Abstract
Background:
Knowledge of the normal MRI appearances of the ankle ligaments and tendons is particularly important in the diagnosis of ankle sprains. In most clinical practices, the ankle is imaged in a neutral position with standard imaging planes and sequences. The purpose of our study was to investigate whether passive positioning influences the MRI appearances of the ligaments of the ankle.
Methods:
The axial and coronal T1- weighted MR images obtained from 10 subjects were reviewed by two musculoskeletal radiologists. The following imaging planes were used: dorsiflexion with inversion, dorsiflexion with neutral, dorsiflexion with eversion, neutral with inversion, neutral, neutral with eversion, plantarflexion with inversion, plantarflexion with neutral, and plantarflexion with eversion. A subjective rating system was used to determine the optimal imaging plane and position for individual ligaments in each volunteer. Each ligament was rated on a scale (of 1 to 6).
Results:
There were significant differences in the appearances of the anterior talofibular (p = 0.0002), calcaneofibular (p < 0.0001), and posterior talofibular (p 0.0001) ligaments between the optimal and least optimal ankle positions in the axial plane, and in those of the (plantar calcaneonavicular) spring (p 0.0001), tibiocalcaneal (p 0.0001), posterior tibiotalar (p = 0.0087) and posterior talofibular (p = 0.0213) ligaments in the coronal plane.
Conclusions:
Kinematic MRI of the ankle is feasible and appears to improve visualization of ankle ligaments compared to MRI.
INTRODUCTION
Knowledge of the normal MRI delineation of ankle ligaments and tendons is particularly important in making a diagnosis of an ankle sprain. In most clinical practices, the ankle is imaged in a neutral position with standard imaging planes and sequences. However, most ankle ligaments are difficult to identify on conventional axial and coronal MRI, because the ligaments often course obliquely. Complex oblique MRI planes have been advocated to image the foot in orthogonal planes. 12 The oblique axial images of the normal lateral collateral ligament complex were obtained using T1-weighted spin-echo (SE) imaging. 4 The spectrum of normal and abnormal appearances of the medial collateral ligament complex was illustrated using three-dimensional Fourier transform (3DFT) gradient-recalled echo (GRE) MRI. 6 Orthogonal imaging of the anatomic structures is an important concept in the ankle, because there is an abrupt shift in orientation of most tendons and ligaments from vertical to horizontal at the tibiotalar joint.
Kinematic or cine MRI has been reported for examinations of the knee, 9–10,14–15 the shoulder, 1,3,13,19 and the temporomandibular joints. 2 Some investigators reported the usefulness of kinematic MRI of the ankle for the evaluation of tendons and ligaments. 8,16 MRI with the ankle in full dorsiflexion and full plantarflexion has been shown to affect the location and configuration of the peroneal and posterior tibial tendons (magic angle phenomenon) as well as the appearance of the ankle ligaments. 11,16 –18,20 Recently, the usefulness of an incremental passive positioning device in the detection of peroneal tendon subluxation has been reported. 16 Additionally, Farooki et al. 5 reported the use of a passive positioning device to view the MRI. However, there are few studies examining the appearance of ligaments in the ankle with changes in ankle position.
The purpose of this study was to investigate whether passive positioning influences the MRI appearances of the ligaments in the ankle.
MATERIALS AND METHODS
Subjects
Ten healthy volunteers (five men, five women; ranging in age from 22 to 54 years; mean age 36 years) with no history of previous ankle sprains were imaged with a 0.35 Tesla superconducting magnet (OPART, Toshiba Medical, Tokyo) using a commercial flexible neck coil. Free and unrestricted plantarflexion and dorsiflexion and inversion and eversion were allowed using the neck coil. The lower extremity was placed in a commercially-available positioning device (CHAMCO, Inc., Florida), which was nonferromagnetic and calibrated (Figure 1). Two adjustable handles on the apparatus allowed precise prescription of varying angles of plantarflexion and dorsiflexion and inversion and eversion. The volunteers were imaged in a supine position. Tape, towels, and foam pads were used to immobilize the ankle and the coil. Our institution did not have an ethics committee when the present study was performed. However informed consent was obtained from all volunteers participating in the present study and our research was performed according to the Declaration of Helsinki principles.
The MRI protocol consisted of T1-weighted (TR/TE = 467/15) conventional spin-echo images with a slice thickness of 4-mm and 1.6-mm interslice gap. The field of view was 18 cm, and the data acquisition matrix was 192 × 304. The number of excitations was two. The acquisition time was 3 min and 2 sec. The following imaging planes were used: (1) 10 degrees of dorsiflexion with 20 degrees of ankle inversion, (2) 10 degrees of dorsiflexion with neutral position, (3) 10 degrees of dorsiflexion with 20 degrees of eversion, (4) neutral position with 20 degrees of inversion, (5) neutral position, (6) neutral position with 20 degrees of eversion, (7) 20 degrees of plantarflexion with 20 degrees of inversion, (8) 20 degrees of plantarflexion with neutral position, and (9) 20 degrees of plantarflexion with 20 degrees of eversion. Axial and coronal images in these nine positions were ultimately evaluated.
Positioning device made of nonmetallic material for kinematic MRI of the ankle. The ankle joint motion device enables joint motion in any combination of three planes of rotation: flexion and extension, abduction and adduction, and inversion and eversion, and provides repeatable imaging of the ankle through incremental stepped motion.
Imaging Analysis
MRI scans were retrospectively reviewed for the subjective appearance of the ligaments by two musculoskeletal radiologists (OT, HA) who were aware of the purpose of the study. Ratings were assigned by consensus opinion. The localizer sequence was not available to the readers during review of the images. The images obtained from the 10 volunteers were reviewed randomly by both readers. The appearance of each anatomic structure was evaluated separately for each ankle position obtained. The images from 10 different subjects were not compared with each another. The following major structures in the ankle were evaluated on the axial plane: anterior talofibular, posterior talofibular, calcaneofibular, and anterior tibiofibular ligament. The following structures were evaluated in the coronal plane: posterior talofibular, calcaneofibular, posterior tibiofibular, posterior tibiotalar, tibiocalcaneal, and (plantar calcaneonavicular) spring ligament. A subjective rating system was used to determine the optimal imaging plane and position for individual ligaments in each volunteer. Each ligament was rated on a scale from 1 to 6, with 6 being the highest rating (optimal imaging). The ratings were based on several criteria overall: homogeneity of signal intensity, length of visualized course, configuration, and definition of the ligament from surrounding structures.
Statistical Analysis
As a rating system was used, nonparametric analysis was performed with a Friedman test to determine the significance of differences in the ratings for viewing of a particular ligament. Fisher's protected least significant difference analysis was used to determine the significance of differences between the various imaging positions, when there was a significant difference in the ratings for viewing a particular ligament.
RESULTS
There were significant differences in the appearances of the anterior talofibular (p = 0.0002) (Figure 2), calcaneofibular (p < 0.0001) (Figure 3), and posterior talofibular ligament (p < 0.0001) (Figure 4) ligaments between the optimal and least optimal ankle positions on the axial plane (Table 1), and in those of tibiospring (p < 0.0001), tibiocalcaneal (p < 0.0001), posterior tibiotalar (p = 0.0087) (Figure 5, A and B) and posterior talofibular ligament (p = 0.0213) (Figure 5, C and D) ligaments in the coronal plane (Table 2). There were no significant differences in the appearances of the posterior tibiofibular (p = 0.1243) ligament (Figure 5, C and D) between the optimal and least optimal ankle positions in the coronal plane.
Axial T1-weighted images (TR/TE = 467/15) in a 52-year-old man.
Axial Plane
The anterior talofibular ligament was optimally seen in a neutral dorsiflexion/plantarflexion position with eversion (Figure 2, A) and the least optimal imaging position of the ligament was dorsiflexion with inversion (Figure 2, B). The second best imaging position of the ligament was plantarflexion with neutral position.
The calcaneofibular ligament was optimally seen in plantarflexion with neutral eversion/inversion position (Figure 3, A) and the least optimal imaging position was dorsiflexion with inversion (Figure 3, B). The second best imaging position was plantarflexion with eversion.
The posterior talofibular ligament was optimally seen in dorsiflexion with eversion (Figure 4, A) and the least optimal imaging position was plantarflexion with neutral eversion/inversion position (Figure 4, C). The second best imaging position was neutral position.
Coronal Plane
The talocalcaneal ligament and posterior tibiotalar (Figure 5, A) were optimally seen in neutral eversion/inversion position with dorsiflexion. The least optimal imaging position of the tibiocalcaneal ligament was plantarflexion with inversion, and that of the posterior tibiotalar (Figure 5, B) was plantarflexion with eversion. The second best imaging position of the tibiotalar ligament was dorsiflexion with eversion, and that of the posterior tibiotalar ligament was a neutral position with eversion.
The spring and posterior talofibular (Figure 5, C) ligaments were optimally seen in dorsiflexion with eversion and the least optimal imaging position for both ligaments (posterior talofibular; Figure 5, D) was plantarflexion with inversion. The second best imaging position for both ligaments was dorsiflexion with neutral position.
DISCUSSION
The difficulty in seeing the ligaments and tendons in the ankle on MRI is most likely explained by an inadequate appreciation of the three-dimensional orientation of each ligament, the extent of its bone attachments and its relations to overlying tendons, tendon sheaths, and retinacula. Hence, MRI of the ankle in a simple neutral position should be avoided. Some investigators have advocated using a reformatted oblique MRI to image ankle and foot structures. 6–7 However, considerable experience and detailed knowledge of the normal anatomy of the ankle are required to make a reformatted image.
Kinematic MRI techniques have been developed and applied to assess the functional movement of various joints, including the temporomandibular joint, 2 patellofemoral joint, 14–15 shoulder, 1,3,13,19 and ankle. 8,16 Kinematic MRI using an incremental passive positioning device might be useful to evaluate specific tendinous and ligamentous structures in the ankle 5 and can be performed without detailed knowledge of the normal anatomy of the ankle.
Axial T1-weighted images (TR/TE = 467/15) in a 36-year-old man.
The anterior talofibular ligament courses from the anterior margin of the lateral malleolus to a talar attachment just anterior to its fibular articular surface. This ligament is sometimes composed of two bands: a larger upper band that may reach the origin of the anterior tibiofibular ligament and a smaller lower band that may reach the origin of the calcaneofibular ligament. In the present study, the anterior talofibular ligaments were optimally seen in a neutral position with eversion. The anterior talofibular ligament could be lax and easily seen in eversion of the ankle. Not only could the anterior talofibular ligament be confused with the anterior margin of the lateral malleolus, but also the lateral portion of the anterior talofibular ligament could be confused with the origin of the anterior tibiofibular and calcaneofibular ligament with dorsiflexion of the ankle. In fact, the least optimal imaging position of the anterior talofibular ligament was in dorsiflexion with inversion. In neutral position or with plantarflexion of the ankle, the anterior talofibular ligament could displace away from the anterior margin of the lateral malleolus. Our findings concur with those of Farooki et al. 5 but differ from those of Schneck et al. 17 who found that 10 to 20 degrees of dorsiflexion provided optimal imaging of the anterior talofibular ligament. This discrepancy in findings may relate to subtle differences in positioning within the full range of motion and taping of the foot to the table for positioning. 17
Axial T1-weighted images (TR/TE = 467/15) in a 30-year-old woman.
The calcaneofibular ligament descends vertically and obliquely from the a depression in front of the apex of the lateral malleolus and inserts into a tubercule on the lateral surface of the calcaneus. The calcaneofibular ligament is difficult to identify on conventional axial and coronal MRI, because of the oblique course of the ligament. We found that plantarflexion with neutral eversion/inversion position allowed optimal demonstration of the calcaneofibular ligament in the axial plane. The peroneus longus and brevis muscles and tendons cross superficial to the calcaneofibular ligament. Inversion of the ankle displaces the calcaneofibular ligament away from the peroneus longus and brevis muscles and tendons; however, this position places the calcaneofibular ligament results in contact with the posterolateral cortical border of the calcaneus (Figure 3, B). Conversely, eversion of the ankle displaces the calcaneofibular ligament away from the lateral cortex border of the calcaneus, causing confusion of the calcaneofibular ligament with the peroneus longus and brevis muscle-tendon units could occur (Figure 3, C). Plantarflexion with neutral position of the ankle in neutral eversion/inversion demonstrates full-length imaging of the calcaneofibular ligament, because the calcaneofibular ligament descends vertically and obliquely from anterior to posterior. In fact, our findings were identical to those of other investigators. 5,7,17
Coronal T1-weighted images (TR/TE = 467/15) in a 52-year-old man.
Optimal and least optimal positions of the structures in the ankle on axial plane
ATFL = anterior talofibular ligament; PTFL = posterior talofibular ligament; CF = calcaneofibular ligament; Bold type = optimal position. Italics = worst position.
p < 0.0001;
p = 0.0002.
Optimal and least optimal positions of the structures in the ankle on coronal plane
PTFL = posterior talofibular ligament; TC = tibiocalcaneal ligament; TS = tibiospring ligament; PTT = posterior tibiotalar ligament. Bold letter = optimal position. Italics = worst position,
p < 0.0001;
p = 0.0213;
p = 0.0087.
The posterior talofibular ligament arises from the medial aspect of the distal portion of the fibula close to the origin of the calcaneofibular ligament. It courses horizontally and medially and then inserts into the posterolateral talar tubercle. We found that dorsiflexion with eversion allowed optimal demonstration of the posterior talofibular ligament on both the axial and coronal planes. Imaging with dorsiflexion of the ankle could separate the posterior talofibular ligament from other ligamentous structures, such as the posterior tibiofibular and deep transverse ligaments (Figure 5, C). 7 Our findings of optimal imaging in the axial plane during plantarflexion of the ankle concur with those of Schneck et al. 17 In addition, eversion of the ankle could avoid confusion of the posterior talofibular ligament (Figure 4, B and C) with the cortical border of the posterolateral talar tubercle.
The deltoid ligament is composed of superficial and deep layers. The tibiocalcaneal and spring ligaments are superficial layers of the deltoid ligament. The deltoid ligament is further divided into three sets of fibers, which in the anterior to posterior direction, are the tibionavicular, tibiocalcaneal, and posterior tibiotalar ligaments. 6 The spring ligament is between the tibionavicular and tibiocalcaneal ligament and is the only component of the deltoid ligament that does not have two bony attachments. Superiorly, the ligament attaches to the anterior colliculus of the medial malleolus. Inferiorly, the ligament attaches to the superior portion of the calcaneonavicular ligament, just anterior to the tibiocalcaneal ligament. The tibiocalcaneal ligament attaches the anterior colliculus of the medial malleolus to the sustentaculum tali of the calcaneus. On the coronal plane, we found that a component of dorsiflexion allowed good demonstration of the spring and tibiocalcaneal ligaments, because both ligaments descend obliquely from the anterior to posterior. Our findings concur with those of Schneck et al. 17 In contrast, plantarflexion allowed poor imaging of these ligaments. Confusion of the spring and tibiocalcaneal ligaments with the posterior tibial muscle, the flexor digitorum longus muscle, the tendons of the muscles, and the tendon sheaths also may be attributable to poor imaging of both ligaments during plantarflexion of the ankle. Additionally, plantarflexion of the ankle may cause the spring ligament to overlap the tibiocalcaneal ligament.
The anterior tibiotalar and posterior tibiotalar ligaments are parts of the deep layer of the deltoid ligament. A previous study 17 of the anterior tibiotalar ligament in cadavers found that this ligament could be markedly thin or absent. Hence, the ligament was not evaluated in the present study. The posterior tibiotalar ligament extends from the tip of the medial malleolus to the entire nonarticular medial surface of the body of the talus. It is the thickest and strongest element of the deltoid ligament. We found that a component of dorsiflexion best demonstrated the posterior tibiotalar ligament in the coronal plane, because the posterior tibiotalar ligament descends obliquely from the anterior to posterior. Our findings concur with those of Schneck et al. 17
The present study has several limitations, including the small sample size, the lack of clinical cases, no independent or blinded readers, and ankles not being placed in extreme plantarflexion or extreme dorsiflexion. Despite our small sample size, significant differences were recognized for imaging of the anterior talofibular, calcaneofibular, and posterior talofibular ligaments between the optimal and least optimal ankle positions in the axial plane and in that of the spring, tibiocalcaneal, posterior tibiotalar and posterior talofibular ligaments in the coronal plane.
In conclusion, kinematic MRI of the ankle is feasible and appears to improve imaging of ankle ligaments compared with standard MRI. Position-dependent pathology can be overlooked with conventional imaging. Consequently, kinematic MRI may play a vital role in evaluation of the ankle ligaments.