IS A MAGIC ANGLE EFFECT OBSERVED IN THE COLLATERAL LIGAMENTS OF THE DISTAL INTERPHALANGEAL JOINT OR THE OBLIQUE SESAMOIDEAN LIGAMENTS DURING STANDING MAGNETIC RESONANCE IMAGING?

Authors


  • Abstract presented at the 16th Annual Scientific meeting, ECVS, Dublin, 2007.

Address correspondence and reprint requests to Meredith A. Smith, at the above address. E-mail: meredith.smith@aht.org.uk

Abstract

Collagen fibers oriented at 55° to the static magnetic field (B0) are characterized by an artifactual increase in signal intensity due to the magic angle effect. We hypothesized that there would be increased signal intensity in the collateral ligaments of the distal interphalangeal joint and oblique sesamoidean ligaments when these ligaments were at angles approaching 55° to a horizontal B0 during standing magnetic resonance (MR) imaging. MR imaging was performed on four cadaver forelimbs in a 0.27 T standing system. Transverse and dorsal images were obtained using various sequences, with limbs angled at 0°, 4°, 8°, and 12° to the vertical. Images were analyzed and the angle of each ligament to B0 determined. Mean signal intensity in the ligament and cortex of the adjacent phalanx was measured and ratios calculated. With subjective interpretation, there was increased signal intensity in the collateral ligaments of the distal interphalangeal joint and oblique sesamoidean ligaments over ranges of angles of 60–78° and 57–69°, respectively, to B0. In fast spin echo (FSE) sequences, with a long echo time (72 ms), the effect was less pronounced. FSE sequences can help determine the significance of increased signal intensity within tissues. In spite of limited positions of a limb during standing MR imaging compared with horses under general anesthesia, deviation from a vertical posture sufficient to cause a magic angle effect can still occur in both ligaments tested. Conformation may contribute to the occurrence of the magic angle effect during standing MR imaging. Effort should be made to position horses squarely and to minimize leaning during image acquisition.

Introduction

Tendons and ligaments are composed of dense, highly organized collagen. As a result, movement of hydrogen nuclei is restricted and these tissues have low signal intensity on magnetic resonance (MR) images.1 Increased signal intensity within collagenous tissue is generally considered to be due to pathological change. However, when collagen fibers are oriented at 55° to the static magnetic field (B0), an artifactual increase in signal intensity occurs due to the magic angle effect.2–4 To avoid misinterpretation of the magic angle effect, it is important to understand how this artifact appears with different MR imaging systems. In high-field systems with the static magnetic field parallel to the long axis of the limb, the magic angle effect is frequently observed at the insertion of the deep digital flexor tendon.1 In a low-field system with the static magnetic field perpendicular to the long axis of the limb, there has been little investigation into structures affected by the magic angle effect.

Our objectives were (1) to determine the relative signal intensity in the medial and lateral collateral ligaments of the distal interphalangeal joint and the oblique sesamoidean ligaments in a range of sequences, when the limbs were positioned at different angles to the static magnetic field during low field standing MR imaging; (2) to identify which of the six commonly used sequences were more susceptible to the magic angle effect; and (3) to compare both the medial and lateral collateral ligaments and the medial and lateral oblique sesamoidean ligaments on images obtained with the limb in a vertical axis and with the limb tilted.

Materials and Methods

Cadaver forelimbs (two left fore and two right fore) from four horses with no history of forelimb lameness, humanely destroyed for reasons other than this study, were used. All limbs were sectioned at the level of the proximal aspect of the third metacarpal bone and frozen at −20°C within 8 h of euthanasia. Before scanning, each limb was thawed for 12 h and then underwent MR imaging using a 1.5 T Signa Echospeed short bore magnet to confirm the absence of ligament abnormalities and to obtain reference images of the structures of interest. Standard clinical positioning with the limb parallel to the static magnetic field and a standard sequence protocol were used for this preliminary imaging.5

The four limbs were then imaged in a 0.27 T system designed for standing horses. Each limb was fixed vertically with the metacarpal region proximally, in a custom-designed support stand. The stand was tilted relative to the vertical using wooden wedges positioned under one side. The stand was tilted from the left side for two limbs and from the right side for two limbs. The degree of tilting of the stand relative to the vertical was determined using a spirit level and an inclinometer. The angle of the stand to the vertical was measured before, during, and after scanning to ensure no movement had occurred. Six sequences were obtained in a transverse plane for each limb within the stand tilted at 0°, 8°, and 12° for the foot and at 0°, 4°, and 8° for the pastern (Table 1). These angles were considered to encompass the range of angles clinically achievable in a live horse positioned within the magnet. In addition, T2 fast spin echo (FSE) dorsal images were obtained.

Table 1.   Values for Echo Time (TE), Repetition Time (TR) and Flip Angle for Six Transverse Sequences; T1 Spin Echo (SE), T1 2D Gradient Recalled Echo (GRE), T1 3D GRE, T2* 2D GRE, T2* 3D GRE, and T2 Fast Spin Echo (FSE)
SequenceTE (ms)TR (ms)Flip Angle (°)
  1. Note that the value for the FSE sequence is the effective TE.

T1 SE2050090
T1 2D GRE89775
T1 3D GRE72340
T2* 2D GRE1313030
T2* 3D GRE133422
T2 FSE72250090

One of the four limbs was used to assess a wider range of angles of the collateral ligaments of the distal interphalangeal joint with respect to the static magnetic field to test whether the peak alteration in signal intensity occurred with a collateral ligament at 55° to the static magnetic field. The standing MR imaging system had an internal width of 21 cm; therefore, the extent of tilting possible using the entire distal limb fixed within the stand was limited. To expand the range of angles assessed, the foot was removed from the limb at the level of the metacarpophalangeal joint and the lateral and medial aspects of the hoof were trimmed. This enabled the foot to be angled to up to 35° from the vertical using wooden blocks. For all limbs at each position, the angle of the ligaments to the static magnetic field was determined using the techniques described in Figs. 1A–D.

Figure 1.

 (A) To determine the angle of each collateral ligament of the distal interphalangeal joint to the static magnetic field (dashed line), and the angle relative to the vertical, dorsal plane T2 fast spin echo (FSE) images were acquired for each limb. A line bisecting the phalanges was drawn freehand along the long axis of the limb on a dorsal image (line 1). A second line (line 2) was drawn perpendicular to line 1. Lines parallel with the collateral ligaments of the distal interphalangeal joint were drawn intersecting line 2 and the obtuse angles measured. The angle of each ligament relative to the long axis of the limb was calculated by subtracting 90° from the measured angle. In this image, the limb is positioned vertically with no tilting. The lateral collateral ligament of the distal interphalangeal joint is at 75° to the static magnetic field and 15° (105°−90°) from vertical and the medial collateral ligament is at 76° to the static magnetic field and 14° from vertical (104°−90°). (B) The limb is positioned within the stand with the medial aspect raised by α°. The lateral collateral ligament of the distal interphalangeal joint is now at 75°+α° to the static magnetic field (B0) and 15°−α° from vertical and the medial collateral ligament is at 76°−α° to the static magnetic field and 14°+α° from vertical. (C) To determine the angle of the proximal aspect of each oblique sesamoidean ligament to the static magnetic field, and the angle relative to vertical, dorsal plane T2 FSE images were acquired for each limb. A line was drawn freehand along the long axis of the limb on a dorsal image bisecting the proximal phalanx (line 1). A second line (line 2) was drawn perpendicular to line 1. Lines parallel with the proximal third of the oblique sesamoidean ligaments were drawn intersecting with line 2 and the acute angles measured. The angle of each ligament relative to the long axis of the limb was calculated by subtracting the measured angle from 90°. In this image, the limb is positioned vertically with no tilting. The lateral oblique sesamoidean ligament is at 67° to the static magnetic field and 23° (90°−67°) from the vertical and the medial oblique sesamoidean ligament is at 65° to the static magnetic field and 25° from vertical (90°−65°). (D) The limb is positioned with the lateral aspect raised by α°. The lateral oblique sesamoidean ligament is now at 67°+α° to the static magnetic field (B0) and 23°−α° from vertical and the medial oblique sesamoidean ligament is at 65°−α° to the static magnetic field and 25°+α° from the vertical.

Region of interest analysis was performed for selected MR images using efilm software. Mean signal intensity was determined using ellipses of fixed cross-sectional area (0.1 cm2) drawn onto transverse images of each collateral ligament of the distal interphalangeal joint, or oblique sesamoidean ligament, at the same predetermined slice level for each sequence. To enable comparison among horses, the ratios of signal intensity within each ligament to the signal intensity within a fixed cross-sectional area of the cortex of the middle or proximal phalanx, measured as above, on the same slice were calculated for each sequence.

Fourteen additional limbs from clinically sound horses that were humanely destroyed for reasons other than this study were dissected by one clinician (M.A.S.) following a standardized protocol. The size, orientation, and symmetry of the collateral ligaments of the distal interphalangeal joint and the oblique sesamoidean ligaments were subjectively evaluated.

To determine whether maximal signal intensity of the collateral ligaments of the distal interphalangeal joint occurred at the expected angle of 55°, the difference between signal intensity of the tested collateral ligament and the collateral ligament on the opposite side was plotted against the angle of the tested ligament to the static magnetic field. The angle at which peak difference in signal intensity was observed was identified (Fig. 2).

Figure 2.

 Difference in signal intensity (SI) in the lateral and medial collateral ligaments of the distal interphalangeal joint (DIPCL) as a function of the angle of the ligaments to the static magnetic field. The peak signal intensity occurred when the ligaments were at 55° to B0, due to the magic angle effect.

To compare pulse sequences, the range of signal intensity ratios measured for the corresponding range of angles of the collateral ligament and the oblique sesamoidean ligament to the static magnetic field were determined for each sequence (Table 2)

Table 2a.   Range of the Angle of the Collateral Ligament of the Distal Interphalangeal Joint (DIPCL) to the Static Magnetic Field, the Corresponding Range of Signal Intensity Ratio (Ratio SI) and the Angle of the Collateral Ligament of the Distal Interphalangeal Joint to the Static Magnetic Field at the Maximal Signal Intensity (MSI) for Each Sequence
SequenceRange of
SI Ratio
Angle (°) of
DIPCL at MSI
  1. SE, spin echo; GRE, gradient recalled echo; FSE, fast spin echo.

T1 SE0.6–5.963
T1W 2D GRE2.2–13.672
T1W 3D GRE1.7–7.767
T2W 2D GRE1.7–8.368
T2W 3D GRE1.3–9.568
T2 FSE0.6–1.968
Table 2b.   Range of the Angle of the Oblique Sesamoidean Ligament (OSL) to the Static Magnetic Field, the Corresponding Range of Signal Intensity Ratio (Ratio SI), and the Angle of the Oblique Sesamoidean Ligament to the Static Magnetic Field at the Maximal Signal Intensity (MSI) for Each Sequence
SequenceRange of SI RatioAngle of OSL at MSI (°)
  1. SE, spin echo; GRE, gradient echo; FSE, fast spin echo.

T1 SE1.9–14.259
T1W 2D GRE3.6–27.559
T1W 3D GRE2–20.363
T2W 2D GRE1.2–10.363
T2W 3D GRE1.4–15.667
T2 FSE0.9–3.557

Transverse MR images for each pulse sequence were evaluated for symmetry in collateral and oblique sesamoidean ligament signal intensity and homogeneity, clarity of tissue margins, and relationships with other tissues. An interpreter experienced in MR imaging evaluated 144 images and graded the appearance of the ligaments on each image as either (a) symmetric in signal intensity, (b) increased signal intensity in the ligament on the right side of the image compared with the left, or (c) increased signal intensity in the ligament on the left side compared with the right. Before image interpretation, 30 of 144 images were randomly selected and graded three times in random order to assess repeatability.

Results

When the limb was tilted away from vertical at angles of 0–35°, the angles of the collateral ligaments of the distal interphalangeal joint ranged from 45° to 80° to the static magnetic field for the first limb. There was a peak difference in signal intensity between the two collateral ligaments of the distal interphalangeal joint when the lateral collateral ligament was oriented at 55° to the static magnetic field (Fig. 2).

When the four limbs were positioned vertically, the angles of the collateral ligaments of the distal interphalangeal joint ranged from 11° to 18° (lateral collateral ligament) and 5° to 16° (medial collateral ligament) from the vertical. When the four limbs were tilted away from the vertical axis by 8° and 12°, the angles of the collateral ligaments ranged from 13° to 30° from the vertical. These angles corresponded to a range of 60–85° to the static magnetic field. The angles of the oblique sesamoidean ligaments ranged from 17° to 25° (lateral oblique sesamoidean ligament) and 19° to 23° (medial oblique sesamoidean ligament). As a result of tilting by 4° and 8°, the angles of the ligaments ranged from 21° to 33° from the vertical. These angles corresponded to a range of angles of between 57° and 73° to the static magnetic field. There was least alteration in the signal intensity ratio for the collateral ligaments in T2 FSE sequences and most for T1 2D gradient recalled echo (GRE) (Table 2a). A broad range for the signal intensity ratio was seen in spin echo (SE), T1 2D GRE, T1 3D GRE, T2* 2D GRE, and T2* 3D GRE sequences; however, the signal intensity ratio varied minimally in an FSE sequence (Table 2a).

The range of the signal intensity ratio for the oblique sesamoidean ligament was less in a T2* 2D GRE sequence (Table 2b). Signal intensity varied little on an FSE sequence (Table 2b). FSE sequences were least likely to have increased signal intensity in ligaments at angles approaching 55° to the static magnetic field (Table 2).

On transverse images, there was symmetric signal intensity within the collateral ligaments of the distal interphalangeal joint for all pulse sequences, when the limbs were positioned vertically and these ligaments were within a range of 72–85° to the static magnetic field. The ligaments had homogeneous low signal intensity and the margins were clearly defined with respect to the surrounding tissues. There was symmetric signal intensity within the oblique sesamoidean ligaments when the limbs were positioned vertically and these ligaments were within a range of 65–73° to the static magnetic field. The ligaments had a fascicular pattern of intermediate signal intensity and the margins were clearly defined by tissue boundaries of high signal intensity for these ranges of angles.

When the limb was tilted by 12° from the vertical, one collateral ligament of the distal interphalangeal joint was within a range of 60–73° to the static magnetic field. For T1 SE, T1 2D GRE, T1 3D GRE, T2* 2D GRE, and T2* 3D GRE there was marked asymmetry in signal intensity among the collateral ligaments. The collateral ligament at 60–73° to the static magnetic field had relatively increased signal intensity (Fig. 3), and poorly defined margins from the surrounding soft tissues, which also had moderately high signal intensity. However, this asymmetry was not present in T2 FSE images. The size of the ligament did not change in images from any pulse sequence. The angle of the lateral collateral ligament of the distal interphalangeal joint varied between 13° and 18° from the vertical, and in any given limb was consistently at a greater angle from the vertical than the medial collateral ligament, which varied from 5° to 16° to the vertical.

Figure 3.

 (A) A transverse T2* gradient recalled echo (GRE) magnetic resonance (MR) image obtained with the limb vertical [repetition time (TR) 34 ms, echo time (TE) 13 ms, flip angle 22°]. The lateral collateral ligament (arrow) of the distal interphalangeal joint is at approximately 77° to the static magnetic field. The signal intensity within the lateral and medial collateral ligaments of the distal interphalangeal joint is comparable and the ligaments are symmetric in appearance. (B) Transverse T2* GRE MR image of the same limb as in part A (TR 34 ms, TE 13 ms, flip angle 22°). The limb was tilted so that the lateral collateral ligament of the distal interphalangeal joint (arrow) was at approximately 65° to the static magnetic field. The signal intensity within the lateral collateral ligament is greater than that of the medial collateral ligament and the ligaments are asymmetric in appearance, with loss of definition of the lateral collateral ligament.

When the limb was tilted by 8° from the vertical, the oblique sesamoidean ligament on the lower side of the limb was within a range of 57–65° to the static magnetic field. There was marked asymmetry in signal intensity among the oblique sesamoidean ligaments on T1 SE, T1 2D GRE, T1 3D GRE, T2* 2D GRE, and T2* 3D GRE sequences. The oblique sesamoidean ligament at 57–65° to the static magnetic field had relatively increased signal intensity (Fig. 4), resulting in poorly defined margins from the surrounding soft tissues, which also had moderately high signal intensity. This asymmetry was almost unapparent on T2 FSE images.

Figure 4.

 (A) A transverse T1 spin echo (SE) magnetic resonance (MR) image obtained with the limb vertical [repetition time (TR) 500 ms, echo time (TE) 20 ms]. The medial oblique sesamoidean ligament (arrow) is at approximately 73° to the static magnetic field. The signal intensity within the lateral and medial oblique sesamoidean ligaments is comparable and the ligaments are symmetric in appearance. (B) A transverse T1 SE MR image of the same limb as in part A (TR 500 ms, TE 20 ms). The medial oblique sesamoidean ligament (arrow) is at approximately 65° to the static magnetic field. The signal intensity within the medial oblique sesamoidean ligament is greater than that of the lateral oblique sesamoidean ligament and the ligaments are asymmetric in appearance, with loss of definition of the medial oblique sesamoidean ligament.

Repeatability of subjective image interpretation for a randomized selection of 30 of 144 images interpreted three times was 100%. Upon subjective assessment of transverse images of the lateral and medial collateral ligaments of the distal interphalangeal joint, signal intensity was symmetric in 100% of images of limbs positioned vertically (72–85° to the static magnetic field). Increased signal intensity was identified in the collateral ligament at ranges of 64–78° (8° tilt) and 60–73° (12° tilt) to the static magnetic field, for all sequences in all the four limbs, except the T2 FSE sequence. For all four limbs, images from the T2 FSE sequence were graded as symmetric at all angles.

With subjective interpretation of transverse images of the lateral and medial oblique sesamoidean ligaments, signal intensity was symmetric in 18 of 24 (75%) images of limbs positioned vertically (65–73°) to the static magnetic field. Mild asymmetry among the ligaments was seen in 6 of 24 images. Increased signal intensity was identified in the oblique sesamoidean ligament at ranges of 61–69° (4° tilt) and 57–65° (8° tilt) to the static magnetic field for all sequences in three limbs with the exception of the T2 FSE sequence. In the fourth limb, the ligaments were symmetric in five of six sequences at ranges of 61–69°, and symmetric in three of six sequences at ranges of 57–65° (8° tilt) to the static magnetic field. Increased signal intensity in the ligament at 57–65° was seen in the other sequences except for the T2 FSE. For all four limbs, images from the T2 FSE sequence were graded as symmetrical at all angles.

Upon dissection of 14 additional limbs, the lateral collateral ligament of the distal interphalangeal joint was consistently at a greater angle from vertical compared with the medial collateral ligament. The lateral hoof wall was consistently more sloping than the medial aspect. The oblique sesamoidean ligaments were not all oriented at the same angle, and there was divergence of fibers within the proximal third of the ligament. There was variation in cross-sectional diameter of the oblique sesamoidean ligaments in some dissected limbs, with one ligament being noticeably larger than the other in four limbs, in spite of a normal gross appearance.

Discussion

The magic angle effect in the collateral ligaments of the distal interphalangeal joint has already been described.6 In this work, we report the occurrence of the magic angle effect in oblique sesamoidean ligaments and in ligaments of horses undergoing standing imaging. We found a repeatable increase in signal intensity within the collateral ligaments of the distal interphalangeal joint and the oblique sesamoidean ligaments positioned at a range of angles to the static magnetic field. Upon subjective comparison of images there was increased signal intensity in the collateral ligaments over a range of 60–78° to the static magnetic field and in the oblique sesamoidean ligaments over a range of 57–69° to the static magnetic field in a range of sequences. Consideration of the magic angle effect is therefore essential during the interpretation of images obtained with a standing low-field MR system.

The maximal increase in signal intensity was seen at angles approaching 55° to the static magnetic field. As expected, the effect was less pronounced in sequences with a longer echo time. We also observed increased signal intensity at angles greater than 55° to the static magnetic field, which we cannot fully explain. Factors that may contribute to regions of increased signal intensity include the fit of the radiofrequency coil and the distance of the structure from the arm of the magnet. We observed that when a radiofrequency coil is pressed tightly against one part of the limb, there is increased signal intensity in the tissues adjacent to the coil. This additional artifact may have been contributing to a magic angle effect at some angles in this study.

Collagen-containing tissues are most likely to be affected by the magic angle effect because these structures are highly ordered, which restricts the motion of water molecules bound to the collagen. Reduced motion of molecules shortens T2 relaxation time. Generally, this explains how restricted motion of water molecules causes low signal intensity in tissues, because T2 relaxation does not contribute to the echo. However when the tissue is aligned at the magic angle to B0, T2 decay is slowed, resulting in increased signal intensity.3 The magic angle phenomenon is most commonly seen on short echo time SE or GRE sequences.7 We found that both T1- and T2*-weighted images can be affected at echo times in clinical use for standing MR imaging. However, with a long effective echo time of 72 ms in an FSE sequence, the magic angle effect was not detectable during subjective image interpretation.

It is likely that the increased signal intensity seen in the proximal aspect of the oblique sesamoidean ligaments is partly due to the magic angle effect and partly due to divergence of fibers and a partial volume effect. A lesser-increased signal intensity in the distal portions of the oblique sesamoidean ligaments is likely due to magic angle effect alone. Upon measurement of the angles of the proximal third of the ligament from MR images, this angle may vary from approximately 17° to 25° to the vertical for the lateral oblique sesamoidean ligament and 19° to 23° for the medial oblique sesamoidean ligament. Consideration of the approximate angle of the structure of interest to the static magnetic field should help determine which artifact is present.

The lateral collateral ligament of the distal interphalangeal joint was consistently at a greater angle from the vertical than the medial collateral ligament. This may reflect the conformational variation in the shape and flare of the hoof wall among individual horses. The medial hoof wall was found to be steeper in 80 of 100 horses.8 During clinical standing MR imaging, horses commonly stand in a base wide posture, with ligaments and tendons on the lateral aspect of the distal limb, therefore at greater risk of being affected by the magic angle effect. Having the pastern and hoof be offset laterally with respect to the long axis of the distal limb will affect the position of the ligaments of the metacarpophalangeal joint within the magnet, resulting in increased magic angle effect in the medial oblique sesamoidean ligament. Additional FSE sequences can help prevent false positive diagnoses.

In one high-field study of 199 horses with lameness localized to the foot, 62 (31%) had injuries of a collateral ligament of the distal interphalangeal joint.9 Significant findings included alteration in ligament size, definition and shape, increased signal intensity within the ligament and/or periligamentous tissues, reduced signal intensity in bone reflecting mineralization and/or increased signal intensity within bone at the origin and/or insertion, and osseous cyst-like lesions.10 Increased signal intensity within or surrounding either of these ligaments without other obvious abnormality may reflect the magic angle effect. We found variations in size and orientation of the oblique sesamoidean ligaments in dissected foot specimens. Asymmetry in cross-sectional area of the oblique sesamoidean ligaments has also been seen in 24% of normal horses.11

In a clinical setting, we recommend that every effort be made to stand the horse squarely on all four limbs and to minimize leaning during image acquisition. A careful record of conformation and foot balance should be made before image interpretation. By viewing the images during acquisition, a judgment can be made as to whether asymmetry in signal intensity is due to the magic angle effect or a lesion. A combination of FSE and SE or GRE sequences may be optimal to help rule out magic angle effect seen on SE or GRE sequences, while still ensuring acquisition of images with sufficient soft tissue contrast.12

We have shown that in spite of the relatively limited possible positions of the distal limb during standing MR imaging compared with a horse under general anesthesia, deviation from a vertical posture sufficient to cause a magic angle effect can still occur, and is possible within different ligaments. Caution must be used when interpreting increased signal intensity alone as representative of a lesion in either the proximal portion of an oblique sesamoidean ligament or a collateral ligament of the distal interphalangeal joint. Alteration in ligament size, definition, and shape, as well as increased signal intensity within and around the ligament and abnormal mineralization and/or fluid within bone at the origin and insertion10 must be used to make a diagnosis.

ACKNOWLEDGMENTS

Meredith Smith is funded by The Horse Trust. Financial assistance for this research was provided by Hallmarq Veterinary Imaging Ltd (Guildford, UK).

Ancillary