British Journal of Radiology (2004) 77, 641-647
© 2004 British Institute of Radiology
doi: 10.1259/bjr/31774546
Contrast-enhanced MRI of the menisci of the knee using ultrashort echo time (UTE) pulse sequences: imaging of the red and white zones
P D Gatehouse, PhD
1
T He, PhD
2
B K Puri, PhD
3
R D Thomas, FRCS
2
D Resnick, MD
4 and
G M Bydder, FRCR
4
1 The Cardiac Magnetic Resonance Unit, Royal Brompton Hospital, Sydney Street, London SW3 6NP, 2 The Department of Surgery and the 3 Imaging Sciences Department, MRC Clinical Sciences Centre, Imperial College Faculty of Medicine, Hammersmith Hospital, Du Cane Road, London W12 0HS, UK and the 4 Department of Radiology, University of California, San Diego, USA
Correspondence: Dr G M Bydder, Department of Radiology, UCSD Medical Center, 200 West Arbor Drive, San Diego, CA 92103-8756, USA
 |
Abstract
|
|---|
The objective of this study was to demonstrate the red and white zones of the meniscus of the knee using MRI. Ultrashort echo time (UTE) pulse sequences with an initial TE of 0.08 ms and later echoes at 5.95 ms, 11.08 ms and 17.70 ms were used to image the meniscus of the knee in two normal subjects before and after intravenous administration of gadodiamide. Difference images were formed by subtraction of later echo images from the first. The difference images showed obvious enhancement in an area consistent in location and dimensions with the red zone of the meniscus. Regions of interest placed within this area, central to it (corresponding to the white zone), and peripheral to it (corresponding to perimeniscal tissue) all showed increases in signal intensity after intravenous contrast administration. The greatest change in signal intensity in these regions of interest was seen with the shortest TE and in perimeniscal tissue on the original images. The increase in signal intensity was greatest in the red zone on the difference images. Using UTE pulse sequences and difference images derived from them, it is possible to visualize enhancement selectively in the red zone of the meniscus. Less obvious but significant changes in signal intensity were also present in the white zone.
|
Introduction
|
|---|
There is considerable interest in identifying the vascular (red) and avascular (white) zones of the meniscus of the knee with MRI. This reflects the fact that tears in the vascular portion are more likely to heal than those in the avascular region. As a result, meniscus-preserving surgical techniques may be appropriate treatment for tears in the red zone while debridement is generally more suited for those in the white zone [1–3]. The white zone is situated centrally and the red zone occupies the more peripheral 10–33% of the meniscus with perimeniscal structures and the joint capsule more peripheral still [4, 5]. The location and relative sizes of the zones are shown schematically in Figure 1
.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 1. Diagram of the menisci of the knee seen in cross section. The central avascular white zone is shown as white, and the vascular red zone as dark.
|
|
During the 16 years following its first use for imaging the knee, no difference in appearance was reported between the red and white zones of the meniscus with MR. However, in an investigation by Chan et al [6] published in 1998, increased signal intensity was observed peripherally in the menisci of cadaveric knee specimens using conventional proton density and T2 weighted pulse sequences. This increased signal was thought to arise from the red zone. The authors of this study advanced the concept that the low signal region seen with conventional MRIs, and generally thought to be the whole meniscus, was in fact only the central white zone and that the red zone was more peripheral to this. In a subsequent study, detailed measurements in patients and cadavers established that the low signal area seen with conventional images is the entire meniscus and that the area of higher signal seen in the earlier study was probably perimeniscal tissue and not the red zone [7]. In this later study the red and white zones could not be distinguished with MRI and signal enhancement following contrast administration was not observed in the normal meniscus in cadavers or patients.
When the normal meniscus is examined with pulse sequences having echo times (TEs) of 9–20 ms as used in previous studies, it typically shows very low or absent signal intensity due to its short T2, although non-zero signal may be seen in the posterior horn of the lateral meniscus as a result of the magic angle effect [8]. When the meniscus is imaged with ultrashort TE (UTE) sequences with TEs of 0.08–0.20 ms, i.e. about 100 times shorter than those used in the previously reported studies, moderate or high signal is seen within it [9, 10]. With UTE sequences, MR signal is detected before it decays to the very low level seen with conventional longer TE pulse sequences.
We were interested in the possibility that the presence of detectable signal might allow contrast enhancement to be observed in the red and white zones when UTE sequences were used. In addition, the perimeniscal tissue, associated blood vessels, and joint fluid all have longer T2 signals than menisci. It might also be possible to subtract an image obtained with a later echo using a longer TE from the first UTE echo image and so obtain a difference image in which contrast enhancement was obvious in the short T2 meniscus, but that in tissues or fluids with a longer T2 (such as perimeniscal tissue) was much reduced and therefore not a source of confusion. The third general concept we wished to apply, was to acquire the initial UTE image concurrently with precisely registered later echo images with TEs in the range normally used for diagnostic purposes. This was to ensure that any contrast enhancement seen on the UTE or difference images could be correctly assigned to the meniscus or perimeniscal tissue by direct reference to the longer TE images in which the meniscus was of low signal. We also planned to use a higher intravenous dose of gadolinium chelate than in previous clinical studies and employ a non-ionic agent rather than an ionic one since the former might show greater uptake into the fibrocartilage and fibrous connective tissue of the meniscus. In order to assess the value of these approaches we have performed a pilot study in two normal subjects.
|
Subjects and methods
|
|---|
Institutional review board approval was obtained for this study. Two normal male subjects (aged 21 and 58 years) were studied after informed consent had been obtained. Neither had a history of knee injury or disease and both were presumed to have normal menisci. All studies were performed on a 1.5T Sonata MR system (Siemens, Erlangen, Germany) using an 18.5 cm flexiloop surface receiver coil. The UTE sequence employed a half radiofrequency excitation pulse with radial imaging of k-space from the centre followed by another half-excitation with the polarity of the slice selection gradient reversed and repeated radial imaging. The k-space data from the two half-excitations acquired in this way were added to produce a radial line of k-space. The process was repeated through 360° in 512 steps [9, 10]. The data were then mapped on to a 512 square grid and reconstructed by 2D Fourier transformation.
Using this technique, four echoes of the same type were obtained at 0.08 ms, 5.95 ms, 11.08 ms and 17.70 ms. Three difference images were formed by subtracting each of the later TE images from the first image. 10 to twenty 4 mm multislice images were obtained in the sagittal plane with a 280–340 mm field of view, TR of 500 ms, nominal flip angle (for long T2 components) of 80°, slice gap of 20–100% and scan time of 8.5–17 min. Gadodiamide (Amersham Healthcare, Amersham, UK) 0.3 mmol kg–1 was administered intravenously after the baseline images were obtained. Contrast enhanced images were obtained approximately 5–10 min later. The scans were observed for differences in signal intensity before and after contrast enhancement. Regions of interest were used to obtain signal intensity values and calculate T2*.
|
Results
|
|---|
The pre-enhancement image (e.g. Figure 2a) showed the meniscus as isointense to articular cartilage and perimeniscal tissue (similar results were seen in both subjects). The later echo image at TE=5.95 ms (Figure 2c) showed the meniscus with a generally lower signal with the outer aspect slightly lower again. The pre-enhancement difference image (Figure 2e) showed the meniscus as moderate signal intensity with a slightly higher signal in its outer aspect.
View larger version (73K):
[in this window]
[in a new window]
|
Figure 2. A 21-year-old man with normal menisci: (a) Pre ultrashort echo time (UTE), (b) post UTE, (c) pre TE=5.95 ms echo image, (d) post TE=5.95 ms image, (e) pre difference image (c subtracted from a) and (f) post difference image (d subtracted from b). In (a) the menisci and the outer layer of the articular cartilage are almost isointense. Joint fluid has a slightly lower signal. In (c) the menisci are seen as low signal areas with the outer aspects slightly lower in signal intensity. In (e) (the subtracted image) the menisci are seen with moderate signal intensity with slightly higher signal on their outer aspects. In (b) enhancement is seen but it is not possible to define the extent of the menisci. In (d) the menisci are of low signal except for a small linear area (arrow) with features consistent with a collagen tie. In (f) generally higher signal is seen in the menisci together with obvious enhancement (arrows). This area of enhancement was designated the intermediate region, the lower signal area within the meniscus central to it was designated as the central region and the area just peripheral to it was designated as the peripheral region. Regions of interest were then placed in these regions on (f) and transferred to (b) and (c) etc. for corresponding measurements of signal intensity.
|
|
After contrast administration (Figure 2b
) there was obvious enhancement but the meniscus could not be distinguished from other tissues. The later echo (TE=5.95 ms) image (Figure 2d) showed a low signal within the meniscus except for a small linear higher signal enhancing area (arrow) which was probably a collagen tie.
The contrast enhanced difference image (Figure 2f) showed obvious enhancement (arrows) which by reference to Figure 2d, was within the meniscus. This enhancing region was designated as the intermediate region, the area central to it within the meniscus was designated as the central region, and the area peripheral to the intermediate region was designated as the peripheral region for the purpose of measuring signal intensities. Regions of interest defined from the contrast enhanced difference image (Figure 2f) were placed in the designated regions as well as in the corresponding areas on other images acquired at the same time (e.g. Figures 2b, d). Signal intensity values were measured in these regions in both anterior and posterior locations and used to calculate T2* values by fitting them to a mono-exponential decay function. Four slices were chosen in each subject in regions where partial volume effects were not a major problem.
The T2* values are shown in Table 1. The intermediate region has a slightly shorter mean T2* (6.5 ms) than the central region (7.6 ms). There was little difference in T2* before and after enhancement.
The signal intensity values for the pooled results of both subjects were analysed using the ANOVA program SPSS for Windows, version 10.13 (SPSS Inc., Chicago, IL). The value of the signal intensity for the four different TE values differed significantly, taking into account subject; anterior or posterior (location); central, intermediate or peripheral (site); and pre- or post-contrast status (F=18.9; df=3, 87; p<0.0001). Plots of the mean values of the signal intensity for the pre- and post-enhancement central region (Figure 3), pre- and post-enhancement intermediate region (Figure 4), as well as the pre- and post-enhancement peripheral region (Figure 5) are shown. The greatest change in signal intensity was seen with the shortest TE which showed increases of 46%, 76% and 86% in central, intermediate and peripheral regions, respectively. However, on the difference images the intermediate region showed the greatest change (109%) followed by the peripheral region (67%) and then the central region (46%) (Figure 6).
View larger version (8K):
[in this window]
[in a new window]
|
Figure 3. Plot of mean (±standard deviation) signal intensity against echo time (TE) (a) before and (b) after contrast administration for the central region. The differences in pre- and post-enhancement signal intensity were significant at each TE (p<0.001). The differences in signal intensity are greatest at the shortest TE. There is an increase in signal in (b) following contrast administration (46% at TE=0.08 ms).
|
|
View larger version (8K):
[in this window]
[in a new window]
|
Figure 4. Plot of mean signal intensity against echo time (TE) (a) before and (b) after contrast enhancement for the intermediate region. The differences in signal intensity are significant at each TE (p<0.001) with the greatest at the shortest TE. There is a marked increase in signal in (b) (76% at TE=0.08 ms).
|
|
View larger version (8K):
[in this window]
[in a new window]
|
Figure 5. Plot of mean signal intensity against echo time (TE) (a) before and (b) after contrast enhancement for the peripheral region. The differences in signal intensity were significant at each TE (p<0.001) with the greatest at the shortest TE. There was an obvious increase in signal in (b) (86% at TE=0.08 ms).
|
|
View larger version (12K):
[in this window]
[in a new window]
|
Figure 6. Mean difference in signal intensity (echo time (TE)=0.08 ms minus TE=5.95 ms images) for central, intermediate and peripheral regions pre (lower) and post (upper) contrast administration. The intermediate region shows a slightly higher signal intensity than the other two regions before enhancement (lower). This becomes greater after enhancement (upper). These signal intensity values parallel the appearances shown in Figure 2f.
|
|
The ratio of the width of the intermediate region to the sum of the widths of the central and intermediate regions was measured in both subjects. The mean value was found to be 26.8±7.3%.
|
Discussion
|
|---|
In both subjects an obvious increase in signal intensity was seen in the intermediate region on the difference images. Direct comparison with later echo images in which the meniscus was of low signal intensity showed that this area of enhancement was in the meniscus rather than in perimeniscal tissue. The enhancement had a well defined inner margin and extended to the outer margin of the meniscus as well as to its superior and inferior margins in a distribution consistent with anatomical descriptions of the location of the red zone. In addition, the fraction of the whole meniscus that showed enhancement (26.8±7.3%) corresponded with published data (10–33%) for the fraction of the meniscus occupied by the red zone even without allowance for partial volume effects due to the 4 mm thick slices used in this study. The differences in enhancement were also consistent with the known differences in vascularity between the red and white zones. We therefore concluded that the intermediate region represented the red zone of the meniscus, the central region represented the white zone and the peripheral region represented perimeniscal tissue. We have referred to them as such in the remainder of this paper.
The mean T2* values of the meniscus were short (6.5 to 7.8 ms). The red zone was just detectable without contrast enhancement (e.g. Figure 2c, e). This may have reflected its slightly shorter T2* (Table 1). The T2* values of perimeniscal tissue were significantly longer (12.6 to 14.4 ms). The lack of significant change in T2* after contrast administration suggests that there were no major susceptibility effects due to contrast administration.
The signal intensity in the red and white zones as well as the perimeniscal region showed an increase after intravenous contrast administration. This increase was most marked at short values of TE and less at longer values. The TEs employed in conventional sequences (e.g. 9–20 ms) are in the longer range and would be expected to show much lower signal increase with contrast administration than UTE images. This may help account for the fact that contrast enhancement has not been previously observed in the normal meniscus.
The use of a higher dose of contrast agent was also likely to have increased the level of tissue enhancement in comparison with previous studies. It is also possible that the use of a non-ionic agent rather than an ionic one may have increased transport of the contrast agent into the fibrocartilage of the meniscus and so increased the level of enhancement. This is a phenomenon that has been documented in the intervertebral disc [11] which is another fibrocartilagenous structure, as well as in hyaline articular cartilage [12].
The changes seen on the difference images reflect both the increase in the signal intensity seen on the original images and tissue T2* weighted images. Perimeniscal tissue showed the greatest increase in signal on the original images. In spite of this increase its long T2* resulted in much less increase in signal on difference images (Figure 6). The red zone showed the next greatest change in signal intensity on the UTE images. However, its shorter T2* resulted in a higher signal increase than perimeniscal tissue on the difference images. The white zone showed the least change in signal intensity on the original image, but had a short T2*. The change apparent in this tissue on the difference images was similar to that of perimeniscal tissue (Figure 6).
Of interest was the observation of enhancement in collagen ties. These structures have previously been reported on high resolution images with TE=9 ms in cadavers. Contrast enhancement was not observed [8]. The ties showed enhancement which was mainly in the red zone on the first and later two echoes. This may reflect local vascularity in the tissue around the ties.
Enhancement in the white zone was manifest as an increase in signal intensity rather than an obvious change on the difference images. Because the white zone is avascular, the contrast agent must have diffused into this region from the red zone with a possible or probable contribution from the joint fluid.
The influence of disease is of interest since diseases which increase the T2 of the meniscus may make enhancement more obvious on the longer TE images but less obvious on difference images.
UTE sequences are not generally available at the present time and there are difficulties with their use since they require rapid switching of transmit and receiver coils. In this study the standard knee coil was unsatisfactory in this regard, and so a 18.5 cm diameter flexiloop surface receiver coil was used. While this coil met the switching requirements, the direction of its B1 field when the knee was placed inside it was only about 30–40° away from that of Bo rather than perpendicular to it. As a result, its performance was undoubtedly reduced. In addition, local areas of high signal intensity were seen in the areas of the knee adjacent to this coil. These problems could be avoided by use of a dedicated knee coil which met the switching requirements.
In summary, contrast enhancement was observed in regions anatomically consistent with the red and white zones of the meniscus. This enhancement was also consistent with the known differences in vascularity between the zones, the MR properties of the pulse sequences used, the T2* values of the meniscus, as well as the dosage and the biological behaviour of the contrast agent.
|
Acknowledgments
|
|---|
We wish to acknowledge the help of the Diagnostic Investigations of Spinal Conditions (DISCS) and the Arthritis Research Council.
Received for publication June 18, 2003.
Revision received October 20, 2003.
Accepted for publication December 1, 2003.
|
References
|
|---|
- Messner K, Gao J. The menisci of the knee joint: anatomical and functional characteristics and a rationale for clinical treatment. J Anat 1998;193:161–78.
- Arnoczky SP, Warren RF. Microvasculature of the human meniscus. Am J Sports Med 1982;10:90–5.[Abstract/Free Full Text]
- Newman AP, Daniels AU, Burks RT. Principles and decision making in meniscal surgery. Arthroscopy 1993;9:33–51.[Medline]
- Day B, Mackenzie WG, Shim SS, Leung G. The vascular and nerve supply of the human meniscus. Arthroscopy 1985;1:58–62.[Medline]
- Scapinelli R. Studies on the vasculature of the human knee joint. Acta Anat (Basel) 1968;70:305–31.[Medline]
- Chan PH, Kneeland JB, Gannon FH, Luchetti WT, Herzog RJ. Identification of the vascular and avascular zones of the human meniscus using magnetic resonance imaging: correlation with histology. Arthroscopy 1998;14:820–3.[Medline]
- Hauger O, Frank LR, Boutin RD, et al. Characterization of the "red zone" of knee meniscus: MR imaging and histologic correlation. Radiology 2000;217:193–200.[Abstract/Free Full Text]
- Pertify CG, Janzen DL, Tirman PF, van Dijke CF, Pollack M, Genant HK. "Magic angle" phenomenon: a cause of increased signal in the normal lateral meniscus on short-TE MR images of the knee. AJR Am J Roentgenol 1994;163:149–54.[Abstract/Free Full Text]
- Gold GE, Pauly JM, Macovski A, Herfkens RJ. MR spectroscopic imaging of collagen: tendons and knee menisci. Magn Reson Med 1995;34:674–54.
- Gatehouse PD, Bydder GM. Magnetic resonance imaging of short T2 components in tissues. Clin Radiol 2003;58:1–19.[CrossRef][Medline]
- Ibrahim MA, Haughton VM, Hyde JS. Enhancement of intervertebral discs with gadolinium complexes: comparison of an ionic and a non ionic medium in an animal model. AJNR Am J Neuroradiol 1994;15:1907–10.[Abstract]
- Gray ML, Burstein D, Xia Y. Biochemical (and functional) imaging of articular cartilage. Semin Musculoskelet Radiol 2001;5:329–43.[CrossRef][Medline]
Top
Abstract
Introduction
