone. Previous results indicate, however, that the orientation of the collagen fibers is preserved also in detached cartilage disks, but the collagen network density is considerably JOURNAL OF ORTHOPAEDIC RESEARCH 2006

decreased.29 Although the collagen network and thus the mechanical properties of cartilage may have been altered while detaching disks, unconfined compression is considered to provide an accurate technique to measure intrinsic tissue properties. The only measurement geometry applicable for intact articular surface is indentation. It is sensitive to cartilage thickness, for example, and necessitates complex modeling approaches to extract material parameters. After isolation from bone and surrounding cartilage, samples occasionally experience swelling and curling,30 which may alter the macromolecular density and other material properties and thus affect the results, especially in the degenerated samples. Mechanical testing of the deformed cartilage samples under unconfined conditions increases uncertainties, and some of the samples had to be excluded for this reason. Further, a thin layer of the deep-most cartilage may have been left attached to the bone surface in sample preparation, which could bias the biomechanical measurements. Nonetheless, even though detaching cartilage disks from subchondral bone could affect the results, it was feasible technique in this study to prepare samples for mechanical and high-field high-resolution MRI measurements. In conclusion, T2 and dGEMRIC appear to provide feasible tools for assessing cartilage properties also at clinical field strength, although these parameters cannot fully characterize the mechanical properties of cartilage. T2 mapping can reflect some of the topographical variation in mechanical properties of human cartilage at both field strengths, while dGEMRIC shows significant correlations with mechanical parameters within individual topographical locations at high field strength. These results will likely have relevance when interpreting quantitative MRI measurements in vivo.

ACKNOWLEDGMENTS The authors thank National Bio-NMR Facility, University of Kuopio, Kuopio, Finland. Financial support from the Academy of Finland (grant No. 205886), Kuopio University Hospital (EVO 5031329), and Sigrid Juselius Foundation is acknowledged.

REFERENCES 1. Mow VC, Fithian DC, Kelly MA. 1990. Fundamentals of articular cartilage and meniscus biomechanics. In: Ewing JW, editor. Articular cartilage and DOI 10.1002/jor


2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

knee joint function: basic science and arthroscopy. New York: Raven Press Ltd.; p 1–18. Mow VC, Ratcliffe A, Poole AR. 1992. Cartilage and diarthrodial joints as paradigms for hierarchical materials and structures. Biomaterials 13:67–97. Benninghoff A. 1925. Form und Bau der Gelenkknorpel in ihren Beziehungen zur Function. Erste Mitteilung: die modellierenden und formerhaltenden Faktoren des Knorpelreliefs. Z Anat 76:43– 63. Bader DL, Kempson GE. 1994. The short-term compressive properties of adult human articular cartilage. Biomed Mater Eng 4:245–256. Korhonen RK, Laasanen MS, Toÿra¨s J, et al. 2003. Fibril reinforced poroelastic model predicts specifically mechanical behavior of normal, proteoglycan depleted and collagen degraded articular cartilage. J Biomech 36:1373–1379. Froimson MI, Ratcliffe A, Gardner TR, et al. 1997. Differences in patellofemoral joint cartilage material properties and their significance to the etiology of cartilage surface fibrillation. Osteoarthritis Cartilage 5:377–386. Maroudas A, Evans H, Almeida L. 1973. Cartilage of the hip joint. Topographical variation of glycosaminoglycan content in normal and fibrillated tissue. Ann Rheum Dis 32:1–9. Burstein D, Gray M. 2003. New MRI techniques for imaging cartilage. J Bone Joint Surg Am 85A(Suppl 2):70–77. Nieminen MT, Toÿra¨s J, Rieppo J, et al. 2000. Quantitative MR microscopy of enzymatically degraded articular cartilage. Magn Reson Med 43:676–681. Xia Y, Moody JB, Burton-Wurster N, et al. 2001. Quantitative in situ correlation between microscopic MRI and polarized light microscopy studies of articular cartilage. Osteoarthritis Cartilage 9: 393–406. Nieminen MT, Rieppo J, Toÿra¨s J, et al. 2001. T2 relaxation reveals spatial collagen architecture in articular cartilage: a comparative quantitative MRI and polarized light microscopic study. Magn Reson Med 46:487–493. Bashir A, Gray ML, Hartke J, et al. 1999. Nondestructive imaging of human cartilage glycosaminoglycan concentration by MRI. Magn Reson Med 41:857–865. Nieminen MT, Rieppo J, Silvennoinen J, et al. 2002. Spatial assessment of articular cartilage proteoglycans with Gd-DTPA-enhanced T(1) imaging. Magn Reson Med 48:640–648. Nieminen MT, Toÿra¨s J, Laasanen MS, et al. 2004. Prediction of biomechanical properties of articular cartilage with quantitative magnetic resonance imaging. J Biomech 37:321–328. Nissi MJ, Toÿra¨s J, Laasanen MS, et al. 2004. Proteoglycan and collagen sensitive MRI evalua-

DOI 10.1002/jor

16.

17.

18.

19.

20. 21.

22.

23. 24.

25.

26.

27.

28.

29.

30.

9

tion of normal and degenerated articular cartilage. J Orthop Res 22:557–564. Wayne JS, Kraft KA, Shields KJ, et al. 2003. MR Imaging of normal and matrix-depleted cartilage: correlation with biomechanical function and biochemical composition. Radiology 228:493–499. Samosky J, Burstein D, Grimson E, et al. 2005. Spatially-localized correlation of dGEMRIC-measured GAG distribution and mechanical stiffness in the human tibial plateau. J Orthop Res 23:93–101. Kurkija¨rvi J, Nissi M, Kiviranta I, et al. 2004. Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) and T2 characteristics of human knee articular cartilage: topographical variation and relationships to mechanical properties. Magn Reson Med 52:41–46. Nieminen MT, Menezes N, Williams A, et al. 2004. T2 of articular cartilage in the presence of Gd-DTPA(2-). Magn Reson Med 51:1147–1152. Abragam A. 1961. The principles of nuclear magnetism. London: Oxford University Press. Toÿra¨s J, Rieppo J, Nieminen MT, et al. 1999. Characterization of enzymatically induced degradation of articular cartilage using high frequency ultrasound. Phys Med Biol 44:2723–2733. Jurvelin JS, Buschmann MD, Hunziker EB. 1997. Optical and mechanical determination of Poisson’s ratio of adult bovine humeral articular cartilage. J Biomech 30:235–241. Brown H, Prescott R. 2001. Applied mixed models in medicine. New York: John Wiley & Sons, Inc. Buckwalter JA, Mankin HJ. 1997. Articular cartilage, part II: degeneration and osteoarthrosis, repair, regeneration, and transplantation. J Bone Joint Surg Am 79:612–632. Xia Y. 1998. Relaxation anisotropy in cartilage by NMR microscopy (muMRI) at 14-microm resolution. Magn Reson Med 39:941–949. Maier CF, Tan SG, Hariharan H, et al. 2003. T2 quantitation of articular cartilage at 1.5 T. J Magn Reson Imaging 17:358–364. Mendlik T, Faber SC, Weber J, et al. 2004. T2 quantitation of human articular cartilage in a clinical setting at 1.5 T: implementation and testing of four multiecho pulse sequence designs for validity. Invest Radiol 39:288–299. Dunn TC, Lu Y, Jin H, et al. 2004. T2 relaxation time of cartilage at MR imaging: comparison with severity of knee osteoarthritis. Radiology 232: 592–598. Keinan-Adamsky K, Shinar H, Navon G. 2005. The effect of detachment of the articular cartilage from its calcified zone on the cartilage microstructure, assessed by 2H-spectroscopic double quantum filtered MRI. J Orthop Res 23:109–117. Setton LA, Tohyama H, Mow VC. 1998. Swelling and curling behaviors of articular cartilage. J Biomech Eng 120:355–361.
