| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Department of Radiology, University Hospitals Leuven, Herestraat 49, 3000 Leuven, Belgium (I.V.B., H.T.C.B., S.D.P., P.P.M.A.B., E.A.G., G.J.M.); Department of Organic Chemistry, Centre Hospitalier Universitaire, Mons, Belgium (L.V.E.); and ESAT-PSI (Centre for the Processing of Speech and Images, Department of Electrotechnics, Faculty of Engineering), Catholic University Leuven, Leuven, Belgium (F.M.). Received June 17, 2003; revision requested August 27; final revision received March 12, 2004; accepted March 24. Address correspondence to I.V.B. (e-mail: iwan.vanbreuseghem@uz.kuleuven.ac.be).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
© RSNA, 2004
Index terms: Cartilage, MR, 452.121419 • Knee, MR, 452.121419 • Magnetic resonance (MR), technology, 452.121411, 452.121413, 452.121419
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
|
Materials and Methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
|
T1 and T2 estimation was then performed with the turbo mixed sequence (I.V.B., H.T.C.B.). The following parameters were used: for the SE acquisition, repetition time (TR) of 1945 msec and TE of 7.4 and 70.0 msec (1945/7.4, 70.0); for the IR acquisition, TR of 1545 msec, TE of 7.4 and 70.0 msec, and inversion time of 400 msec (1545/7.4, 70.0/400); turbo factor of 12; matrix of 304 x 304; acquisition percentage, 75%; field of view of 140 x 140 mm. The measured pixel size was 0.46 mm in the readout direction and 0.61 mm in the phase-encoding direction. Reconstructed pixel size was 0.27 mm in both directions.
Human Volunteers and Imaging
Approval for this study was obtained from our institutional review board, and written informed consent was obtained from all participants. Twenty-five adult volunteers (age range, 23–45 years; mean age, 30.8 years), including 17 men (age range, 25–39 years; mean age, 29.9 years) and eight women (age range, 23–45 years; mean age, 32.5 years), were randomly recruited for the study. Criteria for inclusion were absence of medical history of disease and no current use of any medication. Imaging was performed in both knees unless there was a history of knee trauma or surgery. Criteria for exclusion included presence of swelling, stiffness, restricted mobility either passive or active, and/or pain in or around the knee joint. Since there is evidence that T2 may be increased in asymptomatic senescent patients (2), we categorically excluded volunteers older than 45 years. Of a potential sample of 50 knees, a total of 46 were imaged. Four knees (in four volunteers) were excluded for the following reasons: Partial meniscectomy had been performed in two knees: One knee was affected by a minor but recent sports-related injury, and one knee was omitted because of technical problems.
MR imaging was performed (I.V.B., S.D.P.) by using the turbo mixed sequence with the same parameters in all volunteers. Twenty-four sections were obtained in an oblique sagittal direction with a section thickness of 3 mm and a 0.01-mm gap between sections. With this method, the entire femorotibial joint space was imaged. Two signals were acquired within a total acquisition time of 8 minutes 57 seconds. The choice to generate only two images for calculation of T2 and T1 was based on the results of a few preliminary acquisitions for optimization of bandwidth, echo train length, in-plane resolution, and total acquisition time. T2 maps were displayed by using the available software tools provided by the MR imager manufacturer, which are based on least-squares algorithms.
A standard commercially available quadrature receive-only knee coil (Philips Medical Systems) was used for all sequences. All knees were positioned supine with the femorotibial joint space at the gradient isocenter of the coil. To account for diurnal variations in cartilage relaxation times, all experiments were performed during the early morning hours (7:00–9:00 AM). We asked our volunteers to limit their physical activity prior to the examination, to rule out possible T2-altering effects.
Image Evaluation
The MR imaging data for T2 mapping were transferred to a workstation for further postprocessing with a proprietary application created by using commercially available software (IDL, version 5.5; RSI, Boulder, Colo). A color-coded look-up table corresponding to the calculated T2 relaxivity values was used to color the images. To facilitate interpretation, images were presented after manipulation through a median filter and by using a linear interpolation technique. This method, though artificial, enabled differentiation between the superficial and deep areas of cartilage.
Two-dimensional regions of interest (ROIs) were determined in the deep and superficial areas of cartilage as follows: A total of 276 ROIs with a mean surface area comprising 1700 pixels were drawn on the obtained images (I.V.B.): 50% of the ROIs were in the deep cartilage, and the other 50% were in the superficial cartilage. The ROIs were drawn in the central weight-bearing part of the cartilage that is bordered by the outer meniscal margins. Six ROIs were drawn for each of the 46 knees. The ROIs were equally distributed among different knee compartments, with two ROIs (one in deep cartilage and one in superficial cartilage) per compartment: The medial femoral condyle and the medial tibial plateau were investigated in 34 knees (136 ROIs), and the lateral femoral condyle and the lateral tibial plateau were investigated in 35 knees (140 ROIs). An additional 12 ROIs were drawn in the central posterior portion of either the lateral (n = 6) or the medial femoral condyle (n = 6) in three volunteers, for use as controls. T2 mapping was successfully performed in all volunteers included in the study. To test for intraobserver variability, measurements were repeated in six ROIs in each of three knees (18 ROIs) (I.V.B.).
We graphed the profile of T2 values measured in the central segment of the medial compartment in one knee, in both the femoral and the tibial cartilage plates, to demonstrate T2 as a function of location in cartilage.
T1 maps also were calculated as a result of our application of the turbo mixed technique. No analysis of these images was performed, however, because T1 measurements were tangential to the purpose of our study.
Statistical Analysis
The correlation between T1 and T2 relaxation times determined with spectroscopy and those estimated with turbo mixed imaging in the phantom study was assessed with calculation of Pearson correlation coefficients. Relative differences between determined and estimated values were calculated separately for T1 and T2 by using the equation RD = [ABS(TM – SP)/SP] · 100, where RD is relative difference and ABS(TM – SP) is the absolute value of the difference between the relaxation times at turbo mixed imaging (TM) and at spectroscopy (SP), respectively. T1 and T2 ranges, means, and standard deviations in the phantom experiment also were determined to evaluate the reliability of measurement with spectroscopy.
ROI measurements of T2 were averaged for the superficial and the deep cartilage areas separately, and the range of T2 values in each of the two areas of cartilage was determined. T2 in both superficial and deep cartilage was measured in milliseconds, and overall results were calculated as means ± standard deviations. Statistical analysis of the obtained values was performed with a paired Student t test. A P value of less than .001 was considered to indicate a statistically significant difference.
To test for intraobserver variability, the intraclass correlation coefficient was calculated for absolute agreement between the repeated ROI measurements in three knees. The 95% confidence intervals also were calculated by using statistical software (SPSS for Windows, release 9.0.0; SPSS, Chicago, Ill).
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
|
|
|
|
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
This sequence has the potential to generate cartilage maps based on the use of calculated ratios and a nonlinear least squares algorithm to solve the mathematic equations of both SE and IR measurements for T1 and T2 (8). Calculation of both T1 and T2 from a single measurement not only is very efficient but also reduces problems of misregistration between T1- and T2-weighted images. In a recent article (9), an alternative technique for combined T1 and T2 mapping was presented that is based on gradient-echo acquisitions. The applicability of this technique for T1 and T2 mapping of cartilage, however, remains to be proved.
Despite our use of a very different method, the results of our study are in very close agreement with those of other studies. Results of studies with ex vivo T2 mapping at high field strength showed T2 relaxation time ranges of 51–77 msec (10) and 20–55 msec (11). Extended studies also were done with quantitative in vivo MR imaging for T2 mapping in human subjects. Various investigators measured T2 relaxation times in patellar cartilage at 3.0 T and obtained ranges of 32–67 msec (3), 34–60 msec (12), and 45.3–67 msec (2). In the latter study, T2 relaxation times also were measured in femorotibial cartilage at 3.0 T, and a range of 45.5–55.7 msec was found (2). Similar results were obtained in the pediatric knee with use of a 1.5-T clinical MR imaging system (13).
As has been reported in the literature, T2 values of cartilage are affected by the orientation of collagen (the magic angle effect) and the concentration of collagen (1–4). Concerning collagen concentration, in accordance with the results of previous studies, our results demonstrated increasing T2 values in the transitional zone, with the highest T2 values near the articular surface. In agreement with the findings of Mosher et al (12), the T2 variation in the superficial cartilage (41–65 msec) was greater than that in the deep cartilage (30–45 msec). Concerning the magic angle effect, an extensive literature describes the focal T2 increase in areas in which collagen fibers are oriented at a 55° angle relative to the applied static magnetic field (B0) (14–17). The importance of the magic angle effect lies in the potential for diagnostic error in the presence of early cartilage damage, which also is associated with a focal increase in T2. The magic angle effect has been widely studied and discussed on the basis of ex vivo examinations and, more recently, in vivo tests (18). There is still debate, however, about the effect of collagen fiber orientation on T2 relaxation times in cartilage (19–21). Therefore, the magic angle effect, albeit inevitably present in our study, was deliberately not taken into account.
A major source of error with the turbo mixed imaging technique is that data are acquired at only two points on the T2 (and T1) decay curve, whereas this curve is generally fitted by using at least three data points. The decision to acquire only two data points is based on an assumption about signal intensity at baseline (t = 
). This assumption results in a measurement that does not take baseline noise into account, and it thus creates a systematic error. For future clinical use, however, this shortcoming is not a major drawback, because, at the first examination of a patient, relative measurements (eg, measurements that allow comparison of pathologic areas of cartilage with surrounding normal cartilage, or evaluation of focal cartilage repair over time) will be performed.
Other inherent sources of error also must be considered. First, the image with the longest TE will include a substantial data contribution from stimulated echoes, and this contribution may be greater than that present on the short-TE image. However, because the basic sequence is a fast-SE technique with a turbo factor of 12, some stimulated echoes may influence the short-TE image, as well. Quantification of this effect was beyond the scope of this study. Second, the use of a multiecho, multisection sequence introduces magnetization-transfer contrast into the images. As the two effects are opposite, the resulting T2 measurements may seem more accurate, but the evidence for such a conclusion is purely circumstantial (22). Third, the lengthy period between the echoes (62.6 msec) introduces substantial diffusion weighting into the T2 measurements. Diffusion weighting, however, is unlikely to produce a substantial error in measurements in healthy cartilage, because water mobility there is low (23). The potential source of error in diseased cartilage, in which diffusion is increased, needs to be determined.
These potential sources of error in the turbo mixed imaging technique may account for the differences in the T1 and T2 measurements between turbo mixed imaging and spectroscopy in the phantom experiment. One should note, however, the good correspondence between measurements with turbo mixed imaging and with spectroscopy for T1, as well as for T2, in the expected relaxation ranges. This correspondence is directly related to the imaging sequence parameters (8).
This preliminary report about our turbo mixed technique has several limitations. One drawback is related to the study design: Only healthy adult volunteers were included, and therefore a straightforward extrapolation of our results to patients is not possible. A second drawback is the lengthy imaging session of about 9 minutes. To avoid motion-induced artifacts, stringent immobilization of the knee is required. Alternatively, to reduce the time required for imaging, smaller volumes could be imaged. This alternative may be possible in selected patients. The reduction in signal-to-noise ratio that would result from the reduction in the number of phase-encoding lines for the smaller field of view probably could be counteracted by using small coils that provide excellent performance. Third, as discussed earlier in this article, major errors are introduced with the turbo mixed technique, and these errors affect the estimated T2 (and T1) values. Further research is mandatory to evaluate the feasibility of the technique in clinical applications and to optimize the acquisition parameters.
On the other hand, the turbo mixed sequence has some promising features. First, we performed T2 mapping by using a standard clinical 1.5-T MR imager, whereas prior investigations of T2 mapping primarily were performed with 3.0-T imagers. The agreement between our results and those of previous studies could extend the application of this technique. Second, the reconstructed pixel size in our study was 0.27 x 0.27 mm. In practice, approximately 10 voxels can be distinguished between the femoral and tibial cartilage plates. This spatial resolution would enable the detection of small focal changes in T2 relaxation. Third, the turbo mixed technique may be used for the simultaneous mapping of T1 and T2 values in articular cartilage, which would allow direct comparison of this information. The usefulness of T1 mapping after contrast agent injection remains to be proved. If contrast-enhanced T1 mapping is applicable, the use of the present sequence would lead to a drastic reduction in overall imaging time compared with the time required for high-spatial-resolution imaging for separate T1 and T2 mapping. Currently, the lengthy imaging time for separate mapping of T1 and T2 relaxation times in cartilage is an important drawback for routine application of cartilage mapping. Although T1 maps were available in our volunteer study, no analysis of these images was performed, because unenhanced T1 mapping of articular cartilage gives no additional information regarding biochemical content (5).
In conclusion, we have shown that application of the turbo mixed sequence is feasible for T2 mapping in human cartilage with a clinical MR imaging system.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, I.V.B., G.J.M.; study concepts, I.V.B., H.T.C.B., F.M., L.V.E.; study design, I.V.B.; literature research, I.V.B., S.D.P., H.T.C.B.; experimental studies, I.V.B., H.T.C.B., L.V.E.; data acquisition, I.V.B., S.D.P., E.A.G., L.V.E.; data analysis/interpretation, F.M., I.V.B., P.P.M.A.B., L.V.E.; statistical analysis, F.M., I.V.B.; manuscript preparation and definition of intellectual content, I.V.B., H.T.C.B.; manuscript editing, E.A.G., S.D.P.; manuscript revision/review, all authors; manuscript final version approval, I.V.B., G.J.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and Methods ResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. Saupe, C. W. A. Pfirrmann, M. R. Schmid, T. Schertler, M. Manestar, and D. Weishaupt MR Imaging of Cartilage in Cadaveric Wrists: Comparison between Imaging at 1.5 and 3.0 T and Gross Pathologic Inspection Radiology, April 1, 2007; 243(1): 180 - 187. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
