AZURVET Referral Veterinary Centre, Saint-Laurent-du-Var, France
Stifle injuries are increasingly recognized as a major cause of hind limb lameness. Ultrasonography is well suited to image the stifle and has gained popularity as it is easily available, may be used in the standing patient, and will provide access to most parts of this joint. It has tremendously improved our ability to confirm or rule out stifle lesions, which, in most cases, primarily involve soft tissue structures.
Radiography may be useful to detect bony lesion such as fractures, osteochondrosis, cyst-like lesions, etc., but most frequently, only late degenerative changes will be recognized. Magnetic resonance imaging (MRI) and computed tomography (CT) are increasingly used for the diagnosis of stifle conditions in horses but access to these modalities remains limited. CT-arthrography (images acquired after intra-articular injection of iodinated contrast medium) of the stifle has been described in recent years. It procures useful information on joint surfaces and the outlines of menisci and ligaments. However, the parenchyma is poorly visualized and the presence of synovial thickening can alter the appearance of intra-articular structures. MRI has been little used so far because of difficulties placing the stifle in most magnets and long acquisition times, although it may be increasingly used in coming years.
The equine stifle is a large, complex region and its ultrasonographic examination requires a thorough knowledge of the anatomy. A comprehensive, standardized approach is recommended. A systematic approach to the joints is paramount to avoid overlooking subtle changes, which may have strong clinical repercussions, in particular meniscal tears or focal cartilage or subchondral trauma. Many artifacts can impair the sensitivity and specificity of the examination, so it is particularly important to acquire experience and a thorough knowledge of the gross and ultrasonographic anatomy in order to optimize stifle ultrasonography.
Preparation and Scanning Technique
The stifle region should be finely clipped from distal to the tibial tuberosity up to the stifle skin fold, and all around the limb, then prepared as usual. Most of the examination is performed with the limb weightbearing, although, in very lame horses, examination can be adequately performed with the limb partially flexed. To scan the cranial aspect of the femorotibial joints, however, the limb must be flexed in order to expose the joint surfaces and intra-articular structures. It may be useful to use a wooden block to stabilize the foot. It is not necessary to achieve full flexion, which is often not tolerated by the injured horse. Intra-articular analgesia will dramatically interfere with evaluation of the joint, as it causes hemorrhage, inflammation, and effusion, and because air may be introduced into the joint space or periarticular soft tissues. Gas can persist in the joint for up to 2 weeks; it is therefore preferable to perform an ultrasonographic examination prior to diagnostic intra-articular analgesia whenever possible.
A high-frequency (8–15 MHz) linear probe is best to evaluate the femoropatellar joint and the medial aspect of the stifle. A standoff pad may be used but can be fiddly and is not usually necessary. To examine the cranial, lateral, and caudal aspects of the stifle, 4–10 MHz convex to micro-convex, curved array transducers are preferred, as they are easier to align to deeper structures not parallel to the skin.
Comprehensive examination of the stifle may be time-consuming, although with experience it may be performed in under 15 minutes. Typically, it is performed in 5 stages (Figure 7.1 ).
Starting with the limb extended and weightbearing (if the horse allows), the femoropatellar articulation is examined with a linear transducer from a cranial approach, using both transverse and longitudinal sections of the limb and spanning the area from the fold of the stifle proximally to the tibial tuberosity distally and half of the circumference of the femur from medial to lateral (Figures 7.1a).
The examination is continued, with the limb in the same extended position, over the medial femorotibial joint region with the same probe. The examination is easiest using longitudinal sections (i.e., with the probe placed in a frontal plane, perpendicular to the tibial plateau (Figure 7.1b)). Transverse section images may be useful to evaluate the meniscal parenchyma over their whole width, notably to look for vertical tears (Figure 7.1c).
The lateral aspect of the joint is approached with a linear probe and, if necessary, with a convex or microconvex probe. Longitudinal (frontal) plane images are preferred and it may be necessary to slightly tilt the probe downward to improve the images (Figure 7.1d).
The caudal aspect of the stifle is assessed through the caudal thigh muscles using a convex transducer (Figure 7.1e).
Finally the limb is flexed (~45 degrees) and the cranial aspect of the femorotibial articulations is examined using a convex or microconvex transducer (Figure 7.1f). Each structure is imaged in both longitudinal and transverse planes, which requires an accurate knowledge of the topography of each anatomical structure (i.e., joint surfaces, tendons, and ligaments)
Ultrasonographic Anatomy
The anatomy of the stifle region has been described elsewhere and the reader is encouraged to study anatomy texts in detail to better understand the 3-D arrangement of this complex joint.
Femoropatellar Joint
The femoropatellar joint is best imaged with the joint extended, as most of the trochlea is hidden by the patella when the limb is flexed. In the extended position the patella points craniolaterally, and the large medial trochlear ridge is palpable under the skin cranially. Starting in transverse planes, the probe is placed cranioproximally over the quadriceps muscles and examination is continued from proximal to distal, down to the tibial tuberosity. The quadriceps muscle is relatively homogeneous with a typical hypoechogenic muscular pattern. There is no insertion tendon as such, so that there is a sharp interface between the muscle belly and the proximal outline of the patella (Figure 7.2). The cranial surface of the patella forms a sharp, smooth hyperechogenic interface. Distally, it forms a shallow groove at the origin of the middle patellar ligament (PL) (see below). Medially, it is prolonged by the hypoechogenic, crescent-shaped parapatellar fibrocartilage, which curves around the medial trochlear ridge and terminates via the thin medial collateral patellar ligament (Figure 7.3). It also gives rise to the medial PL (see below).
Distal to the apex of the patella, the trochlear groove appears as a wide, U-shaped bone interface covered by anechogenic cartilage (Figures 7.4 and 7.5). In many horses, the latter is irregular in the center of the groove, but this is considered to be a normal feature. The medial trochlear ridge is broad, smooth, and rounded, and is covered by relatively thin cartilage (0.8–1 mm thickness). This becomes irregular medially. The lateral trochlear ridge is much narrower and triangular in cross-section. Its cartilage is thicker (2–4 mm). In longitudinal sections (parasagittal), the trochlear ridges form a smooth, convex bone interface, topped by anechogenic cartilage, which should be perfectly regular in thickness. The lateral PL is seen as a striated structure lying directly over the cartilage of the lateral ridge. The capsule is tightly applied against the cartilage surface, and no fluid is normally visible over the trochlea.
Anatomically, the patellar ligaments actually correspond to the quadriceps tendon and they are, therefore, similar in appearance to digital flexor tendons (Figure 7.6). Their borders are, however, ill defined due to poor contrast with the surrounding, hyperechogenic infrapatellar fat pad (see Figure 7.4). The middle PL is round to oval in cross-section and is the largest of the three PLs. It runs from the apex of the patella to the cranial-most part of the tibial tuberosity, within the fat pad and in the center of the trochlear groove (Figure 7.6). Distally, it becomes more triangular in section and always contains thin, hypoechogenic lines, giving it a webbed appearance. This should not be mistaken for a tear. Tilting of the probe to create an off-incidence artifact will cause the ligament to become hypoechogenic, enhancing its outline within the fat pad (Figure 7.7). The medial PL is triangular in cross-section and, in the extended limb, lies over the medial aspect of the medial trochlear ridge, 4 or 5 cm caudal to the apex of the ridge (Figure 7.8). Proximally, it becomes more heterogeneous as its fibers spread out into the parapatellar fibrocartilage (Figure 7.9). Distally, it inserts on the medial aspect of the tibial tuberosity, approximately 2 cm medial to the middle PL and receives a tendinous branch from the sartorius muscle. The lateral PL is thinner, flatter, and crescent-shaped in cross-section (Figure 7.4). It caps the apex of the lateral trochlear ridge, outlining the cartilage (Figure 7.5). It inserts on the tibial tuberosity, immediately lateral to the middle PL’s insertion. There is often a small to moderate amount of anechogenic synovial fluid in the lateral femoropatellar joint recess, caudal to the LPL, and over the lateral surface of the lateral trochlear ridge. A small amount of fluid may be present caudal to the medial PL, although this is less prominent.
Medial Femorotibial Joint: Medial Aspect
A linear transducer is best suited, as the joint surfaces lie very close to the skin. The topography of the medial aspect of the joint and organization of the medial meniscus are reviewed in Figure 7.10. The proximal edge of the tibial condyle and the round, smooth surface of the medial femoral condyle form a triangular space filled by the echogenic medial meniscus (Figure 7.11). The latter is separated from the condyles by thin, hypoechogenic cartilage. Linear, anechogenic artifacts running perpendicular to the probe interface through the meniscal parenchyma are caused by refraction at the insertion of the capsule on the outer surface of the meniscus. These are easily mistaken for tears but they should remain perpendicular to the probe when the latter is tilted. The capsule inserts over the whole abaxial aspect of the meniscus, so that fluid can only accumulate proximal to it, over the abaxial surface of the medial femoral condyle. Discrete distension of the joint is common cranial to the collateral ligament but the synovial membrane should remain thin. Small villi may be seen in the pouch in normal horses. Occasionally, very small gas bubbles, forming multiple, punctate hyperechogenic interfaces casting strong acoustic shadows, may be seen within the joint fluid. These represent nitrogen released by articular depression (vacuum phenomena). This is a normal, physiological occurrence. Transverse plane images of the menisci are obtained by rotating the probe by 90 degrees. The scan plane should be parallel to the edge of the tibial plateau. These images are useful to better evaluate the configuration of certain tears.
The medial collateral ligament (MCL) is a strong, flattened structure, approximately 5–6 mm in thickness. As in most joints, it is made up of two poorly differentiated branches (Figure 7.10). It is necessary to rotate the probe to image each branch individually, as they run obliquely to each other, crossing over the meniscus. The superficial branch runs vertically; it originates on the femoral epicondyle several centimeters proximal to the joint, and inserts on the abaxial surface of the tibia (Figure 7.12). It receives part of the insertion of the adductor muscle proximally. The deep branch originates caudal to it, and runs in a craniodistal direction to insert on the tibia, cranial to the superficial branch. It is strongly adherent to the abaxial surface of the meniscus.
Lateral Femorotibial Joint: Lateral Aspect
As for its medial counterpart, the general topography is presented in Figure 7.13. The joint is covered by a thicker layer of muscle and the bone surfaces are very oblique to the skin. It is consequently difficult to obtain sharp images using a linear transducer. Convex or micro-convex probes, allowing the beam to be angled relative to the skin, may provide better images. Craniolaterally, the tendon of origin of the long digital extensor and peroneus tertius muscles originates on the lateral femoral epicondyle and runs within the extensor groove of the proximal tibia (Figure 7.14). The groove is covered by cartilage and a synovial recess extends from the lateral femorotibial joint (LFT), between the bone surface and the tendon. Similarly to the MCL, the lateral collateral ligament (LCL) is divided into two branches, superficial and deep, although they are more difficult to tell apart. They insert respectively on the tibia and the fibular head. The popliteal tendon originates from a small depression on the lateral femoral epicondyle, immediately cranial to the LCL and caudal to the long digital extensor tendon. It runs obliquely in a caudodistal direction, passing over the meniscus and underneath the LCL. Because of its oblique course, it will be located proximal to the meniscus at the level of the collateral ligament and directly adjacent (abaxial) to the meniscus caudolaterally. It is triangular in cross-section at the level of the collateral ligament and should not be mistaken for a torn portion of the meniscus. It becomes rectangular in shape over the caudal horn. The tendon runs within the joint space cranially and in the capsule caudally. Its hypoechogenic muscle belly fans out over the caudal proximal tibia. Fluid may be seen proximal to the meniscus and cranial or caudal to the collateral ligament, but this is less common than in the medial compartment.
Cranial Aspect of the Femorotibial Joints
Imaging the cranial aspect of the femorotibial articulation requires that the stifle be flexed, in order to expose the femoral condyles, cranial horn of both menisci, intercondylar space, and the cruciate ligaments (Figure 7.15). Full flexion will expose most of the articular surfaces of the femoral condyles but the cruciate ligaments will flatten out against the tibial eminence and horses with stifle disease may resent this degree of flexion unless sufficient analgesia is provided. A 30–45-degree flexion is sufficient to image the cranial meniscal attachments and cruciate ligaments. A convex array or micro-convex transducer must be used to allow for angulation of the beam in relation to the skin. The surface of the femoral condyles, overlying cartilage, and cranial horns of the menisci are easily visualized dorsomedially (medial femoral condyle) or laterally (lateral femoral condyle), deep to the infrapatellar fat pad. The round surfaces of the condyles can be assessed for cartilage or subchondral defects as flexion exposes their distal-most aspect (Figure 7.16).
To image the cranial cruciate ligament (CrX), the ultrasound beam must be aligned perpendicular to the ligament fibers (Figure 7.17): the probe is placed immediately distal to the patella, over or medial to the middle PL. The probe is angled downward in the sagittal plane to image the surface of the tibial eminence. The beam is then rotated 15–20 degrees clockwise in the left limb and anticlockwise in the right limb until the linear pattern of the ligament is recognized. The CrX originates caudally on the axial (medial) aspect of the lateral femoral condyle, and inserts in a dip between the medial and lateral prominences of the tibial eminence. It may be useful to approach the joint immediately medial to the middle PL and point the probe slightly toward the lateral side to image the axial (medial) surface of the lateral femoral condyle, before rotating very slightly toward the intercondylar space to find the ligament, which curves around the condyle. Cross-sectional images can be obtained by rotating the probe 90 degrees in the same position. The CrX is hyperechogenic with a regular striated pattern. As the ligament curves around the caudal cruciate ligament, it is difficult to image its entire length on one scan image. Cross-sections may be obtained but as the ligament is surrounded by connective tissue and fat, its margins are ill-defined and the ligament contrasts poorly with intercondylar tissue.
The cranial (femoral) origin of the caudal cruciate ligament (CaX) is visualized by directing the beam upward, with the probe placed between the lateral and middle PLs, or directly over the latter, immediately proximal to the tibial tuberosity (Figure 7.18). The surface of the intercondylar space of the femur is identified between the condyles and distal to the trochlear groove. The CaX runs distocaudally in the sagittal plane, directly over the bone surface of the femur between the condyles. It crosses over the medial aspect of the cranial cruciate ligament within the intercondylar fossa. The ligaments are slightly curved, so that it is difficult to obtain perfect alignment of the fibers throughout their lengths and both ligaments are often visible on the same scan image (Figure 7.19 ).
Finally, the cranial tibial insertions of the menisci (cranial meniscotibial ligaments – CMTL) are imaged from the craniolateral and craniomedial aspects respectively. The cranial horn of the meniscus is visualized as a wedge-shaped structure between the femoral and tibial condyles cranioabaxially. Each meniscus is followed cranially by following the curved contour of the ipsilateral femoral condyle and then along the cranial aspect of the tibial eminence. To visualize the ligaments, the probe must be angled downward to keep the proximal surface of the tibia perpendicular to the beam (Figure 7.20). Both longitudinal and transverse images of the ligaments may be obtained from this position by rotating the transducer by 90 degrees, revealing a regularly striated, echogenic structure running between the cranial meniscal horn and the cranial aspect of the tibial eminence. This requires some practice but reproducible images of the CMTL and cranial meniscal horn can easily be obtained.
Caudal Aspect of the Femorotibial Joints
The topography is reviewed in Figure 7.21. The leg is examined extended and weightbearing. Because of the large caudal muscle mass, the depth of the joint varies from 5–12 cm. For this reason, lower frequency (e.g., 5–9 MHz) transducers may be necessary. Convex or micro-convex probes are used to angle the probe up or down in relation to the skin to image the various ligaments. Adequate images of the caudal cruciate ligament are obtained with the transducer placed on the caudal aspect of the distal thigh muscles, 10–15 cm proximal to the junction thigh/leg. The probe is pushed into the semimembranosus/semitendinosus muscle mass and angled downward at 15–30 degrees (i.e., perpendicular to the tibia) (Figure 7.1e). From a caudal approach, the femoral condyles are round and smooth with an even anechoic cartilage. The caudal meniscal horns are wedge-shaped and curve around the caudal aspect of each condyle (Figure 7.22). The joint capsule is only obvious when highlighted by a joint effusion which may be voluminous in the caudal aspect of the stifle.