1.10 Imaging

There are a number of imaging modalities currently available to the practising orthopaedic surgeon, including digital radiography (DR), multidetector computed tomography (MDCT), magnetic resonance imaging (MRI), ultrasound, nuclear medicine studies, and studies where iodinated contrast is injected into the extremity joints, such as conventional arthrography or magnetic resonance arthrography.

The radiographic examination is the best and most available imaging study for the initial evaluation of patients with orthopaedic problems. Generally, radiography should precede any complex imaging studies such as computed tomography (CT) or MRI.

X-rays used in medical imaging are either generated in x-ray tubes or emitted from radioactive isotopes. Those generated in x-ray tubes are used in radiography, fluoroscopy, and CT. X-rays emitted from radioactive isotopes are used in nuclear medicine and are called γ-rays. In the x-ray tube (Figure 1.10.1), x-rays are generated when a fast stream of electrons is suddenly stopped by the target or anode (positive terminal). The electrons originate on the negative terminal of the tube which is also known as the cathode or filament. The area of the target bombarded by electrons is known as the focal spot., which should be as small as possible, preferably 1 mm or less. The quantity of x-rays emitted from the x-ray tube is proportional to the number of electrons flowing from the filament (cathode) to the target (anode); this is measured in milliamperes (mA) and is preselected by the technologist. The quality, or penetrating properties, of the x-ray beam is determined by the energy of the electrons striking the target. This is determined by the kilovoltage setting (kV_p). Modern x-ray tube casings are designed with filters to remove the low energy radiation from the x-ray beam. Filtration is an essential technique utilized to change the composition and improve the quality of the x-rays beam. High-energy x-rays generate significant scatter radiation which results in foggy images and diminished tissue contrast on radiographs. The use of a fixed or a reciprocating (motorized) grid is the most commonly applied technique for controlling scatter in medical radiography.

For most orthopaedic problems, two orthogonal radiographic views of the structure under investigation are considered a basic requirement. Depending on the anatomical region, additional views are often needed. In the trauma setting, two orthogonal views may be difficult to obtain. When a non-displaced fracture of the navicular is suspected it is advantageous to put the long axis of the navicular parallel to the film (Figure 1.10.2). Small avulsion fractures at the dorsum of the triquetrum are best detected on mildly pronated lateral views of the wrist (Figure 1.10.3). The pisiform, pisitriquetral joint, and the hook of the hamate are best visualized on mildly supinated lateral views as well as carpal tunnel views (Figure 1.10.4).

Standing AP views of the knees are recommended for patients with osteoarthritis, to demonstrate the extent of articular cartilage loss (Figure 1.10.5). Many authorities believe that radiography for ankle injuries is overutilized, and several criteria have been suggested to limit unnecessary examinations; the Ottawa criteria are probably the best known.

It is not necessary for orthopaedic surgeons to know all the appropriate views that can be used. X-ray departments will arrange for appropriate views, provided that the diagnosis under consideration is clearly stated in the request form.

Neck pain can be initially evaluated with an anteroposterior (AP) and lateral views of the cervical spine. The American College of Radiology has developed appropriateness criteria for the use of imaging studies in the trauma setting. If the upper thoracic spine or lower cervical spine is obscured by the shoulders, a swimmer’s view is usually quite helpful. For the lumbar spine, AP and lateral views are typically sufficient. Occasionally, oblique views are used to look for defects in the pars interarticularis, but these oblique views add significant radiation dose to the patient, increase the cost of the examination, and should not be used routinely.

Gonadal shielding should always be used where appropriate, and care taken to minimize exposure to other patients and staff. Appropriate personnel radiation-monitoring devices should be worn by health professionals whenever there is a possibility of exposure to radiation.

The introduction of helical CT and MDCT has transformed the value of CT in acquiring volumetric data sets, improving resolution and allowing very thin slices (0.3mm). It also enables reconstruction of images in any plane as well as three dimensions (Figures 1.10.6–1.10.8). Other advantages of MDCT are increased speed and increased total volume covered. This is valuable for multitrauma patients, with head, chest, spine, and abdominal injuries, as all can be scanned at the same time. The whole body from the head to below the hips can be imaged in less than 30s.

MDCT scanners currently generate hundreds or even thousands of images on each patient. It has, therefore, become impractical to print images or view studies one image at a time. Postprocessing workstations are now a necessity, and the task of interpreting CT studies from printed images has disappeared. Parallel advances in picture archiving and communication (PACS) technology have taken place, allowing real-time viewing, manipulation of large stacks of images, and distribution of images throughout the medical centre.

Magnetic resonance imaging (MRI) is now the primary imaging technique for internal derangement of the knee, disc disease, muscle and tendon injuries, early detection of osteonecrosis, and evaluation and staging of many soft tissue and bone neoplasms.

The basic principle of clinical MRI is that any nucleus with an unpaired proton or neutron, such as hydrogen atoms, will react to external magnetic fields. When a subject is placed in a strong magnetic field, some of the protons align with the magnetic field while spinning and precessing around the axis of the magnetic field. This is analogous to the wobble of a spinning top (Figure 1.10.9). Hydrogen atoms in different tissues (fat versus muscle) will have different constants and precess at different frequencies.

The equilibrium of spin alignment can be disturbed by applying a radiofrequency pulse at the frequency of precession. As the protons come back into alignment, they emit electromagnetic radiation in the radiofrequency spectrum that is collected and turned into magnetic resonance images.

A pulse sequence is a precisely defined pattern of radiofrequency pulses and listening intervals. The most commonly used pulse sequence in the study of the musculoskeletal system is the spin–echo sequence (Figure 1.10.10).

A spin–echo sequence can be T₁ weighted, accentuating the T₁ properties of tissue, or T₂ weighted, accentuating the T₂ properties of the tissue. In general, T₁-weighted sequences depict anatomy better and T₂-weighted images show pathology to advantage.

On T₁-weighted images, tissues with a short T₁ have high signal intensity and are bright (Box 1.10.3). An example of a substance with a short T₁ is fat. Tissues with long T₁, such as cerebrospinal fluid, have low signal intensity. On T₂-weighted sequences, tissues with a short T₂ have low signal intensity. Examples include tendon and ligament. Tissues with a long T₂, such as cerebrospinal fluid, are bright on T₂-weighted sequences. Inflamed, oedematous tissue has more extracellular water than normal and will therefore be lower in signal intensity on T₁-weighted images and will be higher in signal on T₂-weighted images than normal tissue. Most tumours appear relatively dark on T₁-weighted images, less than fat and similar to muscle. On T₂-weighted images, most tumours show increased intensity, but not usually as bright as cerebrospinal fluid.

Conventional spin–echo images are the mainstay of MRI, but the T₂-weighted images take a long time to acquire and are susceptible to motion artefacts. Other pulse sequences have been developed that generate images in much less time; one such sequence is called fast spin–echo. Imaging times can be cut significantly because up to 32 lines of the image can be acquired in a single repetition. Besides reducing time of acquisition, fast spin–echo images have less distortion of the image when hardware or metal is present in the scanned area (Box 1.10.4). One disadvantage to fast spin–echo sequences is that oedema is harder to detect than with spin–echo images. Spin–echo T₂-weighted images have traditionally been used to detect oedema that is often present as a sign of many pathological processes. These T₂-weighted sequences are sensitive for detecting oedema because signal intensity of fat is low and signal intensity of water is high on T₂-weighted images. However, with fast spin–echo T₂-weighted images, the signal from fat is not suppressed and oedema can be masked. To overcome this limitation, methods to suppress the signal from fat are often employed.

The two currently used methods of suppressing fat signal are frequency selective chemical presaturation and inversion recovery. Chemical presaturation is a technique that is applied to other sequences such as fast spin–echo, whereas inversion recovery is a pulse sequence itself, and cannot be combined with other sequences. When using chemical presaturation, the machine takes advantage of the fact that the hydrogen protons in fat have a slightly faster precessional rate than water hydrogen protons. By placing a very selective radiofrequency pulse that only deflects fat hydrogens immediately before the main pulse sequence, signals from fat can be nullified, and signals from all other tissues unaffected. This is usually used with fast spin–echo T₂-weighted sequences to increase the conspicuity of oedema or fluid. Chemical presaturation can also be used with T₁-weighted images after the administration of contrast. Increases in signal from contrast accumulation are more conspicuous when surrounding fat (bright on T₁-weighted images) is dark.

Inversion recovery sequences can effectively decrease signals from fat, but they have the disadvantage of a long repetition time, and take a relatively long time to acquire. Newer inversion recovery pulse sequences include a fast inversion recovery sequence, which still suppresses fat signal and is very sensitive for oedema, but is acquired rapidly, similar to fast spin–echo. One important point is that if fat suppression is desired for images obtained following contrast administration, chemical saturation should be used and inversion recovery avoided. Inversion recovery will suppress the signal changes caused by the contrast, whereas chemical saturation will not.

Gadolinium is a transition element which has paramagnetic activity, shortening both the T₁ and T₂ relaxation times of nearby hydrogen nuclei. The T₁ shortening produces a higher (brighter) signal on T₁-weighted images. Enhancement of tissues is roughly proportional to blood flow to the tissue, and results in increased signal on T₁-weighted images. Therefore, following contrast administration, T₁-weighted sequences are typically obtained and compared to T₁-weighted images before contrast.

Contrast is most commonly used as an intravenous agent, but can also be instilled into joints, resulting in an MR arthrogram. This has been applied to the shoulder for improving the characterization and detection of labral abnormalities and in the knee for evaluating osteochondral fragments and postoperative menisci.

Tendon degeneration, tendinopathy, partial and full thickness tears of the rotator cuff all show-up as areas of increased signals on the T₂-weighted images. Fat suppression and magnetic resonance arthrography will increase the conspicuousness of cuff abnormalities (Figure 1.10.11).

MR arthrography extends the capabilities of conventional MRI because contrast solution distends the joint capsule, outlines intra-articular structures, and leaks into abnormalities.

By demonstrating the inferior labral–ligamentous complex, MR arthrography makes a major contribution to the evaluation of patients with suspected glenohumeral instability (Figure 1.10.12). Although MR arthrographic images have demonstrated greater than 90% accuracy in the detection of anteroinferior glenoid labral tears, diagnostic confidence may be further increased when the shoulder is images in abduction and external rotation (ABER). This involves flexing the elbow and placing the patient’s hand posterior to the contralateral aspect of the head or neck. An anteroinferior glenoid labral tear that is non-displaced when the shoulder in a neutral position has a greater likelihood of being displaced from the glenoid rim and becoming more conspicuous when the shoulder is in the ABER position.

The multiplanar capabilities of MRI enable the elbow joint to be imaged in true sagittal and coronal planes, facilitating more accurate diagnosis of ligamentous injuries. The superior soft tissue contrast of MRI provides simultaneous evaluation of bone and soft tissue, allowing for assessment of all the static and dynamic stabilizers.

The use of MRI to evaluate the wrist has lagged behind that of larger joints, because of technical limitations of spatial resolution. Thin and contiguous slices are needed for adequate MRI of the wrist. Therefore, high-resolution MRI is essential to evaluate normal and abnormal features of the hand and wrist.

MRI or MR arthrography can localize and characterize tears of the triangular fibrocartilage complex (Figure 1.10.13), helping to identify those patients that would benefit from surgery. MR arthrography combines the advantages of arthrographic depiction of anatomic perforation with the direct visualization of marrow, cartilage, and soft tissues allowed by MRI and can be performed with single-, double, or triple-compartment injection.

MRI is an effective modality for diagnosing tears of the anterior cruciate ligament, and is also accurate for detection of meniscal tears. Normal menisci have low signal intensity on all pulse sequences, and tears are seen as linear increased signal that contacts a surface of the meniscus on two consecutive images. Another sign of meniscal tears is abnormal contour of the meniscus, with truncation being a sign of a displaced fragment (Figure 1.10.14). Following partial meniscectomy, retears are difficult to diagnose by MRI. Intra-articular contrast is a method to improve the characterization of the postoperative meniscus.

A tear of the anterior cruciate ligament can be shown on MRI by detecting abnormalities of the ligament itself, such as increased signal within the ligament, or by detecting secondary signs that the anterior cruciate ligament is torn (Figure 1.10.15). Primary signs include abnormal increased signal of the anterior cruciate ligament, and abnormal orientation of the ligament. Secondary signs result from laxity of the knee including anterior translation of the tibia, abnormal posterior position of the posterior horn of the lateral meniscus, and abnormal orientation of the posterior cruciate ligament such as hyperbuckling. A prior severe injury may show as bone contusions, medial collateral ligament injury, and haemarthrosis.

Any increase in signal in the posterior cruciate ligament is abnormal. Conversely, the anterior cruciate ligament often has some strands of increased signal within it, and this can be problematic for detecting subtle anterior cruciate ligament tears, and decreases the accuracy of MRI for evaluating partial anterior cruciate ligament tears. Medial collateral ligament injuries produce, abnormal signals both deep and superficial to the medial collateral ligament, and if the ligament is focally disrupted or not visualized at all, then complete tear is diagnosed.

In a patient with persistent pain following trauma in whom plain films are negative, MRI can show contusions or occult fractures. It can also be used to evaluate patients with suspected injury to the patellar ligament In patellar tendinitis or partial tearing, the ligament may show increased signal intensity on T₁-weighted, T₂-weighted, and proton density images; it also shows increased AP diameter proximally and the margins of the affected ligament become indistinct, especially posterior to the thickened segment.

One artefact that commonly occurs with musculoskeletal MRI is the magic-angle phenomenon (Figure 1.10.16). Tendons can have a falsely increased signal when they are oriented approximately 55 degrees out of the main magnetic field. The supraspinatus tendon, just proximal to its insertion on the greater tuberosity, commonly shows this artifact. Other tendons that commonly show magic-angle effects are the patellar ligament, the foot flexor tendons such as flexor hallucis longus, and, occasionally, the peroneal tendons. If increased signal is seen on a short echo time (TE) sequence, it is important to confirm the abnormality on a long TE sequence.

For the Achilles tendon, axial and sagittal planes are sufficient for evaluation (Figure 1.10.17). However, the other tendons that cross the ankle joint do not run in a straight course, but curve around the ankle and hindfoot roughly parallel to the posterior facet of the subtalar joint. Therefore, in order to image these tendons in cross-section, oblique imaging planes should be used. When studying the foot flexors and extensors, sagittal sequences are followed by oblique coronal sequences that are perpendicular to the posterior facet of the subtalar joint (Figure 1.10.18).

From an orthopaedic standpoint, MRI is very useful for evaluating patients with possible disc disease, metastatic disease, congenital anomalies, and trauma.

MRI is very sensitive for detecting abnormal signal and morphology of intervertebral discs (Figure 1.10.19). Normally the disc is lower in signal intensity than vertebral bone marrow on T₁-weighted images, and higher in signal on T₂-weighted images. With disc degeneration, signal remains low on the T₂-weighted images.

While MRI is very sensitive for detecting abnormal discs, care must be taken when interpreting the study to correlate with the patient’s symptoms. In an asymptomatic population, over one-third have disc abnormalities at more than one level by MRI criteria. MRI with contrast enhancement is the best technique to study individuals with the failed back syndrome, and differentiating recurrent disc herniation from epidural scarring following surgery.

For evaluating patients with suspected metastases to the spine, MRI is also the procedure of choice. It shows the extent of bony involvement, degree of compression of the cord, and extraspinal extension of the tumour. It is more specific than bone scan for characterizing metastatic deposits, and the entire spine can be surveyed in the right clinical setting. Metastases replace the normal fatty bone marrow and therefore appear as focal low-signal areas on T₁-weighted images. MRI is also helpful for distinguishing benign from metastatic aetiologies of compression fractures. If a part of the affected vertebral body marrow is intact, then the fracture is likely to be benign. Metastatic compression fractures typically show diffuse complete replacement of the marrow signal. One caveat to this rule is in the setting of acute compressions. The oedema associated with acute compression fracture leads to diffuse replacement of the marrow signal, and resembles metastatic deposit. In this setting, the best approach is to repeat the MRI in 6 weeks if the clinical suspicion for metastatic disease is low, or to biopsy the collapsed vertebra in the setting of higher suspicion.

Patients with vertebral osteomyelitis usually present with back pain. The typical plain radiographic features of adjacent vertebral endplate destruction with loss of intervertebral disc height may lag for several weeks. Often, the white blood cell (WBC) count and erythrocyte sedimentation rates are non-specific, and no fever is present. In this setting, MRI is a very sensitive method to detect vertebral osteomyelitis and discitis. The classic findings are adjacent vertebra with abnormal replacement of marrow signal with oedema, low signal on T₁-weighted and high signal on T₂-weighted images. The intervertebral disc usually shows increased signal on T₂-weighted images as well.

For most tumours, there are no tissue-specific features on MRI that allow definitive diagnosis. Notable exceptions include lipoma and soft tissue haemangioma. The MRI appearance of a benign lipoma is characteristic. However, differentiating the low-grade liposarcoma from a benign lipoma with some internal septations can be quite difficult. It is generally agreed that interpretation of MRIs on patients with suspected bone tumours should not be done without the plain films at hand. Use of intravenous contrast is also controversial.

The main use of MRI in the evaluation of patients with tumours is staging and postoperative follow-up. MRI is superior to plain film for detecting the extent of disease for surgical planning. It can be difficult to separate peritumoural oedema from actual tumour. When evaluating a patient with primary bone malignancy, it is important to scan the entire extremity involved to detect any occult skip metastases. For the postoperative patient, MRI can be used to detect recurrent disease. However, haematoma or seroma in the surgical bed are often present and can make exclusion of recurrent or residual tumour difficult. Follow-up scans will typically show stable or decreasing size of seromas and haematoma whereas tumours generally increase in size.

Bone scans are most commonly done using technetium-99m labelled phosphonates (^99mTc MDP, ^99mTc HEDP). When ^99mTc decays (half-life 6h), gamma rays are emitted, which are then detected to produce an image. ^99mTc-labelled phosphonates accumulate preferentially in the mineral phase of newly forming bone. Areas of increased bone remodelling, whether it is due to infection, trauma, or tumour, will appear ‘hot’ on the bone scan compared to normal bones. Therefore bone scan abnormalities are usually not specific for a disease process. The primary strength of bone scans is its relatively high sensitivity when trying to rule out bone disease.

Bone scans are routinely done 3h after injection of ^99mTc-labelled phosphonates. The 3-h interval allows clearance of the radiopharmaceutical from the soft tissue to reach optimal bone to soft tissue contrast. Planar whole body scans are ideal for screening of the entire skeleton. They may not, however, show the optimal contrast for lesion detection in the spine because of uptake in the normal parts of the vertebrae anterior and posterior to the lesion. Spinal lesions are significantly better detected with tomographic imaging (single photon emission computerized tomography, SPECT) because of the better spatial resolution. Exact localization of the abnormality in the vertebra on SPECT imaging also helps with differentiation of benign versus malignant lesions. To evaluate the blood flow and vascularity of a lesion, a three-phase bone scan is performed, which includes immediate flow and blood pool images followed by bone scan at 3h.

Although many bone metastases are lytic on x-rays, they stimulate osteoblastic activity and new bone formation. This accounts for the increased uptake of bone tracers in osseous metastases. The typical pattern is the presence of multiple focal areas of increased uptake predominantly in the axial skeleton.

Bone scans are more sensitive than x-rays in detection of metastatic bone disease, and are used for assessment of treatment response to chemotherapy and hormonal therapy. Primary malignant bone tumours generally demonstrate intense uptake of bone tracers. Osteosarcomas typically appear as areas of markedly increased uptake on the bone scan. Bone scan cannot be used to assess the extent of the Ewing sarcoma because the area of increased uptake is usually larger than the extent of the tumour. Bone scans in osteosarcoma and Ewing sarcoma are used to evaluate metastatic disease at initial staging and follow-up (Figure 1.10.20). Bone scans are of limited value in assessing response to neoadjuvant chemotherapy because of persistent increased uptake related to bone remodelling.

Many benign bone tumours also show intense uptake of the tracer on bone scan. Therefore bone scans cannot be used to differentiate between benign and malignant lesions. Bone scans can be used to screen for polyostotic disease in fibrous dysplasia and enchondroma.

Bone scans are highly sensitive for diagnosis of occult fractures in patients with pain and a negative x-ray. Bone scans are done most frequently to evaluate fractures in the wrist, hips and spine, but become positive within 24h after any fracture in 90–95% of patients. In elderly patients, bone scans may only become positive after some time and repeat imaging at 72h is suggested if the initial bone scan is negative. Bone scans may remain positive for up to 3 years after a fracture.

Bone scans can be also helpful to evaluate post-traumatic complex regional pain syndrome, type 1 (CPRS-1; reflex sympathetic dystrophy). The typical pattern of CRPS-1 is diffuse increased bone tracer uptake of the involved extremity, most prominent in periarticular regions. The flow and vascularity may be increased in the initial phase of disease, but may be negative in the late phase of the disease after 60 weeks of onset of symptoms. Bone scans are also used in the management of heterotopic ossification. There is intense soft tissue uptake of the radiopharmaceutical during formation of heterotopic ossification, but this normalizes or shows only mild uptake in mature heterotopic ossification.

Early stress fractures are not usually visible on plain radiographs. A negative bone scan excludes stress fracture lesions even in clinically suspected cases. The typical finding of a stress fracture on bone scan is focal fusiform increased uptake at the site of injury (Figure 1.10.21). Bone scintigraphy is also helpful to differentiate stress fractures from shin splints. The typical presentation of shin splints on the bone scan is a linear uptake along the posteromedial aspect of the middle third of the involved tibia (Figure 1.10.21).

A positive bone scan in the pars region indicates spondylolysis as the cause of low back pain and correlates with good outcome after fusion surgery. Tomographic images of bone scintigraphy (SPECT) of the lumbar spine need to be obtained as more than 50% of active spondylolysis may not be detected with routine planar bone scans (Figure 1.10.22).

A three-phase bone scan is the initial test of choice for diagnosis of osteomyelitis after a negative/inconclusive x-ray. Increased bone tracer uptake reflecting bone remodelling may, however, also be seen after fracture, surgery and hardware placement. In these cases labelled WBC scans are used to complement bone scans.

Labelled WBCs, which can be labelled with ¹¹¹In or ^99mTc, do not accumulate in fractures or sites of orthopaedic surgery in the absence of infection (Figure 1.10.23). Labelled WBCs, however, accumulate in normal bone marrow. If bone marrow distribution has changed secondary to surgery or prosthesis or if there is active marrow proliferation, such as seen in patients with diabetic arthropathy, the labelled WBC scan may show increased/asymmetric uptake in the absence of infection. Therefore labelled WBC imaging should be done in conjunction with a bone marrow scan to differentiate infection from changes in bone marrow.

The most commonly used radiopharmaceutical in clinical positron emission tomography (PET) imaging is the glucose analogue 2-[¹⁸F] fluoro-2-deoxy-D-glucose (F-18 FDG). FDG uptake is enhanced in the malignant cells because of the enhanced transport of glucose (increased number of glucose transporter proteins on cancer cells) and because of the inefficient use of glucose for energy generation in malignant cells. After FDG enters the cells it is phosphorylated in the same way as glucose but does not enter further metabolic pathways and therefore accumulates inside the cell. FDG uptake is not specific for tumours and can be also seen with infections and inflammation.

F-18 FDG decays by positron emission. The emitted positron has a very short range and readily captures an electron when it comes to rest, releasing two photons.

FDG PET has been used for diagnosis and grading of soft tissue sarcomas. FDG PET is also positive in most bone sarcomas. False positive PET is seen in a number of benign bone lesions including giant cell tumours, fibrous dysplasia, eosinophilic granuloma, chondroblastoma, aneurysmatic bone cysts, non-ossifying fibroma, and osteomyelitis.