Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2010 Jan 1.
Published in final edited form as: Magn Reson Imaging. 2008 Sep 17;27(1):131–136. doi: 10.1016/j.mri.2008.05.004

Examiner repeatability of patellar cartilage T2 values

Matthew F Koff a, Sebastien Parratte a, Kimberly K Amrami b, Kenton R Kaufman a,*
PMCID: PMC2675164  NIHMSID: NIHMS87547  PMID: 18801631

Abstract

Aim

The purpose of this study was to evaluate the intra- and interexaminer resegmentation precision of patellar cartilage T2 mapping measurements in healthy subjects.

Materials and Methods

T2-weighted images of patellar cartilage for 10 subjects were acquired. Two individuals manually segmented patellar cartilage at each slice location twice, once on each of two separate days. Bulk average and zonal T2 values for the superficial, middle, and deep layers of cartilage were calculated. The root mean square (RMS) and coefficient of variation (COV) were calculated using the repeated measurements of each slice of each subject by each examiner.

Results

The intraexaminer bulk T2 differences were 0.2±1.0 ms, with an RMS error of 0.7 ms and a COV of 1.9%. The differences of interexaminer bulk T2 values was 1.0±1.4 ms, with an RMS error of 1.2 ms and a COV of 3.3%. The superficial zone of cartilage had the highest zonal variability of T2 values. The average interexaminer T2 values for the superficial, middle and deep zones were 42.2±5.6, 38.1±5.3 and 31.9±4.6 ms, respectively.

Conclusion

The interexaminer variability of calculated T2 values highlights the difficulty of interpreting significant differences of T2 values which are similar in magnitude. The repeatability measurements of patellar cartilage T2 values were less than reported intersession T2 repeatability.

Keywords: Reproducibility of results, MRI, Cartilage, Patella

1. Introduction

The World Health Organization has declared the 2000–2010 decade as the bone and joint decade with the goal of improving the understanding, the prevention and the treatment of musculoskeletal disorders [1]. Osteoarthritis, one of the most common musculoskeletal diseases, is a main target of this program. In 1997, an estimated 16% of the US population had a form of arthritis, and its prevalence is expected to further increase in the near future [1]. Therapeutic innovations developed for management of the early stage of osteoarthritis (OA) as well for surgical management of advanced cartilage lesion [2] require an objective quantitative method to confirm their validity and usefulness [3–7]. Clinical scores can be used as outcome measures for surgical procedures, but these methods only provide an indirect and subjective evaluation of the articular cartilage [8].

New methods of medical imaging may provide an objective quantitative method for evaluation of the cartilage status. Magnetic resonance imaging (MRI) is a powerful noninvasive evaluative tool for imaging cartilage, especially the evolving MRI technique of T2 relaxation time mapping [3–7,9–12]. The MRI time constant T2 represents the rate of internuclear dephasing between excited hydrogen dipoles [13]. The stratification of T2 through the depth of articular cartilage reflects the arrangement of the Type II collagen fibers within the collagen matrix as well as water content of the tissue [5,12]. T2 mapping has been demonstrated to be sensitive to degenerative changes of cartilage [3,5,12,14]. A current goal is to use T2 relaxation time for the clinical evaluation of regional cartilage heterogeneity and for discrimination between healthy and diseased cartilage [3,4,9,10].

Determining the precision error and reproducibility of an imaging method is a fundamental step before use of the imaging technique in routine clinical practice. Eckstein et al. [15] proposed a multifaceted analysis of the reproducibility of a new imaging technique, including: (a) technical precision — defined as repeated analysis of an identical dataset within one session; (b) resegmentation precision — defined as repeated image analysis of an identical dataset in different image analysis sessions, days or weeks apart; (c) interscan test–retest precision — defined as analysis of two datasets with repositioning between the scans; (d) long-term acquisition precision — defined as analysis of at least two datasets acquired in different imaging sessions, days or weeks apart; (e) overall precision — representing the error of (b)+(c) or (b)+(d); (f) interobserver precision — defined as image analysis performed by different observers; and (g) interscanner precision — defined as image analysis with similar evaluated the interscan test–retest precision aswell as long-term acquisition precision of cartilage T2 values [4]; however, the resegmentation precision has not been evaluated.

Resegmentation precision of T2 values assesses the reproducibility of the measurement methodology due to changes in technical or human performance during the segmentation process. This type of precision is especially necessary in evaluating the applicability of T2 mapping within a clinical setting, as different radiologists may produce varying results for quantitative cartilage analysis (Fig. 1). The resegmentation precision may be further evaluated by intraexaminer reproducibility and interexaminer reproducibility. These different measurements of T2 mapping precision evaluate the induced additional sources of error when the image analysis is performed by different observers over different days. This type of analysis is important since different readers may have different perceptions about what should and should not be considered cartilage during the segmentation process [3]. To the best of our knowledge, the intra- and interexaminer resegmentation precision of cartilage T2 values have not been reported. Therefore, the purpose of this study was to evaluate the intra- and interexaminer resegmentation precision of patellar cartilage T2 mapping measurements in healthy subjects. The specific goals of the study were to (1) evaluate global (bulk) tissue T2 values and depth dependent resegmentation intraobserver precision error of patellar cartilage T2 values, and (2) evaluate global and depth dependent interobserver precision error of patellar cartilage T2 values.

Fig. 1.

Fig. 1

Open in a new tab

(Top) Representative patellar cartilage T2 maps of a slice from one subject as processed separately by Examiners 1 and 2. There is a bulk T2 value difference of 1.9 ms between the examiners. (Bottom) Pixels of each T2 map are shown superimposed on one another. The overlapping portions of the regions of interest defined by Examiners 1 and 2 are shown in yellow. Pixels unique to Examiners 1 and 2 are shown in red and blue, respectively. The image shows that Examiner 1 defined a larger region of interest than Examiner 2, resulting in a different bulk T2 value.

2. Methods and materials

2.1. Subjects

Following local institutional review board approval with informed consent, 10 male volunteers (31±4 years old, range 25–36) were enrolled in the study. All subjects had no pain, stiffness or other symptoms of OA in the knee which was scanned.

2.2. Data acquisition

Magnetic resonance (MR) images of each subject's right patella were obtained using a 1.5-T clinical MRI scanner (GE Health Systems, Milwaukee, WI, USA). Six subjects were scanned using one scanner, and the remaining subjects were scanned using a second scanner. A dedicated transmit–receive knee coil was used for all imaging. An MR-compatible knee positioning device was used in both scanners to ensure consistency of knee position and flexion angle relative to the knee coil. A series of axial T2-weighted images were acquired using a multislice multiecho fast spin echo (FSE) pulse sequence [11]. Images were acquired across 11 slice locations centered on the patella. Imaging parameters were: Repetition time (TR)=1500 ms, slice thickness=2 mm, slice spacing=4 mm, in-plane resolution=0.49×0.49 mm, field of view=12×12 cm, acquisition matrix=256×160 (zero filled to 256×256), Bandwidth (BW)=31.25 MHz, Array Spatial Sensitivity Encoding Technique (ASSET) acceleration=1. Eight echo images were acquired at each slice location. The echo times (TEs) used for the first scanner was TE=7.82 – 62.53 ms, with 7.82-ms increments. The TEs used for the second scanner was TE=8.48 – 67.84 ms, with 8.48-ms increments. The acquisition time for the T2 mapping images was approximately 5 min. A single MR technologist acquired the images for all subjects. The series was acquired as part of an ongoing study in our laboratory, which had a total imaging time of approximately 1 h.

2.3. Data analysis

Two individuals processed the data at each slice location twice, once on each of two separate days. The different data analysis sessions were separated by more than 24 hours. All data processing by the two individuals was performed independently from one another. One individual had experience of analyzing over 800 cartilage T2 maps, and the second individual had experience of analyzing over 100 cartilage T2 maps. Custom written software (MATLAB, Natick, MA, USA) was used to process the image data. Patellar cartilage was manually segmented at each slice location [9]. The first echo image was used to delineate the deep cartilage from the underlying subchondral bone, and the last echo image was used to delineate the superficial cartilage from the surrounding joint space. Image slices at the proximal and distal pole of the patella were not included in the analysis due to potential volume averaging. T2 values were calculated on a pixel-by-pixel basis by fitting the TE to the corresponding signal intensity data (SI) using a monoexponential decay equation:

SI(TE)=So·exp(−TE/T2).

The solution for T2 was calculated using a nonlinear least squares fitting algorithm [11]. A bulk average T2 value was calculated from all pixels within the individual slices and used for bulk repeatability analysis. In addition, a semi-automated program [9] separated the cartilage within each slice into a deep zone (0–30% depth), middle zone (30–70% depth) and superficial zone (70–100% depth). Briefly, the program fit a spline to the most superficial pixels and a spline to the deep pixels within the cartilage region of interest (ROI) defined previously by the examiner. Next, the program individually parameterized the splines to calculate (x,y) coordinates at 80 equally spaced points. Using the (x,y) coordinates of corresponding parametric points on the superficial and deep splines, the program then calculated (x,y) coordinates at 30% and 70% depths on a line which connected the two splines. Therefore, the (x,y) coordinates of the superficial most pixels and the (x,y) coordinates of the calculated points at 30% depth created an ROI representing the deep zone of cartilage, the (x,y) coordinates of the deep pixels and the calculated (x,y) coordinates at 70% depth created an ROI representing the superficial zone of cartilage and the calculated (x,y) coordinates of the points at 30% and 70% depths represented the middle zone of cartilage. An average T2 value within each cartilage zone was used for zonal repeatability analysis. Finally, the number of pixels within an ROI defined by each examiner for each image plane was calculated, as well as the percentage of common pixels within the ROIs defined by both examiners. On average, the data processing took approximately 20–45 minutes for each patella.

2.4. Statistical analysis

Bland–Altman plots [16] were generated to evaluate intra- and interexaminer repeatability of bulk and zonal T2 values. These plots, which display the difference between two measurements on the ordinate and the average of two measurements on the abscissa, are effective at highlighting trends and differences between two independent measurements of a single variable. A repeatability coefficient was calculated from the Bland–Altman analysis as 1.96 times the standard deviation of the differences [16,17].

Additional numerical measures of intra- and interexaminer variability were calculated using the root mean square (RMS) error and coefficient of variation (COV) of the bulk and zonal T2 values. The RMS and COV were calculated using the repeated measurements of each slice of each subject by each examiner, as outlined previously by Gluer et al. [18]. In addition, a two-sample, two-tailed t test was performed to detect significant differences of the number of pixels within the ROIs defined by the individual examiners.

3. Results

A preliminary two-sample t test of the calculated T2 values was performed to evaluate the effect of using different scanners for data acquisition. No significant difference of T2 value was found due to using different scanners for image acquisition (P=.68, mean difference: −0.2±4.3 ms). Subsequently, the data from both scanners was pooled together for repeatability analysis.

A total of 66 slices of data from all subjects were available for analysis. The average bulk T2 calculated by the first examiner was 37.9±4.4 ms (mean±S.D.). The average bulk T2 calculated by the second examiner was 36.9±4.1 ms. The differences of intraexaminer bulk T2 values was 0.2±1.0 ms, with a coefficient of repeatability of 2.0 ms, an RMS error of 0.7 ms and a COV of 1.9% (Table 1, Fig. 1). The differences of interexaminer bulk T2 values was 1.0±1.4 ms, with a coefficient of repeatability of 2.8 ms, an RMS error of 1.2 ms and a COV of 3.3% (Table 1). The Bland–Altman plots for the intra- and interexaminer repeatability of bulk patellar cartilage T2 values show a random distribution of the data points indicating no underlying relationship between the method of segmentation by each examiner (Fig. 2).

Table 1.

Repeatability of patellar cartilage T2 values

Repeatability measurement Cartilage zone T2 Difference (mean±S.D.) (ms) RMS error (ms) Coefficient of repeatability (ms) Coefficient of variability
Intraexaminer Bulk 0.2±1.0 0.7 2.0 1.9%
Superficial 0.2±1.9 1.3 3.7 3.1%
Middle 0.1±1.0 0.7 1.9 1.8%
Deep 0.3±1.1 0.8 2.1 2.5%
Interexaminer Bulk 1.0±1.4 1.2 2.8 3.3%
Superficial 0.5±3.0 2.1 5.8 5.0%
Middle 0.9±1.5 1.2 2.9 3.2%
Deep 1.0±1.3 1.2 2.5 3.7%
Open in a new tab

Fig. 2.

Fig. 2

Open in a new tab

Bland–Altman plots for bulk T2 value intraexaminer repeatability (left) and interexaminer repeatability (right). The mean difference of each plot is shown as a solid line and the upper and lower limits of agreement are shown as dashed lines. The intraexaminer coefficient of repeatability is 2 ms, and the interexaminer repeatability is 2.8 ms.

The repeatability of zonal T2 values was dependent upon the cartilage zone of interest. The superficial zone of cartilage had the highest variability of all three zones of tissue analyzed. The average T2 values for the superficial, middle and deep zones were 42.2±5.6, 38.1±5.3 and 31.9±4.6 ms, respectively. The inter-examiner Bland–Altman plots for each cartilage zone display larger differences of data between the two examiners for the superficial zone, as compared to differences for the middle and deep zones (Fig. 3). The intraexaminer coefficient of repeatability of the superficial zone was more than 1.5 ms greater than the middle and the deep zones (Table 1). Similarly, the interexaminer coefficient of repeatability of the superficial zone was almost 3 ms greater than the middle and deep zones (Table 1). The superficial zone also had the largest COV for intra- and interexaminer repeatability, 3.1% and 5.0%, respectively. In general, the middle zone had the best repeatability and COV than the other two zones. Overall, the bulk T2 evaluation tended to result in better repeatability, RMS and COV than the zonal T2 evaluation.

Fig. 3.

Fig. 3

Open in a new tab

Bland–Altman plots of depth dependent T2 value interexaminer repeatability for the superficial (left), middle (center) and deep (right) zones of cartilage. The mean difference of each plot is shown as a solid line and the upper and lower limits of agreement are shown as dashed lines. The middle and deep zones had greater measures of repeatability than the superficial zone.

There was no significant difference between the number of pixels segmented by each examiner for all analyzed image slices (P=.4). On average, the first examiner defined 534.6±241.8 pixels, 86.0±7.6% of which were common to the pixels defined by the second examiner. The second examiner defined 510.5±227.3 pixels, 89.2±7.1% of which were common to the pixels defined by the first examiner. Visual inspection of the ROIs did not indicate anatomical areas where pixels were commonly over- or undersampled by an individual examiner.

4. Discussion

This study evaluated the intra- and interexaminer resegmentation variability and repeatability of patellar cartilage T2 values. Understanding the resegmentation repeatability of T2 values is important since changes of T2 values over time may be influenced by examiners defining different ROIs within image data sets in addition to the natural progression of pathology.

The results of the current study aid in interpreting significant differences and changes of T2 in experimental protocols. The current analysis found the range of interexaminer coefficient of repeatability to be between 2.0 and 5.8 ms for bulk- and depth-dependent comparisons. This coefficient of repeatability is similar to previous reports of T2 changes due to joint loading [19,20], producing changes from <1 to approximately 4 ms, or differences due to gender [21], producing differences of approximately 2 ms. Furthermore, the coefficient of repeatability was slightly less than the change of cartilage T2 through the depth of the tissue as reported in previous studies [9,22,23]. Therefore, based on the current interexaminer coefficient of repeatability, if a study finds a significant difference between two sample groups and the mean difference is at most 6 ms, then the results may depend upon the individual who processed the data. The data set should be reprocessed by a second individual to verify the significant difference which was initially found. If a study finds a significant difference between two sample groups and the mean difference is greater than 6 ms, then the differences are likely due to genuine changes of the biological composition of the tissue rather than the individual processing the data set.

The current analysis found better repeatability in the middle and deep zones of cartilage than in the superficial zone of the tissue. We attribute this result to our method of manual image segmentation. Our method used the first echo image to delineate the articular cartilage from the underlying subchondral bone surface. Defining the subchondral bone surface was facilitated since the low T2 value of bone produced an obvious dark boundary with the articular cartilage. The last echo image was used to differentiate the articular surface from the surrounding joint space. However, defining the articular surface proved to be difficult in several slices due to focal degeneration of the articular surface. The degeneration produced an indistinct boundary between the articular surface and surrounding joint space. As a result, the ease of defining the deep zone combined with the difficulty of defining the articular surface and the semiautomated program for defining the middle zone produced the variability of the depth dependent T2 values seen in the study.

Using manual segmentation, the intra- and interexaminer errors (COV and RMS) tended to be less in magnitude than a previous study which reported intra- and intersession analysis of bulk, zonal and regional T2 values [4]. In evaluating session dependent T2 values, Glaser et al. [4] acquired a series of 3D T1-weighted fast low-angle-shot water excitation images and a series of T2 mapping images. The water excitation images were registered to the T2 mapping images to optimize the overlap of anatomic structures. The transformed water excitation images were then used to define cartilage ROIs for T2 calculations. This potentially time-intensive methodology resulted in a range of bulk patellar T2 COV values of 2.76% to 5.37%. In addition, the zonal dependent patellar T2 COVs were calculated as 4.68%, 3.89% and 3.92% for the superficial, middle and deep zones of cartilage, respectively. The zonal dependent patellar T2 RMS errors were calculated as 1.95, 1.43 and 1.02 ms for the superficial, middle and deep zones of cartilage, respectively. These values calculated by Glaser et al. tend to be larger than our current observer dependent COV and RMS values. Interpretation of T2 data with knowledge of intersession repeatability and interexaminer repeatability is important for the application of T2 in a clinical setting. Combining our results with the results of Glaser et al. indicates that comparing scans across scanning sessions may have a greater effect on the variability of calculated T2 values than the effect of different examiners defining the cartilage ROIs for analysis.

The T2 COV values compare well to studies which have used other imaging techniques to quantify OA. The COV for the current T2 repeatability analysis ranged between 1.8% and 5.0%. Previous studies which used radiographs to evaluate joint space width measurements in the knee produced COVs ranging from of 0.86% to 12% [24–27]. The COV of repeated cartilage volume measurements using MRI data has been reported to be 1.0–6.7% [28]. The similar magnitudes of intra- and interexaminer T2 repeatability COV with these previous quantitative studies indicates the potential of using T2 mapping to accurately track the progression of OA within diarthrodial joints.

The analysis of pixels common to each examiner indicated that one examiner typically defined slightly larger ROIs during image analysis. The average pixel count difference between the two examiners was relatively low — approximately 24 pixels. The pixel count difference may be attributed to the level of experience of the two examiners. We do not believe the pixel count difference adversely affected the T2 repeatability analysis, since a majority of the pixels (>86%) were common to both examiners.

The current sample of subjects was composed of young healthy subjects asymptomatic for OA. It is likely that the resegmentation error would change if subjects with confirmed OA were evaluated. We would anticipate larger values for the coefficient of repeatability, COV and RMS errors for bulk and depth dependent T2 values if OA were present within the joint [28].

This study provides one measurement of bulk and zonal T2 mapping repeatability. The current results indicate good agreement between measurements by a single observer and between two observers. Additional studies which evaluate T2 mapping repeatability of cartilage during OA or after cartilage repair are needed in order to provide support for T2 mapping in the clinical environment.

Acknowledgments

The authors would like to thank Paul W. Weishaar for his assistance for performing the scanning and David Stanley and Dan Rettman for their assistance in modifying the protocol on the scanner required for the experiment. Funding for this study was provided by National Institutes of Health — NIAMS: F32AR053430 (MFK) and R01AR048768 (KRK, KKA).

References

  • 1.Brooks PM. The burden of musculoskeletal disease–a global perspective. Clin Rheumatol. 2006;25:778–781. doi: 10.1007/s10067-006-0240-3. [DOI] [PubMed] [Google Scholar]
  • 2.Richter W. Cell-based cartilage repair: illusion or solution for osteoarthritis. Curr Opin Rheumatol. 2007;19:451–456. doi: 10.1097/BOR.0b013e3282a95e4c. [DOI] [PubMed] [Google Scholar]
  • 3.Eckstein F, Ateshian G, Burgkart R, Burstein D, Cicuttini F, Dardzinski B, et al. Proposal for a nomenclature for magnetic resonance imaging based measures of articular cartilage in osteoarthritis. Osteoarthritis Cartilage. 2006 doi: 10.1016/j.joca.2006.03.005. [DOI] [PubMed] [Google Scholar]
  • 4.Glaser C, Mendlik T, Dinges J, Weber J, Stahl R, Trumm C, et al. Global and regional reproducibility of T2 relaxation time measurements in human patellar cartilage. Magn Reson Med. 2006;56:527–534. doi: 10.1002/mrm.21005. [DOI] [PubMed] [Google Scholar]
  • 5.David-Vaudey E, Ghosh S, Ries M, Majumdar S. T2 relaxation time measurements in osteoarthritis. Magn Reson Imaging. 2004;22:673–682. doi: 10.1016/j.mri.2004.01.071. [DOI] [PubMed] [Google Scholar]
  • 6.Glaser C, Horng A, Mendlik T, Weckbach S, Hoffmann RT, Wagner S, et al. T2 relaxation time in patellar cartilage–global and regional reproducibility at 1.5 Tesla and 3 Tesla. Rofo. 2007;179:146–152. doi: 10.1055/s-2006-927203. [DOI] [PubMed] [Google Scholar]
  • 7.Mendlik T, Faber SC, Weber J, Hohe J, Rauch E, Reiser M, et al. 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. 2004;39:288–299. doi: 10.1097/01.rli.0000119196.50924.f3. [DOI] [PubMed] [Google Scholar]
  • 8.Roos EM, Dahlberg L. Positive effects of moderate exercise on glycosaminoglycan content in knee cartilage: a four-month, randomized, controlled trial in patients at risk of osteoarthritis. Arthritis Rheum. 2005;52:3507–3514. doi: 10.1002/art.21415. [DOI] [PubMed] [Google Scholar]
  • 9.Koff MF, Amrami KK, Kaufman KR. Clinical evaluation of T2 values of patellar cartilage in patients with osteoarthritis. Osteoarthritis Cartilage. 2007;15:198–204. doi: 10.1016/j.joca.2006.07.007. [DOI] [PubMed] [Google Scholar]
  • 10.Glaser C. New techniques for cartilage imaging: T2 relaxation time and diffusion-weighted MR imaging. Radiol Clin North Am. 2005;43:641–653. vii. doi: 10.1016/j.rcl.2005.02.007. [DOI] [PubMed] [Google Scholar]
  • 11.Li X, Hornak JP. T2 calculations in MRI: linear versus nonlinear methods. J Img Sci Tech. 1994;38:154–157. [Google Scholar]
  • 12.Nieminen MT, Rieppo J, Toyras J, Hakumaki JM, Silvennoinen J, Hyttinen MM, et al. T2 relaxation reveals spatial collagen architecture in articular cartilage: a comparative quantitative MRI and polarized light microscopic study. Magn Reson Med. 2001;46:487–493. doi: 10.1002/mrm.1218. [DOI] [PubMed] [Google Scholar]
  • 13.Majumdar S, Orphanoudakis SC, Gmitro A, O’Donnell M, Gore JC. Errorsin the measurements ofT2 using multiple-echo MRI techniques. II. Effects of static field inhomogeneity. Magn Reson Med. 1986;3:562–574. doi: 10.1002/mrm.1910030410. [DOI] [PubMed] [Google Scholar]
  • 14.Wayne JS, Kraft KA, Shields KJ, Yin C, Owen JR, Disler DG. MR imaging of normal and matrix-depleted cartilage: correlation with biomechanical function and biochemical composition. Radiology. 2003;228:493–499. doi: 10.1148/radiol.2282012012. [DOI] [PubMed] [Google Scholar]
  • 15.Eckstein F, Glaser C. Measuring cartilage morphology with quantitative magnetic resonance imaging. Semin Musculoskelet Radiol. 2004;8:329–353. doi: 10.1055/s-2004-861579. [DOI] [PubMed] [Google Scholar]
  • 16.Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307–310. [PubMed] [Google Scholar]
  • 17.Wyss CA, Koepfli P, Mikolajczyk K, Burger C, von Schulthess GK, Kaufmann PA. Bicycle exercise stress in PET for assessment of coronary flow reserve: repeatability and comparison with adenosine stress. J Nucl Med. 2003;44:146–154. [PubMed] [Google Scholar]
  • 18.Gluer CC, Blake G, Lu Y, Blunt BA, Jergas M, Genant HK. Accurate assessment of precision errors: how to measure the reproducibility of bone densitometry techniques. Osteoporos Int. 1995;5:262–270. doi: 10.1007/BF01774016. [DOI] [PubMed] [Google Scholar]
  • 19.Nag D, Liney GP, Gillespie P, Sherman KP. Quantification of T2 relaxation changes in articular cartilage with in situ mechanical loading of the knee. J Magn Reson Imaging. 2004;19:317–322. doi: 10.1002/jmri.20000. [DOI] [PubMed] [Google Scholar]
  • 20.Liess C, Lusse S, Karger N, Heller M, Gluer CC. Detection of changes in cartilage water content using MRI T2-mapping in vivo. Osteoar-thritis Cartilage. 2002;10:907–913. doi: 10.1053/joca.2002.0847. [DOI] [PubMed] [Google Scholar]
  • 21.Mosher TJ, Collins CM, Smith HE, Moser LE, Sivarajah RT, Dardzinski BJ, et al. Effect of gender on in vivo cartilage magnetic resonance imaging T2 mapping. J Magn Reson Imaging. 2004;19:323–328. doi: 10.1002/jmri.20013. [DOI] [PubMed] [Google Scholar]
  • 22.Xia Y, Moody JB, Burton-Wurster N, Lust G. Quantitative in situ correlation between microscopic MRI and polarized light microscopy studies of articular cartilage. Osteoarthritis Cartilage. 2001;9:393–406. doi: 10.1053/joca.2000.0405. [DOI] [PubMed] [Google Scholar]
  • 23.Dardzinski BJ, Mosher TJ, Li S, Van Slyke MA, Smith MB. Spatial variation in cartilage T2 of the knee. Radiology. 1997;205:546–550. doi: 10.1148/radiology.205.2.9356643. [DOI] [PubMed] [Google Scholar]
  • 24.Grochowski SJ, Amrami KK, Kaufman K. Semi-automated digital image analysis of patellofemoral joint space width from lateral knee radiographs. Skeletal Radiol. 2005;34:644–648. doi: 10.1007/s00256-005-0944-5. [DOI] [PubMed] [Google Scholar]
  • 25.Buckland-Wright C. Protocols for precise radio-anatomical positioning of the tibiofemoral and patellofemoral compartments of the knee. Osteoarthritis Cartilage. 1995;3 Suppl A:71–80. [PubMed] [Google Scholar]
  • 26.Boegard TL, Rudling O, Petersson IF, Jonsson K. Joint space width of the tibiofemoral and of the patellofemoral joint in chronic knee pain with or without radiographic osteoarthritis: a 2-year follow-up. Osteoarthritis Cartilage. 2003;11:370–376. doi: 10.1016/s1063-4584(03)00030-x. [DOI] [PubMed] [Google Scholar]
  • 27.Piperno M, Hellio Le Graverand MP, Conrozier T, Bochu M, Mathieu P, Vignon E. Quantitative evaluation of joint space width in femorotibial osteoarthritis: comparison of three radiographic views. Osteoarthritis Cartilage. 1998;6:252–259. doi: 10.1053/joca.1998.0118. [DOI] [PubMed] [Google Scholar]
  • 28.Eckstein F, Burstein D, Link TM. Quantitative MRI of cartilage and bone: degenerative changes in osteoarthritis. NMR Biomed. 2006;19:822–854. doi: 10.1002/nbm.1063. [DOI] [PubMed] [Google Scholar]

RESOURCES