Radiography provides a familiar “global” perspective that is particularly advantageous when assessing congenital vertebral anomalies and vertebral instability using dynamic views (e.g., atlantoaxial subluxation; Figure 28-2), or when needing accurate landmarks for surgical planning such as rib identification, as the presence of transitional vertebrae may be difficult to discern on tomographic images. Additionally, radiographs have the highest spatial resolution (i.e., the ability to resolve fine detail) of all the modalities and therefore may be required to further elucidate subtle lysis or periosteal new bone and fracture healing response (see Figure 28-1), even after identification with CT or MRI.

Figure 28-2 A, Static neutral midsagittal T2W magnetic resonance (MR) image and (B) dynamic flexed lateral radiograph of the cranial cervical vertebral column of a dog. On the MR image, compression of the cervicomedullary aspect of the spinal cord by a dorsally displaced dens is evident (arrow) and suggests atlantoaxial subluxation; however, the cause is further elucidated on the radiographs in a flexed position. Caudal displacement of the spinous process of the axis (white arrow) and dorsal displacement of the dens (black arrow) are compatible with atlantoaxial instability.

Myelography is a radiographic examination of the spinal cord and vertebral column performed after the injection of contrast medium into the subarachnoid space. Abnormalities must be inferred by alterations in the contrast medium–filled subarachnoid space. Because it is invasive and is associated with adverse effects,141,273 including the intensification of preexisting neurologic signs, myelography has relatively few indications if MRI or CT is available. As with routine radiography, the perspective and resolution of dynamic cervical myelography may be necessary to aid in the diagnosis of cervical vertebral malformation/instability. Myelography is also occasionally performed with spinal CT, particularly when abnormalities identified on CT images need further clarification.

Tomography: computed tomograpHy and magnetic resonance imaging

CT and MRI produce images that are tomographic or slice oriented, thereby eliminating the summation effects and depth perception losses associated with radiography. In addition, CT and MR images are digital; therefore, they depict differences in tissue composition more faithfully than radiographs and can be manipulated interactively with appropriate software to produce views in any desired plane or in three dimensions. These advantages often make spatial relationships easier to conceptualize, and thus aid in surgical planning (see Figure 28-1).

Choosing the Most Appropriate Tomographic Imaging Modality

Contrast Resolution: Given its superior ability to discriminate between soft tissues, including gray and white matter (Figure 28-3), MRI is the gold standard imaging modality for evaluation of the central and peripheral nervous systems. Likewise, its exceptional contrast resolution permits unprecedented evaluation of soft tissue structures outside of the nervous system (Figure 28-4), thus making MRI an increasingly popular choice for a growing number of disorders.

Figure 28-3 T2W fast spin echo (A) and postcontrast T1W (B) MR images of the brain of a dog with a cerebral mass suggestive of round cell neoplasia. Note the exceptional intracranial contrast resolution. The marrow fat of the calvaria is very T1 hyperintense and moderately T2 hyperintense. The dense cortical bone has no signal (black) because it lacks hydrogen protons. Normal brain tissue is light gray, and cerebrospinal fluid in the right lateral ventricle is T1 hypointense and very T2 hyperintense. In the T2W images, the normal myelinated white matter of the right cerebrum (arrow) is hypointense to the gray matter. Peritumoral vasogenic edema is present in the left cerebral white matter. This has prolonged T1 and T2 relaxation times, making the white matter T1 hypointense and T2 hyperintense. As expected for vasogenic edema, the T2 hyperintensity tracks along the centrum semiovale and into the corona radiata, extending into individual gyri of the cerebrum, along the internal capsule, and into the diencephalon. In A, the borders of the mass are difficult to distinguish from the vasogenic edema in the left cerebrum. However, in B, marked gadolinium enhancement of the mass occurs, consistent with vascularization and lack of a functional blood-brain barrier.

Figure 28-4 A, Transverse T2W magnetic resonance (MR) image and (B) corresponding postcontrast computed tomography (CT) image (displayed with soft tissue [left] and bone [right] windows) of the sixth cervical vertebra in a dog with metastatic carcinoma to muscle. In A, greater conspicuity of the metastatic tumor (arrow) is evident because of the superior contrast resolution of MRI. In B, only faint iodinated contrast enhancement (black arrow) is seen; however, the fine periosteal reaction (white arrow) and the intact underlying cortical bone are revealed, caused by spatial resolution of CT that is superior to MRI.

Although superior to radiography, the contrast resolution of CT is inferior to MRI (see Figure 28-4). Likewise, intracranial CT suffers from several artifacts such as beam hardening that affect the ability to image areas such as the caudal cranial fossa. This can be remedied somewhat by use of thick section reformatting of thin slices.²⁰⁶ Ultimately, the use of CT to evaluate intracranial structures is limited (see below, Additional Advantages and Disadvantages of CT and MRI).

Within the vertebral column, resolution of modest differences in attenuation with CT is limited by the small size of the vertebral canal in dogs and cats. Within the vertebral canal, contrast is largely provided by epidural fat (Figure 28-5). When pathology reduces epidural fat, identifying the underlying lesion or its relationship with the meninges and spinal cord may be difficult. Loss of contrast can be compensated for by the use of myelography with CT; added contrast between the subarachnoid space and the spinal cord helps to delineate intradural and extradural structures (Figure 28-5, C). In some instances, pathology may increase contrast in the epidural space by the addition of tissue with greater attenuation. For example, acute hemorrhage and mineralization are usually readily identified, as they are relatively hyperattenuating (bright) in comparison with the spinal cord (see Figures 28-1 and 28-5). For this reason, CT can be used in the evaluation of chondrodysplastic dogs with suspected intervertebral disc disease, with the assumption that the herniated disc material will contain mineral or hemorrhage.¹⁸⁶ It is critically important to recognize that the choice to use CT is based on the validity of two assumptions: that the cause of the clinical signs is the result of intervertebral disc disease, and that the herniated disc material will be mineralized. When these two assumptions can be made with a high degree of certainty (i.e., when the patient is a chondrodysplastic dog with an acute onset of clinical signs), CT may be an efficient and effective imaging modality to use. However, in nonchondrodysplastic dogs or in chondrodysplastic dogs in which the most likely diagnosis is not intervertebral disc disease, CT may not resolve the lesion. In such cases, MRI or CT with myelography should be considered.

Figure 28-5 A–B, Transverse noncontrast computed tomography (CT) (soft tissue window) and (C) CT myelogram (bone window) images of a dog with a fracture-subluxation at the T12-T13 intervertebral space. In A, the normal spinal cord is surrounded by hypoattenuating (minus 100 Hounsfield units) epidural fat (arrows). In B, at the level of the T12-T13 intervertebral disc, the epidural fat is effaced, and hyperattenuating material (85 to 110 Hounsfield units) is noted in the ventral half of the canal, compatible with mineralized disc material and possible hemorrhage. The addition of subarachnoid iodinated contrast medium (myelogram) permits delineation of intradural and extradural structures and confirms extradural spinal cord compression. Subluxation of the dorsal articulation and fracture fragments ventral to the vertebral body are evident.

Spatial Resolution: CT affords a relatively detailed evaluation of bone, as its spatial resolution is superior to MRI (see Figure 28-4). On an MR image, tissues devoid of hydrogen protons will have no signal; dense cortical bone is therefore very hypointense (black) on an MR image (see Figure 28-3; Figure 28-6), making subtle fracture lines, cortical lysis, or periosteal new bone (see Figure 28-4) difficult to resolve. This does not mean, however, that MR images have no utility in bone evaluation. To the contrary, when pathologic soft tissues infiltrate and replace the normal signal void of bone, bony abnormalities become apparent (see Figure 28-6). The same applies when the hyperintense (bright) signal of fat within bone marrow is altered by pathology (see Figure 28-6). Thus, although CT is generally thought to be a superior bone imaging modality, the opposite may be true with an infiltrative type of pathology.

Figure 28-6 Two examples of bone alteration seen on magnetic resonance imaging (MRI). A, Transverse postcontrast T1W image of a cat with a left cerebral meningioma. Hyperostosis of the adjacent calvaria (black arrows) is evident. The calvaria appears thickened, and the normally hyperintense signal in the fat within the diploe (white arrow) is not evident, indicating replacement of the marrow fat with dense mineral, which is largely devoid of hydrogen atoms. The broad attachment of the mass along the calvaria combined with the change identified in the overlying bone suggest an extra-axial location of the mass. The well-demarcated margins and the strong uniform enhancement pattern are consistent with meningioma. Note the marked mass effect with displacement of the left cerebrum and diencephalon created by the meningioma. In addition, note the normal enhancement of the pituitary gland related to the lack of a complete blood-brain barrier (red arrow). B, Sagittal T2W image of the humerus of a dog with osteosarcoma. A fat saturation technique has been applied, so if normal, the entire humerus should appear dark (because of suppression of marrow fat), similar to the humeral condyle and scapula (white arrows). Instead marked T2 prolongation of nearly the entire humerus and adjacent soft tissues is evident. A portion of the cranial cortex (underlying periosteal new bone; arrowheads) appears irregular with faintly increased signal, indicating tumor infiltration.

With CT, sagittal and dorsal plane images must be retrospectively reformatted or reconstructed from the transverse ones, usually resulting in images with inferior spatial resolution. Nevertheless, these reformatted images are often satisfactory, especially for bone and other high contrast structures (e.g., contrast-filled subarachnoid space). This is especially true for images produced with multidetector scanners.

With MRI, images from each anatomic plane are obtained using separate acquisitions. Because they are acquired directly, the spatial resolution of sagittal and dorsal plane MR images is generally superior to that of reformatted CT images. Although this makes an MRI study more time consuming, the added perspective gained by direct multiplanar MR images is invaluable for elucidating the precise localization or extent of disease (Figure 28-7).

Figure 28-7 A, Transverse postcontrast T1W with fat saturation and (B) sagittal T2W images of the cranial cervical vertebral column of a dog. In A, a large contrast-enhancing intradural mass (arrow) is seen. In the sagittal plane orientation (B), a “golf tee sign,” created where the hyperintense cerebrospinal fluid in the dorsal subarachnoid space abuts the mass (arrows), is evident, indicating extramedullary localization. The mass is compatible with meningioma or a nerve sheath neoplasm.

Additional Advantages and Disadvantages of CT and MRI: Because CT readily identifies acute hemorrhage and fracture, it is often considered the first line for imaging in patients with neurologic trauma (see Figures 28-1 and 28-5; Figure 28-8). Moreover, the time needed to acquire CT images is generally shorter than for MRI; this may be advantageous in the immediate posttrauma setting, especially if monitoring or access to the patient inside the magnet gantry is limited. In addition, because each MR sequence takes several minutes, patient and physiologic motion can degrade image quality. Likewise, metal in the scan field (e.g., orthopedic implants, drill bit shavings⁶⁵ [Figure 28-9], identification microchips, gunshot pellets) can render all or parts of an MR image nondiagnostic. Despite all of these drawbacks, MRI may still be necessary if the contrast resolution of nervous tissue is insufficient with other modalities.

Figure 28-8 Transverse noncontrast computed tomography (CT) image of a dog with head trauma. A biconvex hyperattenuating mass (74 Hounsfield units) is noted overlying the region of the right parietal lobe of the cerebrum. On postcontrast images (not illustrated), the enhancing dura was noted on the mesial side of the mass. Findings are compatible with an epidural hematoma. The lesion is identifiable as a result of the difference in attenuation properties compared with normal brain (30 to 40 Hounsfield units).

Figure 28-9 Transverse noncontrast T1-weighted magnetic resonance (MR) image of a dog following craniotomy for meningioma resection. Multiple signal voids surrounded by hyperintense rims (arrows) are noted within the extracranial tissues; one obscures the craniotomy site (white arrow), the other occurs dorsally in the subcutaneous tissue (longer white arrow). These represent artifacts caused by metallic drill bit shavings. The tiny shavings were not visible radiographically. Also noted is enlargement of the left lateral ventricle, likely a compensatory effect of tissue resection.

Both CT and MR imaging use contrast media to help improve contrast resolution and provide insight into the vascularity and perfusion of pathologic processes. Intravenous administration of both iodinated CT and gadolinium-based MRI contrast medium agents is associated with moderate hemodynamic alterations but no adverse systemic effects in dogs ²⁰³ and cats,²⁰⁴ whereas serious side effects are known to occur in people.^142,170,215

Angiography may be performed without selective catheterization with both CT and MRI, and in some MRI studies, without the need for contrast media. Dynamic imaging of the vertebral column (images acquired in different anatomic positions [flexed, extended, or distracted]) can be performed with MRI or CT, although the ability to obtain direct sagittal images usually makes this more feasible with MRI (Figure 28-10).¹⁹⁵ MRI can provide physiologic information through the use of specialized pulse sequences; these include brain perfusion, brain diffusion, spectroscopy, and brain activation (also known as functional MRI, or fMRI).^8,55,101,213 Finally, MRI does not use ionizing radiation; nevertheless, its magnetic field and radiofrequency energy pose safety concerns for both the patient and personnel.

Figure 28-10 Top, Static neutral and (bottom) dynamic flexed T2W magnetic resonance (MR) images of the lumbosacral vertebral column of a dog. In the top image, a hypointense protrusion of the annulus fibrosus or ligamentous hypertrophy/buckling with compression of the cauda equina is evident. In the bottom image, near-complete resolution indicates a substantial ligamentous component compatible with a diagnosis of lumbosacral instability.

Ultrasonography and Scintigraphy

Ultrasonography has a limited role in neurologic system imaging because of the inability of sound waves to penetrate bone. Nevertheless, sonography can be used to evaluate the brain and spinal cord using a natural acoustic window (e.g., bregmatic fontanelle, foramen magnum) or intraoperatively and postoperatively if a window is created with a craniotomy/craniectomy or laminectomy.* Sonography can be used to visualize pathology of the brachial plexus, lumbosacral plexus, and peripheral nerves, and to guide tissue core or fine needle aspiration biopsy if the acoustic window is sufficient.^† Abdominal sonography (along with thoracic radiography) may also be warranted in patients with intracranial disease before advanced diagnostics and therapeutics, given the high rate of unrelated neoplasia in dogs with primary intracranial tumors.²²⁹ Perhaps the largest role that ultrasonography and scintigraphy have is in the identification of portosystemic shunting in animals with hepatoencephalopathy.^41,238 Although not commonly performed in the clinical setting, single photon emission computed tomography (SPECT) and positron emission tomography (PET) are also used to detect alterations in brain perfusion and cellular metabolism.^{56,137,156,198}

Fundamentals of CT and MRI

Regardless of the imaging modality used, when an electronic image is viewed on a monitor, light emits from the screen with varying intensities of brightness. Depending on the imaging modality, the brightness may be described as opacity (conventional radiography), attenuation or density (CT), or intensity (MRI).

CT

As the patient table is moved through the donut-shaped housing or “gantry” of the CT unit, images are constructed as an x-ray tube within the gantry rotates around the patient while emitting x-ray photons.²²⁵ Similar to conventional radiography, attenuation of photons occurs as they pass through the patient and is largely related to electron density. Directly opposite the x-ray tube, electronic detectors (as opposed to film) absorb the remaining x-rays and convert them into a digital signal. Through a process called back-projection, computer-generated images of each slice of the scanned anatomy are constructed. Depending on the sophistication of the CT unit, detectors may be arranged in single or multiple rows. Images can be obtained as a discrete transverse section or slice, after which the table is advanced and another image is acquired. Alternatively, data can be acquired continuously as the patient table is advanced, resulting in a spiral acquisition. From this volume of data, images are then displayed as discrete slices.

Tissue Contrast in CT

Contrast in CT arises from tissues that differ in their ability to attenuate x-rays. Attenuation values are visually discriminated in the image as different shades of gray but are assigned Hounsfield unit measurements after standardization with water (Table 28-1). Thus, by using region of interest tools, the x-ray attenuation of a specified volume of tissue can be quantified. Quantification provides a guide to the composition of the tissue and serves as an aid for determining whether a tissue is normal (see Figures 28-1, 28-5, 28-8). Hounsfield unit measurement may be useful if tissue composition is uncertain based on subjective visual assessment of the gray level. However, the Hounsfield unit is not specific for a substance and therefore should serve only as a guide, as many combinations of various tissue types, theoretically, will yield the same Hounsfield unit value.²⁴⁸ For example, a Hounsfield unit of 90 could indicate acute hemorrhage or soft tissue combined with mineralization or less common combinations.

Table • 28-1

Hounsfield Unit Measurements for Various Substances

SUBSTANCE	HOUNSFIELD UNIT
Air	−1000
Fat	−100 (approximately)
Water	0
Brain (white matter slightly lower than gray matter)	30 to 40
Acute to subacute clotted blood	60 to 100
Mineral and bone	Variable (e.g., 100 to >1000)
Metal (e.g., iodine)	Variable, depends on dilution (e.g., 100 to >3000)

Although Hounsfield units and their corresponding gray shades have a direct relationship (i.e., the higher the Hounsfield unit, the greater the brightness level on the monitor), it is not always beneficial to apply the entire selection of gray shades, or the gray scale, to the entire range of attenuation values across the image.²⁴⁸ A process called windowing allows the operator to manipulate the gray scale by first selecting the tissue of interest (window level) and then determining what range of tissues around it (window width) will appear gray. For evaluation of neuroparenchyma or other soft tissues, a narrow window width is used to apply the gray shades only to soft tissues, thereby improving their discrimination. (Substances with values above and below the range will appear white and black, respectively.) For bone or lung evaluation, a wide window width is needed to encompass the wide range of tissues potentially found in these organs.

Substances that can be discriminated on a CT image (in order of increasing monitor brightness) include gas, lung, fat, water or other fluids such as cerebrospinal fluid, normal and abnormal soft tissues, mineral, dense bone, and metal, including iodinated contrast medium. Soft tissues that appear less opaque than normal (or are “hypodense” or “hypoattenuating”) may be cystic or fluid filled, necrotic, or edematous, and may have fatty infiltration or contain gas (see Figure 28-1). Soft tissues that appear more opaque than normal (or are “hyperdense” or “hyperattenuating”) may contain hemorrhage (because of the globin in hemoglobin, fibrin, or clot retraction; see Figure 28-8), mineral (see Figure 28-5), or metal (see Figure 28-5, C), or may be densely cellular (particularly those tissues with cells that have a high nuclear-to-cytoplasmic ratio) or densely fibrotic.^248,278

MRI

As with most electronic imaging modalities, information about the composition of tissues is conveyed by an electrical signal. With MRI, these signals are produced by mobile hydrogen atoms, which are extremely abundant in tissues in the form of water and lipid molecules. In addition to moving with the random motion of its host molecule, the positively charged hydrogen atom (proton) spins around its axis and creates a small magnetic field with opposing poles. Although the magnetic field of an individual proton is tiny, the summed field of many protons is significant. In fact, it is this net magnetic field that ultimately creates the MR signal and therefore is responsible for image formation.

The MRI process is described as follows 54,86,168,205,248:

1. During the scan, the patient lies within the bore of a very strong magnet. Alongside the body part of interest are specialized coils for transmitting and receiving radiofrequency waves. When the hydrogen protons of the body experience the strong field of the magnet, they align with the field, somewhat akin to the needle of a compass placed next to a bar magnet. (By convention, the direction of the magnetic field of the scanner [often denoted as B₀] is in the Z plane of a three-dimensional Cartesian coordinate system, with the magnetic fields represented by vectors denoting direction and magnitude. The coils are positioned in the X-Y plane.)

2. Unlike the response of the compass needle to a bar magnet, the protons do not simply align parallel to the magnetic field, but instead begin to wobble or “precess” around the axis of the field. This is analogous to a spinning top that responds to gravity’s pull by wobbling downward rather than simply falling over. Protons and spinning tops precess because they have angular momentum—a property that keeps spinning objects in motion in response to an applied force. Thus, rather than simply aligning with the force, the object rotates around its axis. In addition, some of the protons (or “spins”) are observed to align in the direction of B₀ (spin-up), and some in the opposite direction (spin-down). Because an excess number of spins is observed to align with the field, a net vector of magnetization (longitudinal magnetization, M₀) is created in the “up” direction (Figure 28-11, top image).

Figure 28-11 90-degree radiofrequency excitation. Top image, the gray cone represents the spin excess precessing around the strong field of the magnet, B₀, which is directed along the Z axis. The summed magnetic force of the spins is represented by a net vector of longitudinal magnetization, M₀ (red arrow). Middle image, radiofrequency energy (blue electromagnetic wave) is applied in the X-Y plane; its magnetic field (radiofrequency B₁) precesses around the main magnetic field, B₀, at precisely the same frequency (the Larmor frequency) as M₀. When M₀ experiences the magnetic force of the radiofrequency pulse (radiofrequency B₁), it begins to precess about this force while continuing to precess about the strong field of the magnet, B₀. This simultaneous precession causes a downward spiraling or “nutation” of M₀ toward the X-Y plane, and generates a precessing field of magnetization (transverse magnetization, M_XY), while losing longitudinal magnetization (bottom image). As soon as the radiofrequency pulse is removed, M_XY immediately begins to decay, as protons lose phase coherence or dephase. (Note that the radiofrequency pulse must precess in resonance, i.e., at the same frequency as M₀; otherwise, the two would be moving in and out of sync with each other, and the precession of M₀ around radiofrequency B₁ and subsequent nutation would not be possible. RF, Radiofrequency.)

3. The rate at which a spinning object precesses is proportional to the strength of the applied force, such as the magnetic field produced by the MRI unit. Thus, the spins precess with a frequency (Larmor frequency) that is proportional to the strength of their local magnetic environment. Although the main determinant of a spin’s Larmor frequency is the strength of the magnet itself, other static magnetic fields (in the form of applied encoding gradients—additional magnets that add to or subject from the main magnetic field in a graded manner) and magnetic substances within tissues (e.g., hemoglobin), as well as the randomly fluctuating magnetic fields of neighboring protons, also contribute.

4. The spins are then perturbed by radiofrequency pulses through a process called excitation. Because radiofrequency waves are a form of electromagnetic energy, they create both electrical and magnetic fields. With the application of radiofrequency energy, M₀ now experiences a second magnetic field (denoted as B₁) and begins to precess about this force while simultaneously precessing about the strong field of the magnet, B₀ (see Figure 28-11, middle image). This simultaneous precession causes a downward spiraling or “nutation” of M₀ toward the X-Y plane. Thus, radiofrequency excitation serves to generate a precessing field of magnetization in the X-Y plane (transverse magnetization, M_XY; see Figure 28-11, bottom image). It should be emphasized that B₁ of the radiofrequency pulse must precess in resonance (i.e., at the same frequency as M₀); otherwise, the two would be out of sync with each other, and the precession of M₀ around B₁ and subsequent nutation would not be possible. For this reason, the resonance phenomenon is critical to radiofrequency excitation, hence the name magnetic resonance imaging.

5. As soon as the desired amount of transverse magnetization is generated, the radiofrequency pulse is removed. According to the law of electromagnetic induction, which states that a magnetic field that is changing in strength and direction will generate an electrical voltage in a coiled wire, this precessing (time varying) transverse magnetization creates alternating current in the receiver coils positioned in the X-Y plane as stored radiofrequency energy is released. This measured current is the MR signal. This signal immediately begins to decay as protons relax and (1) return to a state of equilibrium, allowing longitudinal magnetization to recover (described by a time constant called T1 recovery), and (2) stop precessing in unison or “dephase” (described by a time constant called T2 decay). Dephasing occurs because the Larmor frequencies of the spins making up M_XY are heterogeneous; this heterogeneity, in turn, is influenced by inhomogeneities in the magnetic field, as described in item 3 above.

6. T1 and T2 relaxation are different for different types of substances and are influenced by characteristic properties, namely, the interaction of the proton with its environment and with neighboring protons, respectively. Thus, when the sequence of radiofrequency excitation followed by relaxation (in a process called a pulse sequence) is repeated numerous times, subsequent signals from different tissue types will have different strengths. This variation in signal intensity, which varies widely between tissues, provides contrast within the image and can be accentuated by the MRI technician by manipulating certain parameters in the pulse sequence through a process called weighting.

7. To create a meaningful image, the signal intensities from various tissues within the body must be localized first to an image slice, then to a position within that slice. By a process called spatial encoding, additional magnetic fields (encoding gradients) are superimposed over B₀ to predictably alter the Larmor frequencies of spins in a graded linear fashion along one of three orthogonal axes. Because of the resonance phenomenon, only those spins whose Larmor frequencies are in sync with the transmitted radiofrequency frequency will undergo excitation, as described in item 4 above. The desired slice location therefore can be chosen by turning on the slice selection gradient while transmitting radiofrequency pulses at a selected frequency. Within that slice, additional gradients are applied to alter the phase and frequencies of spins, again in a predictable graded fashion, but this time while the signal is being measured. This permits a direct correlation between frequencies in the measured signal and the position of spins along these gradients. By a process called Fourier transform, various signal amplitudes or strengths for each of the frequencies can then be deciphered. These amplitudes or signal intensities determine the brightness or gray levels for each picture element (pixel) located within the image slice.

Pulse Sequences and Weighting

A pulse sequence is simply the series of timed events by which a radiofrequency pulse is used to create signal.²⁰⁵ Because the signal following the excitation of protons is short-lived, an “echo” of that signal is created by the application of additional pulses by one of two basic methods: spin echo or gradient echo.

T1 and T2 relaxation occur simultaneously following the radiofrequency excitation pulse, regardless of the type of pulse sequence. Nevertheless, by altering the timing between successive radiofrequency excitation pulses, (echo time) or the timing of the echo (repetition time), the relative contribution of T1 or T2 relaxation to image contrast can be controlled. Images are referred to as T1-weighted (T1W) or T2-weighted (T2W) (see Figure 28-3). Alternatively, images can be weighted based largely on the density of protons in the tissue (proton density–weighted). If the operator selects a short TR, the differences in signal intensity between tissues that differ in T1 relaxation times will be accentuated, and the image will be T1W. For T2-weighting, the operator selects a long echo time to maximize T2 contrast in the image.

In general, T1W and proton density–weighted images provide good anatomic detail, and T2W images are considered to be more sensitive to pathology. One exception to this rule is that T1W images more readily depict pathologic tissues having short T1 (e.g., lesions containing fat, methemoglobin, gadolinium contrast medium). Signal intensities on proton density–weighted images usually mirror those on T2W images, although images can be additionally weighted to suppress cerebrospinal fluid, permitting better visualization of periventricular white matter lesions.

Despite the vast number of vendor-specific names, abbreviations, and acronyms, it is important to remember that spin echo and gradient echo are the only pulse sequences upon which all others are built. Other sequences are variations intent on optimizing desired effects. For example, the basic spin echo and gradient echo pulse sequences can be modified to create sequences that reduce scan times, accentuate tissue contrast or anatomic detail, suppress signal from certain tissues, identify hemorrhage, or emphasize physiology or function.8,21,55,101,213 A taxonomy of pulse sequences is illustrated in the schematic in Figure 28-12.²¹