| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
How I Do It |
1 From the Department of Radiology, Division of Neuroradiology, University of Utah, Salt Lake City (K.M.W., J.S.T., J.R.H.); and Department of Radiology, University of Washington, Seattle (C.E.H.). From the 2000 RSNA scientific assembly. Received September 15, 2000; revision requested October 24; revision received January 30, 2001; accepted February 6. Supported by National Institutes of Health grants R01HL48223 and R01HL53596. J.S.T. supported in part by GE Medical Systems. C.E.H. supported in part by Pathway MRI (formerly Ultraimage). Address correspondence to K.M.W., Department of Radiology, Naval Medical Center, 620 John Paul Jones Circle, Portsmouth, VA 23708 (e-mail: kmwelker@pnh10.med.navy.mil).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
Index terms: Brain, MR, 10.12141, 10.12142 • Magnetic resonance (MR), coils • Magnetic resonance (MR), technology • Radiology and radiologists, How I Do It
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
In recent years, there have been numerous technologic advances in RF coil design and a proliferation in the number of specialized coils available to the imaging professional. Such coils demonstrate a wide range of functional characteristics ranging from large-volume general use head coils to small phased-array surface coils for niche applications.
Despite their importance to image quality, appropriate RF coil selection is often discounted during MR imaging of the brain and skull base. This may be partially related to lack of information regarding the strengths and weaknesses of various coil configurations. However, much of the widespread neglect of the nuances of coil selection may be attributed to a relative paucity of literature in which the logical rules for deriving an appropriate RF coil choice for a specific imaging situation are described.
In this article, we will first describe the general characteristics of various coil designs as they apply to imaging of the brain and skull base. Next, we will describe a logical algorithm that may be used to narrow the choice of available coils to those most suited to a particular imaging situation. Finally, we will provide a number of specific examples regarding the application of this algorithm to brain and skull base imaging. This section will also provide data regarding coil selection practices at our institution and our expected levels of spatial resolution for each imaging application.
|
CHARACTERISTICS OF MR RF COILS |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
In general, RF coils perform most efficiently when physically aligned to receive signals that emanate perpendicular to the magnetic field B0 of the imager (3). In other words, coils operate best when a vector that is geometrically normal to the loop of the coil-receiving element is oriented perpendicular to the main magnetic field B0. This necessarily suggests some physical constraints on coil orientation within an MR imager. For instance, coils designed for superconducting electromagnetic imagers may not operate properly in a permanent-magnet or resistive imagers, especially if these imagers are designed with a vertically oriented B0 magnetic field. Therefore, our discussion will be limited to the use of RF coils as applied to a standard superconducting MR imager.
RF coils are typically classified into two broad categories: volume coils and surface coils (4). These categories differ in their reception characteristics and are used for different imaging scenarios.
Volume coils provide homogeneous transmission and reception over a large anatomic region. This region is typically the cylinder or ellipsoid circumscribed by the coil. For neuroimaging applications, this sensitive volume typically includes the entire calvarium and contents. However, because noise reception is nonlinearly proportional to the volume of tissue being interrogated, the overall SNR of these coils is generally lower than that of surface coils (5). Nevertheless, the superb reception homogeneity provided by volume coils in conjunction with their relatively larger imaging volume make them the most frequently used coils for MR imaging of the brain and skull base. Because volume coils are designed to create homogeneous RF fields, they are often used as both transmit and receive coils. A birdcage coil is a typical example of a volume coil used for neuroimaging applications.
Quadrature reception is a technique applied with some volume coils to improve signal-to-noise characteristics. This method is also known as circularly polarized detection. Two separate phase-sensitive detectors are used to detect signal components that are 90° out of phase (4). These signal components are then separately correlated and summated to produce a final output signal. Because noise arriving from the imaging sample is random and does not correlate between the two detectors, this technique increases the SNR of the coil by a factor of 
2 with respect to nonquadrature detection techniques (1,6).
In contrast to volume coils, surface coils are designed to operate efficiently over a limited region of interest. These coils, also known as "local coils," provide high signal-to-noise reception over a small geometric area immediately adjacent to the coil. This improvement in SNR is achieved at the cost of RF reception declining dramatically beyond the region adjacent to the coil (7). Therefore, surface coils inherently create an inhomogeneous reception pattern. The useful depth of reception for a circular surface coil is approximately equivalent to the radius of the coil (3). Because of their spatially inhomogeneous properties, surface coils are usually not suitable for RF pulse transmission (5).
Surface coils can be constructed in either flexible or rigid designs. In general, flexible surface coils can be applied more closely to the contour of the anatomic structure under consideration, with the potential of improved SNR from the anatomic region of interest. In addition, flexible coils are generally more comfortable for the patient and may lead to a reduction in motion artifact (8). Flexible coils, however, have less durability than rigid coils because of the repeated stress on the receiving elements. Moreover, because of the fixed relationship of the internal circuit elements with each other, the operating characteristics of rigid-design surface coils are more reproducible.
Phased-array coils combine features of both volume coils and surface coils. They consist of two or more geometrically aligned surface coils used in conjunction to extend the high SNR characteristics of a single surface coil over a larger field of view (FOV) (9). Surface coils in juxtaposition interact through mutual inductance with both signal and noise being transferred from one coil to the other. Most of this problem can be overcome by physically overlapping the coils in a precise geometric configuration and using low impedance preamplification on the output of each individual coil (9).
The geometric relationship of coil reception elements is critical to the performance of phased-array surface coils. A marked reduction in SNR may occur if the reception elements of phased-array coils are positioned orthogonal to the B0 magnetic field or if elements from contralateral arrays in a dual phased-array coil system are overlapped. Similarly, if the reception elements of a flexible phased-array coil are flexed to an excessive degree, optimal coil geometry is lost and there is a consequent decline in SNR.
Phased-array coils typically require an individual hardware channel for each reception element and, therefore, place greater demands on the RF reception and computer hardware of the MR imager. To account for the reception inhomogeneity of phased-array surface coils, a computerized intensity-correction algorithm may be used to postprocess the images (10,11). In short, phased-array coils demand more imager computer resources in the form of memory, array processors, and processing time.
|
SPECIFIC COILS FOR IMAGING OF BRAIN AND SKULL BASE |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
Standard Birdcage Head Coil
Birdcage coils are volume coils that represent an improvement over saddle coils in terms of field homogeneity. These coils consist of two circular or elliptic conducting end rings joined by conducting "rungs" or "legs" (Fig 1a). This electronic configuration simulates a parallel-line transmission wire being draped over a cylinder (13). Trim capacitors on the end rings and legs are used to equalize self inductance and mutual RF coupling of the legs. This geometry creates an RF field distribution with markedly improved homogeneity with respect to a saddle coil.
|
|
Quadrature detection is generally used with birdcage coils to obtain a 2 improvement in SNR (6) over that which would be obtained if the same coil was used in a linear mode. Commonly, the two orthogonal signal ports within the coil are combined to create a single hardware output channel. One drawback of birdcage coils is the presence of multiple conducting rungs over the patient’s face. These may obstruct vision and cause claustrophobia in some patients.
Reduced-Volume Birdcage Head Coil with Reflecting End Cap
The reduced-volume birdcage with a reflecting end-cap coil is a high-performance investigational coil that improves on the design of standard birdcage coils (Fig 1b) (16). This head coil includes a metallic conducting end cap that replaces the superior conducting end ring of the birdcage. The diameter of the end cap is specifically designed to exceed that of the birdcage. The advantage of the end cap is that it increases coil sensitivity throughout the imaging volume and improves homogeneity at the end of the coil adjacent to the end cap. The conducting end cap contains numerous slits that divide it into a pielike configuration. These slits serve to reduce eddy currents generated by the MR imager gradient coils. Further signal-to-noise improvement is achieved by reducing the overall volume of the coil, particularly in the z axis (superior to inferior).
These two modifications markedly enhance the SNR in the region of the brain. Because this coil has a limited z-axis dimension, however, it demonstrates a marked drop-off in reception homogeneity below the foramen magnum in most patients. The high-performance design of this coil is partially achieved at the expense of patient comfort. The small volume and reflecting end-cap of this coil contribute to the patient’s sense of confinement. Owing to its excellent imaging characteristics, however, it is expected that this coil design will undergo further modification to improve patient comfort prior to commercial production.
Neurovascular Phased-Array Coil
Neurovascular phased-array coils use multiple phased-array reception elements in an effort to provide useful imaging of both the head and neck with a single coil. The name is derived from the fact that these coils are often used to acquire both brain MR images and cervical carotid artery MR angiographic images in a patient. Alternatively, neurovascular coils have been referred to as dedicated head and neck coils. These hybrid coil models typically consist of quadrature head coil elements in a modified saddle or birdcage design coupled with anterior and posterior cervical reception elements (Fig 2). The goal of this design is to extend the range of useful signal coverage into the region of the neck. The head and neck coil elements can be used separately or in conjunction with each other. The major advantage of this coil configuration is that it allows the technologist to acquire studies of both the head and neck regions without exchanging coils (17). Because of this timesaving benefit, neurovascular phased-array coils are among the most commonly used coils for neuroimaging applications. Neurovascular coils are useful for evaluating pathologic conditions that extend beyond the boundary of the skull base. For instance, combined MR angiography of the neck and circle of Willis is an application that is frequently accomplished by using a neurovascular coil (18). Often, neurovascular coils are manufactured in a helmet-type configuration that can cause considerable anxiety in the claustrophobic patient.
|
|
|
It should be noted when considering the use of these dual phased-array coils that the degree of signal loss near the center of the head is dependent on the anatomic thickness of the head. The amount of RF signal attenuation near the center of the skull can be markedly less in children or adults with narrow heads.
Dual Single-Circular-Element Coil
The dual single-circular-element coil is a commercially available (from several manufacturers) circular surface coil with a reception element diameter of 7.5 cm (Fig 4). This coil is often referred to as a temporomandibular joint coil. Each circular reception element can be used individually or can be paired with the contralateral element as a single-channel reception unit. Alternately, the elements can be operated in a two-channel phased-array configuration, allowing simultaneous acquisitions of bilateral paired structures. The dual single-circular-element coil is useful for imaging superficial structures that are geometrically limited in both anteroposterior and superoinferior extents. For anatomic objects meeting these constraints, this coil provides exceptionally high signal-to-noise efficiency. Although typically used for imaging of the temporomandibular joint, this coil finds application in imaging the inner ears, internal auditory canal, and orbits.
|
|
|
COIL SELECTION ALGORITHM |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
|
The anatomic region of interest defines the requisite imaging volume. For optimal imaging, this region of interest should be defined as small as is reasonably possible. A small region of interest will permit customized coil, FOV, and pulse sequence selection. Unfortunately, nonspecific symptoms or generalized disease conditions necessitate evaluation of larger anatomic regions.
Once an anatomic region has been defined, the clinical history is then used to derive a desired spatial resolution. Mathematically, the spatial resolution may be defined as the tissue volume assigned to each voxel (volume element) in the final output display. This measurement is expressed in terms of cubic millimeters of tissue per voxel. Different imaging techniques are needed to achieve the higher spatial resolution required to image small anatomic objects with confidence. Superficially, one might conjecture that imaging with the maximal spatial resolution that is technically achievable represents a universal imaging goal. However, increases in spatial resolution can exert a negative effect on imaging times and permissible FOV size. In addition, it should be remembered that the SNR is linearly proportional to voxel volume (20). Small voxels can consequently have a negative effect on SNR (21). Finally, increases in spatial resolution exacerbate the effects of motion artifact on the image.
The desired spatial resolution may vary for any particular imaging axis. For instance, spatial resolution in terms of section thickness may be sacrificed to improve the in-plane spatial resolution for each section. Conceptually, one can imagine that although the volume of a voxel is constrained, the lengths of the particular sides of the voxel can still be varied in an infinite number of combinations. Nevertheless, if three-dimensional (3D) imaging is performed for the purpose of obtaining multiplanar reformations, isotropic voxels must typically be used (22).
The final important parameter that is dictated by clinical history is the desired contrast resolution. Contrast resolution is defined as the ratio of signal intensity differences of objects on the image to background noise. This value is inseparably linked to the MR pulse sequences that will subsequently be used to demonstrate inherent differences in tissue composition (23). For example, a fluid-attenuated inversion-recovery (FLAIR) sequence provides much greater contrast resolution for differentiating multiple sclerosis plaques from white matter than does a T1-weighted sequence. Conversely, T1-weighted images provide excellent contrast resolution for the demonstration of fat or methemoglobin. Also, contrast enhancement techniques such as spectroscopic fat saturation or magnetization transfer pulses may be included in the pulse sequence to increase lesion conspicuity. The MR pulse sequences selected to satisfy the demands of contrast resolution place important constraints on RF coil choice. This occurs because of the markedly distinct imaging times and efficiencies associated with various pulse sequences.
Anatomic Constraints on Coil Selection
Requisite anatomic coverage is the initial constraining factor used to limit the available set of coils for any particular imaging application. It represents a numeric refinement of the region of interest extracted from the patient’s history. For two-dimensional imaging, the specification of anatomic coverage requires definition of an explicit two-dimensional FOV, as measured with the distances on the frequency- and phase-encoding axes. The FOV is further assigned an orientation, such as transverse or coronal. This information is coupled with a proposed imaging extent in the orthogonal direction to define an imaging volume that will encompass the anatomic structures important to the clinical scenario.
The interplay of spatial resolution and anatomic coverage specifications merits careful consideration. With other factors remaining equal, a requirement for high spatial resolution will substantially limit the permissible dimensions of the anatomic imaging volume. To preserve spatial resolution (volume per voxel) while increasing the total imaging volume, the number of voxels in the acquisition must be increased. An increase in the number of voxels necessitates expansion of the acquisition matrix size or the number of sections acquired. Such measures typically cause a secondary increase in imaging time (2). Later, the importance of limiting the imaging time will be discussed in detail.
Once an acceptable imaging volume has been defined, coil selection can be limited to those coils that provide sensitivity throughout the requisite anatomic volume. For example, both the standard quadrature birdcage head coil and the temporal dual phased-array coil provide suitable anatomic coverage for imaging of a temporal lobe. However, the dual single-circular-element coil is unsuitable, given that the 3.3-cm radius of sensitivity for this coil does not satisfy coverage requirements. From a slightly different standpoint, a reduced-volume birdcage end-cap coil cannot be used for evaluation of a lesion that extends a substantial distance below the skull base. Although the overall volume enclosed by that coil may be adequate to encompass the lesion, the requirement to spatially center the imaging volume at the skull base precludes use of that particular coil.
Imaging Protocol Constraints on Coil Selection
The next major constraint on coil selection arises from a group of issues related to imaging protocols. Imaging protocol constraints are based on the premise that the operating efficiency of the coil must be sufficient to produce images with a useful SNR during a time-limited imaging sequence. We have selected 10 minutes as an arbitrary limit for any single imaging sequence, given that this is a reasonable amount of time for the average patient to lie motionless in the MR unit. In reality, the duration of most imaging sequences falls well below this 10-minute limit.
Aside from the reception efficiency of the coil, meeting the 10-minute limit also depends on the data processing ability of the computer system of the imager. The specific number of data channels that can be simultaneously processed, the processing speed, and the available memory play an important role in determination of imaging times. An RF coil with superior efficiency may generate average images if the data processing capabilities of the imager fall well below the data collection capabilities of the coil. Finally, it should be noted that limiting of imaging times is an important patient throughput consideration.
To achieve the desired spatial resolution, compatible acquisition matrix size and section thickness are selected by using the following relationship for two-dimensional Fourier transform imaging: SR = FOVphase / MSphase x FOVfrequency / MSfrequency x ST, where SR is the spatial resolution, in cubic millimeters per voxel; MS is the matrix size; and ST is the section thickness, in millimeters (subscripted "phase" and "frequency" refer to the phase- and frequency-encoding directions, respectively); FOV is expressed in millimeters. For 3D imaging sequences, partition thickness is substituted for section thickness. Noncontiguous sections can be used to effectively extend the imaging extent while maintaining the intrinsic spatial resolution of each section. Gaps between sections are also useful for the elimination of cross talk between sections (22).
As previously elaborated, the contrast resolution mandated by the clinical history dictates the imaging sequences used. Each sequence has an associated repetition time. For a spin-echo (SE) or fast SE multisection acquisition, the repetition time of the pulse sequence multiplied by the number of phase-encoding steps in the image matrix and divided by the echo train length determines the imaging time for each section of a single-excitation acquisition (2). Depending on the reception characteristics of a given RF coil, a single excitation may be insufficient to gather enough signal to construct a clinically useful image, especially if a small voxel size has been specified. Given that the SNR is proportional to the square root of the number of signals acquired, the number of signals may have to be increased to provide sufficient SNR to create an acceptable MR image (21,22). If the total imaging time for a given coil to meet this signal requirement exceeds the 10-minute per sequence limit, that particular coil is disqualified as a viable candidate for the imaging application at hand.
By way of example, a standard quadrature birdcage head coil meets anatomic constraints for generating high-spatial-resolution images of the inner ear. However, it does not have the necessary signal gathering efficiency to generate a two-dimensional T2-weighted fast SE image of the inner ear with a desired spatial resolution of 0.3 mm3/voxel. After 10 minutes of imaging, the SNR of any images acquired with that particular coil at the prescribed spatial resolution falls well below acceptable limits.
Patient Factors Constraining Coil Selection
The final set of limitations on the selection of an appropriate coil relate to patient body habitus, comfort, and anxiety level. First, it may not be possible to physically accommodate some patients with a given coil owing to head and neck size considerations. This is particularly true of the reduced-volume birdcage end-cap coil, which may not fit on the head of larger patients. Rarely, patients may require brain imaging with the body coil due to their inability to fit within any of the available specialized coils for brain and skull base imaging. It has previously been stated that the temporal and carotid dual phased-array coils demonstrate an accentuated loss of signal in the central brain region in patients with a thick head. Therefore, this factor may weigh into the decision of whether to substitute with a volume coil.
It should be noted that serious motion artifact on an image can eliminate any of the gains in image quality achieved through optimal coil selection; therefore, patient comfort becomes an important consideration. Typically, the use of a head coil has a positive effect on reducing motion, given that it physically stabilizes the patient’s calvarium. However, given that head coils can substantially add to the sense of confinement that is inherent in being positioned within the bore of an MR imager, the use of a particular coil in an anxious patient may cause unacceptable claustrophobia with secondary patient motion. Although from a physics standpoint, a particular coil may be optimal for the imaging at hand, it may have to be exchanged for a less efficient coil in an effort to quell patient anxiety.
Finally, if a functional MR study is being performed, the coil must physically allow introduction of the visual, auditory, or tactile patient stimulus as required by the functional MR paradigm.
|
EXAMPLES OF RF COIL APPLICATION |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
Our in-house selection of available coils is outlined in Table 2. We routinely have access to three coils not yet approved by the U.S. Food and Drug Administration (FDA): the reduced-volume birdcage end-cap coil, the temporal dual phased-array coil, and the carotid dual phased-array coil. These are investigational coil designs that are used at our hospital with institutional review board approval. A commercial production version of the temporal dual phased-array coil has recently received FDA approval. It is anticipated that refined versions of the other two coils will soon become commercially available.
|
Table 3 summarizes typical or baseline geometric imaging parameters used in MR protocols at our institution. It must be emphasized that the recommended FOVs, matrix sizes, and other imaging selections outlined in Table 3 represent a generic starting point for the assignment of imaging parameters. Further improvement in imaging quality and spatial resolution can routinely be accomplished by tailoring the examination to the nuances of the particular patient and clinical question at hand. Moreover, future advances in MR imager, RF coil, and computer designs will provide the capability of achieving additional gains in spatial resolution. It is important to understand that any specific spatial resolution can be achieved by using a number of different combinations of matrix size, section thickness, and FOV.
|
The only coils meeting the anatomic coverage criterion are the standard quadrature birdcage head coil and the neurovascular coil. For most screening studies, the reduced-volume birdcage end-cap coil is unsuitable due to the rapid decline in signal below the foramen magnum. Both the standard birdcage head coil and the neurovascular coil will satisfy imaging protocol constraints. When using the recommended parameters, the longest imaging sequence is transverse FLAIR imaging. This requires about 4
minutes to complete. Given that the standard quadrature head coil is generally more comfortable for the patient, it is preferentially used over the neurovascular coil unless concomitant cervical imaging is indicated.
Brainstem or Cranial Nerve MR Examination
Thorough evaluation of the brainstem and/or associated cranial nerves requires improved spatial resolution in comparison with generalized brain screening, because small structures are being assessed. A spatial resolution of 0.8 mm3/voxel is desirable. The imaging volume for the brainstem and cranial nerves can be small, given that the majority of the cerebrum can be excluded from the study. Geometric parameters based on a 14-cm isotropic FOV and 2-mm section thickness are outlined in Table 3. Typical sequences include sagittal T1-weighted, transverse T2-weighted fast SE, transverse FLAIR, transverse T1-weighted without and with gadolinium enhancement, and coronal gadolinium-enhanced T1-weighted sequences. Fat saturation improves contrast on the T2-weighted and gadolinium-enhanced T1-weighted sequences.
The standard head coil, the neurovascular coil, and the temporal dual phased-array surface coils all meet anatomic coverage requirements. Because of the precipitous decline in signal sensitivity below the foramen magnum, the reduced-volume birdcage end-cap coil is generally not used when assessing the lower cranial nerves because the extracalvarial extent of these nerves will be suboptimally demonstrated with that coil. Similarly, the anatomic range offered by the smaller phased-array surface coils will not cover the full extent of the brainstem and cranial nerves.
With regard to imaging protocol constraints, the standard birdcage head coil and the neurovascular coil suffer from diminished SNR over the imaging volume selected for brainstem assessment. This is because their larger reception volumes introduce additional noise. However, the temporal dual phased-array surface coils are able to gather sufficient signal from the specified volume, often with a single signal acquisition for each sequence. When using these coils and the recommended imaging geometries, the maximum imaging time required for any given sequence is about 5 minutes, the time necessary to accomplish transverse FLAIR imaging. Examples of clinical imaging of the cranial nerves with the dual temporal phased-array coils are provided in Figure 6. In a patient with a very wide head, the temporal lobe dual phased-array coils may manifest signal drop-off at the center of the image. Should this prove problematic, the standard quadrature head or neurovascular coil may be used as an alternative. However, this coil substitution will necessitate a partial sacrifice of spatial resolution or an unwanted prolongation of imaging time.
|
|
Anatomic coverage of the orbits can be provided by all of the RF coils in our armamentarium. However, MR imaging of the orbit creates a particular set of challenges from an imaging protocol perspective, owing to an inherent difficulty in controlling the motion of the globes (24). Whereas imaging in most other regions of the head can be governed by a time limit of 10 minutes per sequence, successful orbital imaging often requires imaging times of less than 4 minutes, given that this represents the maximum amount of time that most patients can fix their gaze (25).
Surface coils play an important role in orbital imaging, given the superficial location of most of the orbital contents. Because of their increased signal sensitivity within the anterior orbit, the use of phased-array surface coils can substantially decrease imaging time, with a secondary—but critical—reduction in motion artifact. Although the carotid phased-array coils and dual single-circular-element coils are suitable for evaluation of the anterior orbit, the larger temporal dual phased-array coils provide greater depth of signal reception and are, therefore, better suited for evaluation of the posterior orbit (Fig 7) and orbital apex. The dual phased-array coils are placed on each side of the head, with an approximately 4-cm separation between the anterior edges of the imaging elements. The calculated SNR for phased-array surface coils has been shown to slightly exceed that for standard head coils in imaging of the anterior optic pathways up to the region of the optic chiasm (26). Owing to diminishing surface-coil reception sensitivity near the center of the head, however, a volume coil may be necessary for satisfactory imaging of orbital pathologic conditions that extend into the retrochiasmatic region (27). Alternatively, if imaging of the globe is the sole consideration, the carotid dual phased-array coils provide superior SNR over the small requisite FOV.
|
|
Most of our coils provide the required anatomic coverage. Two notable exceptions are the dual single-circular-element and carotid phased-array coils. Because of their restricted diameters, the RF reception fields of these smaller surface coils do not penetrate the central head with sufficient sensitivity to allow assessment of the sella turcica.
In most individuals, volume coils provide greater signal reception at the center of the head when compared with that of the temporal dual phased-array coil. Specifically, the reduced-volume birdcage end-cap coil will most efficiently extract signal from this region. In the event that this coil is unavailable or cannot accommodate the patient, the standard quadrature birdcage head coil or neurovascular coil is used as an alternative. It should be noted, however, that in a patient with a narrow head, the temporal dual phased-array coil may provide a paradoxical improvement in SNR over the previously discussed volume coils.
Cortical Dysplasia and Gray Matter Heterotopia
The search for subtle malformations of the cerebral cortex necessitates high spatial resolution. Concomitantly, one must achieve excellent contrast resolution between gray matter and white matter. Moreover, the ability to review images in multiple planes improves sensitivity for the detection of subtle cortical dysplasia (29).
Our targeted spatial resolution for evaluation of cortical dysplasia is 0.7 mm3/voxel. The pulse sequences we use to generate the required tissue contrast include 3D T1-weighted fast spoiled gradient-recalled-echo (SPGR) with imaging geometry as outlined in Table 3. T2-weighted fast SE and FLAIR imaging also provide useful contrast resolution.
If a specific region of cortex has already been identified with electroencephalographic data or previous MR images, excellent dedicated local imaging of the cerebral cortex is accomplished with phased-array surface coils (30,31). In this situation, we use the temporal dual phased-array coils (Fig 8). As previously stated, these coils are suitable for dedicated evaluation of most regions of the cerebral cortex with the exception of the vertex.
|
|
|
Requisite anatomic coverage is provided by both the volume coils and the temporal dual phased-array coils. However, only the temporal dual phased-array coils provide sufficient signal-to-noise characteristics to complete a coronal SPGR sequence, as prescribed, within the 10-minute limit. When this specialized coil is used, the 3D SPGR sequence requires approximately 9 minutes of uninterrupted imaging time. Clinical examples of high-spatial-resolution imaging of the hippocampus are provided in Figure 10.
|
|
All of our available coils provide sufficient anatomic coverage for the internal auditory canal. Owing to the excessive size of their sensitive volumes, however, the majority of these coils introduce an inordinate amount of noise from the patient. Because of the small-FOV requirements and the need to image bilateral structures, imaging is most efficiently accomplished with the dual single-circular-element coil (Fig 11). By using this coil, a transverse T2-weighted 3D fast SE study is acquired in approximately 8 minutes. It should be noted that the dual single-circular-element coil has a limited depth of coverage due to its small diameter. If full evaluation of the cerebellopontine angle is of primary concern, a larger phased-array surface coil, such as the carotid or temporal phased-array coil, may provide improved depth of signal reception (37).
|
|
Our high-spatial-resolution 3D time-of-flight circle of Willis imaging protocol achieves a spatial resolution of 0.2 mm3/voxel. To accomplish this, the transverse FOV is set at 22 x 16 cm (frequency encoding by phase encoding). Sixty sections of 0.9-mm thickness are obtained. To preserve spatial resolution, a 512 x 256 (frequency encoding by phase encoding) acquisition matrix is used. The single imaging sequence is a vascular time-of-flight SPGR with single-slab technique and magnetization transfer. Postprocessing of the data set is routinely performed by using zero filling in all three dimensions. A z-buffer continuity algorithm may be used to enhance maximum intensity projection data (38).
Intuitively, one would select a volume coil for most MR angiography applications, given the anatomic extent of the intracranial vasculature. Of the volume head coils, only the reduced-volume birdcage end-cap coil possesses sufficient reception efficiency to satisfy the previously elaborated imaging protocol constraints (Fig 12). It is, therefore, the coil of choice for high-spatial-resolution MR angiography.
|
|
For simultaneous combined imaging of the cervical and cranial vasculature, as might be required during dynamic gadolinium-enhanced MR angiography, the neurovascular coil must be used. This coil alone can provide the requisite length of anatomic coverage in the z axis.
Last, it should be noted that the temporal dual phased-array coil is useful for the occasional niche application of imaging a suspected or previously diagnosed aneurysm of the distal portion of the middle cerebral artery. Because of the peripheral nature of such a lesion, surface-coil technology can be used to advantage.
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
This article was written by Kirk M. Welker, Lt Comdr, MC, USNR, while a fellow at the University of Utah. The views expressed in this article are those of the author and do not reflect the official policy or position of the Department of the Navy, Department of Defense, nor the U.S. Government.
J.S.T. is a stockholder in Pathway MRI. C.E.H. is a part owner of and stockholder in and receives consulting fees from Pathway MRI.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONCHARACTERISTICS OF MR RF...SPECIFIC COILS FOR IMAGING...COIL SELECTION ALGORITHMEXAMPLES OF RF COIL...SUMMARYREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
