A Cause of Increased Signal in the Normal Lateral Meniscus on Short-TE MR Images of the Knee

Transcription

1 X/94/ American Roentgen Ray Society harles G. Peterly1 Dennis L. Janzenl,2 Phillip F. J. Tirman3 ornelis F. van Dijke1 Michael, Harry K. Genant1 Received December 7, 1 993: accepted after revision February 3, Presented at the annual meeting of the Amencan Roentgen Ray Society, New Orleans, April Department of Radiology, University of alifornia, San Francisco, 55 Pannassus Ave., San Francisco, A Address correspondence to. G. Peterfy. 2Present address: Department of Radiology, Vancouver General Hospital, 855 W. 12th Ave., Vancouver, B.. V5Z 1M9. 3San Francisco Magnetic Resonance enter, 3333 alifornia St., Ste. 15, San Francisco, A Present address: Oklahoma Diagnostic Imaging, 236 NW. 62nd St., Oklahoma ity, OK Magic-Angle Phenomenon: A ause of Increased Signal in the Normal Lateral Meniscus on Short-TE MR Images of the Knee.. :? :.. OBJETIVE. Increased signal intensity is frequently present in the upsloping, medial segment of the posterior horn of the normal lateral meniscus on routine short- TE MR images of the knee. This attribute can mimic or obscure abnormalities in this portion of the meniscus. In the present study, we examined whether this appearance could be accounted for by the magic-angle phenomenon resulting from the angular orientation of this meniscal segment relative to the static magnetic field. SUBJETS AND METHODS. Fifty-eight consecutive knee MR examinations were studied. Sixteen were excluded because of frank evidence of preexisting abnormality of the lateral meniscus. In the remaining 42 examinations, the signal intensity in the medial segment of the posterior horn of the lateral meniscus on short-te (<2 msec) sequences was correlated with the angular alignment of this meniscal segment relative to the static magnetic field. In five asymptomatic volunteers, meniscal signal intensities were monitored as the leg was positioned in various degrees of abduction. RESULTS. Increased signal was present in the medial segment of the posterior horn of the lateral meniscus in 31 (74%) of the 42 patients. In 25 (81%) of these, this meniscal segment was oriented at 55-6#{176}. Increased signal intensity was also present in this meniscal segment in three (6%) of five asymptomatic knees imaged in the neutral position. In each of these, abduction of the leg decreased the meniscal signal by 52-8%. Pathologic evaluation of two menisci showed numerous concentrically arranged collagen fibers. NLUSION. Increased signal in the upsloping portion of the posterior horn of the lateral meniscus on short-te images often is due to the magic-angle phenomenon rather than to meniscal degeneration or tear. AJR 1994;163: MR imaging has become widely accepted as an accurate method of assessing meniscal disease [1, 2]. Meniscal fibrocartilage normally exhibits a uniformly low signal intensity on all pulse sequences [3] owing to the extremely rapid T2 relaxation resulting from the high intrameniscal content of collagen [4-7]. Increased signal within a meniscus on short-te (<2 msec) images generally is considered to be indicative of a tear or mucoid/eosinophilic degeneration [3, 8-12]. High intrameniscal signal on long-te sequences is a more specific but a considerably less sensitive sign of meniscal tear. Reported nonpathologic causes of increased intrameniscal signal include truncation artifact [13] and misinterpretation of interfaces between the meniscus and adjacent anatomic structures, typically the transverse intermeniscal ligament anteriorly or the meniscofemoral ligament posteriorly [14]. We frequently observed increased signal within the medial segment of the postenor horn of the lateral meniscus (MPLM) on short-te MR images that could not be accounted for by any of these explanations. As meniscal fibrocartilage contains a dense network of longitudinally organized collagen fibers, we postulated that the abnormal meniscal signal was a manifestation of the magic-angle phenomenon previously described in tendons and other collagen-containing tissues [15-18]. The magic-angle phenomenon refers to the increase in signal intensity that occurs when collagen fibers in these tissues are oriented at 55#{176} relative to the static magnetic field (B). At this particular angle, dipole-dipole interactions that contribute to

2 15 PETERFY ET AL. AJR:163, July 1994 T2 relaxation among water protons constrained by the collagen fibers are nulled. This increases the T2 and accordingly the signal intensity on short-te sequences. In the present study, we investigated this hypothesis by examining the relationship between meniscal orientation and signal intensity on short-te MR imaging of the knee. Subjects and Methods Fifty-eight consecutive knee MR examinations performed at our institution between June 23 and July 2, 1993, were studied. Sixteen cases were excluded because of previous lateral meniscectomy (n = 3), discoid lateral meniscus (n = 2), or clear evidence of tear involving the majority ofthe posterior hom of the lateral meniscus (n = 11). The remaining study population consisted of 42 patients (28 men, 14 women) 2-64 years old (mean, 39 years). The indications for MR imaging of the knee included evaluation of possible internal derangement (n = 4) or mass lesions near the knee (n = 2). MR imaging was performed with a 1.5-T clinical whole-body MR scanner with a commercial transmit-receive volume coil for the knee (General Electric, Milwaukee). The MR imaging protocol in all patients consisted of these five sequences: (1 ) sagittal Ti -weighted spin echo, 1/ 13 (TRITE), with a 16-cm field of view, a 256 x 1 92 matrix, a 4-mm slice thickness, a i-mm interslice gap, and two acquisitions; (2) sagittal T2-weighted fast spin echo, 34/9 (TR/ effective TE [TEeff]), with an echo-train length of eight, a 256 x 256 matrix, a 4-mm slice thickness, a i-mm interslice gap, and two acquisitions; (3) coronal Ti -weighted fast spin echo, 7i6/i5 (TR / TEeN), with an echo-train length of four, a 256 x 256 matrix, a 5-mm slice thickness, a i.5-mm interslice gap, and two acquisitions; (4) coronal fat-presaturated T2-weighted fast spin echo, 28/42 (TR/ TEeN), with an echo-train length of eight, a 256 x 192 matrix, a 5- mm slice thickness, a i.5-mm interslice gap, and two acquisitions; (5) axial fat-presaturated T2-weighted fast spin echo, 28/9 (TR/ TEeff), with an echo-train length of eight, a 256 x 1 92 matrix, a 5- mm slice thickness, a 2-mm interslice gap, and two acquisitions. The Ti -weighted coronal images were used to determine the slices of interest and the angular orientation of the meniscal segments. The sagittal slice locations were superimposed on a coronal image through the posterior portion of the lateral meniscus. The most medial sagittal slice of the lateral meniscus was excluded to avoid partial volume averaging of the meniscus with adjacent fatty tissue in the intercondylar notch. The next three slices through the lateral meniscus were designed as segments a, b, and c (from central to lateral). On the coronal image, a line tangent to the segment of meniscus included in each slice was drawn at the computer console, and the angle between this line and the B axis was measured (Fig. ia). B was oriented along the long axis of the bore of the magnet (z-axis) and corresponded to the vertical axis of the sagittal and coronal MR images. Both quantitative and semiquantitative evaluations of meniscal signal were made. Signal intensities were objectively measured by manually placing a region-of-interest cursor in the posterior horn of the lateral meniscus on the sagittal slices corresponding to segments a, b, and c (Fig. i B). are was taken to exclude the cleft sometimes seen between the posterior horn of the lateral meniscus and the origin of a meniscofemoral ligament (ligaments of Humphry or Wrisberg). The meniscal-segment signal-intensity ratio (SIRa b, on c) was defined as Sla b, orc/sin, where Sla, b, onc is the signal intensity of the meniscal segment of interest and SIN is the signal intensity of a normal black segment of the anterior horn of the lateral meniscus. For example, for segment a of the posterior horn of the lateral meniscus, SIRa Sla/SIN. An abnormal increase in the signal intensity of the posterior horn of the lateral meniscus on Ti -weighted sagittal images was also subjectively evaluated by consensus opinion of two observers using a semiquantitative grading system: signal intensities of lateral meniscal segments a, b, and c were graded as normal when homogeneous low signal intensity was present, as mildly increased when the signal intensity was greater than normal but less than that of adjacent cartilage and joint fluid, and as markedly increased when the signal intensity was equal to that of adjacent cartilage and joint fluid. Subjective assessments were made without knowledge of findings on the coronal images and angular orientation of the menisci. In addition to the 42 consecutive MR examinations, the knees of five asymptomatic volunteers were imaged before and after realignment of the knee within the magnet bore. The volunteers included three men and two women years old (mean, 3 years). The knees were initially imaged in routine positioning with the long axis of the leg parallel to B. Ti -weighted coronal and sagittal images were obtained with a 5-in. (i 3-cm) surface coil placed behind the knee. The angular orientation and signal intensities in the posterior horn of the lateral meniscus were measured in the manner described. The leg was then abducted approximately i 52O from B, and imaging and measurements were repeated. In one healthy volunteer, the knee was imaged in six increasingly abducted positions, and SIR and meniscal angle measurements were obtained in each position. Fig. 1.-Measurement of signal intensity and angulation of medial sagment of posterior horn of lateral meniscus on Ti-weighted spin-echo MR images of knee. A, oronal MR image shows location of three sagittal slices (a, b, and c) through upsloping portion of posterior horn of lateral meniscus that were used in this study. Most central slice (x) was discarded to avoid partial volume averaging of tissue in notch. Meniscal angle () was taken as that subtended by z- axis of image (Be) and a line tangent to meniscal segment. B, Sagittal MR image through lateral meniscus, corresponding to silos a (medial segment of posterior horn of lateral meniscus) in A, shows increased intrasubstance signal intensity and loss of distinctness of superior margin of poetenor horn (arrow). Manually posftioned cursor (1) was used to measure meniscal signal intensity in region of interest.

3 AJR:163, July 1994 NORMAL LATERAL MENISUS ON SHORT-TE MR 151 Fig. 2.-A and B, Sagittal Tiweighted spin-echo MR Images of two different knees show magic-angle phenomenon producing markedly increased signal intensity in medial sagment of posterior horn of lateral meniscus in one knee (arrow, A) and mild magic-angie effect in medial sagment of posterior horn of lateral meniscus in another knee (arrow, B). Meniscal signal is heterogeneous in both affected menisci. Two normal lateral menisci obtained from amputated knee spedmens were sectioned longitudinally and stained with hematoxylin and eosin to show the orientation of the collagen fibers. omparisons of the meniscal angle and SIR for the different segments of the lateral meniscus were made with Student s t-test for unpaired data. Values are expressed as means plus or minus standard deviation. Results Increased signal was subjectively observed in at least one of the three image slices through the MPLM in 31 (74%) of 42 patients. Most often, the increased meniscal signal exhibited marked intrasegmental heterogeneity (Fig. 2). In 25 (81 %) of these 31 patients, the high-signal meniscal segment was oriented at 55-6#{176}relative to B (Table 1). The mean angular orientation of the 67 segments with normal meniscal signal intensity was 74.1 ± 6.7#{176}, whereas that of the 26 segments with mildly increased signal intensity was 66.9 ± 6.4#{176} and that of the 33 segments with markedly TABLE 1 : omparison of Subjective Meniscal Signal Intensities and Angular Orientation of Meniscal Segments Relative to the Static Field A Meniscal Angle (#{176}) Segment Normal Signal IncreasedSignal Increased Signal ----i- 72±6.4(13)64.9 ± 4(9)a (2) b 71.6±6.8 (19) 65.9±5.4 (ii)c 61.1 ±2.4 (i2)a c 75.9±6.9 (35) 71.7±7.4 (6) 59 (1) Note-Values are mean angles ± standard deviation; the numbers of meniscal segments are in parentheses. Segment a designates the most medial section through the medial segment of the posterior horn of the lateral meniscus; segments b and c refer to adjacent lateral sections. astatistically significant (based on Student s t-test) compared with meniscal angles for normal signal intensity (p <.1). bstatistically significant (based on Student s t-test) compared with meniscal angles for normal intensity (p <.1). cstatistically significant (based on Student s t-test) compared with meniscal angles for normal signal intensity (p <.5). B increased signal intensity was 59.9 ± 3.5#{176}. Increased signal intensity was seen most frequently in the most medial segment of the MPLM. All 25 meniscal segments oriented at 55-6#{176}had subjectively abnormal signal intensity; 2 (8%) were graded as markedly increased (Fig. 2A) and five (2%) were graded as mildly increased (Fig. 2B). An orientation of 55-6#{176}was observed in 2 (48%) of the most medial segments (segment a) and in five (12%) ofthe adjacent segments (segment b). An orientation within the magic angle range of 55-6#{176}was not found in any of the most lateral MPLM segments (segment c). The objectively measured meniscal SIR increased in meniscal segments oriented at or near 55#{176} relative to B (Fig. 3). In the 25 meniscal segments oriented 55-6#{176}relative to B, the mean SIR was 2.45 ±.98, whereas the 11 meniscal segments oriented at more than 6#{176} relative to B had a mean SIR of 1.39 ±.55 (p <.1) >. 2..) 1.5,) E 1. 5) Meniscal Angle (#{176}) Fig. 3-Graph shows relative signal intensities of 123 segments of medial segment of posterior horn from 41 knees plotted against angular orientation of each segment to static magnetic field. Values are means ± standard errors. Note marked increase in signal-intensity ratio in sagments within 1 O of magic angle of 55)

4 152 PETERFY ET AL. AJR:163, July 1994 Meniscal signal and angular orientation were measured in five asymptomatic volunteers before and after realignment of the knee within the magnet bore. When the leg was positioned in the usual manner, with the long axis of the leg aligned with the long axis of the magnet bore (Figs. 4A and 4), increased signal was present in the MPLM in three (6%) of five subjects and in six (4%) of 15 meniscal segments. After abduction of the knee (Figs. 4B and 4D), increased signal remained in the MPLM in only one (2%) of five subjects and one (7%) of 15 segments. The mean SIR of the six meniscal segments with abnormal signal decreased after realignment from 3.7 ±.78 to 1.3 ±.39 (p <.1). This represented a mean reduction of 66% (range, 52-8%). Realignment of the knee changed the angular orientation of the six high-signal-intensity meniscal segments from 66.3 ± 6.7#{176} to 8.2 ± 4.8#{176}. Meniscal signal intensity and angular orientation were measured in one asymptomatic volunteer with the knee positioned in six progressively abducted stations (Fig. 5). When the leg was held in the neutral position, an SIR of up to 3.55 was observed in lateral meniscal segments that were oriented between 5#{176} and 6#{176}. When the leg was abducted, cc > 5). ) /) ci Fig. 4.-A-D, Ti-weighted spinecho MR images of knee in asymptomatic volunteer. oronal (A) and saglttal () images of conventionally p51- tioned knee show markedly increased signal intensity in upsioping medial segment of posterior horn (arrows). oronal (B) and sagittal (D) images of same knee using parameters identical to those in A and, but with leg abducted, show uniform low signal intensity in medial segment of posterior horn and sharply defined meniscal margins with no evidence of tear or other abnormality..5 1 I I I I Meniscal Angle (#{176}) Fig. 5.-Graph shows changes in relative signal intensity within medial segment of posterior horn of knee in asymptomatic volunteer as angulation of medial segment relative to static magnetic field was increased by progressively abducting volunteer s leg.

5 AJR:163, July 1994 NORMAL LATERAL MENISUS ON SHORT-TE MR 153 the angulation of the MPLM increased to as high as 9#{176}, and all portions of the lateral meniscus exhibited uniformly low signal intensity (SIR < 1.5). As adduction of the leg lowered the meniscal angle below 55#{176}, the SIR again began to decrease. However, because of the constraints of the magnet bore, the leg could not be adducted past a meniscal angle of 4#{176}. Gross and histologic examination of the lateral meniscus from two amputated knee specimens showed the collagen fibers in the periphery of the meniscus to be arranged in a predominantly parallel, circumferential orientation (Fig. 6). Discussion This study shows that meniscal fibrocartilage, like many other collagen-containing tissues, including tendons, ligaments, and articular cartilage, exhibits the magic-angle phenomenon on routine MR images. This phenomenon manifests as increased signal intensity in meniscal segments oriented near 55#{176} relative to B and can potentially mimic or obscure meniscal abnormality on short-te MR images. The magic-angle phenomenon is a manifestation of the anisotropic behavior of collagen on MR imaging [15-18]. Under normal circumstances, collagen, by virtue of its highly organized structure, tends to restrict the mobility of local water protons and promote dipole-dipole interactions between them. These dipolar interactions contribute to T2 relaxation and are responsible for the uniform low signal intensity exhibited by normal tendons on even short-te sequences. The internuclear vector for this dipolar interaction is oriented along the same axis as the collagen fiber and reduces to zero when 3(cos2 ) - 1 =, where is the angle between the internuclear vector and B. This condition is satisfied when equals 54.7#{176}. Therefore, when tendons are oriented at approximately 55#{176} to B (B is aligned along the z-axis [i.e., the long axis of the magnet bore] in superconducting systems), T2 relaxation time increases sufficiently (-1-fold) to allow signal to emerge on short-te images [15, 16]. This phenomenon can mimic tendinitis and rupture, and it is especially common among the ankle tendons as they curve around the malleoli to enter the foot [19]. The magicangle effect might also be responsible for some cases of increased signal intensity in the critical zone of the supra- Fig. 6.-A, Gross pathologic specimen of disarticulated knee shows lateral meniscus (LM) cleaved horizontally and its upper half(u) peeled back to isveal concentric arrangement of eellagen fibers (arrows) in peripheral half of meniscus. f = free edge. B, Photomicrograph shows light microscopic appearance of lateral meniscus: a dense, parallel array of collagen fibers running concentrically in periphery of meniscus. Numerous split lines (solid arrows) between collagen fibers in peripheral half of meniscus illustrate preferred path for meniscal tears. Toward free edge of meniscus (open arrow), arrangement of collagen becomes less concentric. (H and E, original magnification xioo) spinatus tendon of the normal shoulder on short-te images [15, 16]. This phenomenon is not, however, exclusive to tendon imaging, and has been seen in other tissues that contam parallel arrangements of collagen, including the extensor retinaculum of the wrist and hyaline articular cartilage in the human ankle and bovine knee [15, 18]. Human meniscal fibrocartilage also contains abundant collagen (75% by weight) [4]. Ninety to ninety-eight percent of this collagen is type I, which is the form typically found in tendons. Articular cartilage, which also shows magic-angle effects [15, 18], contains primarily type II collagen fibrils [2]. ollagen fibers within the meniscus are spatially arranged to withstand the tensile stresses generated during weight bearing [5]. Fiber orientation, as shown by histologic examination, polarized light microscopy, and X-ray diffraction studies, is predominantly circumferential, particularly in the peripheral half of the meniscus [4-6]. This arrangement accounts for the propensity of tears to run longitudinally, as they follow the split lines between the collagen fibers rather than traversing fibers (Fig. 6). Some radial fibers are present centrally and at the meniscal surfaces as crossties to resist longitudinal splitting [4]; however, the majority of fibers in the meniscus run longitudinally. The light microscopic appearance of the parallel collagen fibers in the peripheral half of the meniscus (Fig. 6) closely resembles that of tendons [5]. This organization of collagen predisposes the meniscus to the magic-angle effect. The posterior horn of the normal lateral meniscus slopes upward as it ascends from the lateral tibial plateau to its insertion on the posterior portion of the tibial eminence, and often achieves an angle of about 55#{176} relative to the long axis of the leg and therefore B. Increased signal in this portion of the lateral meniscus on short-te sequences, therefore, may result from either magic-angle phenomenon, meniscal degeneration or tear, or maceration. As the remainder of the lateral meniscus and the entire medial menisci are oriented about 9#{176} to B, increased intrameniscal signal in these regions on short-te sequences cannot be accounted for by the magic-angle phenomenon and must be attributed to an intrinsic meniscal abnormality. The susceptibility of meniscal tissue to the magic-angle effect was demonstrated directly in this study by showing the

6 154 PETERFY ET AL. AJR:163, July 1994 emergence of signal in normal, otherwise signalless menisci when they were deliberately angled at 55#{176}. The high prevalence (74%) of increased signal intensity in the upsloping portion of the lateral meniscus in the clinical cases surveyed furthermore suggested that the magic-angle phenomenon was probably the most common cause of abnormal signal in this particular meniscal segment on routine MR images of the knee. Despite our efforts to exclude other causes, such as frank tear, it is possible that some of the menisci exammed in this study harbored small tears or degeneration within this segment. This might in part have accounted for the 19% of menisci with abnormally high signal in the MPLM that were angulated outside the range in which the magicangle phenomenon is known to occur, although this was not verified directly. Heterogeneity of the angle-dependent signal changes in the meniscus (Fig. 2) contributed to the nonspecific appearance of this process. We speculate that this heterogeneity was due to nonuniformities in the alignment of circumferential collagen fibers in the meniscus (Fig. 6). In tendons or hgaments, magic-angle effects can be differentiated from inflammation and rupture by their appearances on long-te images: magic-angle effects become less conspicuous as TE is lengthened beyond about 4 msec, whereas tendinitis and rupture become increasingly salient under these conditions. Prolonging TE is not a practical option in the menisci, however, because meniscal tears are frequently not visible on long-te images [3]. MR evaluation of the menisci, therefore, must rely on the short-te images to exclude meniscal abnormality. This poses a significant potential problem in the MPLM. Although the present study did not attempt to determine the actual impact of this caveat on the diagnostic accuracy of MR imaging in the knee, we have anecdotally encountered cases in which such tears were overlooked on MR images. One method of eliminating magic-angle effects in the lateral meniscus is to image the knee in slight abduction, which, as shown in this study (Fig. 4), alters the orientation of the MPLM relative to B sufficiently to escape this phenomenon, but does not angle the medial meniscus enough to evoke the magic-angle effect there. With the commercial volume coil used in the present study, the leg could be abducted approximately 15#{176} by shifting the coil off isocenter. Greater abduction probably would require detaching the coil from its base and angling it directly. Reorienting the coil in this fashion should not significantly affect the performance of vertically polarized volume coils, such as the one used in this study, but might result in a small degree of signal loss (theoretically about 6% for 2#{176} angulation, that is, 1 - cos 9) with horizontally polarized volume coils [21]. Abduction of the leg will also alter slightly the cross-sectional appearance of the knee on sagittal and axial images, although it is not anticipated that this would cause any significant problems with respect to image interpretation. In fact, sagittal imaging of the knee in slight abduction should facilitate visualization of the anterior cruciate ligament, which would be more vertically oriented with this positioning. In conclusion, it is important to recognize that the magicangle phenomenon is not a finding isolated to tendons, but can occur in any collagen-containing tissue. The implications of altered T2 relaxation in imaging each of these tissues must therefore be considered. In the lateral meniscus, the magicangle phenomenon occurs in the naturally upsloping portion of the posterior horn and can potentially mimic or conceal tears and maceration on short-te images of the knee. AKNOWLEDGMENTS We thank Sidney Dent for technical assistance and the San Frandisco Magnetic Resonance enter for providing some of the imaging time for this study. REFERENES 1. rues JV, Mink J, Levy TL, Lotysch M, Stollen DW. Meniscal tears of the knee: accuracy of MR imaging. Radiology i987;164: Reicher MA, Hartzman 5, Duckwiler G, Bassett LW, Anderson IJ, Gold RH. Meniscal injuries: detection using MR imaging. Radiology 1986; 159: , Lotysch M, Mink J, rues JV, Schwartz SA. Magnetic resonance imaging in the detection of meniscal injuries. Magn Reson Imaging 1 986;4: Renstrom P, Johnson RJ. Anatomy and biomechanics of the menisci. lin Sports Med i99;9: Aspden RM, Yarker YE, Hukins DWL. ollagen orientations in the meniscus of the knee joint. JAnat i985;14: Bullough PG, Munuera L, Murphy J, Weinstein AM. The strength of the menisci of the knee as it relates to their fine structure. J Bone Joint Surg Br i97;52-b: Peto 5, Gillis P. Fiber-to-field angle dependance of proton nuclear magnetic relaxation in collagen. Magn Reson Imaging 1 99;8: Stoller DW, Martin,, rues JV Ill, et al. Meniscal tears: pathologic comelation with MR imaging. Radiology i987;1 63: Hodler J, Haghighi P, Pathnia M, Trudell D, Resnick D. Meniscal changes in the elderly: correlation of MR imaging and histologic findings. Radiology i992;184: Komnick J, Trefelner E, Mcarthy 5, Lange R, Lynch JK, JokI P. Meniscal abnormalities in the asymptomatic population at MR imaging. Radiology i99;177: Dillon EH, Pope F, JokI P, Lynch JK. Follow-up of grade 2 meniscal abnormalities in the stable knee. Radiology i99i 181: Kaplan PA, Nelson NL, Garvin KL, Brown DE. MR of the knee: the significance of high signal in the meniscus that does not clearly extend to the surface. AJR i99i;156: Tumen DA, Rapoport Ml, Erwin WD, McGould M, Silvers RI. Truncation artifact: a potential pitfall in MR imaging of the menisci of the knee. Radiology : Vahey TN, Bennett HT, Amnington LE, Shelboumne KD, Ng J. MR imaging of the knee: pseudotean of the lateral meniscus caused by the meniscofemoral ligament. AJR i99;1 54: Erickson SJ, Prost RW, Timins ME. The magic angle effect: background physics and clinical relevance. Radiology 1 993;188: Erickson SJ, ox IH, Hyde JS, amera GF, Strandt JA, Estkowski LD. Effect of tendon orientation on MR imaging signal intensity: a manifestation of the magic angle phenomenon. Radiology : Fullerton G, ameron I, Ord V. Orientation of tendons in the magnetic field and its effect on T2 relaxation times. Radiology 1 985;1 55: Rubenstein JD, Kim JK, Morava-Protzner I, Stanchev PL, Henkelamn RM. Effects of collagen orientation on MR imaging characteristics of bovine cartilage. Radiology 1 993; 188: Petenty, Linanis R, Steinbach L. Recent advances in magnetic resonance imaging of the musculoskeletal system. Radiol lin North Am 1 994;32: Eyre DR, Wu JJ, Woods P. artilage-specific collagens: structural studes. In: Kuettnen K, Schleyerbach R, Peyron JG, Hascall V, eds. Articular cartilage and osteoarthritis. New York: Raven, i 992: Kneeland JB, Hyde JS. High-resolution MR imaging with local coils. Radiologyi989;171 :1-7