General Considerations

Nuclear medicine is a relatively recent advance in diagnostic imaging of the horse, pioneered by Ueltschi¹ in Europe and Twardock and Devous^2,3 in the United States. Seeherman and colleagues⁴ and Lamb and Koblik⁵ further demonstrated the value of bone scintigraphy in the evaluation of sports horses. For detailed description of procedures and interpretation of scintigraphic images, the reader is referred to Equine Scintigraphy.⁶

Bone scintigraphy is a way to reach within the horse and extract clinically useful and relevant information and helps to answer many lameness questions that we previously could not answer. Scintigraphy is used to assess the current status of known radiological abnormalities, pursue diagnosis in horses with negative or equivocal radiographs, screen horses with obscure forelimb or hindlimb lameness, and evaluate horses with poor performance. Scintigraphy can be useful to corroborate functional information gained with other advanced imaging modalities such as magnetic resonance imaging (MRI). However, a gamma camera is not an answer machine, and it is critically important to correlate scintigraphic findings with clinical lameness examination findings. Scintigraphy can play an important role as a screening tool but should never replace diagnostic analgesia. In fractious or highly strung horses with hindlimb lameness, scintigraphic examination may be used to provide a diagnosis, but clinical relevance still should be confirmed.

Nuclear medicine involves the in vivo or in vitro use of radioisotopes in the diagnosis and management of clinical disease. Several terms are used synonymously with nuclear medicine, including nuclear scintigraphy, bone scintigraphy, and gamma scintigraphy, and although these terms differ slightly, for horses most clinicians refer to bone scintigraphy, the technique most commonly performed. Bone scintigraphy is highly sensitive compared with radiography. It can detect as little as 10⁻¹³ g of radiopharmaceutical in bone, whereas changes measured in grams must occur before a lesion can be detected using radiography.³ Important factors that decrease sensitivity include time from radiopharmaceutical administration to image acquisition, body part–to-camera distance, shielding, motion, time from injury to image acquisition, ambient temperature and peripheral perfusion, and amount of background radiation. For example, the combination of negative factors including high background radiation, motion, endogenous shielding, and distance reduce the sensitivity of pelvic and other axial skeletal scintigraphic imaging and that of upper limb imaging in both the forelimb and hindlimb. High sensitivity (94.4%) in horses with extremity fractures was found, but a lack of sensitivity in pelvic imaging was suggested.⁷ Specificity is low compared with other modalities because disparate diseases can similarly alter blood flow and binding sites in bone. Direct trauma (osteitis) and fracture; stress-related bone injury, including fracture and osteoarthritis (OA); infection (infectious osteitis and less frequently osteomyelitis); osteochondrosis; enostosis-like lesions; and neoplasia are in theory difficult to differentiate scintigraphically. Accuracy can be improved by acquiring several images from different perspectives, minimizing factors affecting sensitivity, and knowing the history. For example, a focal area of increased radiopharmaceutical uptake (IRU) in the caudolateral tibial cortex of a 2-year-old Thoroughbred (TB) filly would undoubtedly represent stress-related bone injury rather than a rare bone tumor.

Radiographs depict activity in bone that has already occurred in the past several days to years. Scintigraphy is a functional evaluation of bone at the time of imaging. Scintigraphic evidence of bone activity means active bone formation, bone modeling, is occurring that might take weeks to be visible radiologically. Therefore a major advantage of scintigraphy is early detection of bone injury. Scintigraphy will unlikely accurately reflect changes in bone that occurred longer than 3 to 4 months before imaging. Horses given substantial rest before examination are not good candidates unless the examination is a follow-up to assess healing.

Scintigraphy and Other Imaging Modalities: Where do They Fit?

Recent advances in imaging such as digital radiography (Chapter 15), computed tomography (CT; Chapter 20) and MRI (Chapter 21) have led to questions about the role of scintigraphy in lameness diagnosis. We are often asked questions such as, “Would you use scintigraphy or MRI for diagnosis?” and “Should I put MRI or scintigraphy into my practice?” We would not choose to be without the ability to use scintigraphy in lameness diagnosis. We believe that the combined and concurrent use of scintigraphy and MRI yields the most useful and relevant information and that seldom, given strict case selection, is information redundant. At the origin of the suspensory ligament in the proximal aspect of the third metacarpal and metatarsal bones, a complex area from which pain commonly originates, a combination of imaging modalities is most useful to characterize soft tissue and bony injury. Likewise, lameness abolished with palmar digital analgesia is complex, and the concurrent use of advanced imaging modalities provides complementary information. Nuclear scintigraphy has proved useful in the investigation of foot pain, specifically from the podotrochlear apparatus, and has helped determine clinical significance of lesions discernible with MRI, because a positive correlation between scintigraphy and MRI grades in lesions of the navicular bone was found.^8,9 We believe that both scintigraphy and MRI are critical for addressing complex lameness questions.

Radioisotopes and Radiopharmaceuticals

A radioisotope emits radiation (particles) that is captured using a scintillation camera. Radioisotopes such as indium-111 (¹¹¹In) are used occasionally, but the most common and useful radioisotope is technetium-99m (^99mTc). ^99mTc, a short-lived (metastable) radioisotope with a half-life of 6 hours, is ideal for radiation safety and horse retention. ^99mTc is excreted almost entirely through the kidneys, so containing and monitoring urine is extremely important. ^99mTc is produced when molybdenum-99 (⁹⁹Mo) decays to ⁹⁹Tc. The metastable (^99mTc) radioisotope gives off a gamma ray (140 keV) that is used for imaging. Commercially, ⁹⁹Mo and ^99mTc generators can be purchased for use in large hospitals, but an alternative, cost-effective method is the daily purchase of individual doses, a practice that obviates the need to house generators. Directly from the generator, ^99mTc is in the ionic form of sodium ^99mTc pertechnetate (Na^99mTcO₄) that can be injected or mixed with a bone-seeking agent or pharmaceutical. Radiation is measured in curies (Ci) or millicuries (mCi); becquerels (Bq), megabecquerels (MBq), and gigabecquerels (GBq) are also used. One millicurie is equal to 37 MBq. The recommended dose of ^99mTc is 0.4 to 0.5 mCi (14.8 to 18.5 MBq)/kg, totaling 150 to 200 mCi (5.5 to 7.4 GBq) per horse. Low doses reduce radiation exposure but prolong acquisition time or may result in inadequate image quality if insufficient counts are obtained. High doses may actually reduce overall radiation exposure by limiting exposure time.

Na^99mTcO₄ can be injected intravenously directly, but only flow and pool phase studies can be performed. For most equine studies, Na^99mTcO₄ is mixed with a pharmaceutical. For bone, ^99mTc is bound to methylene diphosphonate (MDP), or hydroxymethane or hydroxymethylene diphosphonate (HDP, HMDP). MDP is slightly less expensive and is the most common pharmaceutical used worldwide, but we prefer HDP because the kidneys clear it slightly faster, allowing early acquisition of delayed phase images. Henceforth MDP and HDP are used interchangeably.

The exact mechanism of binding of ^99mTc-MDP to bone remains unclear. ^99mTc-MDP is thought to bind to exposed sites on the inorganic hydroxyapatite crystal. Binding sites are exposed under normal and pathological conditions in areas of actively remodeling bone or in soft tissues undergoing mineralization.^{10,11 99m}Tc-MDP uptake occurs by the processes of chemical adsorption onto, and by direct integration into, the crystalline structure.¹² Other possible mechanisms to account for increased uptake include incorporation into the organic matrix or local hypervascularity.¹¹ In rat models, ^99mTc was found to be incorporated into the organic matrix rather than the inorganic portion of newly formed bone.^13,14 Radiopharmaceutical may dissociate with incorporation of ^99mTc and MDP individually into the organic and inorganic phases, respectively.^{15 99m}Tc-MDP adsorption might depend on pH and the presence of phosphates, calcium compounds, and other cations.¹⁶

Accumulation of ^99mTc-MDP is not simply the result of changes in local blood flow, although blood flow is likely increased in sites of actively remodeling bone. Although increased blood flow does not significantly affect a bone scan,¹⁰ adequate blood flow is necessary to deliver radiopharmaceutical to available binding sites in bone. Poor correlation was found between perfusion index and radiopharmaceutical uptake (RU) in delayed images when evaluating people undergoing distraction osteogenesis.¹⁷ This study suggested that blood flow is not closely linked to bone metabolism and that delayed images most accurately predicted osteogenesis.¹⁷ Rather than measuring blood flow, delayed images reflect changes in bone metabolism. When three-phase bone scintigraphy was used in people with osteonecrosis of the jaw secondary to bisphosphonate administration, delayed images were all positive, but increased perfusion (and increased blood pool) was found in 9 of 12 patients.¹⁸ A three-phase bone scan can detect increased perfusion and soft-tissue inflammation if present, but it is metabolic activity of bone that influences delayed phase images. Decreased blood flow, caused by infarction or ischemia, can greatly affect a bone scan but is an unusual clinical problem (see discussion of photopenia, page 226). However, decreased peripheral blood flow in old horses or those imaged in cold weather or on days with high diurnal temperature change can adversely affect image quality (see discussion of poor-quality bone uptake, page 222).

The most important aspects of ^99mTc-MDP binding relate to timing of the scan and the stage of modeling (formation). In actively remodeling bone, osteoclastic activity predominates during bone resorption, whereas osteoblastic activity dominates during bone modeling. Modeling occurs independently or in conjunction with remodeling in cancellous and cortical bone. Histological and scintigraphic findings were evaluated in a rat tibial evacuation model, and ^99mTc-MDP was found to bind to sites of active calcification, most prevalent 12 days after injury.^{19 99m}Tc-labeled phosphonates were identified during bone formation, and ongoing resorption was not necessary for increased uptake to occur.^{12 99m}Tc-MDP accumulated in areas of calcification or in fixed bone fragments,^12,20 and accumulation was mediated by osteoblastic activity.¹² The high sensitivity of bone scintigraphy is attributed to increased osteoblast activity that precedes morphological changes visible radiologically.²¹ Other mechanisms may exist, however, because positive bone scan results may be seen in people with diseases such as osteomalacia, in which high bone matrix turnover and failure of calcification occur.¹⁹

Site and stage of binding are important from a clinical perspective. Binding sites for ^99mTc-MDP are created by osteoblast activity during bone modeling, and maximal IRU occurs 8 or 12 days after bone injury.^19,22 An acute fracture caused by direct trauma may not be scintigraphically evident for several days. Acute, traumatic injury differs from stress-related bone injury, because the latter, particularly common in racehorses, results from a continuum of bone changes that might lead to stress or catastrophic fractures and OA. Microfracture, periosteal callus, and subchondral bone damage precede the development of stress or complete fracture in the dorsal cortex and distal articular surface of the third metacarpal bone (McIII), the third metatarsal bone (MtIII), the humerus, the tibia, and the pelvis.^23-25 In horses with acute lameness from stress-related bone injury, bone scan findings usually are immediately positive because bone modeling is ongoing. In horses with stress-related bone injury a bone scan result is likely to be positive long before catastrophic fracture occurs, an important advantage of scintigraphy compared with radiography. In horses with traumatic injury, such as an acute pelvic or other upper limb fracture, a false-negative scan result may occur from lack of modeling. Other factors resulting in false-negative results include distance, shielding, high background activity, and motion. For example, a horse developed acute hindlimb lameness during hospitalization, and, suspecting a pelvic fracture involving the acetabulum, clinicians performed scintigraphic examination on day 2, but the scan result was negative. On day 9 faint IRU appeared, consistent with fracture.

In horses with stress-related bone injury and traumatic injury a considerable decrease in RU occurs within 6 to 8 weeks after injury. Decreased intensity after fracture varies with the specific bone and fracture type. The ideal time to image, particularly in horses with pelvic or other axial skeletal trauma, is 10 days to 8 weeks after injury, although positive bone scan findings are often obtained in horses with distal extremity fractures 72 to 96 hours after known trauma.

Given the short half-life of ^99mTc, it is the most widely used and versatile radioisotope when imaging a lame horse and for most nuclear medicine procedures in horses and people. The same radioisotope injected intravenously, ^99mTc, can be combined with different drugs (pharmaceuticals) to target different tissues. For example, for lung perfusion images to be obtained, macroaggregated albumen is labeled with ^99mTc; for brain imaging, glucoheptanate is labeled with ^99mTc. Standard ^99mTc-HDP is used for sinus, dental, and temporomandibular joint imaging and can be used to evaluate peripheral vasculature in first-pass studies after intravenous injection or by injection directly into distended peripheral veins for evaluation of venous return. To obtain high levels of radiopharmaceutical in the distal extremities for specific imaging or to perform experimental studies, Na^99mTcO₄ can be administered using intraosseous or intravenous regional limb perfusion techniques.²⁶ Our preliminary results with this technique show uneven distribution within distal limb tissues and persistence of radioisotope in large, distended veins. Peripheral lymphatic fluid dynamics can be studied using ^99mTc-sulfur colloid injected directly into distended lymphatics or subcutaneous tissues.²⁷ Soft tissues are most commonly imaged using three-phase bone scintigraphy with ^99mTc-HDP (MDP), but recently, radiolabeled biotin was found to be safe and useful to detect soft tissue inflammation without concurrent uptake of the radiopharmaceutical in bone.²⁸ We have limited experience using radiolabeled white cells to image osteitis, osteomyelitis, and cellulitis in horses, and although the technique is occasionally useful, our results do not appear as promising as those previously reported.²⁹ Briefly, white blood cells harvested from affected horses or foals are labeled with ^99mTc and hexamethylpropyleneamine oxime (^99mTc-HMPAO) and then reinjected to allow accumulation at sites of infection. Infection imaging is an important topic in human medicine, and advances in equine imaging using novel agents such as antibodies or antimicrobial agents may hold promise.³⁰

Imaging Equipment

Scintigraphy can be performed in two ways, but the most common and useful method is acquisition of two-dimensional images with a gamma camera. Alternatively, a handheld probe is used to acquire count density computer-generated charts, or graphs are created to determine IRU. Known as probe point counting, this form of scintigraphic examination was first introduced in 1984.^31,32 Some probes are like miniature gamma cameras built with a single photomultiplier tube. Crystal probe detectors are more expensive than the photomultiplier detectors, but they are small and easily used on the body surface, or per rectum in horses with suspected pelvic fracture.³² Probe point counting is done with a substantially lower radioisotope dose, minimizing cost and radiation exposure. Probe point equipment is inexpensive. Because the probe is placed directly on the skin surface, body part–to-probe distance is minimal. A gamma camera is often difficult to get close to the body surface, a problem that can prolong acquisition time and decrease image quality. A good correlation appears between results of probe point counting and gamma camera imaging in acute injuries in young horses, but in older horses the technique has limited sensitivity. However, quantitative information is available only as a histogram, and an actual image allowing qualitative assessment is not obtained.

Gamma camera images are obtained in analog or digital form, and permanent hard copies of a two-dimensional scintigraphic image are generated. In people, one or several gamma cameras operating simultaneously can be used to generate a cross-sectional image similar to a CT or MRI scan. This is known as single photon emission CT (SPECT) and photon emission tomography (PET). Although SPECT and PET imaging are theoretically possible in anesthetized horses, we are not aware that the techniques have been performed.

A reconditioned gamma camera is perfectly satisfactory, less expensive than new equipment, and durable, often being useful for at least 15 to 20 years or more, with proper maintenance. Gamma cameras are either large field of view (LFOV) or small field of view (SFOV), based simply on the size (Figures 19-1 to 19-3). LFOV cameras can be rectangular or circular in shape, but crystals for rectangular cameras are more costly. In adapting equipment for use in horses, an important consideration is how to move the camera safely and easily to the horse and vice versa. Most standard gantry (support structure of the gamma camera) designs are unsuitable for equine imaging, because good-quality images can be obtained only by having the body part close to the camera. Resolution is inversely proportional to the distance between the body part and camera. To quickly obtain dorsal and lateral, and lateral and plantar, images of the forelimbs and hindlimbs, respectively, the camera must be able to be lowered below floor level (see Figure 19-3). Alternatively, the horse must be positioned on an elevated platform. Limbs can be held manually near the camera, but this practice increases radiation exposure to personnel and should be avoided whenever possible. Suitable gantries are commercially available or can be custom made. A lead collimator is also required.

Figure 19-1 The nuclear medicine facility at New Bolton Center, University of Pennsylvania, has large field of view and small field of view cameras. A custom-made central column overhead gantry holds the rectangular large field of view camera, which can be lowered below floor level to image the distal extremities.

Figure 19-2 The small field of view camera in the floor allows easy acquisition of solar images of the foot. After a wooden cover is applied, the horse stands over the small field of view camera for the solar image. A lead-lined wrap placed around the coronary band region shields the foot and camera from radiation emanating from the more proximal aspects of the affected limb.

Figure 19-3 Construction of a recess or pit into which the large field of view camera can be lowered allows imaging of the distal extremities with minimal movement of the horse. The lateral aspect of the hoof touches the face of the camera (clear polycarbonate protects the collimator). This position minimizes distance and improves resolution, improves image quality, and decreases acquisition time. A curved, lead-lined shield protects the camera and left forelimb from radiation emitted by the right forelimb.

Image Acquisition

For flow and pool phase images, horses are sedated and positioned in front of the camera, and ^99mTc-HDP (or MDP) is injected intravenously. Delayed phase images are then obtained 2 to 4 hours later. Scintigraphic images are produced when the horse, now the radiation source, emits gamma rays from normal and abnormal bone. The gamma rays must pass through overlying tissues and traverse the distance between the body part and the camera. An area of IRU emits more gamma rays than does adjacent normal bone and contributes more counts to the scan. Sites farther from the camera, those being shielded by overlying soft tissue or adjacent bone, or those in areas of high background activity may not be visible, because they may not contribute enough gamma rays. Gamma rays strike a sodium iodide crystal, and scintillations produced are detected and amplified by photomultiplier tubes that subsequently transmit information through electronic circuits. The image is displayed on an oscilloscope or sent directly to a film processor, called analog imaging. Alternatively, information is transmitted to a computer, stored digitally, manipulated, and subsequently sent to a film processor or printer.

An important recent advancement is the ready availability of user-friendly computer programs to acquire, analyze, store, and process images. Systems based on Apple, Windows, and Unix are currently available and are straightforward. Motion correction software is the latest, exciting advance; it improves upper limb and axial skeleton image quality by negating the effects of motion and allowing higher count numbers than with conventional software (Figure 19-4). Modern software also allows postprocessing of images—for example, masking out the bladder, which might otherwise steal counts.

Figure 19-4 Delayed phase scintigraphic images of an 8-year-old Thoroughbred with neurological signs consistent with compression of the cervical spinal cord. A, Corrected (left) and uncorrected (right) right lateral images (cranial is to the right) showing marked improvement in image quality using motion correction techniques. Note there is normal radiopharmaceutical uptake in the intervertebral joint between the third (Cr3) and fourth cervical vertebrae. B, Laterolateral digital radiographic image (cranial is to the right) of the cervical spine showing marked osteoarthritis in this intervertebral joint (arrows). Based on lack of cervical stiffness or pain and normal scintigraphic activity, this radiological change was considered old and inactive. C, Right lateral scintigraphic image of the caudal cervical spine showing narrowing (arrow) of the vertebral canal between the fifth and sixth cervical vertebrae, a finding validated using myelography. Useful anatomical information can be obtained from scintigraphic images.

Images are obtained in either static or dynamic mode. Static images, most commonly obtained, are acquired using a predetermined number of counts per image. For example, because motion and soft tissue covering are limited in the distal limbs, good-quality images can be obtained using 100,000 to 150,000 counts per image. In general, increasing counts per image improves image quality, but for more counts to be acquired, time and motion become factors. Motion correction can add flexibility.

A fundamental principle of image interpretation is comparing images of one limb with those of the contralateral limb, but the clinician should keep in mind that both limbs, and for that matter all limbs, can be abnormal. To compare limbs accurately using this technique, body part–to-camera distance and camera position relative to the limb must be standardized between limbs. Time to acquire each image ranges from 30 to 90 seconds depending on radioisotope dose, type and age of horse, and ambient temperature. Time rather than count number can be standardized, assuming that limb perfusion is adequate and symmetrical. For example, lateral images of both metacarpal regions are obtained for 90 seconds and compared.

The number of images to obtain is based on the body part being examined, but usually at least two images are required. In some areas this may not be possible, but the clinician should be aware that false-negative results can occur. A dorsal scintigraphic image gives detailed information about only the dorsal aspect of the limb, and abnormal palmar regions might be missed. The image and information are not the same as depicted in a dorsopalmar radiographic image. Similarly, areas of IRU on the medial side of the limb might not be visible on a lateral image. Areas more distant from the camera contribute substantially less radiation than those closer because of the inverse square law. Bone interposed between a medial lesion and the camera can effectively shield the site. A lateral image of a right forelimb failed to reveal a fracture of the medial aspect of the distal phalanx (Figure 19-5). Radiographic examination and dorsal and solar scintigraphic images, however, revealed an incomplete fracture that we completely missed in the lateral scintigraphic image. Additional images may be necessary for accurate diagnosis.

Figure 19-5 Delayed phase scintigraphic images of a horse with a fracture of the medial aspect of the right forelimb (RF) distal phalanx. The lateral (LAT) image of the right forelimb does not show the area of increased radiopharmaceutical uptake, and little difference is noted between both front feet. The dorsal and solar images, however, clearly show increased radiopharmaceutical uptake. It is extremely important to obtain more than one scintigraphic image. LF, Left forelimb.

We prefer to localize lameness, because this allows for detailed examination of a specific area. For instance, routine screening images of the front digit include lateral and dorsal delayed images, but in horses in which lameness is abolished using palmar digital analgesia, pool phase images and lateral, dorsal, solar, and occasionally medial, palmar, and flexed delayed images are obtained. Diagnostic accuracy can be improved by acquiring many different images, and pinpointing a lesion using scintigraphy can allow a focused radiographic, ultrasonographic, CT, or MRI study to be performed. Combined imaging can yield additional useful information.^8,9 When augmenting the routine dorsal and plantar images of the metatarsophalangeal joint with flexed lateral and sometimes flexed dorsal and medial images, instead of saying, “There is IRU in the fetlock joint,” the diagnostician can say, “There is focal IRU involving the distal, plantarolateral aspect of the MtIII,” a much more accurate description (Figure 19-6).

Figure 19-6 Lateral (left), dorsal (second from left), flexed lateral (second from right), and flexed dorsal delayed phase scintigraphic images of the right forelimb (RF) of a 2-year-old trotter with subchondral bone injury. The flexed lateral and dorsal images best show the focal area of increased radiopharmaceutical uptake involving the distal, medial aspect of the third metacarpal bone. Although increased radiopharmaceutical uptake appears in other images, the flexed dorsal image allows differentiation of increased radiopharmaceutical uptake involving the third metacarpal bone and the proximal phalanx.

Dynamic acquisition can be used before motion correction in delayed images or to evaluate blood flow. One- or 2-second per frame images are generated sequentially and can be evaluated individually or combined into a single composite image, with or without motion correction. First-pass angiography can be used to assess blood flow in the aorta, iliac, and femoral arteries³³ in horses with suspected thromboembolism (Figure 19-7) or to assess blood flow in the distal limb.

Figure 19-7 Composite scintigraphic image generated after dynamic acquisition of 32 sequential images (2 seconds per frame) with the gamma camera positioned over the terminal aorta and branches. Imaging started 20 seconds after administration of ^99mTc-hydroxymethane diphosphonate in the external jugular vein (cranial to the left, right is to the top). A, In a normal horse the external iliac arteries show a V-shaped divergence and a linear caudal course of the internal iliac arteries is seen. B, In this horse with aortoiliac thromboembolism, the direction and contour of the external iliac arteries is disrupted and the internal iliac vessels lack recognition.