Computed Tomography of the Head, Cerebral Vessels, Neck, and Spine

Head and Face

CT is valuable for assessing intracranial pathological conditions, especially in the acute stage. It is performed to guide further investigations and therapy. It is fast and causes no security concerns, in contrast to magnetic resonance imaging (MRI). (Patients must be screened before being positioned in the MRI machine, which causes delays.) Motion can cause artifacts, but it is less of an issue than with MRI. (Terminology commonly found on imaging requests is written in italics, to help the student identify conditions. Not all the conditions mentioned later are illustrated.)

A common indication for CT scanning is acute neurological dysfunction, often presenting as paralysis (stroke) or transient ischemic attack (TIA). The neurological deficit can be caused by a cerebral infarction from blockage of a blood vessel in the neck or head, by a spontaneous brain hemorrhage caused by hypertension or an underlying blood vessel abnormality such as an aneurysm, by an infection (viral encephalitis or bacterial abscess), or by a primary or metastatic tumor.

Cerebral infarction is the result of loss of adequate blood supply to a portion of the brain. In the early stages, it may be difficult to see the location of the infarcted tissue on CT. Early edema (as seen in Fig. 17-1, A) causes loss of gray-white matter differentiation, which is the earliest CT sign of infarction. As the infarct matures, it becomes more obvious and better defined. Because of the increased water content of the infarcted tissue, the density is decreased (i.e., dark) (Fig. 17-1, B).

Hemorrhage in the brain substance (Fig. 17-2, A) can occur spontaneously, either because of an underlying vascular anomaly, such as a tumor or an arteriovenous malformation (AVM) or because the patient’s hypertension has weakened the walls of the capillary vessels. Hemorrhage can also occur in the subarachnoid space (Fig. 17-2, B), resulting in bloody cerebrospinal fluid (CSF) (from a leaking aneurysm) or between the brain and its meninges (epidural and subdural spaces). Acutely extravasated blood is of increased density (i.e., white). The degree of mass effect caused by the hematoma is reflected in the amount of brainstem compression from the shift of the midline structures. This affects the patient’s level of consciousness, which is measured by the Glascow Coma Scale (range 3-15). A normal score is 15 and the lowest score is 3. Delay in scanning any patient with a score less than 13 or 14 after a head injury should be avoided because this score can change in less than 5 minutes in the presence of an expanding intracranial hematoma. The technologist should always advise the medical personnel immediately if a dramatic change in level of consciousness is seen while the patient is in the scanner. (This is one type of patient who should not be left unsupervised, while waiting for transport, for example.)

Hematomas evolve in density over time, becoming less and less dense and finally approaching CSF in density. The clot will be jellylike in the acute phase and will eventually liquefy and shrink. Chronic subdural hematoma is the most common condition followed up by repeated CT scanning to document its complete disappearance. Reaccumulation of a chronic subdural hematoma is also common, requiring repeated drainage.

When a traumatic brain or facial injury is suspected, CT provides rapid information about contusions (brain bruise) and hematomas (blood clot) in the brain and in the spaces between skull and brain (epidural and subdural hematomas). Facial and skull fractures are easily seen (Fig. 17-3, A), except linear cranial vault fractures coursing in the plane of section. The depth of depressed bone fragments can be measured (Fig. 17-3, B). The results of corrective surgery and the location of metallic hardware in the face, as well as the degree of healing callus, are easily appreciated on follow-up examinations.

The most common brain tumor is a metastasis; metastases are often multiple (Fig. 17-4, A). Common primary brain tumors are meningioma and glioblastoma multiforme (GBM) (Fig. 17-4, B). Their location and pattern of enhancement is used to predict their histological grade preoperatively.

Stereotaxic procedures by various methods are commonplace in neurosurgery. They are used to accurately localize the lesion before biopsy or resection. The preoperative studies can be obtained with fiducial markers applied to the scalp or with the head in a stereotaxic frame. Different systems are available, with some frames attaching to the CT table, in place of the head holder. Newer systems are frameless, using a combination of computer programs to superimpose the images and correlate them to the patient’s anatomy.

In the case of meningitis or encephalitis, it is useful to analyze the CSF for the presence of cells and various compounds, such as glucose, proteins, antibodies, and bacteria and viruses. A lumbar puncture (LP) can only be safely performed if the possibility of an intracranial mass (hydrocephalus [Fig. 17-5, A], tumor or abscess) has been ruled out. Many physicians prefer the certainty of CT before proceeding with the LP because signs of an intracranial mass are not always present. Brain atrophy can also result in large ventricles and must be distinguished from hydrocephalus (Fig. 17-5, B).

Many congenital brain abnormalities are diagnosed in childhood, whereas others are only recognized in adulthood. Some types of seizure disorders are caused by localized arrested brain development (for example, cortical dysplasia). Surgery may alleviate the seizures. Some brain anomalies remain asymptomatic and are an imaging curiosity (for example, cavum of the septum pellucidum, cisterna magna).

Apart from the brain, conditions affecting other components of the head, such as the sinuses, temporal bones, and orbits are commonly evaluated by CT scanning.

Sinus disease (sinusitis, tumor) is readily seen with CT (Fig. 17-6, A). Sinus anatomy shows many variations important to the ear, nose, and throat surgeon. To avoid complications such as entering the brain or orbit inadvertently, sinus surgery is often performed with direct correlation with the CT images, a procedure known as functional endoscopic sinus surgery, or FESS.

The mandible, being a curved bone, is not optimally assessed for pathological conditions with plain films; it is easily evaluated with CT by using two-dimensional and three-dimensional (3D) reformatted images; a 3D reformatted mandible is shown in Figure 17-6, B.

Orbital CT scan images should always be available in at least two planes (transverse and coronal) because of the pyramidal shape of the bony orbit. The intraorbital fat delineates the optic nerve, extraocular muscles, and superior ophthalmic veins. For example, CT accurately localizes radiopaque foreign bodies for the surgeon. Wood is not radiopaque and may mimic air, so bone windows should be available (shown in Fig. 17-3, C). Although orbital ultrasonography is used widely to study the ocular globe, CT is better at characterizing orbital mass lesions and identifying the cause of proptosis (displacement of the globe).

Other uses of CT include looking for pituitary microadenoma in the pituitary gland situated in the sella turcica (cause of many endocrine problems) and masses in the middle ear (cholesteatoma) and internal auditory canal (acoustic neuroma) as causes for hearing loss (Fig. 17-7).

Cerebral Blood Vessels

Helical scanning makes it possible to image structures during the continued intravenous (IV) injection of a bolus of iodinated contrast medium. Multislice CT scanners extend the length of the vascular tree that can be imaged with one contrast injection; because of the speed of image acquisition, a pump is used to inject the contrast. A greater scan length (i.e., length of travel of the tabletop) is possible with a greater number of simultaneous slices. At this time, the artifacts caused over the carotid arteries by metal in the mouth (dental fillings, caps, braces) cannot be completely eliminated. Magnetic resonance angiography (MRA) or catheter angiography may be necessary to rule out or confirm a vascular abnormality suggested by CT angiography (CTA).

CTA can outline the cerebral blood vessels from their origin on the aortic arch to their various intracranial branches, as shown in Figure 17-8, A and B. Because of the ease and convenience of CTA, the number of patients being investigated for vascular conditions has multiplied. The vessels supplying the brain originate from the aortic arch (the innominate or brachiocephalic trunk, left common and left subclavian arteries). Both the subclavian and the vertebral arteries, the common carotid artery bifurcations (site of most stenoses), the internal carotid arteries, the intracranial arteries (anterior, middle and posterior cerebral, basilar, and cerebellar) arteries (aneurysm sites) are readily visible (Fig. 17-9).

A delay of a few seconds between injection and acquisition of images of the entire brain increases the visualization of the jugular veins and dural sinuses (sagittal and transverse sinuses), sites of possible thrombosis. This is known as a CT venogram (CTV) (not shown).

An important use of CTA is in suspected acute cerebral infarction because rapid intervention may result in lessening the extent of brain damage. After a non contrast CT of the brain to rule out hemorrhage, CT perfusion (CTP) at the level of the suspected infarct is obtained (in some centers) to delineate the amount of brain at risk for permanent damage; CTA of the cerebral vessels follows to identify the presence of clot in a major intracranial artery (usually the middle cerebral or the basilar artery) before a “clot-busting” drug, known as a thrombolytic agent (such as thrombin plasminogen activator, or TPA) is administered.

Patients with a hemorrhage caused by, for example, a suspected aneurysm or an AVM can undergo a CTA to confirm the diagnosis before being transferred to a more specialized treatment center. This is particularly useful in smaller or more remote localities where a neurosurgeon or neurologist is not available.

Spine and Neck

In many centers CT has replaced plain films of the cervical spine in cases of trauma because it shows injuries not detectable otherwise (see examples of fractures in Fig. 17-10). A certain amount of “pathology” overlap is seen in the cervical region because conditions affecting the vessels, the bones, and the soft tissues are often found when one type of cervical examination is being carried out. This is most evident when the study is a CTA covering the neck; in addition to bone and vessels, abnormalities of the thyroid gland, the salivary glands, lymph nodes, and neck muscles can be appreciated.

In the case of trauma, the carotid or vertebral arteries may be injured (resulting in a dissection or occlusion) during the violent neck motion that caused a spinal fracture or dislocation. Patients with carotid stenosis caused by arteriosclerosis are usually older; the cervical spine often shows disk herniation or osteophytes causing nerve root compression or spinal stenosis. The converse is also true. Studies looking for cervical disk herniation may also show carotid bifurcation calcification, a possible cause of symptomatic carotid artery disease and a potential cause of stroke or TIA.

Masses can occur along the chains of lymph nodes in the neck (for example, lymphoma, Hodgkin’s disease), along nerves (such as vagus nerve schwannoma), or adjacent to the vessels (carotid body tumor; Fig. 17-11, A). Tumors occurring in the salivary glands, nasopharynx, tongue, and floor of the mouth can metastasize to the neck nodes, changing the tumor grade and treatment protocol and the survival rate (Fig. 17-11, B).

Preservation of the voice can be achieved in localized laryngeal cancers when treatment consists of a partial laryngectomy. This emphasizes the need for accurate delineation of tumor.

Enlargement of the thyroid gland as the result of a goiter is a frequent incidental finding when the neck is imaged for any reason (Fig. 17-11, C). When large, a goiter can plunge into the superior mediastinum and cause airway obstruction. Thyroid hormone production can be abnormal, resulting in various symptoms.

Many infections start in the disk space (diskitis, osteomyelitis) at any level of the spine; they can spread to the epidural space (epidural abscess), resulting in spinal cord compression and paralysis. Other infections extend into the prevertebral or paraspinal tissues; in the neck, they can obstruct the airway or cause difficulty swallowing and drooling. An acutely swollen neck can be life threatening because of the potential for obstruction of the airway, particularly in children.

Tumors affecting the spine or spinal cord can originate as primary tumors or metastasize from other sites to bone or soft tissue, affecting the vertebral bodies, the pedicles, laminae, and facet joints. Their treatment often requires removal of much bone, resulting in possible instability of the spine. Bone grafts and metallic instrumentation are frequently inserted to preserve patient mobility. Metallic instrumentation may fail, requiring frequent follow-up by imaging (plain films/CT) and occasionally surgical revision.

Instrumentation of the spine is also used in treating atlantoaxial (C1-2) instability in rheumatoid arthritis and in treating scoliosis, spine fractures, spinal stenosis and spondylolisthesis. It requires accurate placement of pedicle screws, hooks, wires, and plates. CT accurately depicts the position of the instrumentation. In some cases, screws may traverse a foramen and impinge on a nerve root or a vertebral artery; a neurologic deficit or pain can ensue (Fig. 17-12, A).

Another type of spinal procedure is vertebroplasty. Radiopaque cement (methylmethacrylate) is injected under fluoroscopic control into spinal compression fractures to reduce pain and in some cases to decrease the kyphosis.

Postprocedure CT scans of the many types of spinal interventions described previously are common to assess the results (Fig. 17-12, B and C).

Head

Neck

When an IV contrast medium is used to enhance images, major vascular structures such as the carotid arteries and jugular veins can be identified. Both the superficial mucosa and the submandibular glands enhance. The parotid gland, with its high fat content, demonstrates an intermediate density between those of fat and muscle (Fig. 17-20).

Thin sections of the larynx show the relationship of the cartilages to the adjacent soft tissue. The level of the true vocal cords is defined by the position of the arytenoid cartilages (Fig. 17-21).

Spine

The sectional anatomy of the lumbar and cervical spines is illustrated in Figures 17-22 and 17-23, respectively. Spine images are acquired in the transverse plane, either helically or by varying the gantry tilt. Coronal or sagittal images can now be obtained easily with multiplanar CT scanners. CT differentiates the disk from the adjacent ligamentum flavum, thecal sac, intraspinal fat, and bone. The nerve roots are easily identified as they exit through the intervertebral foramina. Epidural veins can be seen, particularly in the lumbar spine, but are better visualized with contrast medium enhancement.

• Metallic objects: The area to be scanned should be free of all jewellery, hairpins, metal gown snaps, wires attached to the patient, and other devices. They can be moved out of the field when they are noted on the scout (scanogram, topogram) images, without having to repeat the scout in most cases. Glasses are removed but contact lenses remain in place for brain scanning.

• Fasting is not required. However, if IV contrast is being used, a large meal immediately before the CT scan is discouraged in case the contrast causes nausea or vomiting.

• An empty bladder is important, especially if the time in the CT scan may be prolonged, as for a CT-guided biopsy or nerve root block.

• Patient anxiety may be eased by the technologist’s calm demeanor. After confirming the patient’s identity, the technologist introduces himself or herself, gives a short explanation of the procedure and the reason for using contrast, and describes expected sensations (heat, noise, table movement, need for staying still).

• Contrast injection requires a review of known drug/contrast allergies and of the renal function to minimize contrast reactions and kidney impairment resulting from contrast-induced nephropathy. (This step may be done before the study is scheduled, but the technologist should be aware of this information before contrast is administered.) The appropriate needle size and injection site for the requested study can then be chosen.

• Cardiac medications may indicate poor cardiac function. This is relevant information for CTA.

• A comfortable position on the scanning table, with holding straps and warm blankets or cushions under the knees and elbows, may result in less motion, reducing image repetition and radiation dose.

• The eyes should be closed when the laser positioning light is used near the eyes.

• Combative or uncooperative patients (because of, for example, head injury, psychiatric problems, dementia, and alcohol) require special precautions. More restraints than usual may be needed to obtain a diagnostic study. Use of rapid helical scanning (for example, 800 milliseconds, instead of the usual 1000 to 2000 milliseconds) reduces the resolution between gray and white matter, but the study will be of sufficient quality to demonstrate life-threatening lesions, such as an intracerebral hematoma and the degree of midline shift it has caused, necessitating urgent surgical intervention. The technologist must be very vigilant to avoid personal injury.

• Body fluid protection should always be observed (the minimum is gloves when dealing with a trauma patient). When appropriate, the patient or the technologist will need to wear a mask and gown (e.g., carriers of antibiotic-resistant bacteria, termed MRSA (for methicillin-resistant Staphylococcus aureus) or VRE (for vancomycin-resistant Escherichia coli).

Head and Its Contents

Two transverse scanning planes are widely used in brain imaging. In adults, the orbitomeatal baseline (OM line, also called Reid’s baseline or RBL) joining the infraorbital rim to the superior border of the external auditory meatus is favored because it results in better images of the orbits, sella turcica, posterior fossa, and brainstem. Some centers use the canthomeatal line (CM line), which joins the lateral canthus of the eye to the center of the external auditory canal. The angle measured between these two lines is 10 degrees.

In children, a steeply angled plane, commonly 20 to 25 degrees to the OM or CM line is used to scan the brain to avoid radiating the ocular lens leading to early cataracts (see Fig. 17-26, A, later in this section).

The transverse plane is still the most used plane for image interpretation of the skull and its contents. If reformatted images are required, it is better to acquire thin slices initially and to reformat the images. This technique is useful when scanning the orbits; for example, the lenses of the eyes are scanned only once, dental amalgam artifact is eliminated, and images in true coronal plane (perpendicular to the hard palate) are generated. Patient comfort is also increased compared with direct coronal imaging, when the neck must be hyperextended to achieve a near coronal position and the gantry tilted appropriately. Coronal sections are particularly useful when structures that run parallel to the transverse plane are assessed, such as the floor and roof of the orbits, the sella turcica and adjoining cavernous sinuses, and the skull base and vertex.

Exceptions to this approach exist. Direct coronal images of the temporal petrous bones/internal auditory canals and sella turcica are preferred because the resolution of the 0.6 to 1.25 millimeter (mm) slices is greater than with reformatted images and the orbits are avoided. Prone direct coronal images are also preferred in some centers for the study of paranasal sinuses because air-fluid levels tend to pool away from the path of drainage, known as the ostia, resulting in a better study than the supine transverse or hyperextended (supine coronal) positions. Thin transverse images with sagittal and coronal reformatted images are used when extension of the neck is not possible.

Sagittal images have not been used routinely in the past but are gaining in popularity. They are particularly useful when assessing midline structures and midline abnormalities of the brain and spine.

The algorithm (or kernel) chosen will affect the resolution of the images. In some cases, it may be possible to acquire diagnostic images using a low milliamperage, as with paranasal sinuses or facial bones. Fractures, sinusitis, and nasal polyps can easily be assessed this way, using only a bone algorithm or equivalent kernel. Temporal (petrous) bones, on the other hand, need a higher milliamperage to get accurate detail of the densest bone in the body and its small structures, such as the cochlea, semicircular canals, and ossicles.

Contrast enhancement is not visible on bone algorithm or bone windows. When contrast is used, images on soft tissue algorithms (standard or equivalent) are also required to see the enhancing tissue, such as with sinus tumors and chronic sinus infections.

Neck, Spine, and Cerebral Blood Vessels

Traditionally, the transverse plane was the preferred plane for interpretation. However, multiplanar reconstructions in coronal, sagittal, and oblique planes greatly enhance the radiologist’s ability to diagnose conditions affecting these areas, which contain bony curves (cervical lordosis, mandible), soft tissue curves (carotid bifurcation, cervical disk), and many overlapping organ systems (salivary glands, airway, lymph nodes).

To avoid artifacts on soft tissue algorithm caused by dental amalgam and capped teeth, when possible, transverse angled images on either side of the teeth can be obtained by tilting the gantry. The artifacts are not as distracting if only a bone algorithm (kernel) is needed, for example, to study only the bones; in that case, a single set of images may be acquired. Newer computer software allows reformation of images scanned with a different gantry tilt. This software may be on the scanner workstation or on the PACS (picture archiving communications systems) unit used for reporting.

Milliamperage, Kilovoltage, Helical Pitch, and Rotation Time

The preset techniques and protocols supplied by manufacturers are helpful. However, the technologist must be able to manipulate milliamperage, kilovoltage, helical pitch, scan time, and slice thickness to suit the circumstances.

Increasing the milliamperage and the kilovoltage will improve the signal-to-noise ratio. Sufficiently high technique is necessary for contrast resolution of soft tissue in the brain (gray-white matter differentiation). Similarly, spine imaging in large patients can be improved by increasing the technique (by using the maximum kilovoltage and increasing the scan time).

Increasing the scan rotation time will result in better resolution because each image is generated from more projections.

Increasing the helical pitch results in a shorter scan time but reduces the overall image quality because of more interpolation in the scan data and vice versa.

Longer scan times can result in longer tube cooling and the possibility of patient motion (often a result of back pain).

Slice Thickness

Slice thickness should be adapted to the area of interest. Studies requiring reformatted images are acquired helically, with thin slices. Sequential image acquisition is preferred for brain, petrous bones, and neck masses, for better resolution.

Thin transverse sections show the base of the skull well, giving excellent definition of the temporal bones, pituitary fossa, facial bones, and foramina. Thicker slices increase the definition of gray and white matter in the brain.

The low density of various brain tissues, such as fat (in the myelin of white matter) and CSF, outlines components of the brain such as frontal and temporal lobes, basal ganglia, and ventricles. The high density of calcium is found in many normal structures inside the skull (pineal gland, choroid plexus) and must be differentiated from pathological calcium (tumor, AVM).

Spatial resolution is improved with thinner slices, at the cost of a decreased signal-to-noise ratio. Increasing the milliamperage and kilovoltage will improve the signal-to-noise ratio but will also increase the radiation dose.

The average slice thickness has decreased over the years, with the old 10-mm slice of the brain now being 2.5 to 5 mm. Our current practice for a routine head scan consists of 2.5- to 3-mm sequential slices in the posterior fossa and 5-mm slices for the rest of the head.

Orbits are studied with 1- to 2-mm helical slices and a small FOV (13-15 cm).

Temporal bone images are sequential 0.6 to 1 mm thick with a 10-cm field of view.

CTA image acquisition is 1-mm helical. CTP, on the other hand, uses cine mode and 5- to 10-mm thick slices; the tabletop may toggle back and forth during CTP image acquisition, depending on the scanner.

The cervical spine is imaged with helical 1- to 2-mm slices every 1 to 2 mm when good bone detail is needed. On the other hand, 2.5- to 3-mm sequential slices overlapped every 2 mm are used to look for a disk herniation. Wide shoulders present a challenge. Enlarging the field of view and increasing the technique may not always be successful in preventing overrange artifacts from the humeral heads. A swimmer’s position is used in some settings to decrease these artifacts.

Lumbar images consist of 3- to 5-mm sequential slices every 3 to 4 mm. One method uses overlapping thicker slices (for example, 7-mm slices every 4 to 5 mm) to improve visualization of herniated lumbar disks, for example, when scatter radiation is increased because of a large girth.

Matrix Size and Reconstruction Algorithms

Matrix size is predetermined at 512 × 512 on most current equipment. Some systems have a 1024 × 1024 matrix (e.g., Philips).

Many reconstruction algorithms are available. The three basic algorithms used for the brain, neck, and spine are the standard, the bone, and the detail algorithms (General Electric) or the Siemens kernels (31, 70, and 45 seconds).

The standard algorithm (and its equivalents) is chosen for most soft tissue imaging, such as the brain and the intervertebral disk, because it provides good contrast resolution. It is important to distinguish gray from white matter when looking for acute cerebral infarction.

The detail algorithm is useful when studying the soft tissues of the neck. It increases edge definition, especially when good definition of lymph nodes and fat planes is required.

The bone algorithm is used to increase the sharpness of the bony detail. This algorithm optimizes spatial resolution; contrast resolution is poor. This is why it should not be used in isolation when IV contrast is administered because tissue enhancement is not visible on bone windows or bone settings. However, bone algorithm is the preferred technique when subarachnoid (intrathecal) contrast has been injected by LP. This is true whether the imaging is of the spine (post myelography CT) or of the brain (CT cisternography).