Abstract
General pointers for feline radiography
Given the size of our feline patients, a grid is generally not required. A table-top technique using a high detail film–screen system is suitable for all but the most obese cats, as the scattered radiation produced during an exposure is typically not enough to significantly affect film quality. Radiographic exposures are such that all x-ray machines used today in veterinary practices should be more than adequate to obtain good quality images of cats.
Although some of our more relaxed feline patients may allow reasonable positioning for thoracic and abdominal radiography without use of chemical restraint, as a rule sedation or general anaesthesia will be required to achieve diagnostic quality radiographs of the musculoskeletal system. The starting point should always be orthogonal views of the area of interest (Figure 1), and, when radiographing the limb, it is helpful if similar views of the contralateral limb are taken for comparison (Figure 2).
(a) The lateral view of the stifle in this skeletally immature cat shows a mild joint effusion, but is otherwise unremarkable. The popliteal sesamoid bone can be seen just caudal to the proximal tibial epiphysis. (b) The orthogonal view identifies an oblique fracture through the lateral femoral condyle
(a,b) A transverse fracture through the left calcaneus can be seen on the dorsoplantar and lateral views of the left hock (arrow) (c,d) Dorsoplantar and lateral views of the right hock allow careful comparison of an anatomically normal joint. This may identify further, more subtle changes in the injured hock, and/or help avoid overinterpretation of normal structures in this complex joint
Compared with dogs, the shape and size of different cat breeds is less variable, so developing a radiographic technique chart should be a relatively straightforward exercise. A successful technique chart for feline imaging requires a constant anode–film distance, the shortest possible exposure time (achieved by selecting higher mA values), consistency of film type, and a preference for detailed intensifying screens. Although there are a number of fairly complex methods for developing such a chart, a simple listing of the correct exposures found through trial and error would suffice for feline radiography in most instances. Optimal darkroom technique (or correct post-processing if a digital system is used) will then ensure that the information contained on any exposed film will be best presented for interpretation.
Before attempting to evaluate any radiographs, the clinician should always critique the films with one question in mind: ‘are they of diagnostic quality?’ If the answer is no, then the reason should be identified and corrected while the patient is still in the radiography room.
When evaluating conventional film–screen radiographs, the background lighting in the viewing room should be dimmed. A proprietary radiographic viewing box should always be used to assess films; ceiling lights and windows are unsuitable light sources for viewing radiographs. Hot-lights are helpful when assessing darkened areas of the film, and the use of a magnifying glass can improve evaluation of feline joint and skull radiographs. When evaluating digital radiographs, the choice of viewing monitor can affect image quality. Medical-grade monitors are preferred over standard PC monitors. 1
The skull
Interpreting radiographs of the feline skull can be quite challenging, and single views are often unhelpful. Standard views would include lateral (right or left) and ventrodorsal (or dorsoventral) projections (Figure 3a–c). The author prefers the dorsoventral view over the ventrodorsal view, as it is easier to position the skull flat on the x-ray plate without any axial rotation, and often the temporomandibular joints are better visualised.
(a) Lateral view of the skull. The feline tympanic bulla, unlike its canine counterpart, is divided into craniolateral and caudomedial compartments by an incomplete bulla septum Dorsoventral (b) and ventrodorsal (c) views of the skull give the best visualisation of the external ear canals, and allow direct comparison between left and right tympanic bullae. The temporomandibular joints are more consistently seen on the dorsoventral view (arrows) (d) Open-mouth rostrocaudal view for assessment of the tympanic bullae. The patient is placed in dorsal recumbency with its nose pointing towards the x-ray tube. The endotracheal tube anchors the tongue in the midline by being tied to the mandible. Rope ties placed around the upper and lower canine teeth are used to pull the mouth open. The increased opacity over the left tympanic bulla could be caused by otitis media or a polyp within the bulla, although an overlying tongue could also produce this change. Any abnormality seen in one view of the skull should be confirmed in other views (e) Rostroventral–caudodorsal oblique view of the skull and (f) lateral oblique view of the left tympanic bulla in a cat that had been hit by a car, showing a compression fracture of the ventral bulla wall and an increase in opacification of the bulla, presumed to be haemorrhage. The temporomandibular joints (both normal) can be seen just cranial to the bulla on the lateral oblique view. The cat had sustained a mandibular symphyseal fracture at the time of the accident, and a wire had been placed around the base of the mandibular canine teeth prior to the rostroventral–caudodorsal oblique view (e) being taken
Supplementary views are chosen to target particular pathology or anatomical regions (Figure 3d,e), 2 and include:
Left and right lateral 20° oblique views for the teeth, bullae, mandibles and temporomandibular joints.
Open-mouth rostrocaudal view for the bullae.
Rostroventral–caudodorsal oblique view for the bullae. 3
Intraoral dorsoventral view for the nasal chambers.
Although the shape of the feline skull is such that adequate exposure factors for one view should be suitable for all others, repeating the lateral view after increasing the exposure by 5 kVp may improve visualisation of the nasopharynx if this is a region of particular interest.
When positioning the lateral oblique views, both left (L) and right (R) markers are used to identify which dentition, bulla, mandible or temporomandibular joint is being profiled, irrespective of whether or not the profiled side is nearer the x-ray plate. The appropriate marker should be placed adjacent to the surface that is being skylined. Therefore, if the right tympanic bulla is being radiographed, then the slight skull rotation required means that the left dorsal skull is also being skylined. The L marker is placed above the skull, and the R marker is placed below the tympanic bulla (Figure 4a). If the left tympanic bulla is being radiographed, the right dorsal skull is being skylined and the markers are reversed (Figure 4b). The use of two markers confirms that an oblique view has been taken.
(a,b) Lateral oblique views of the tympanic bullae and temporomandibular joints. Both radiographs were taken with the patient in right lateral recumbency. In (a), the right marker below the bulla confirms that the right tympanic bulla is being skylined. The left marker above the nose confirms that it is an oblique radiograph. In (b), the skull has been repositioned to skyline the left tympanic bulla, and now the left marker sits below the bulla while the right marker is positioned above the nose
The increasing availability of CT and MRI in veterinary practice has done much to highlight the limitations of skull radiography, particularly in trauma patients. Both allow markedly improved visualisation of normal anatomy, with CT preferred for the assessment of bony structures and MRI preferred for soft tissue (Figure 5). Both enable slice-by-slice assessment of the skull without superimposition of adjacent structures. With modern CT scanners slices can be as thin as 0.6 mm, which allows reconstruction in multiple planes without significant loss of image quality. MRI slices tend to be thicker (2–4 mm), and most sequences used routinely cannot be reconstructed in multiple planes. Fractures that are not visible on radiographs become clear on a CT scan (Figure 6), and any confusion over whether asymmetry on a radiograph represents a true lesion or a positioning artefact is removed. 4
Standard projections of the skull include lateral and ventrodorsal or, as is the author’s preference, dorsoventral views.
Comparison axial CT (a) and MRI (b) slices through the tympanic bulla and inner ear. The CT image clearly shows the fine bony septum separating each bulla into craniolateral and caudomedial compartments. The magnetic resonance (MR) image clearly shows the hyperintense perilymph within the semicircular canals and cochlea (the so-called ‘two ducks’, arrow), but the detail of the underlying bullae is less than that provided by CT. CT is preferred for evaluation of the bony structures of the skull, and can be used in cases of acute head trauma to assess for intracranial haemorrhage. MRI gives far superior soft tissue detail, and has a primary role in neurological imaging
Axial CT slice through the temporomandibular joint showing a small fracture fragment associated with the medial aspect of the left condylar process of the mandible. This fragment was not visible on skull radiographs. The scan also shows a fracture through the right mandible ventral to the temporomandibular joint, which was seen on conventional radiographs
Both CT and MRI can be used for the assessment of patients with acute head trauma. The duration of a CT scan is much shorter than that of an MR image, and CT can be performed without the need for a general anaesthetic, so can be a safer procedure for the patient. MRI gives far superior soft tissue detail, and is the modality of choice when assessing neurological disease, although CT can be useful in demonstrating peracute intracranial haemorrhage as well as CNS oedema (Figure 7).
Compared with radiography, CT and MRI allow markedly improved visualisation, with CT preferred for assessment of bony structures and MRI preferred for soft tissue.
(a) Volume-rendered threedimensional CT image of the skull of a cat that had been shot with an air gun. The entry hole of the pellet can be seen over the dorsal surface of the right frontal sinus (arrow). The pellet had passed through the floor of the frontal sinus into the right frontal lobe of the cerebrum. The paired linear tracks crossing the surface of the skull just caudal to the entry hole were an artefact caused by the metallic pellet below (b) An axial CT scan through the skull of the same cat as Figure 7a, just caudal to the pellet. There is a hypoattenuating region within the right cerebrum (arrow) that is consistent with an area of cerebral oedema (c) Transverse T2 MR image through the brain of a cat that had been attacked by a dog. The linear hyperintense streaks within the left temporal muscle (arrow) had characteristics on other sequences typical of intramuscular haemorrhage. (d) Transverse T2 flair MR image of the same cat as in (c), through the tympanic bullae and cerebellum. There are compression fractures of the right tympanic bulla and left temporal/occipital region. A small amount of haemorrhage and oedema can be seen within the underlying cerebellum (arrow)
The spine
The feline spine has seven cervical, 13 thoracic, seven lumbar and three sacral vertebrae. The number of coccygeal vertebrae is variable. Breeds such as the Manx cat may have as few as a single coccygeal vertebra, whereas most breeds have over 18. Unlike the dog, the spinal cord in the cat extends into the sacrum (Figure 8). The feline spine is also much more flexible than the canine spine, which accounts for the wide variation in its appearance that can occasionally be seen on thoracic and abdominal radiographs (Figure 9). 5 The feline vertebral bodies are more rectangular than the dog’s, appear relatively more elongated, and the spinous processes are thinner and more tapered.
Sagittal T1 MRI sequence of the lumbosacral spine showing the spinal cord extending to the lumbosacral joint
(a,b) These two lateral radiographs of the thoracic spine demonstrate the flexibility of the normal spine in a conscious feline patient. Anaesthesia allows for more consistent positioning when radiographing the spine
General anaesthesia is preferred when positioning a cat for spinal radiography. When radiographing the entire spine it is simplest to consider it as three regions (cervical, thoracic and lumbar), taking a series of lateral and ventrodorsal views centred over C4, T7 and L4, respectively, and coning down the primary beam to the region of the spine being radiographed. A small foam wedge should be placed between the cervical spine and the cassette to ensure that the long axis of the cervical spine remains parallel to the surface of the x-ray table. Further radiographs centred over an area of interest can improve the diagnostic yield (Figure 10). ‘Cat-o-grams’ are not suitable for assessment of the feline spine.
(a) This lateral radiograph of a feline spine shows an osteolytic lesion of the T13 vertebral body. The radiograph is slightly underexposed. (b) The same spine is radiographed again with the primary beam centred over the lesion. The exposure factors have been corrected and the T13 lesion is more clearly seen
Although conventional radiography remains the first step in the evaluation of spinal disease in many of our feline patients, MRI allows far superior assessment of the spinal cord and associated nerves, as well as the paravertebral soft tissue structures. 6
‘Cat-o-grams’ are not suitable for assessment of the feline spine, which requires a series of lateral and ventrodorsal views centred over C4, T7 and L4.
The pelvis
When compared with the canine pelvis, the feline pelvis is narrower, the obturator foramina are more elongate, and there appears to be less acetabular coverage of the femoral heads on the ventrodorsal view (Figure 11), resulting in a lower Norberg angle in the normal cat. 7 Lateral and extended ventrodorsal radiographs centred at the level of the coxofemoral joints are considered the standard radiographic projections when imaging the pelvis, although occasionally additional views such as a frog-legged ventrodorsal or oblique ventrodorsal radiograph may identify lesions that are not clearly seen on the standard views.
Lateral and extended ventrodorsal views centred over the coxofemoral joints are standard for the pelvis.
Ventrodorsal radiograph of a feline pelvis (a) and canine pelvis (b). The feline pelvis is narrower, the feline obturator foramina more elongate, and the feline acetabulum appears shallower than that of the dog
Care should always be taken when positioning the ventrodorsal view, particularly when radiography is being performed to assess for pelvic trauma. Slight rotation can result in one of the sacroiliac joints being skylined and the other obscured, which may be misinterpreted as unilateral sacroiliac subluxation (Figure 12). If positioned correctly, the lateromedial width of both obturator foramina should be identical.
Slightly oblique, ventrodorsal view of the pelvis of a cat that had been hit by a car. The obliquity of positioning has resulted in skylining of the left sacroiliac joint, giving the false appearance of a sacroiliac subluxation. The obturator foramina are clearly not symmetrical
CT enables more comprehensive assessment of the pelvis in cases of trauma. It will demonstrate fractures that are not radiographically apparent, allows the degree of pelvic canal involvement to be clearly visualised, and the use of three-dimensional volume rendering aids surgical assessment and planning (Figure 13). However, CT imaging may not necessarily alter case management in the majority of pelvic trauma cases, and should not be seen as a substitute for a series of well positioned, high quality radiographs in the first instance. 8
(a) Ventrodorsal radiograph of a cat with an oblique fracture through the right ilium, right pubis and right ischium. There was concern that there might have also been a fracture of the right acetabulum, which may have necessitated surgery. (b,c,d) Three-dimensional, volume-rendered CT images of the same pelvis give improved visualisation of the fractures described in (a). Impingement of the pelvic canal is better demonstrated than on the radiograph, and acetabular involvement is ruled out. (e) Axial CT slice through the sacroiliac joint of the same pelvis. The iliac fracture passes through the right sacroiliac joint, and there is also a nondisplaced fracture of the right wing of S1 that was not visible on the radiograph
The limbs
When radiographing any part of a limb, it is often helpful to take similar views of the contralateral side for comparison. Not every lesion is obvious. Having the ‘perfect normal’ as a reference can help identify some lesions that would otherwise be missed, and may prevent overinterpretation of a normal variant that can also be seen on the other limb. Lateral and craniocaudal views (dorsopalmar if the region of interest is distal to the radiocarpal joint; dorsoplantar if the region of interest is distal to the tibiotarsal joint), centred over the region of interest, can be supplemented with skyline views, flexed and extended views, or stressed views as required.
The scapula is a roughly triangular-shaped bone that can be a challenge to radiograph. It is a particularly mobile bone that superimposes over the thorax and spine on the lateral view, and may require slight body rotation to image on the ventrodorsal view. The spine of the scapula is seen as a distinct radiopaque line running the length of the scapula that ends distally in the acromion (Figure 14). The caudally projecting suprahamate process can be seen just proximal to the end of the acromion, and serves as part of the origin of the deltoideus muscle. The coracoid process projects medially from the supraglenoid tubercle, and its variable shape in cats should not be mistaken for pathology. The clavicle is a mineralised crescent-shaped bone, approximately 2.5 cm in length, located just cranial to the scapulohumeral joint. It is mineralised at birth and increases in size as the animal grows. Unlike dogs, where they are only seen occasionally, the paired feline clavicles are always visible (Figure 15).
(a,b) Cropped images taken from dorsoventral and lateral radiographs of the thorax of a cat that had been hit by a car. Both scapulae are included on each view. Subcutaneous emphysema overlies the normal right scapula on the dorsoventral view. The right scapular spine is seen as a distinct continuous thin radiopaque line on both views. The left scapular spine is disrupted on both views, indicative of a fracture of the left scapula (arrows)
(a) Lateral view of the scapulohumeral joint in a cat showing the clavicle (1), the coracoid process (2) projecting from the supraglenoid tubercle, and the caudally projecting suprahamate process (3). The volume-rendered CT images of the feline scapulohumeral joint show the joint when viewed from a lateral aspect (b) and a medial aspect (c). The suprahamate process can be clearly seen from the lateral aspect, and the coracoid process is more readily identified from the medial aspect
The cat is unique among domestic mammals in having a supracondylar foramen over the medial aspect of the distal humerus, through which the brachial artery and median nerve pass (Figure 16). Unlike dogs, the cat has no supratrochlear foramen, and the medial epicondyle is considerably larger than the lateral epicondyle. Occasionally the small supinator sesamoid can be seen lateral to the elbow joint, and this should not be mistaken for pathology (Figure 17).
Craniolateral–caudomedial oblique radiograph of a feline humerus showing the supracondylar foramen (arrow), through which the brachial artery and median nerve pass
Craniocaudal radiograph of the feline elbow. The small supinator sesamoid bone is occasionally seen lateral to the radial head (arrow), and should not be mistaken for pathology. The feline medial epicondyle of the humerus is more prominent than that of the dog. Periarticular osteophytes can be seen associated with the medial articular margin of both the humerus and ulna
As in the dog, the cat has seven carpal bones arranged in two rows (row 1 – intermedioradial, ulnar and accessory carpal bones; row 2 – first, second, third and fourth carpal bones). On the dorsopalmar view, the small rounded sesamoid within the tendon of the abductor pollicis longus muscle can be seen just medial to the intermedioradial carpal bone (Figure 18).
Dorsopalmar radiograph of the feline carpus. The cat has three carpal bones in the proximal row, and four carpal bones in the distal row. The small sesamoid bone of the abductor pollicis longus tendon can be seen just medial to the proximal row of carpal bones (arrow)
On an extended ventrodorsal view, the femoral head should have at least 50% coverage by the overlying acetabulum. The site of attachment of the ligament of the femoral head (the fovea capitis) is seen as a flattened region over its craniomedial aspect on this view (Figure 19).
Extended ventrodorsal view of the pelvis of a cat with hip dysplasia and secondary degenerative joint disease. There is reduced coverage of both femoral heads by the dorsal rim of each acetabulum. The fovea capitis is seen as the flattening articular surface over the craniomedial aspect of each femoral head (1). The greater trochanter (2) acts as an insertion point for the gluteal muscles. The lesser trochanter (3) acts as the insertion point for the iliopsoas muscle
Sesamoid bones are present within the tendon of insertion of the quadriceps femoris (the patella), the medial and lateral heads of the gastrocnemius muscle (medial and lateral fabellae), and the tendon of origin of the popliteal muscle, although one or more of the last three are not always visible. The patella is sharply pointed on a lateral view, triangular-shaped on a caudocranial view, and sits over a shallow femoral trochlea. Mineralisation of the medial meniscus is an uncommon finding that is considered to be of no clinical significance (Figure 20).
(a) This axial CT slice through the patella clearly shows the normal shallow feline femoral trochlea. The lateral fabella is typically larger than the medial fabella (b,c) Lateral and caudocranial views of the stifle in a cat with mineralisation of the medial meniscus (1). The patella is sharply pointed on the lateral view and has a triangular outline on the caudocranial view. Sesamoid bones are present within the gastrocnemius muscle (2) and tendon of origin of the popliteal muscle (3)
As in the dog, the cat has seven tarsal bones arranged in three rows (row 1 – tibial tarsal bone [talus] and fibular tarsal bone [calcaneus]; row 2 – central tarsal bone; row 3 – first, second, third and fourth tarsal bones). The large fourth tarsal bone articulates at the proximal intertarsal joint. The routine radiographs of the hock are the lateral and dorsoplantar views. Supplementary radiographs may include oblique views and stressed radiographs. In such a complex joint, it is often helpful to take comparison views of the contralateral limb (see Figure 2).
When radiographing any part of a limb, it is often helpful to take similar views of the contralateral side for comparison.
Stressed radiographs are used to assess the extent of any damage to the soft tissue support structures of a joint. Unless an obvious subluxation is present, significant joint laxity could easily be missed on routine survey views. Using rope ties, foam wedges and sandbags, the joint is positioned in such a way as to provide tension to the side of a joint where there is a clinical suspicion of damage. A similar view is taken of the contralateral joint, and then comparisons of joint space width and position of the bones are made.
The clinician should always critique radiographic films with one question in mind: ‘are they of diagnostic quality?’
Role of ultrasound
Ultrasound can be a useful adjunct to musculoskeletal radiography in a limited number of patients. High frequency (>10 MHz) linear probes are preferred as the depth of imaging is so small, although high frequency microconvex probes with an echolucent stand-off may also be used.
Ultrasound may be used to assess tendons and muscles, surfaces of bones, and masses within soft tissue structures, 9 as well as to guide a biopsy needle to facilitate sampling of an area of interest (Figure 21).
Lateral (a) and caudocranial (b) radiographs of the left hind limb of a cat with a swollen, painful left gastrocnemius muscle. (c,d) Ultrasound images of the left gastrocnemius muscle of the same cat show irregular anechoic regions within the muscle. The colour Doppler image shows blood flow within the surrounding muscle. The cat had a reduced factor VIII concentration (haemophilia A), and this lesion was a region of intramuscular haemorrhage
Key points
A feline-focused approach to radiography will reward the clinician with better quality images and a greater likelihood of a correct diagnosis.
Sedation or general anaesthesia is generally required to achieve diagnostic quality radiographs of the musculoskeletal system.
The increasing availability of CT and MRI has done much to highlight the limitations of skull radiography, particularly in trauma patients.
CT allows a more comprehensive assessment of the pelvis in cases of trauma; however, it should not be seen as a substitute for a series of well positioned, high quality radiographs of the pelvis in most cases.
When radiographing any limb, similar views of the contralateral limb can be of great help by providing an anatomical normal for comparison.
Knowledge of normal feline skeletal anatomy reduces the risk of misinterpretation.
Footnotes
Funding
The author received no specific grant from any funding agency in the public, commercial or not-for-profit sectors for the preparation of this review article.
Conflict of interest
The author declares that there is no conflict of interest.