Magnetic resonance imaging of the proximal portion of the tarsal region (proximal tarsal region) is not routinely performed in horses. Consequently, descriptions of the MRI characteristics of that region in clinically normal horses are limited,1–4 and detailed interpretations are lacking. The proximal tarsal region is comprised of the tarsocrural and talocalcaneal joints. The tarsocrural joint is a high-motion joint in which the cochlea of the tibia articulates with the corresponding trochlea of the talus. The talocalcaneal joint is a low-motion joint between the plantar aspect of the talus and the dorsal aspect of the calcaneus. A strong interosseous ligament is present in the space between those 2 bones.
The proximal tarsal region has strong collateral ligaments on both the lateral and medial sides, each of which has long (superficial) and short (deep) components.5–7 The long lateral collateral ligament originates from the lateral tibial malleolus caudal to the extensor groove. Although it has small attachments to the distal aspects of the talus and calcaneus, the main portion of the ligament inserts on the fourth metatarsal bone and dorsal margin of the adjacent third metatarsal bone. The short lateral collateral ligament originates from the cranial aspect of the lateral tibial malleolus. It divides into 2 bundles: one inserts on the talus, and the other inserts on the calcaneus.6–8 Two components of the tibiotalar ligament have been described elsewhere.5 The long medial collateral ligament originates from the distal aspect of the medial tibial malleolus. A small branch of that ligament inserts on the medial tuberosity of the talus; however, the main component continues distally to insert on the second and third metatarsal bones. The short medial collateral ligament originates from the cranial aspect of the medial tibial malleolus. It rapidly divides into a short talar branch and a longer branch that inserts on the medial aspect of the sustentaculum tali. An accessory branch deep to the talar branch is also present and connects the tibia to the medial aspect of the talus.4,6
The proximal tarsal region is anatomically complex and challenging to interpret on both high-field and low-field MRI images. Therefore, a description of the normal anatomic features and potential variants of this region would provide a useful reference for clinicians. The purpose of the study reported here was to use high-field and low-field MRI to describe the anatomy of the proximal tarsal region of nonlame horses that had no history of signs of pain in that region.
Materials and Methods
Sample
The study was approved by the Ethical Review Committee of the Animal Health Trust. A convenience sample of 25 tarsi was acquired for high-field and low-field MRI examinations. The tarsi were acquired from horses that had no history of signs of pain in the proximal tarsal region and were euthanized for reasons unrelated to the study. The tarsi were removed from the body at the level of the proximal third of the tibia, and the distal aspect of the limb was removed at the junction between the middle and distal thirds of the metatarsal region to simplify storage and movement of the limbs. Although both tarsi were harvested from each of 25 horses, only 1 tarsus/horse was randomly selected for image acquisition. All tarsi were frozen within 6 hours after euthanasia and stored at −20°C.
MRI protocol
Each tarsus was thawed at room temperature (approx 20°C) for 24 hours before MRI examination. Then, each tarsus was wrapped with a thin sheath of blue modeling compound,a which helped to improve the signal from the superficial structures of the cadaveric limbs.9
High-field MRI images were acquired by use of a 1.5-T cylindrical short-bore magnetb and a human extremity radiofrequency coil. Each tarsus was positioned in the magnet to mimic that of a limb of a live horse in right lateral recumbency. The tarsus was supported by pads to maintain it in a horizontal position and prevent rotation.
Prior to MRI examination, trial scans were performed with the high-field magnet to obtain pilot images and optimize acquisition parameters and scanning planes for the proximal tarsal region. Dorsal images were obtained parallel to the talocalcaneal joint as determined from the transverse pilot image and aligned with the long axis of the tibia as determined from the sagittal pilot image. Sagittal images were obtained parallel to the lateral border of the calcaneus as determined from the transverse pilot image and parallel to the long axis of the tibia as determined from the dorsal pilot image. Transverse images were obtained perpendicular to the long axis of the tibia as determined from the sagittal and dorsal pilot images. The optimized protocol used for high-field MRI included T1W 3-D SPGR, T2*W 3-D GRE, dual-echo PD, dual-echo T2W FSE, and STIR sequences (Appendix 1). Dual-echo sequences were acquired only in the transverse and dorsal planes; all other sequences were acquired in the sagittal, transverse, and dorsal planes.
Low-field MRI images were acquired by use of a 0.27-T open magnetc and a custom-made radio frequency coil for equine limbs, which was shaped to fit the tarsal region. Each tarsus was positioned in the magnet to mimic the position of a standing horse. The protocol was designed so that it could be applied to live horses; therefore, all images were acquired with motion correction, and no 3-D or high-resolution sequences were obtained. The protocol used for low-field MRI included T1W GRE, T2*W GRE, T2W FSE, PD, and STIR sequences (Appendix 2). The T1W GRE and T2*W GRE sequences were acquired in the sagittal, transverse, and dorsal planes. The T2W FSE and STIR sequences were acquired in the transverse and dorsal planes, and the PD sequence was acquired in the dorsal plane only. To ensure that images of the entire proximal tarsal region from the tuber calcanei to the central tarsal bone were acquired, 3 sets of transverse images had to be obtained separately and the coil moved for each level.
Subjective evaluation of MRI images
All images were subjectively evaluated by the same diplomate of the European College of Veterinary Diagnostic Imaging (MB) after discussion and agreement with an associate of the European College of Veterinary Diagnostic Imaging (SJD). This evaluation included the shape, margination, and signal intensity of the long and short collateral ligaments of the tarsocrural joint; thickness and endosteal margination of the subchondral bone of the tarsocrural and talocalcaneal joints; medial and lateral malleoli of the tibia; proximal tubercle of the talus; and talocalcaneal ligament.
Measurements
The collateral ligaments of the tarsocrural joint were measured on high-field T1W 3-D SPGR sequences only; they were not measured on the T2W FSE sequences because of the poor signal-to-noise ratio for most of the superficial structures. Measurements were obtained after completion of an intraobserver repeatability study in which 5 separate measurements of each collateral ligament for 5 tarsi resulted in a variance of < 5%. Cross-referencing was used to standardize measurement sites.
The dorsoplantar length (depth) and lateromedial length (width) of the long collateral ligaments were measured on transverse images at 2 levels. The transverse images were cross-referenced with a dorsal image, and the long collateral ligaments were measured at the level of the distal aspect of the intermediate ridge of the tibia and at the junction between the middle and distal thirds of the talus.
The lateromedial width of the short lateral collateral ligament (tibiotalar bundle) was measured on transverse images at the level of the proximal third of the talus where that ligament was largest. The lateromedial width of the short medial collateral ligament was measured on transverse images at the level of its insertion on the calcaneus.
Gross and histologic evaluation
Two of the 25 evaluated tarsi were randomly selected for gross and histologic evaluation. Each tarsus was sectioned in the transverse or dorsal plane. The sections were oriented in the same plane as the MRI images and were approximately 5 mm thick. Each section was photographed on both sides, and the photographs were stored and used for comparison with MRI images.
Two tarsi in which the short medial collateral ligament had a hyperintense signal were dissected, and the tarsocrural joint and collateral ligaments were examined for gross abnormalities, such as the presence of tears or superficial discoloration. The collateral ligaments for 1 of those 2 tarsi were harvested, fixed in formalin, and stained with H&E stain for histologic examination.
Data analysis
The data generated were purely descriptive. For the MRI images, normal anatomic structures and variations were determined by comparison with gross anatomic sections and data provided in the published literature.5–8 Comparison of the MRI images with the gross anatomic sections was useful for identification of the exact margins of the tarsal collateral ligaments, especially of the calcaneal bundles of the short lateral and medial collateral ligaments.
Results
Subjective evaluation of high-field MRI images
The long lateral collateral ligament of the tarsocrural joint was identified in all scanning planes; however, the sagittal plane was not useful for subjective evaluation. In all tarsi, the long lateral collateral ligament was positioned plantar to and in close contact with the lateral digital extensor tendon. In 13 of the 25 tarsi, the dorsal border of the long lateral collateral ligament had an indentation that was consistent with the shape of the adjacent extensor tendon just distal to the lateral tibial malleolus. The ligament had a consistent shape in all 25 tarsi; it had a crescent shape at its origin and transitioned to an oval shape distal to the lateral tibial malleolus. The margins of the long lateral collateral ligament were well-defined in GRE and PD sequences up to the level of the talocalcaneal-centroquartal (proximal intertarsal) joint, but then the ligament became elongated and difficult to follow over the distal tarsal bones. Distal to the tarsometatarsal joint, it was consistently observed as a thin, elongated structure. The long lateral collateral ligament had homogeneous low signal intensity in the T2W FSE and STIR sequences for all 25 tarsi. It had an increased heterogeneous signal intensity in the T1W 3-D SPGR sequence for 14 of the 25 (56%) tarsi and a heterogeneous signal intensity in the T2*W 3-D GRE and PD sequences for 6 of those 14 tarsi. The increase in signal intensity was located in the axial portion of the ligament at the level of the distal aspect of the lateral tibial malleolus or just distal to it (Figure 1). A poorly defined hyperintense line was observed traversing the ligament body at the level of the distal aspect of the talus in the T1W 3-D SPGR sequences for 11 of the 25 tarsi. This was interpreted as a division between the ligament's bundles because the axial portion of the ligament inserted on the calcaneus just distal to that level. The origin of the long lateral collateral ligament was identified at the distal caudal aspect of the tibia caudal to the extensor groove (Figure 2). Endosteal irregularities were observed in the extensor groove of 10 tarsi, at the origin of the ligament for 6 tarsi, and at the caudal cortex of the tibia just caudal to the origin of the ligament for 17 tarsi. In 8 of the 25 tarsi, there was a short hyperintense linear structure that penetrated the tibial cortex near the origin of the long lateral collateral ligament, which likely represented a blood vessel, in T1W 3-D SPGR and T2*W 3-D GRE sequences.
Representative transverse T1W 3-D SPGR (A) and T2*W 3-D GRE (B) images obtained by use of a high-field (1.5-T) MRI magnet (ie, high-field images), transverse T1W GRE image (C) obtained by use of a low-field (0.27-T) MRI magnet (ie, low-field image) at the level of the tarsocrural joint, and high-field dorsal T1W 3-D SPGR image (D) of the proximal portion of the tarsus (proximal tarsal region) of nonlame adult horses that depict various anatomic structures and the signal intensity for those structures in the nondiseased state. One tarsus from each of 25 horses that had no history of signs of pain in the proximal tarsal region and were euthanized for reasons unrelated to the study underwent high-field and low-field MRI. The tarsi were removed from the body at the level of the proximal third of the tibia, and the distal aspect of the limb was removed at the junction between the middle and distal thirds of the metatarsal region to simplify storage and movement of the limbs. Notice that the axial portion of the long lateral collateral ligament of the tarsocrural joint (8) has intermediate signal intensity in all sequences. The horizontal dashed line in the dorsal image (D) represents the level at which the transverse images were obtained. Medial is to the left in all images. 1 = Tibia. 1a = Medial malleolus of the tibia. 1b = Lateral malleolus of the tibia. 2a = Medial trochlear ridge of the talus. 2b = Lateral trochlear ridge of the talus. 3 = Calcaneus. 4 = Superficial digital flexor tendon. 5 = Lateral digital flexor tendon. 6 = Medial digital flexor tendon. 7 = Lateral digital extensor tendon. 10 = Long medial collateral ligament of the tarsocrural joint. 12 = Long digital extensor tendon. 13 = Tibialis cranialis tendon. 14 = Fibularis (peroneus) tertius tendon. 15 = Fluid in the tarsocrural joint.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Photograph of an anatomic section of the proximal tarsal region of a nonlame adult horse cut in the transverse plane (A) and corresponding high-field transverse T1W 3-D SPGR (B, D, and E), low-field transverse T1W GRE (C), and high-field dorsal T1W 3-D SPGR (F) images obtained from multiple nonlame adult horses that depict normal variation in thickness and signal intensity of the caudolateral cortex of the tibia (open arrows) and variation in the thickness of the tibial cortex adjacent to the extensor groove (solid arrows). The horizontal dashed line in the dorsal image (F) represents the level at which the transverse images were obtained. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
The short lateral collateral ligament of the tarsocrural joint was observed between the lateral digital extensor tendon and the lateral cortex of the talus in MRI images obtained in the transverse and dorsal planes (Figures 3 and 4). The tibiotalar bundle extended from the cranial portion of the distal aspect of the tibia to the midportion of the talus. The calcaneal bundle of the short lateral collateral ligament appeared as a hypointense band between the ligament and extensor tendon and was clearly identified throughout its length in 12 of the 25 tarsi (Figure 5). Only the distal half of the short lateral collateral ligament was observed in 7 tarsi, and the ligament was poorly visualized in another 6 tarsi. In most tarsi, the body of the tibiotalar bundle had a specific signal intensity pattern that was characterized by an acentric hypointense region in the dorsal aspect surrounded by intermediate signal intensity. That pattern was observed in T1W 3-D SPGR and T2*W 3-D GRE sequences for 16 tarsi, in PD sequences for 13 tarsi, and in all sequences for 1 tarsus. The short lateral collateral ligament had generalized low signal intensity in T1W 3-D SPGR and T2*W 3-D GRE sequences for 8 tarsi, PD sequences for 3 tarsi, and T2W FSE and STIR sequences for 24 tarsi.
Representative high-field transverse T1W 3-D SPGR (A) and T2*W 3-D GRE (B) images, low-field transverse T1W GRE image (C), and high-field dorsal T1W 3-D SPGR image (D) of the proximal tarsal region of nonlame adult horses that depict various anatomic structures and the signal intensity for those structures in the nondiseased state. Notice the bundles of the short medial collateral ligament (11) are difficult to distinguish in all sequences and poor signal intensity compromises interpretation of the long medial collateral ligament (10; arrows) in panel C. 2 = Talus. 2c = Proximal tubercle of the talus. 9a = Tibiotalar bundle of the short lateral collateral ligament of the tarsocrural joint. 9b = Tibiocalcaneal bundle of the short lateral collateral ligament of the tarsocrural joint. See Figure 1 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Photograph of an anatomic section of the proximal tarsal region of a nonlame adult horse cut in the transverse plane (A) and corresponding high-field transverse T1W 3-D SPGR (B), low-field transverse T1W GRE (C), and high-field dorsal T1W 3-D SPGR (D) images obtained from multiple nonlame adult horses. The 2 bundles of the short medial collateral ligament of the tarsocrural joint (11) are difficult to distinguish in both the anatomic section and MRI images. Subchondral bone of the lateral side of the talocalcaneal joint has asymmetric thickness (arrows), which is a normal variant, although the position of the slice may exacerbate that variation in thickness. The distal portion of the fibularis (peroneus) tertius tendon (14) divides into 2 portions; the long lateral tendon inserts on the dorsoproximal aspect of the third metatarsal bone, and the medial tendon extends obliquely over the dorsal portion of the tarsal region to insert on the dorsomedial aspect of the central and third tarsal bones. 16 = Talocalcaneal ligament. See Figures 1 and 3 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Representative close-up high-field transverse T1W 3-D SPGR images of the proximal tarsal region of a nonlame adult horse obtained at the same level as the images in Figure 4 (A) and just distal to it (B). There is an acentric area of low signal intensity (arrowhead) in the tibiotalar bundle of the short lateral collateral ligament (arrows) in panel A, and the distal insertion of the tibiocalcaneal bundle of the short lateral collateral ligament (arrowheads) is visible in panel B. Medial is to the left in both images. See Figure 4 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
The long medial collateral ligament of the tarsocrural joint was easily identified on MRI images obtained in the transverse and dorsal planes. The origin of the ligament was observed just dorsal to the medial head of the deep digital flexor tendon; however, the exact point of origin from the caudodistal aspect of the tibia was poorly defined. The ligament appeared flat and elongated at its origin (Figure 3) and had a trapezoid shape at the midportion of its body (Figure 4). It became thin and elongated at the distal aspect of the talus and was difficult to follow over the distal tarsal bones but was well-defined distal to the tarsometatarsal joint. The long medial collateral ligament had homogeneous low signal intensity along its entire length in T2W FSE and STIR sequences. For 2 tarsi, there was a hyperintense line that obliquely crossed the ligament just distal to its origin in T1W 3-D SPGR sequences. At the midtalus level, the separation between the ligament and the tissue axial to it was unclear, especially in T1W 3-D SPGR sequences for 12 of 25 tarsi. Also at the midtalus level, there was intermediate signal intensity observed at the axial or plantar portion of the ligament in T1W 3-D SPGR, T2*W 3-D GRE, and PD sequences for 14 tarsi and in T1W 3-D SPGR and T2*W 3-D GRE sequences for only 8 tarsi.
The short medial collateral ligament of the tarsocrural joint was identified in MRI images obtained in all 3 (sagittal, transverse, and dorsal) scanning planes. The origin of the ligament was best seen in sagittal images (Figure 6), whereas the body and insertion of the ligament were best seen on transverse and dorsal images. The calcaneal bundle was easily identified in MRI images for all 25 tarsi, and a thin tibiotalar bundle was observed adjacent to the proximal portion of the talus axial to the calcaneal bundle for 11 tarsi. The calcaneal bundle had clearly defined margins, especially distally where it was surrounded by joint fluid. The origin of the short medial collateral ligament was hyperintense in T1W 3-D SPGR, T2*W 3-D GRE, and PD sequences for all tarsi and also in T2W FSE and STIR sequences for 8 tarsi. The dorsal portion of the ligament had a hyperintense signal. The ligament body had a hypointense signal in all sequences for 17 tarsi, a hyperintense signal in T1W 3-D SPGR and T2*W 3-D GRE sequences for 9 tarsi, and a hyperintense signal in T2W FSE, PD, and STIR sequences for 2 tarsi. The insertion of the short medial collateral ligament had a hyperintense signal in the T2*W 3-D GRE, PD, T1W 3-D SPGR, STIR, and T2W FSE sequences for 24, 24, 16, 21, and 5 tarsi, respectively (Figures 7 and 8). The ligament had a central hyperintense line rather than a diffuse hyperintense signal in T2W FSE sequences for 16 tarsi, STIR sequences for 2 tarsi, and all sequences for 1 tarsus.
Representative high-field lateral parasagittal T2*W 3-D GRE (A) and sagittal (B) and medial parasagittal (C) T1W 3-D SPGR images of the proximal tarsal region of a nonlame adult horse. In panel A, the talocalcaneal ligament (16; arrow) has heterogeneous signal intensity in all sequences, which was attributed to infiltration of a mixture of joint fluid and fat between the collagen fibers of the ligament. In panel B, notice that the subchondral bone of the calcaneal portion of the talocalcaneal joint (arrows) decreases in thickness from proximal to distal. The origin of the short medial collateral ligament of the tarsocrural joint (11) is evident in the medial parasagittal image (C). Dorsal is to the left in all images. 17 = Long plantar ligament. See Figures 1, 3, and 4 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Representative high-field transverse T1W 3-D SPGR (A) and T2*W 3-D GRE (B) images, low-field transverse T2W FSE image (C), high-field transverse T2W FSE (D) and STIR (E) images, and high-field dorsal T1W 3-D SPGR image (F) of the proximal tarsal region of nonlame adult horses obtained at the level of the distal insertion of the short medial collateral ligament of the tarsocrural joint (11), which has high signal intensity in all sequences. Notice the proximal portion of the talocalcaneal ligament (16) has heterogeneous signal intensity in panels B, C, D, and E and intermediate signal intensity in panel A, probably because of the presence of joint fluid. See Figures 1 through 4 and 6 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Representative high-field dorsal T2*W 3-D GRE images (A, B, and C) of the proximal tarsal region of a nonlame adult horse that were obtained successively in a dorsal to plantar direction and a low-field dorsal T2W FSE image (D) obtained at the same level as the image in panel C. Notice that the joint fluid of the tarsocrural joint (15) surrounds the distal half of the short medial collateral ligament (11) and highlights its margins. In panel B, there is a decrease in the signal intensity in the proximal half of the medial trochlear ridge of the talus (2a) and mild thickening of the proximomedial subchondral bone of the lateral trochlear ridge of the talus (2b). 3a = Sustentaculum tali. See Figures 1, 3, and 4 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Subchondral bone of the trochlear ridges of the talus was evaluated on both dorsal and transverse images, and subchondral bone with a uniform thin appearance was considered normal. The subchondral bone appeared thickened on the dorsal aspect of the medial trochlear ridge for 2 of the 25 tarsi and had a decrease in signal intensity at the dorsal aspect of the distal third of the lateral trochlear ridge for 1 tarsus. In MRI images obtained in the dorsal plane, the subchondral bone at the proximal portion of the medial trochlear ridge had uniform thickness for 3 tarsi, mild endosteal irregularities for 10 tarsi, and diffuse low signal intensity for 15 tarsi. For 2 of those 15 tarsi, the diffuse low signal intensity extended for more than half the height of the trochlea. The subchondral bone at the lateral trochlear ridge had uniform thickness for 6 tarsi, and a small crescent-shaped thickening in the subchondral bone was identified at the proximolateral aspect of the lateral trochlear ridge of the talus for 19 tarsi (Figure 8).
The subchondral bone of the medial cochlea of the tibia was uniform in thickness in all 25 tarsi. The subchondral bone at the lateral aspect of the lateral cochlea of the tibia had mild irregularities for 4 tarsi. Also at the lateral aspect of the lateral cochlea of the tibia, a small hyperintense area adjacent to the joint surface that slightly extended into the subchondral bone was observed for 15 tarsi.
The lateral malleolus of the tibia had a uniform signal intensity and thin cortex for 17 tarsi. The cortex of the lateral malleolus was slightly thickened for 4 tarsi, and another 4 tarsi had diffuse low signal intensity in the dorsal two thirds of the lateral malleolus. The medial malleolus of the tibia had uniform signal intensity for 20 tarsi, whereas there was a small hypointense area at the distal aspect of the medial malleolus for the remaining 5 tarsi (Figure 9).
Representative high-field dorsal T1W 3-D SPGR images of the proximal tarsal region of 3 nonlame adult horses that depict normal variation in the thickness of the subchondral bone on the medial (2a; asterisk) and lateral (2b; white arrow) trochlear ridges of the talus. The sagittal groove (open arrows) and lateral malleolus (1b; plus sign) of the tibia are also depicted. See Figures 1 and 3 for remainder of key.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
The proximal tubercle of the talus had uniform signal intensity for 17 tarsi. The axial cortex of the proximal tubercle was thickened for 3 tarsi (Figure 10), and the entire proximal tubercle had diffuse low signal intensity for 2 tarsi. The entire tubercle of the talus had diffuse low signal intensity for 3 tarsi, one of which also had a separate hypointense fragment identified plantar to the tubercle (Figure 8).
Representative close-up high-field transverse T1W 3-D SPGR images of the medial aspect of the proximal tarsal region of 2 nonlame adult horses that depict normal variation in the signal intensity within the proximal tubercle of the talus. Notice the lateral cortex of the proximal tubercle of the talus is thickened (black arrows) in both panels. For the tarsus of panel B, there is a diffuse decrease in signal intensity and a separate hypointense bony fragment (white arrow) within the lateral cortex of the proximal tubercle of the talus. Medial is to the left in both images.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
The talocalcaneal joint was best evaluated in sagittal images. The subchondral bone of the talus had uniform thickness for 15 tarsi and was slightly thicker proximally, compared with distally, for the other 10 tarsi (Figure 6). The subchondral bone of the calcaneus decreased in thickness from proximal to distal for 21 tarsi and was uniform in thickness from proximal to distal for the remaining 4 tarsi. In all 25 tarsi, the subchondral bone to the spongiosa junction was smooth and regular. The talocalcaneal ligament had heterogeneous signal intensity and a striped appearance with alternating hypointense and hyperintense lines in all sequences. That pattern was observed throughout most of the ligament, although it was less evident at the proximal aspect of the ligament, especially in the T1W 3-D SPGR sequence.
Objective evaluation of high-field MRI images
Objective measurements for the collateral ligaments of the tarsocrural joint were summarized (Table 1). For both long collateral ligaments, the dorsoplantar depth was greater than the lateromedial width at both the proximal and distal aspects.
Descriptive statistics for objective measurements of the lateral and medial collateral ligaments of the tarsocrural joint for 25 tarsi obtained from horses that had no history of signs of pain in the proximal portion of the tarsal region (proximal tarsal region) and were euthanized for reasons unrelated to the study.
| Structure | Dimension | Level | Mean | Median (range) |
|---|---|---|---|---|
| Long lateral collateral ligament | Width (mm) | Proximal | 8.5 | 8.2 (7.1–10.3) |
| Distal | 6.7 | 6.4 (5.5–9.1) | ||
| Depth (mm) | Proximal | 15.5 | 15.4 (12.8–18.0) | |
| Distal | 19.5 | 19.3 (16.2–23.5) | ||
| Short lateral collateral ligament | Width (mm) | — | 8.1 | 7.9 (6.3–12.2)* |
| Long medial collateral ligament | Width (mm) | Proximal | 5.9 | 5.8 (4.5–7.7) |
| Distal | 7.5 | 7.6 (6.2–10.0) | ||
| Depth (mm) | Proximal | 22.0 | 21.6 (18.1–28.6) | |
| Distal | 23.3 | 23.8 (18.4–27.4) | ||
| Short medial collateral ligament | Width (mm) | — | 5.5 | 5.3 (4.3–8.0) |
All measurements were acquired from high-field MRI images obtained in the transverse plane. Only 1 tarsus/horse was evaluated. The long collateral ligaments were measured at the level of the distal aspect of the intermediate ridge of the tibia (proximal level) and at the junction between the middle and distal thirds of the talus (distal level). The short lateral collateral ligament was measured at the level of the proximal third of the talus where the ligament is the largest, and the short medial collateral ligament was measured at the level of its insertion on the calcaneus.
The upper limit of the range represents a ligament that was subjectively considered enlarged, compared with the adjacent long lateral collateral ligament.
— = Not applicable.
The long collateral ligaments were generally wider than the short collateral ligaments within a similar level. However, for 8 of the 25 tarsi, the width of the short collateral ligament was greater than that of the corresponding long collateral ligament. During subjective evaluation of MRI images for those 8 tarsi, the width of the short collateral ligament was determined to be greater than or similar to that of the corresponding long collateral ligament for 1 and 3 tarsi, respectively.
Subjective evaluation of low-field MRI images
For all 25 tarsi, the structures identified on high-field MRI images could also be identified on low-field MRI images; however, the ability to delineate the margins of ligaments in the low-field images was decreased relative to the high-field images, especially in the T2W FSE and STIR sequences. It was difficult and often not possible to differentiate bundles of the short components of the collateral ligaments in all low-field sequences.
The signal intensity distribution within the ligaments was similar to that described for the high-field MRI images. The typical signal intensity distribution for the short lateral collateral ligament identified in high-field images was not clearly identified in low-field images. The short lateral collateral ligament had diffuse heterogeneous intermediate signal intensity in the T1W and T2*W GRE sequences obtained by low-field MRI for 17 tarsi.
The image quality was variable among slices, with the central slices of each sequence usually having better definition than the outer slices. This was attributed to the distance of the slice from the center of the magnetic field, which caused a decrease in signal and increase in noise, and verified by correlating the slice position with the appropriate pilot image.
In transverse images, the entire skin margin could be observed for only 5 of the 25 tarsi. Subtle loss of signal on the dorsolateral (n = 4), dorsomedial (11), or both dorsolateral and dorsomedial (3) skin surfaces at the level of the distal aspect of the tibia was often identified in GRE sequences and was attributed to close contact between the coil and the limb, which precluded collection of signal from those areas. That signal loss compromised assessment of the proximal portion of the long medial collateral ligament for 4 tarsi (Figure 3). The amount of background noise precluded assessment of both the long lateral and long medial collateral ligaments in STIR and T2W FSE sequences for 9 tarsi, and the short components of the collateral ligaments could also not be assessed for 4 of those 9 tarsi.
Gross and histologic evaluation
Transverse sections obtained from 1 tarsus contained an area of discoloration in the axial aspect of the long lateral collateral ligament just distal to the tibial malleolus. No abnormalities were identified during gross evaluation of the 2 tarsi with a hyperintense signal in the short medial collateral ligament. Histologic examination of sections from 1 tarsus revealed that most of the collagen fibers in the collateral ligaments ran parallel to the long axes of those ligaments, although some collagen fibers coursed obliquely or perpendicular to the long axis (Figure 11). The long lateral and medial collateral ligaments had mild disorganization of the collagen fibers surrounding vascular spaces. The short medial collateral ligament had areas of mucinous and moderately cellular material, which separated the collagen fibers and were consistent with multifocal myxomatous-mucinous degeneration.
Representative photomicrographs of a longitudinal sections of the long (A) and short (B) medial collateral ligaments of the tarsocrural joint of a nonlame adult horse. Notice that most of the collagen fibers of the long medial collateral ligament (A) run parallel to its long axis, although some fibers are oriented at an oblique angle to the long axis (star), and that the short medial collateral ligament (B) had areas of mucinous and moderately cellular material (arrows), which separated the collagen fibers and were consistent with multifocal myxomatous-mucinous degeneration. H&E stain; bar = 100 μm.
Citation: American Journal of Veterinary Research 79, 3; 10.2460/ajvr.79.3.299
Discussion
Magnetic resonance imaging of the proximal tarsal region is indicated when radiographic and ultrasonographic results are inconclusive or when abnormalities are detected on conventional images but the presence of additional lesions is suspected.10 The collateral ligaments of the tarsocrural joint are readily visible by ultrasonography, which is the imaging modality of choice when an injury is suspected, but lesions may not be identified by use of that modality.
Osteochondral lesions that affect the distal aspects of the tibia and talus are fairly common in both young and adult horses.11 Other diseases that affect the proximal tarsal region are less common. For example, in 1 study,12 only 17 horses with primary signs of pain localized to a tarsocrural joint were identified during a 13-year period. Pathological lesions associated with the proximal tarsal region include synovitis of the tarsocrural joint, fractures of the tibial malleolus and talus, osteoarthritis of the talocalcaneal13 and (less commonly) tarsocrural joints, subchondral cyst-like lesions in the tibia or talus,14 and desmitis of the collateral ligaments of the tarsocrural joint.15–21 Computed tomography has been used to diagnose lesions associated with signs of pain localized to the proximal tarsal region when results obtained by radiography and scintigraphy were considered inconclusive. Results of multiple studies3,22–25 indicate that MRI can provide additional diagnostic information for horses with signs of pain localized to the distal portion of the tarsal region (distal tarsal region), and we believe that MRI might be likewise beneficial for the diagnosis of lesions in the proximal tarsal region.
Detailed knowledge of the normal anatomy and variation in signal intensity in healthy horses is crucial for correct interpretation of MRI scans obtained from lame horses. Generally, a collateral ligament should have uniform low signal intensity in all MRI sequences. In the present study, the collateral ligaments of the tarsocrural joint had several areas with hyperintense signal intensity in GRE sequences that followed specific distribution patterns, which were fairly consistent among the tarsi evaluated. That variation in signal intensity was attributed to magic angle effect,26 which affected only the collagen fibers positioned at a 55° angle to the magnetic field. It was curious that the signal intensity distribution within the collateral ligaments was similar between high-field and low-field images because the orientation of the collagen fibers with the magnetic field differed between the 2 techniques. However, the natural curvature of the ligaments likely resulted in fibers that were oriented at the magic angle for both the horizontally (high-field) and vertically (low-field) oriented magnets in a manner similar to that described for the collateral ligaments of the distal interphalangeal joint.26 Spiral fibers within the collateral ligaments of the tarsocrural joint have been described,5 and gross examination of some of the tarsi of the present study confirmed that collagen fibers within the collateral ligaments of the tarsocrural joint were oriented at various angles to the long axis.
Magic angle effect is less prominent in sequences with a long echo time, such as T2W FSE sequences.27 Results of 1 study27 suggest that an echo time of 120 milliseconds is adequate to eliminate magic angle effect in the collateral ligaments of most distal interphalangeal joints, and an echo time of 140 milliseconds results in elimination of magic angle effect but limits the ability to identify small lesions. In the present study, an echo time of 120 milliseconds was used for the high-field T2W FSE sequences; however, that might have been inadequate to eliminate magic angle effect in some locations. In another study,28 magic angle effect was not observed in the collateral ligaments of the distal interphalangeal joint in low-field T2W FSE sequences acquired with an echo time of 72 milliseconds from horses in a standing position.
For some locations of the proximal tarsal region, such as the origin of the short medial collateral ligament of the tarsocrural joint, a partial volume effect may have contributed to the hyperintense signal identified in all sequences.29 When different signal intensities are present in the same voxel, the signal recorded for that voxel represents the mean for the signals detected. Because of the curved interface between the medial malleolus of the tibia and origin of the short medial collateral ligament of the tarsocrural joint, signal might have been collected from different structures in the same pixel, resulting in a high mean signal for the origin of the ligament.
During histologic examination of sections from 1 tarsus, there was evidence of degeneration of the short medial collateral ligament of the tarsocrural joint, which might have affected its signal intensity on MRI images. However, only 1 tarsus was evaluated, and those results should be interpreted accordingly. Also, no such histologic changes were identified in the long lateral collateral ligament of the tarsocrural joint, despite it having a similar hyperintense signal. Although the tarsi evaluated in the present study were obtained from horses with no history or evidence of lameness localized to the tarsus, it was possible some horses had subclinical pathological lesions.
For the tarsi evaluated in the present study, the subchondral bone of the proximal aspect of the talus was thicker on the medial side, compared with that on the lateral side. That differs from the subchondral bone of the distal tarsal region in nonlame horses, which is thicker on the lateral side, compared with the medial side.30 Results of biomechanical studies of the tarsal region of horses indicate that compressive loading is greatest on the medial aspects of the distal portion of the tibia31 and proximal portion of the talus and on the lateral aspect of the proximal portion of the metatarsus.32 Thus, it appears compressive loading is transferred from the medial aspect to lateral aspect of the tarsal region in nonlame horses,30 and the thickness of the subchondral bone of tarsal structures is positively associated with compressive loading forces.
The MRI appearance of the proximal tubercle of the talus varied among the tarsi evaluated in the present study. For some tarsi, the proximal tubercle of the talus had diffuse low signal intensity, which was characteristic of mineralization, with evidence of fragmentation. The proximal tubercle of the talus is an extra-articular bony prominence on which the tarsocrural joint capsule and ligaments insert. Fragmentation of the proximal tubercle has been described as an incidental finding in horses,33 and the pathogenesis of that condition is unknown. However, it was interesting that, for the tarsi of the present study, changes in signal intensity typically affected the entire tubercle or lateral aspect of the tubercle, which is the region where fragmentation is most commonly observed.
The talocalcaneal ligament had heterogeneous signal intensity, which was characterized by a striped pattern (alternating white and black lines), especially at the midportion of its body. The combination of signal intensities observed for the talocalcaneal ligament in the various sequences suggested that a mixture of joint fluid and fat was interposed with the collagen fibers of the ligament.
The acquisition of MRI scans of the proximal tarsal region for live horses in a standing position can be challenging because of the amount of movement of the more proximal aspects of the limb34 and poor tolerance of some horses to the positioning of the magnet between their hind limbs. In the present study, multiple low-field sequences and 3 sets of low-field transverse images were obtained because of the low signal-to-noise ratio at the periphery of the images. Completion of the entire low-field MRI protocol used in this study would be time-consuming in a live standing horse and potentially impossible in a noncompliant horse. Also, the distance of the tarsus from the ground should be considered before low-field MRI is attempted because the magnet can be raised only up to 55 cm, which might preclude its use in large or tall horses. Appropriate positioning of the coil is also challenging. In the present study, we used a custom-made coil that was shaped to fit the tarsal region; however, we are aware that some practices use an upside-down positioned foot coil. We attempted to use an upside-down positioned foot coil during the pilot scans for this study, but it was discarded because complete overlapping of images acquired with that coil above and below the calcaneus was not possible.
In the present study, although the tarsi were accurately positioned within the low-field coil, the dorsal, lateral, and medial surfaces of the proximal tarsal region were often poorly visualized because of the close apposition between the coil and skin surface. In an attempt to improve visualization of the skin surface, we put padding material between each tarsus and the coil, but because the coil was not flexible, it did not improve image quality noticeably, especially for large tarsi, and occasionally precluded accurate evaluation of the long collateral ligaments. When live horses undergo low-field MRI, motion and flow artifacts will be present, which will complicate image interpretation. Delineation of the margins of the collateral ligament margins in low-field MRI images was facilitated by comparison of those images with the corresponding high-field MRI images, and the signal intensities throughout the collateral ligaments were similar between the low-field and high-field images.
Accurate pilot MRI images were crucial for obtaining good-quality images for evaluation, and use of a live pilot to guide centering tarsi within the coils was beneficial in the development of the final MRI protocols used in the present study. During the pilot process, we noticed that image quality decreased as the distance between the tarsus and center of the magnetic field increased, which suggested that more pilots might be required to image the entire region of interest in a live horse.
The present study had some limitations. The sample size was limited to 25 tarsi. Additionally, the horses from which those tarsi were harvested were selected from a hospital population and did not represent a random sample from the general horse population. Therefore, the effect of breed and use on MRI appearance of the proximal tarsal region was not investigated. Gross and histologic examination was performed for only 1 tarsus, and those results cannot be generalized to other tarsi.
In the present study, images obtained in the transverse plane were oriented perpendicular to the tibia to better evaluate the collateral ligaments of the tarsocrural joint and facilitate the acquisition of comparable images by use of both low-field and high-field MRI. For the distal tarsal region, the suggested slice orientation for transverse images is parallel to the distal tarsal joints. For horses that require high-field MRI of both the proximal and distal tarsal regions, a compromise in slice orientation might need to be made on the basis of the suspected likely origin of lameness. In the present study, only 1 tarsus/horse was evaluated, and MRI features of right tarsi were not compared with those of left tarsi. Ideally, both tarsi should be evaluated in clinical patients; this is particularly important because the size of the collateral ligaments of the tarsocrural joint varied among the tarsi evaluated in this study.
Results of the present study indicated that MRI signal intensity varied and followed a specific pattern within the collateral ligaments of the tarsocrural joints of nonlame horses. Knowledge of normal anatomic variation of the proximal tarsal region as described on the basis of the high-field MRI images obtained in this study will assist in the interpretation of MRI images of that region for clinical patients and can be used as a reference for interpretation of images acquired by use of standing low-field MRI because, despite inferior image resolution, the signal intensity distribution was generally comparable between low-field and high-field images.
Acknowledgments
Hallmarq Veterinary Imaging subsidized the cost of low-field image acquisition.
Presented in part as an abstract at the European Veterinary Diagnostic Imaging Conference, Cascais, Portugal, October 2013; and the European Association of Veterinary Diagnostic Imaging (British and Irish Division) Autumn Meeting, Loughborough, England, November 2013.
The authors thank Emma Goodfellow for technical assistance and Jennifer Stewart for performing histologic examinations.
ABBREVIATIONS
| FSE | Fast spin echo |
| GRE | Gradient echo |
| PD | Proton density |
| SPGR | Spoiled gradient echo |
| STIR | Short tau inversion recovery |
| T1W | T1-weighted |
| T2W | T2-weighted |
| T2*W | T2*-weighted |
Footnotes
Play-doh, Hasbro, Pawtucket, RI.
Signa Echospeed 1.5T, GE Medical Systems Ltd, Slough, Berkshire, England.
Standing Equine MRI, Hallmarq Veterinary Imaging, Guildford, Surrey, England.
References
1. Blaik MA, Hanson RR, Kincaid SA, et al. Low-field magnetic resonance imaging of the equine tarsus: normal anatomy. Vet Radiol Ultrasound 2000;41:131–141.
2. Latorre R, Arenciba A, Gil F, et al. Correlation of magnetic resonance images with anatomic features of the equine tarsus. Am J Vet Res 2006;67:756–761.
3. Branch MV, Murray RC, Dyson SJ, et al. Magnetic resonance imaging of the equine tarsus. Clin Tech Equine Pract 2007;6:96–102.
4. Dyson SJ, Murray RC. The tarsus. In: Murray RC, ed. Equine MRI. Oxford, England: Wiley-Blackwell, 2011;221–235.
5. Updike SJ. Functional anatomy of the equine tarsocrural collateral ligaments. Am J Vet Res 1984;45:867–874.
6. Barone R. Articolazioni tarische. In: Barone N, Bortolami R, eds. Anatomia comparata degli animali domestici. Vol 2. 2nd ed. Bologna, Italy: Edagricole, 1998;229–278.
7. Sisson S. Equine syndesmology. In: Sisson S, Grossman JD, Getty R, eds. Sisson and Grossman's the anatomy of the domestic animals. Vol 1. 5th ed. Philadelphia: WB Saunders, 1975;349–375.
8. Whitcomb M. Ultrasonography of the equine tarsus. in Proceedings. 52nd Annu Meet Am Assoc Equine Pract 2006;13–30.
9. Tranquille CA, Parkin TD, Murray RC. Magnetic resonance imaging-detected adaptation and pathology in the distal condyles of the third metacarpus, associated with lateral condylar fracture in Thoroughbred racehorses. Equine Vet J 2012;44:699–706.
10. van Weeren PR. Etiology, diagnosis and treatment of OC(D). Clin Tech Equine Pract 2006;5:248–258.
11. Raes E, Bergman HJ, Van Ryssen B, et al. Computed tomographic features of lesions detected in horses with tarsal lameness. Equine Vet J 2014;46:189–193.
12. Fleck SK, Dyson SJ. Lameness associated with tarsocrural joint pathology in 17 mature horses (1997–2010). Equine Vet Educ 2012;24:628–638.
13. Smith RK, Dyson SJ, Schramme MC, et al. Osteoarthritis of the talocalcaneal joint in 18 horses. Equine Vet J 2005;37:166–171.
14. Montgomery LJ, Juzwiak JS. Subchondral cyst-like lesions in the talus of four horses. Equine Vet Educ 2009;21:629–637.
15. Boero MJ, Kneller SK, Baker GJ, et al. Clinical, radiographic, and scintigraphic findings associated with enthesitis of the lateral collateral ligaments of the tarsocrural joint in Standardbred racehorses. Equine Vet J Suppl 1988;20:53–59.
16. Rose PL, Moore I. Imaging diagnosis—avulsion of the medial collateral ligament of the tarsus in a horse. Vet Radiol Ultrasound 2003;44:657–659.
17. Raes EV, Vanderperren K, Pille F, et al. Ultrasonographic findings in 100 horses with tarsal region disorders. Vet J 2010;186:201–209.
18. Smith MR, Wright IM. Arthroscopic treatment of fractures of the lateral malleolus of the tibia: 26 cases. Equine Vet J 2011;43:280–287.
19. Sherlock CE, Eggleston RB, Peroni JF, et al. Desmitis of the medial tarsal collateral ligament in 7 horses. Equine Vet Educ 2012;24:72–80.
20. Lamb L, Zubrod C, Hague B, et al. Clinical outcome of collateral ligament injuries of the tarsus. Can Vet J 2012;53:518–524.
21. Barker WH, Smith MR, Minshall GJ, et al. Soft tissue injuries of the tarsocrural joint: a retrospective analysis of 30 cases evaluated arthroscopically. Equine Vet J 2013;45:435–441.
22. Biggi M, Zani DD, De Zani D, et al. Magnetic resonance imaging findings of bone marrow lesions in the equine distal tarsus. Equine Vet Educ 2012;24:236–241.
23. Dyson SJ. Lameness associated with mineralisation of the central tarsal bone and a small osseous cyst-like lesion in two sports horses. J Equine Vet Sci 2013;33:51–56.
24. Daniel AJ, Judy CE, Rick MC, et al. Comparison of radiography, nuclear scintigraphy, and magnetic resonance imaging for detection of specific conditions of the distal tarsal bones of horses: 20 cases (2006–2010). J Am Vet Med Assoc 2012;240:1109–1114.
25. Judy CE. Cross-sectional imaging: more ways to look at a problem. Equine Vet Educ 2013;25:624–627.
26. Werpy NM, Ho CP, Kawcak CE. Magic angle effect in normal collateral ligaments of the distal interphalangeal joint in horses imaged with a high-field magnetic resonance imaging system. Vet Radiol Ultrasound 2010;51:2–10.
27. Werpy NM, Ho CP, Garcia EB, et al. The effect of varying echo time using T2-weighted FSE sequences on the magic angle effect in the collateral ligaments of the distal interphalangeal joint in horses. Vet Radiol Ultrasound 2013;54:31–35.
28. Smith MA, Dyson SJ, Murray R. 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? Vet Radiol Ultrasound 2008;49:509–515.
29. Murray RC, Werpy N. Image interpretation and artefacts. In: Murray RC, ed. Equine MRI. Oxford, England: Wiley-Black-well, 2011;101–145.
30. Branch MV, Murray RC, Dyson SJ, et al. Is there a characteristic subchondral bone plate thickness in horses with no history of hindlimb lameness? Equine Vet J 2005;37:450–455.
31. Schneider RK, Milne DW, Gabel AA, et al. Multidirectional in vivo strain analysis of the equine radius and tibia during dynamic loading with and without a cast. Am J Vet Res 1982;43:1541–1550.
32. Schamhardt HC, Hartman W, Lammertink JL. Forces loading the tarsal joint in the hind limb of the horse, determined from in vivo strain measurements of the third metatarsal bone. Am J Vet Res 1989;50:728–733.
33. Espinosa P, Lacourt M, Alexander K, et al. Fragmentation of the proximal tubercle of the talus in horses: 9 cases (2004–2010). J Am Vet Med Assoc 2013;242:984–991.
34. McKnight AL, Manduca A, Felmlee JP, et al. Motion-correction techniques for standing equine MRI. Vet Radiol Ultrasound 2004;45:513–519.
Appendix 1
Optimized protocol used for the acquisition of high-field MRI images of 25 cadaveric tarsi obtained from horses with no history of signs of pain in the proximal portion of the tarsal region (proximal tarsal region).
| Sequence | Repetition time (ms) | Echo time (ms) | Inversion time (ms) | No. of excitations | Matrix size | Slice thickness (mm) | Gap (mm) |
|---|---|---|---|---|---|---|---|
| T1W 3-D SPGR | 8.10 | 3.25 | — | 2 | 256 × 256 | 3 | 0 |
| T2*W 3-D GRE | 7.30 | 3.10 | — | 2 | 256 × 256 | 3 | 0 |
| 256 × 192† | |||||||
| Dual-echo PD | 4,000 | 10 | — | 2 | 256 × 192 | 4 | 1 |
| Dual-echo T2W FSE | 4,000 | 120 | — | 2 | 256 × 192 | 4 | 1 |
| STIR | 10,000 | 24.42 | 120 | 1 | 256 × 192 | 4 | 1 |
Only 1 tarsus/horse underwent MRI examination, and all tarsi were free of radiographic abnormalities.
Used for images acquired in the transverse plane to decrease acquisition time.
— = Not applicable.
Appendix 2
Optimized protocol used for the acquisition of low-field MRI images of the 25 cadaveric tarsi described in Appendix 1.
| Sequence | Repetition time (ms) | Echo time (ms) | Inversion time (ms) | No. of excitations | Matrix size | Slice thickness (mm) | Gap (mm) | Acquisition time (min) |
|---|---|---|---|---|---|---|---|---|
| T1W GRE | 50 | 8 | — | 1 | 256 × 256 | 5 | 0 | 1.51 |
| T2*W GRE | 65 | 12 | — | 1 | 256 × 256 | 5 | 0 | 2.14 |
| T2W FSE | 1,681 | 88 | — | 1 | 256 × 256 | 5 | 0 | 1.41 |
| PD | 1,150 | 22 | — | 1 | 256 × 256 | 5 | 0 | 1.53 |
| STIR | 2,100 | 22 | 60 | 1 | 256 × 256 | 5 | 0 | 2.41 |
See Appendix 1 for key.
