NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

AHCPR Health Technology Assessments. Rockville (MD): Agency for Health Care Policy and Research (US); 1990-1999.

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of AHCPR Health Technology Assessments

AHCPR Health Technology Assessments.

Show details

Surface/Specialty Coil Devices and Gating Techniques in Magnetic Resonance Imaging

Health Technology Assessment Reports, 1990 Number 3

.

Created: .

Foreword

The Office of Health Technology Assessment (OHTA) evaluates the safety and effectiveness of new or unestablished medical technologies that are being considered for coverage under Medicare. These assessments are performed at the request of the Health Care Financing Administration (HCFA). They are the basis for recommendations to HCFA regarding coverage policy decisions under Medicare.

Questions about Medicare coverage for certain health care technologies are directed to HCFA by such interested parties as insurers, manufacturers, Medicare contractors, and practitioners. Those questions of a medical, scientific, or technical nature are formally referred to OHTA for assessment.

OHTA's assessment process includes a comprehensive review of the medical literature and emphasizes broad and open participation from within and outside the Federal Government. A range of expert advice is obtained by widely publicizing the plans for conducting the assessment through publication of an announcement in the Federal Register and solicitation of input from Federal agencies, medical specialty societies, insurers, and manufacturers. The involvement of these experts helps assure inclusion of the experienced and varying viewpoints needed to round out the data derived from individual scientific studies in the medical literature.

After OHTA receives information from experts and the scientific literature, the results are analyzed and synthesized into an assessment report. Each report represents a detailed analysis of the safety, clinical effectiveness, and uses of new or unestablished medical technologies considered for Medicare coverage. These Health Technology Assessment Reports form the basis for the Public Health Service recommendations to HCFA and are disseminated widely. Individual reports are available to the public once HCFA has made a coverage decision regarding the subject technology.

OHTA is part of the Agency for Health Care Policy and Research (AHCPR), Public Health Service, Department of Health and Human Services.

  • Thomas V. Holohan, M.D.
  • Director
  • Office of Health Technology Assessment
  • J. Jarrett Clinton, M.D.
  • Acting Administrator
  • Assistant Surgeon General
  • Copies may be obtained at no charge from:
  • Publications and Information Branch
  • AHCPR
  • 5600 Fishers Lane
  • Parklawn Building, Room 18-12
  • Rockville, MD 20857; (301) 443-4100

Introduction

In 1985 the Public Health Service (PHS), through the Office of Health Technology Assessment (OHTA), published an assessment of magnetic resonance imaging (MRI) (1). The assessment found MRI to be a useful diagnostic imaging modality capable of demonstrating a wide variety of soft-tissue lesions with resolution equal or superior to computed tomography (CT) scanning in various parts of the body. At that time the assessment considered the use of MRI with gating techniques and surface coil devices to be investigational. However, since 1985 many technical developments have occurred in conjunction with MRI procedures. In addition to the developments in surface coil devices and gating techniques, there have been developments in the use of other motion compensation techniques, fast scanning, and magnetic contrast agents (2).

Surface coil devices have been developed to obtain improved resolution by imaging only a small region of the body but with much better signal-to-noise ratio (SNR). In contrast to the original concept of having head coils or body coils in the MRI device encircle the whole object, including the area of interest, smaller radiofrequency antenna coils (surface coils) are directly applied near the surface of the skin close to the tissue (organ) of interest. Gating, another recent adjunct to MRI, is a process that facilitates high-resolution magnetic resonance (MR) images despite the presence of motion. Gating operates by synchronizing the start of the MR pulse sequences to a constant time point of the cardiac or respiratory cycle. Gating techniques and surface coil devices are used in conjunction with MRI procedures to enhance image quality. Although MR is a powerful technique for imaging biological tissues, gating techniques and surface coil devices as well as other techniques and devices have been developed as a means of improving MR image resolution or reducing imaging time. Gating techniques and surface coil devices attempt to improve image resolution by maximizing the signal-to-noise ratio.

This report examines recent developments in surface coil technology and gating techniques and discusses their efficacy for use in MRI studies of various areas of the body.

Surface Coils

Background

MRI is based on the principle that human tissue has weak magnetic properties (3). During an MRI procedure, the hydrogen nuclei in the patient's body are exposed to the system's powerful magnetic field, causing them to align with that field (2). A sequence of radiofrequency pulses briefly applied to the body part being studied creates a transient magnetic field perpendicular to the main field. The hydrogen nuclei absorb energy produced by the radiofrequency pulse, and the direction of their axes of rotation changes. The time required for the nuclei to realign with the magnetic field is called the spin-lattice relaxation time (T1). When the radiofrequency pulse is terminated, energy is emitted from the tissue as a weak radiofrequency signal (4). These radiofrequency signals are received by an antenna within the MRI device and subjected to a series of computer operations that translate the signals into an image.

The antennae that receive the emitted radiofrequency signal are known as coils. A common arrangement is to have two such sets of coils for imaging. Usually, these coils encircle the area of interest and are known as head coils or body coils. When applied to a localized region of the body surface, such as the orbit, knee, or spine, they are known as surface or local coils (2). Surface coils are antenna coils that replace the standard coils for MRI. The same coil may be used to apply the "driving" electromagnetic impulses and receive the signals generated by the tissue. However, a body coil is typically used to transmit the radiofrequency signal, with the surface coil acting only as a receiver.

Currently, multiple surface coils of different geometries are available. Various configurations have been designed to conform to the body part being imaged. Coil configurations are usually simple loops or rectangles that may be flat, curved, or even pliable (to wrap around the extremity) (5). While some of the surface coils are organ specific, others can evaluate various regions in the body.

While some surface coils cost almost $10,000 each, the cost for a set of coils range from $5,000 to $40,000 (6). In comparison, upgraded magnetic gradient coils typically cost $100,000 to $250,000. Some vendors offer surface coils at no additional charge with the sale of an entire MR system. According to a written communication form Newhouse, the same surface coil may be used for countless examinations with little or no additional time necessary for the examination (7). Other investigators believe that the requirements for setup and positioning will increase the preparation time associated with surface coil MRI.

Although safety is not considered an issue, skin burns have occurred during examination of localized regions of tissue with (receiver) surface coils (4). If a body coil is used to transmit the radiofrequency pulse and the surface coil acts only as a receiver, energy given off by the body coil can, by induction, cause dangerously high currents in the receiver coil. When surface coils are coupled to the transmitted radiofrequency field, they resonate with the transmitted power and increase in temperature. Usually the coil is well insulated, with no part directly contacting the body. Insulating materials, such as mylar sheeting, are often placed between the coil and body to avoid local heating of the skin surface. Surface burns can be minimized by care taken by the provider and by some engineering changes in coils to reduce the risk of burns.

To avoid coupling, some surface coils are placed perpendicular to the transmitted radiofrequency field (8). According to Pearce, this passive geometric decoupling is not fail-safe and also severely limits the kinds of views that can be obtained (5). Other surface coils can be alternately tuned and detuned over millisecond intervals with electronic switches. When the radiofrequency field is on, the surface coil is off.

Rationale

As a result of the weak radiofrequency signal emitted by tissues in MRI, image resolution is often limited by a suboptimal signal-to-noise ratio. With surface oil technology, signal reception is increased by proximity of the organ of interest to the receiver antenna (9). Proponents argue that improved radiofrequency coupling between the body structure and the coil can yield substantial signal-to-noise ratio improvements over similar images produced with body coils. They report that the signal-to-noise ratio is two to four times greater when a surface coil is used, although fields of view are correspondingly smaller.

Since a major component of coil noise is the noise from tissues outside the region of interest, proponents contend that a surface coil has decreased noise levels as a result of its decreased field of view. This is considered an advantage when it suppresses artifacts produced by physiological motion of structures outside the field of view of the surface coil. Moreover, proponents have found that the rapid fall-off in sensitivity of surface coils (nonuniform response) allows application of stronger gradients to obtain higher spatial resolution without image wraparound. The increased signal-to-noise ratio obtained with surface coils can be traded for improved spatial resolution or decreased imaging time.

Review of Available Information

MRI procedures in conjunction with surface coil devices have been developing rapidly since Ackerman et al described surface coils in 1980 (10). Numerous studies have appeared in the literature describing the use of surface coils within various protocols of MRI techniques. In 1982 Bernardo et al demonstrated a method that allowed for satisfactory surface coil imaging despite the nonuniformity in the radiofrequency field distribution in the vicinity of the surface coil (11). Bernardo et al's method, tested with phantoms and images of the human back, showed that imaging data taken in the nonuniform field near a surface (local) coil could be satisfactorily displayed by normalizing the intensities (magnetic field) to those of a uniform phantom used as a reference. With these corrections the investigators were able to recognize previously unrecognizable structural features of the back. According to the investigators, surface coils properly designed for a specific application should provide higher resolution and signal-to-noise ratio with fewer artifacts than coils designed to surround the whole object.

In a similar study, Axel used an adapted MRI system with a circular surface coil in place of the conventional receiver coil to image the orbit, neck, chest wall, and lumbar spine (12). The use of a circular surface coil in this study improved the signal-to-noise ratio for regions of interest close to the coil, compared with images obtained with the conventional circumferential receiver coil. According to Axel, who used receive-only surface coils, an improved signal-to-noise ratio provides increased resolution imaging of relatively superficial structures, such as the orbit, neck, chest wall, and lumbar spine. In the studies by Bernardo et al and Axel, at relatively low magnetic field strengths (0.12T), surface coil imaging was used successfully as an alternative to the more conventional circumferential coil imaging (11,12). Axel recommended surface coil imaging for higher resolution images of relatively small superficial structures on parts of the body that would not fit into a head coil (12).

To further expand the use of surface coil MRI, Axel and Hayes investigated the application of four types of surface coils to different anatomical regions at high magnetic field strengths (13). For each type of surface coil and corresponding region of interest studied, they observed increased signal-to-noise ratios and improved spatial resolution. They found surface coils, flat or curved to fit body contours, adequate for most general imaging with a range of coil sizes useful for structures of different sizes or depths. According to the investigators, these coils could be used for imaging relatively superficial structures of the head and trunk. Because objects can be placed in solenoidal surface coils, the authors used this type of coil to image the breast and other relatively protruding structures. To provide better imaging of deeper structures within the limbs, they used smaller local versions of the conventional circumferential coil.

According to Axel and Hayes, the geometry of a surface coil, as well as its size, should be chosen for optimum reach and coverage of the region to be imaged (13). Although a larger surface coil will encompass a correspondingly larger region, there will be some decrease in the signal-to-noise ratio. For optimal imaging of anatomical regions of different sizes or at different depths, the authors recommend utilizing a range of sizes, shapes, and types of surface coils.

In 1985 a subsequent study by Hayes and Axel determined factors affecting surface coil performance (14). They imaged phantoms (1.5T magnetic field) using circular surface coils of various diameters and compared the results with body and head coil images. By using surface coils with 8-, 10-, and 14-cm diameters and varying the depth of the object imaged, the authors showed that the signal-to-noise ratio of surface coil images compared to head and body coil images increased as the coil radius decreased. In this study, the use of surface coils increased signal-to-noise ratio by a factor of four or more for regions of the body close to the surface. Hayes and Axel observed improved signal-to-noise ratio for all three surface coils for regions up ty 6 cm deep in the head and about 12 cm deep in the body. According to the authors, the choice of surface coil radius is primarily determined by the desired field of view extending parallel to the surface. However, the choice of surface coil size must be a compromise between the higer signal-to-noise ratio achieved with smaller coils, the greater depth resolution with larger coils, and the desire to image the entire area of interest without having to move the surface coil. Because multiple surface coils can evaluate the same regions in the body, Kulkarni et al recommended the use of coil-response curves to make a proper (optimal) surface coil selection (15). Phantom studies were used to measure the signal-to-noise ratio response characteristics of surface coils to demonstrate and compare the gain in signal-to-noise ratio over conventional head and body coils. Measurements at various distances (depths) were made at different gradient magnifications and with different numbers of signal averages.

In 1985 Schenk et al studied the utility of surface coil MRI at high magnetic field strengths (16). His group demonstrated improved resolution for neck, inner ear, and lumbar spine images of healthy volunteers when surface coils were used as receiver elements in MRI at 1.5T. According to the investigators, the improved signal-to-noise ratio generated by these surface coils permitted stronger (imaging) gradient fields that resulted in a reduced imaging voxel size with an improved resolution of anatomic detail.

To evaluate the clinical utility of surface coil MRI, Fisher et al studied surface coil images of the orbit, carotid artery, cervical and lumbar spine, thyroid gland, femoral artery, prostate, hip, heart, ankle, knee, wrist, and other structures in 48 patients (17). Image signal-to-noise ratio was calculated to compare surface coil images with images obtained with conventional coils. According to the authors, they obtained a twofold increased for the 10-cm surface coil versus the head coil and a 4.6-fold increase compared with the body coil. With a 20-cm surface coil, they obtained a 2. 3-fold increase compared with the body coil. For surface coils placed around limbs (miniature version of a body coil), the authors calculated 2.8-and 6.4-fold increases versus the head coil and the body coil, respectively. According to Fisher et al, the images obtained with surface coils showed marked improvements in image quality and detail. Improved signal-to-noise ratio allowed the investigators the option of reducing imaging time while maintaining image quality or improving image quality and facilitating the imaging of small and superficial structures.

As part of the study, Fisher et al compared MR images obtained with transmit and receive surface coils with MR images obtained with receive-only surface coils. They found use of the receive-only surface coil resulted in an image with a larger field of view, increased the number of useful sections obtained from one imaging sequence, and provided a more uniform signal within its sensitive volume. Based on these findings, the authors suggested that receive-only surface coils may be the surface coil configuration of choice.

According to Pearce, while a surface coil can be made to transmit radiofrequency power, the emitted field of most designs is far less homogeneous than that produced by the much larger body coil (5). Additionally, because the surface coil is closer to the tissue being excited, the phase angles of processing nuclei can become distorted, causing image artifacts. Alternatively, for high field-strength imaging (2T), Roschmann and Tischler point out the advantages of transmit and receive surface coils (18). Because power absorption is proportional to the patient volume exposed to the radiofrequency field, the spatial confinement of this field should lead to a substantial reduction of the radiofrequency power required. With transmit and receive surface coils, Roschmann and Tischler showed that the required radiofrequency power was reduced by factors of 2 up to 100 with respect to head and body coils. With a transmit and receive single-turn solenoid surface coil (loop-gap resonator), Hornak et al obtained simultaneous images of both breast with a reduced radiofrequency dose to the patient and an improved signal-to-noise ratio (19). Because the single-turn solenoid completely surrounds the breast, the image produced is uniformly intense across the field of view. With methods that use the body coil as the transmitter and a surface coil as the receiver, there is a fall-off in the intensity of the image as the object imaged gets farther away from the coil.

Fitzsimmons et al reported using receive-only surface coils, designed to image the neck region, to provide high-quality MR images for low-field strength systems (0.15T) (20). They demonstrated that in low-field strength systems where the signal strength is limited, increasing the signal-to-noise ratio with surface coils results in substantial gains in image quality for particular areas of the body inadequately imaged with standard coils. Structures imaged by the investigators included the cervical spinal cord, thyroid gland, parathyroid regions, larynx, hypopharynx, and knee.

According to Lufkin et al, planar (flat) and other circumferential surface coils have a marked loss of signal beyond one radius from the center of the coil (21). While these surface coils have a strong coupling and high signal-to-noise ratio to nearby signals, this dropoff in signal intensity limits their effectiveness when imaging deeper structures. To improve MRI of deep structures in anatomic regions such as the neck and knee that do not fit standard head and body coils, the investigators modified the geometry of the planar surface coil to a solenoidal configuration that surrounds the structure. With this coil design, any point in the imaging volume lies within one radius from the coil center.

Studies of phantoms and normal volunteers by Lufkin et al showed a threefold improvement in signal-to-noise ratio for solenoidal surface coils over images obtained with standard body and head coils. According to the investigators, the improved signal-to-noise ratio allows the use of steeper magnetic field gradients resulting in thinner sections and higher spatial resolution. The investigators reported that the higher resolution images obtained with the solenoidal surface coils were clearly superior to images of the same patient obtained without surface coils. Moreover, the problem with rapid dropoff of signal with distance when using planar surface coils was not observed. The investigators suggested that solenoidal surface coil imaging should be very useful in MRI of the neck, extremity joints, and breasts, depending on the orientation of the main magnetic field of the imager. Recently, Hoover et al reported obtaining high-resolution MR scans of the neck region with the use of solenoidal surface coils (22). With a marked improvement in image resolution, they were able to accurately delineate the extent of disease in both the base of tongue and larynx.

In another recent study of MRI of the neck, Kier et al reported using a saddle-shaped surface coil to localize parathyroid tumors (23). Because of the improved anatomic resolution achieved with the saddle coil, the authors suggested reducing the number of excitations per pulse sequence to reduce the time required for the examination. In some MR studies of the neck, investigators combine the use of two surface coils. In performing MRI of the larynx, McArdle et al placed a flat coil under the patient's neck and a second, saddle-shaped coil over the neck (24). The authors found that the combination of these surface coils eliminated signal fall-off in sections away from the coil.

In early studies of surface coil MRI (0.6T) of the lumbar spine, Edelman et al achieved a signal-to-noise ratio more than three to five times greater than achieved with the commonly used whole-body saddle coil (25). Using this high-resolution technique, Edelman et al were able to apply MRI to the detection and characterization of lumbar disk herniation. Moreover, these images were not affected by artifacts caused by respiratory, bowel, or vessel motion.

In 1987 Holtas et al reported that MRI of the spine could be markedly improved if surface coils were used instead of body coils (26). Using a flat surface coil and an MR system with a field strength of 1. 5T, the authors made measurements of the spatial distribution of the signal-to-noise ratio of the surface coil relative to the body coil for thoracic spinal images. They demonstrated that the signal-to-noise ratio of the surface coil can be approximately 2.6 times that of the body coil. They also found that the surface coil images could maintain a signal-to-noise ratio advantage (decreasing) up to 14 cm deep in the body. A comparison of surface coil with body coil images indicated a marked reduction of noise and motion artifacts on the surface coil image from the heart and chest wall or blood flow.

Edelman et al used large circular surface coils (12 to 18 cm in diameter) to image deep abdominal structures in healthy volunteers and patients (27). These structures, including the kidneys, pancreas, spleen, liver, and adrenal glands, are usually in a range of depths from 6 to 15 cm. According to the authors, the twofold to fourfold improvement in signal-to-noise ratio resulted in a marked enhancement in image quality compared with whole-body coil imaging. They found the quality of surface coil images (greater anatomic detail) of the kidneys, pancreas, and spleen superior to the body coil images in all subjects. The region of improved sensitivity of the large circular surface coils extended far enough from the coil to encompass any of the abdominal viscera, except portions of the liver distant from the coil. Additionally, the restricted field of view of the surface coils resulted in images with minimal motion artifacts when compared with body coil images.

Simeone et al compared surface coil images of the pancreas with conventional body coil images in healthy volunteers and patients (28). According to the authors, the pancreatic surface coil images had a twofold improvement in signal-to-noise ratio. They also were able to demonstrate decreased image degradation caused by respiratory motion on the surface coil images. The anatomic resolution was improved on surface coil images because the increased signal-to-noise ratio and decreased motion artifacts allowed the acquisition of images with increased spatial resolution. Alternatively, when the authors used the surface coil to obtain images with the same spatial resolution as the body coil, they were able to shorten examination time.

A report by White et al presented the findings of surface coil MRI of the adrenal glands (29). The adrenal glands were identified with body and surface coils in healthy volunteers and patients. As a quantitative measure of image quality, the signal-to-noise ratio was calculated for surface coil images and compared with the signal-to-noise ratio calculated for body coil images. In T1-weighted studies, the authors found the high-resolution surface coil images showed a threefold improvement in signal-to-noise ratio over body coil images. The surface coil's limited filed to view eliminated respiratory motion artifacts. According to the authors, the surface coil images provided better intrinsic resolution of small adrenal lesions and clearer definition of the extrinsic relationships of large masses to nearby organs.

Although the use of a single circular surface coil will improve the image quality of abdominal viscera, the rapid loss of signal in deeper structures makes the coil less suitable for imaging the bladder and other pelvic organs, according to a 1986 report (30). To image the bladder, Barentsz et al developed a "sandwich type" double surface coil consisting of a combination of two rectangular surface coils placed on the front and back of the patient. They and other investigators have found that the depth-dependent signal loss common to surface coil imaging is minimized if a second surface coil is placed on the opposite side of the region of interest (30,31). With this type of coil the investigators found there is better radiofrequency field homogeneity, resulting in a higher signal-to-noise ratio of all structures in the male pelvis.

Initially, surface coil configurations were simple loops or rectangles; later they became customized for particular imaging purposes. As discussed above, solenoidal surface coils were designed to completely surround the object to be imaged such as the breast, knee, or arm to eliminate the image (signal) dropoff seen with planar coils. MR systems with a vertical field may use solenoidal (circumferential) coils, while MR systems with a horizontal filed must use the slightly less efficient saddle (circumferential) surface coil. Other developments in surface coil design include the double-loop, receive-only coil studied by Akins et al (32). Phantom studies as well as body imaging in patients showed the double-loop coil provided in a merkedly improved signal-to-noise ratio compared with the body coil. Although an oval, flat, conventional surface coil outperformed (higher signal-to-noise ratio) the double-loop coil in the near field, the double-loop coil was superior at depths greater than 9 cm. The cross-coupled, double-loop design provided a higher signal-to-noise ratio and a more homogeneous signal intensity at increasing depths than either the receive-only body coil or the conventional flat surface coil. According to the authors, who also assessed image quality in 75 patient who underwent clinically indicated MRI, the use of this coil substantially enhanced the capability of their MR instrument to image the heart, pelvis, hip, and shoulder. Use of the coil allowed the authors to reduce the pixel volume by using higher field gradients and thinner slices or to shorten imaging time by using fewer signal averages.

Akins et al reported the use of gating to facilitate cardiac imaging. They found the double-loop, receive-only coil could improve signal reception from the heart without significant degradation of images by respiratory motion. However, according to the authors, this type of coil has increased sensitivity to respiratory motion artifacts and, without respiratory gating for imaging in the upper abdomen, there will be an increased likelihood of degradation due to these motion artifacts.

Because the sensitivity of surface coils decreases with increasing distance between the coil and anatomic structure being imaged, investigators have designed surface coils to fit particular anatomical structures (33). In 1986, Doornbos et al reported on the development and application of eight different surface coils used in conjunction with a 0.5T MRI system. According to the authors, the shape and geometry of each coil is determined by the size of the area of interest and its depth below the skin, as well as by the scan plane required. With these coils the investigator were able to obtain high-resolution images of superficially located anatomic structures that included the larynx, scrotum, eyes, shoulder, ankle, knee, and wrist.

The high-resolution images of the larynx obtained by the investigators with the surface coil developed for larynx examination were in sharp contrast to the larynx images of relatively low signal-to-noise ratio usually obtained with head or body coils. The high signal-to-noise ratio obtained with surface coils for the larynx and the other structures studied allowed Doornbos et al to image thin slices and small pixels that resulted in high-resolution images. According to the authors, the construction of surface coils is rather simple, and the electronic equipment needed for their operation is relatively inexpensive and easy to assemble. The authors also noted that the electronic device needed for tuning and matching is interchangeable between various coils.

Discussion

The focus of the surface coil section of the report has been the value of surface coils in enhancing the quality of MR images. The published literature indicates that the development and application of surface coils has been an important aspect of advancing the capabilities of MRI. The value of surface coils to the technical capacity of MRI has been well documented (34). The importance of surface coils lies in the fact that MRI is a noise-limited procedure. To understand the success of surface coils, it is necessary to understand the sources of noise that degrade the MR image. According to Kneeland, the most important source, at field strengths generally used for imaging (> 0. 3T) is the noise that arises from the random motion of electrolytes and other charge-carrying molecules in the patient's body (35). These induce a voltage in the receiver coil (noise) that is superimposed on the signal. A coil (conventional) body or head coil) that receives a radiofrequency signal from a large sample (volume) of tissue also receives noise and artifacts from that sample and superimposes them over the entire image. Thus, images of any small region of interest (tissue) contain noise from the entire volume. Because surface coils receive signals and noise from a small region, the radiofrequency signal is degraded only by the noise from that region and not from tissues outside the region of interest. A review and analysis of recent studies have demonstrated that significantly improved resolution of small anatomic structures is possible when surface coils are used as receiver elements in MRI systems.

Studies of the resolution capability of surface coils have been documented in the literature by Axel and others who have shown definite measurable improvement in resolution using both phantom, and clinical patient imaging (12-14). Hayes and Axel demonstrated increased signal-to noise-ratio in phantom surface coil images by a factor of four or more for regions of the body close to the surface (14). In healthy volunteers, Schenck et al demonstrated improved resolution of small anatomic structures, including the neck and inner ear, with high-field (1.5T) surface coil MRI (16). If the neck is imaged using a body coil, the arrangement yields relatively poor coupling and a suboptimal signal-to-noise ratio. The surface coil technique provides high-resolution images of regions of anatomy not readily accessible to standard MR coils.

An increased signal-to-noise ratio resulting in high-resolution images has been used as the standard measure of improved MRI. Strong inductive coupling of the surface coil to small regions of interest has resulted in increased radiofrequency signal detection. Moreover, because surface coils receives signal from a limited area of the body as discussed above, there is a marked decrease in thermal noise and motion artifacts. Furthermore, the problem of wraparound (aliasing) artifacts encountered on body coil imaging when using increased in-plane magnetic field gradient can be eliminated with surface coils, since no peripheral tissue signal is received to be folded into the (wrapped around) images (27).

Modest gains in signal and decreased in noise result in significantly improved signal-to-noise ratio. Because of the small size of the coil, limited field of view, and the proximity to the area being imaged, a three to five times greater signal-to-noise ratio often is achieved, compared with conventional whole body coils. In the study by Edelman et al, surface coil MR imaging of the lumbar spine achieved a signal-to-noise ratio more than three to five times greater than achieved with a whole-body saddle coil (25). This was attempted with a 0.6T MR system. Holtas et al demonstrated that the signal-to-noise ratio of the surface coil image of the spine can be approximately 2.6 times that of the body coil with a high-field strength (1.5T) MR system (26). The surface coil's limited field of view contributes to high-quality MR images of the spine. Because the surface coil is less sensitive to signals distant from the coil, there is less signal disturbance from motion artifacts. While surface coils have substantially improved the resolution of the spinal column on MR images, the limited field of view requires long segments of the spinal column to be examined with consecutive series of images. According to Haughton, several techniques are being evaluated that may provide optimal resolution through a sufficiently long region of interest without a significant increase in acquisition time (36).

In the surface coil MRI study by Fisher et al, a marked improvement in the images' signal-to-noise ratio was demonstrated when they obtained a twofold increase for a 10-cm surface coil versus the conventional head coil and a 4.6-fold increase compared with the conventional body coil (17). For a 20-cm surface coil, they obtained a 2.3-fold increase in signal-to-noise ratio, compared with the body coil. Included in the structures imaged with surface coils were the kidneys and the parotid, thyroid, and prostate glands.

Initially, Fisher et al performed MRI with surface coils used both for transmission of radiofrequency pulses and for reception of MR signals. Due to disadvantages that included radiofrequency nonuniformity and a smaller field of view, the investigators began imaging with a receive-only surface coil. The radiofrequency pulses were transmitted using a conventional coil, with the surface coil used only for reception. Receive-only coils can take advantage of the good uniformity of the whole-volume transmitter coils. Also, if the radiofrequency magnetic field homogeneity is no longer crucial in the surface coil design criteria, the shape of the surface coil can be adapted to fit the anatomical region and improve patient comfort. Furthermore, close positioning increases the sensivity (37). Imaging was found to be less problematic with these coils (17). Receive-only surface coils, however, require decoupling from the transmitted radiofrequency (35). Boskamp found that the electronic decoupling method enabled the free and easy positioning of surface coils (37). Although some investigators typically use receive-only coils, there are advantages associated with each design.

A number of surface coils have been developed for musculoskeletal imaging. Using the most closely coupled coil possible will maximize the signal-to-noise ratio and achieve the best spatial resolution. Fisher et al reported a strong coupling of saddle-shaped limb surface coils to the ankle, knee, wrist, and elbow (17). They demonstrated 2.8-and 6.4-fold increase in image signal-to-noise ratio compared with the head and body coil, respectively. To provide improved imaging of deeper structures within the limbs, Axel and Hayes recommended smaller versions of the conventional circumferential coils because they could be slipped over the extremity (13). Solenoidal surface coils were recommended by the investigators for body parts such as the foot. Because structures that project from the body can be placed within solenoidal coils, this type of coils has been shown to be suitable for imaging the breast as well. Many designs for dedicated breast coils have been developed and tested. Each design has its own limitation and technical problems. However, surface coil designed for breast imaging provide significant improvements in the quality of MR images of the breast (38,39).

Initially, surface coils were intended for improved viewing of superficial structure and were not consider for improved MRI of deeper structure. Most MR examinations of the upper abdomen have used a conventional whole body coil, which has the advantage of a uniform sensitive volume and large field of view allowing simultaneous examination of the liver, spleen, pancreas, adrenals, and other retroperitoneal organs. Conventional MRI of the abdominal viscera has been limited by low signal-to-noise ratio and artifacts caused by the physiologic motions of respiration and peristalsis. However, studies reported by Edelman, White, Simeone, and their associates indicate that large surface coils contribute to improved MRI of abdominal structures (27-29). They found the imaging quality of relatively deep abdominal structures, such as the liver, pancreas, kidneys, spleen, and adrenals, could be improved. In their study on human subjects, they found a twofold to fourfold increase in signal-to-noise ratio compared with body coils for structures within 12 cm of the coil surface at any given level of spatial resolution. However, because of the limited field of view that can be imaged (examined), Edelman et al recommended that CT or conventional whole body MRI be used first, and that high-resolution surface coil imaging be used to focus on a specific organ in the abdomen to give greater anatomic detail (27).

Based on the findings of their surface coil MR study of adrenal glands, White et al recommended high-resolution surface coil imaging when CT or conventional whole-body MRI has identified an adrenal abnormality but failed to delineate anatomic relationship with the inferior vena cava, kidney, or liver (29). The restricted field of view and additional time required for patient positioning was compensated by the significant improvement in anatomic detail.

A significant improvement in anatomic detail was also reported by Simeone et al for surface coil MRI of the pancreas (28). Placement of the surface coil at the patient's back provided pancreatic images with a twofold improvement in the signal-to-noise ratio compared with body coil images. The surface coil's limited field of view decreased image degradation caused by peristalsis and respiratory motion that normally accompanies conventional whole body MRI. The resulting improvement in anatomic resolution provided the investigators with valuable diagnostic information regarding pancreatic abnormalities. According to Stark, the location of the pancreas in the abdomen limits signal-to-noise ratio improvements obtained with surface coils (40). He claims that difficulties in patient positioning and delays in patient through output discourage clinical use.

In contrast, MR is better suited for scanning in the pelvis because there is relatively little respiratory motion to degrade the image (41). However, there is loss of signal due to the depth of these organs. Signal-to-noise ratio can be improved by the use of surface coils, as has been reported for MRI of the bladder and scrotum (30,42).

Surface coils that provide an increased signal-to-noise ratio can be used to produce the same resolution in a shorter time or improved resolution in the same time as conventional MR techniques. In addition to surface coils, many other technologic considerations affect the quality of an MR image, such as the magnetic field strength and uniformity, radiofrequency coil design and placement, and pulsing sequences. What some imaging systems are able to accomplish with the use of surface coils, others may achieve with selected pulse sequences or magnetic field gradients. The restricted field of view of surface coil imaging reduces physiologic motion articfacts such peristalsis, aortic pulsations, and respiration, because they are usually outside the sensitive region of the coil. However, the limited field of view allows examination only of a relatively small region. Also, surface coil examinations require more preparation to apply the coil to the patient, and additional time may be needed to position the field of view in the desired anatomic region. According to Ehman, many surface coils have important constraints that also must be considered in terms of their orientation and position within the magnet (43).

Consultations. OHTA has been informed by the National Institutes of Health (NIH) that surface coils are a safe, effective, and appropriate means of increasing the signal- to-noise ratio in viewing organs within their volume of sensitivity. According to NIH, surface coils have proved especially useful for improved imaging of structures near the surface of the body such as the spine, the temporomandibular joint, the smaller structures in the neck (such as the parathyroids), the orbit, and the extremities. Also, surface coils are advantageous for imaging small arteries in the head and will improve every extracranial central nervous system of MRI study.

According to NIH, the usefulness of surface coil MRI is not limited to surface structures. Surface coils will improve imaging of organs they are placed close to no matter where in the body these organs may be. The quality of images of relatively deep abdominal structures such as the liver, pancreas, spleen, kidneys, and adrenal glands has been improved with surface coils. In addition, surface coils have been used to increase the sensitivity of signal detection and improve images of MRI studies in the pelvic region, including the bladder, ovary, uterus, prostate, and scrotum.

NIH has informed OHTA that MRI has become the imaging technique of choice for diseases of the nervous system, especially demyelinating diseases, cerebrovascular diseases, spinal cord diseases, and basal ganglia pathology. MRI with surface coil techniques has been found in studies to be better than the conventional MRI because surface coils improve the quality of the examination and its diagnostic information and may result in better patient treatment.

According to NIH there are no known patient safety concerns when surface coils are properly employed. Although surface coils may carry large currents and cause a temperature rise which may burn the skin, insulation around the coil and diode circuitry can protect the patient.

As of September 1989, the Food and Drug Administration (FDA) had approved the MRI devices of 18 companies for marketing in general interstate commerce. Fourteen of the companies are marketing surface coils. According to the FDA, surface coils are appropriate for imaging small areas and improving the signal-to-noise ratio and image quality.

Medical Specialty Responses. Comments received from the American Heart Association stress that surface coils are a useful adjunct to MRI and are particularly suited for imaging small superficial anatomic structures. The Association indicated that surface coils have been widely used for imaging of the spine, joints, and blood vessels, where they are capable of producing images with considerably higher resolution than those produced with conventional whole body coils.

According to the National Electrical Manufacturers Association (NEMA), surface coils provide additional resolution and improved clarity due to higher signal-to-noise ratios when compared with conventional head and body coil MRI. NEMA proposes that improvements in the clarity of an image lead to improved diagnostic confidence for the clinician. NEMA recommended surface coils for imaging the head and neck; orbit; temporomandibular joint; acoustic canal; lumbar, thoracic, and cervical spine; musculoskeletal system; prostate and testicles; kidney; and breast.

The American College of Cardiology has stated that surface coils improve the detail with which the endocardial border and fine structures of the heart can be visualized and are potentially advantageous in imaging the heart and great vessels.

According to the American College of Radiology, there are numerous areas of clinical imaging in which surface coil MRI provides effective and useful information. These include structures within the orbit, the temporomandibular joint, larynx, thyroid and parathyroid, shoulder, elbow, wrist, breast, scrotum, knee, ankle, and spine. The College finds the cost of using surface coils to be minimal, since the coils can be used for countless examinations. It also believes that surface coil MRI involves little or no extra time and that technologists will require only minimal training in the use of surface coils.

The American Association of Neurological Surgeons has indicated that surface coils are used routinely for high-resolution imaging of most areas of the body where a specific restricted portion of the anatomy is being studied. According to the Association, small surface coil designs are used for the orbit, thyroid, larynx, temporomandibular joint, and temporal bone, while slightly larger surface coils are necessary for high-resolution imaging of the cervical and lumbar spine. For cardiac imaging, a saddle-type solenoid coil is used. The Association state that surface coils and gating are necessary for adequate cardiac and abdominal imaging.

The American Society of Colon and Rectal Surgeons has pointed out that surface coils are used quite extensively in imaging the pelvis, back, and spine. The Society recommended surface coil MRI in any application requiring high resolution where the area to be imaged is within 10 cm of the surface.

Summary

Surface coils are localized antennae that receive radio-frequency signals from the region of interest (tissue) during MRI. When the surface coil functions only as a receiver element, the excitation field is produced by the much larger body coil, which provides greater uniformity of the excitation. With the surface coil, the receiver system is brought much closer to the anatomy being scanned, and the resulting signal is picked up only from the area defined by the diameter and shape of the coil. Surface coils are used to improve detection of the signal and reduce interference from motion artifacts outside their sensitive volume (area of interest). The restricted field of view of surface coil imaging reduces physiologic motion artifacts such as peristalsis, aortic pulsations, and respiration. When body and head coils are inefficient because of the poor coupling between the coil and the small size of the region of interest, surface coils have been used to produce relatively high signal-to-noise ratios and greater anatomic detail. Published stuies have reported improved effectiveness (resolution) of surface coils as compared with conventional whole body coils. Although safety is not considered an issue, certain precautions must be taken to avoid skin burns. Surface burns can be minimized by care taken by the provider, and there have been some engineering changes in coils to reduce the risk of burns.

For images obtained in equal time and with equal spatial resolution, surface coils improve image quality compared with body coils by virtue of the improved signal-to-noise ratio, which results in better anatomic detail and tissue contrast. Alternatively, this greater signal-to-noise ratio can be traded for better spatial resolution or reduced scanning time. Surface coil MRI also eliminates the problem of fold-over (wraparound) artifacts since it is insensitive to signals from structures at the extremes of the image. In addition to surface coils, many other technologic considerations contribute to the quality of an MRI image, such as the magnetic field strength and uniformity, radiofrequency coil design and placement, and pulsing sequences. What some imaging systems are able to accomplish with the use of surface coils, others may accomplish with selected pulse sequences or magnetic field gradients.

The use of surface coils in MRI has increased because of the improved to-noise ratio and image quality compared with imaging that uses whole body or head coils. Surface coils may be advantageous for many kinds of imaging procedures, particularly for small structures that require greater spatial resolution. They are now used routinely for high-resolution imaging of portions of the body where a limited region is being studied. Currently, multiple surface coils of different configurations are available, including flat, curved, and pliable coils that conform to the surface anatomy. Surface coils can be applied to various areas of the body such as the neck, vertebral column, breast, scrotum, heart, and extremities. Small coil designs are used for structures such as the orbit, thyroid, larynx, temporomandibular joint, and temporal bone. Although the improved signal-to-noise ratio over standard body and head coils is proportionately decreased by an increased coil diameter, slightly larger configuration surface coils have provided improved resolution imaging of the spine, abdominal viscera, and pelvic region. MR surface coil imaging of the deeper pelvic structures, such as the bladder, have been improved with the advent of double surface coils placed on the front and back of the patient.

Surface coil examinations may require more preparation to apply the coil to the patient, and additional time may be required to position the field of view in the desired anatomic region. For some uses, MR surface coil imaging is considered the examination of choice; for others, it is recommended as a followup to other imaging modalities, such as ultrasound and computed tomography or conventional MRI, when greater anatomic detail is required of a specific structure.

NIH has informed OHTA that surface coils are a safe, effective, and appropriate means of increasing the signal-to-noise ratio in viewing organs within their volume of sensitivity. According to NIH, there are no known patient safety concerns when surface coils are properly used. The FDA reports that 14 companies are marketing surface coils.

Gating

Background

Magnetic resonance images are built up from information that is acquired over a relatively long period of time. Imaging times for MRI generally exceed those required for CT and other imaging technologies. An MRI examination may consume from several minutes to more than one-half hour (1). In MRI, motion has be detrimental effect of causing artifacts and distortions to appear in images, as well as producing signal loss. Since the acquisition of data in MRI takes a relatively long time, physiologic motion in such activities as respiration, cardiac motion, blood flow, swallowing, and peristalsis results in both blurring and motion artifacts. Motion artifacts appear as "ghosts" in the image (44). Gated MR images of the heart and other organs have been made in an effort to overcome the distortions of motion artifacts. Gated images synchronize the scanner with the physiologic cycle (for example, cardiac cycle) so that each repetition of the scanning pulse sequence (repetition time, TR) is applied with the organ in the same phase of the cycle. According to a written communication by Hammond, this essentially "freezes in" the organ at some point in its cycle for the purpose of imaging (45).

Cardiac. In early MRI studies of the heart, there was severe degradation of the image due to cardiac motion (46). With cardiac gating, data are acquired repeatedly at the identical phase in the cardiac cycle in order to reduce artifacts produced by cardiac activity. Cardiac gating divides the cardiac cycle into a number of frames based on the relationship between the signal acquisition and the cardiac cycle. In cardiac gating the collection of data for the image can be activated by R wave detection on the electrocardiograph (EGG), by plethysmographic signal from a blood pressure cuff, or from a laser-Doppler ultrasound signal. EGG-activated data collection is considered the most reliable, and it can be accomplished without interfacing with the MR signal (47). Images are then generated from averages of all the signals received in a particular frame. Each frame gives a "stop action" view of the heart at a certain point in the cardiac cycle (46).

It has been suggested that cardiac gating not only improves cardiac imaging but also may be useful for MRI studies of the chest including the great vessels, mediastinum, and hila (48). Because these structures are adjacent to the heart, motion artifacts from transmitted pulsations must be considered. Below the diaphragm, cardiac gating has been used for imaging the liver, especially the left lobe, which is located beneath the heart and, therefore, also subject to transmitted motion.

Because motion artifacts can occur due to the pulsatile flow of blood or cerebrospinal fluid, cardiac gating also has been used in imaging other structures. Images of the spine and spinal cord have been obtained by performing cardiac gating to reduce noise and ghosting artifacts due to cerebrospinal fluid pulsation. Because of cerebrospinal fluid motion within the ventricles and subarachnoid space, cardiac gating also has been employed in MRI of the brain (45). For spinal and head imaging in which the gating signal need not be as precisely related to the cardiac cycle, the peripheral pulse has been used as a gating signal.

Use of the cardiac gating procedure in conjunction with MRI should not represent any hazard for the patient. However, in cardiac gating which ordinarily employs electrocardiographic monitoring, conducting wires extend from the patient to the ECG amplifier. With the ECG leads immersed in the magnetic field of the scanner and in the vicinity of the radiofrequency excitation pulses, it is conceivable that currents could be induced in these wires. Because low-intensity radiofrequency fields are utilized, these induced currents should only be in the low microampere range. If the length of the ECG lead is an exact multiple of the radiofrequencies in the scanner, larger voltages can be induced. The potential for induced currents can be lowered by reducing the length of the ECG leads exposed to magnetic or radiofrequency fields. Additionally, existing recommendations or standards for ECG systems have provisions for limiting current in patient leads. The American Heart Association recommendations of 1975 and the American Standard on Diagnostic Electrocardiographs of 1983 require that patient lead currents in an electrocardiograph be limited to no more than 10 microamperes and 20 microamperes, respectively. According to Lanzer et al, no patient-related risks appear to be involved when using the ECG gating procedure (47). Furthermore, these investigators believe that the ECG tracing provides a convenient and important monitoring tool for cardiac patients during MRI evaluation.

The capital equipment costs of the cardiac gating instrumentation and software are considered a small fraction (about 2 percent) of the total cost of the complete MR system. The increased scan time involved in cardiac gating could affect patient throughput and operational costs. Additional training of 1 to 2 days has been recommended for technologists using gating procedures (7).

Respiratory. Respiratory gating was introduced as a modification to conventional MRI for the purpose of eliminating artifacts due to the motion of the diaphragm, chest wall, and lungs (49). Diagnostic MRI of the chest and abdomen is hampered by the normal periodic motion of the Diaphragm associated with breathing (50). Respiratory movement causes both image ghosts and a blurring of spatial resolution. Moreover, the greater the movement and magnetic field strength, the stronger the ghost images. With respiratory gating the electromagnetic pulse and subsequent energy release data (radiofrequency signal) are coordinated to occur at a specific point in time in the respiratory cycle to decrease movement-induced artifacts. While the radiofrequency pulse is continually applied to the tissue to maintain a consistent decay in observed signal, a respiratory monitoring device triggers data acquisition to a constant period of the respiratory cycle. MRI studies of the chest and abdomen have focused on gating and thereby data collection at the end of expiration because it is a period of minimum motion. Also, in most patients there is a slight pause in the cycle at this point before initiation of the next inspiration (49). Since the respiratory cycle is much slower than the repetition rates normally used in scanning, this results in a major increase in scan time. Restricting data acquisition to the end of the expiratory portion of the cycle increases the imaging time two to three times.

Various methods are available to monitor the respiratory cycle. A technician may choose to watch the flow of air at the nose or mouth or monitor changes in the diameter of the chest. At times a sensor at the nose or mouth is used to initiate signals by monitoring the temperature changes of the expired and inspired air. If impedance plethysmography is used, the change in electrical resistance between two electrodes placed on the chest is monitored. This change produces a function that also varies with respiration (49). Detecting chest wall motion has been a common method for monitoring the respiratory cycle and generating signals for data acquisition. With this method a bellows-like device placed on the patient's chest will expand and contract as the patient breathes. These changes in air pressure can be detected and used to produce an appropriate on-off signal of data acquisition.

Combined cardiac and respiratory gating has been attempted for MRI of structures affected by both types of motion. To accomplish combined gating, the Q wave of the ECG is used to trigger the radiofrequency pulse, with actual data acquisition occurring only during the appropriate portion of the respiratory cycle (49). Combined cardiac and respiratory gating has been used in conjunction with MRI of the heart and chest.

Rationale

MRI is sensitive to motion due to the movement of protons into or out of the imaging volume between the radiofrequency pulses of the imaging sequence (51). Respiratory and cardiac movement degrades MR images of the chest, abdomen, and other areas by increasing noise through the production of "ghost" artifacts and by decreasing edge sharpness in moving structures. Proponents of gating argue that a technique that synchronizes the MR pulse sequence to a specific phase of the cardiac or respiratory cycle is necessary to provide sharp discrimination of moving structures. Gating makes the position of an anatomical structure constant by enabling image acquisition only when the structure has the same configuration. During MRI, signals (radiofrequency) from the patient (tissue) continue all the time, but the patient's ECG (R wave) or end-expiration will define a "window" (gate) in time during which signals are recorded for image reconstruction.

Proponents believe that gating should be utilized in imaging of structures that display motion that can be correlated to cardiac or respiratory activities. With gating the signal-to-noise ratio is improved with attendant improved image quality. Proponents recommend gated MRI procedures whenever movement of a body structure needs to be eliminated to obtain satisfactory images of the organs under study.

Review of Available Information

Several of the early gated MRI studies demonstrated improved image quality. In 1984 Schultz et al reported the results of MRI studies of the chest and abdomen with respiratory gating only, with cardiac gating only, with both respiratory and cardiac gating, and without gating (52). They considered the gated images of significantly better quality than the nongated images and found cardiac gating provided clearer definition of ventricular chambers and myocardium. With respiratory gating there was better definition of abdominal structures, and much of the image blurring in the chest and abdomen was eliminated. Combined gating of the cardiac system provided improved imaging of the heart valves and coronary arteries. However, combined gating in the abdomen provided no advantage over respiratory gating alone. Although Schultz et al were able to eliminate image blurring with respiratory gating, this improvement was associated with increased imaging times of 33 to 100 percent.

A 1984 study by Runge et al demonstrated improved quality of respiratory gated images in thoracic and abdominal patients as compared with ungated images (53). However, the use of the technique doubled data acquisition time. In the same year Lanzer et al studied the use of gating to improve MR signal intensity and imaging of the heart (47). The investigators used a plethysmographic signal, a laser-Doppler signal, and an ECG signal for acquiring a gating signal that could be used for timing MRI sequences. Although they obtained increases in the anatomic resolution of the cardiac MR images with each gating method, they considered the ECG method of data acquisition synchronization more reliable and efficient. Because of predictable and constant timing of images to specific segments of the cardiac cycle, electrocardiographic gating is considered the standard technique (54). This study by Lanzer et al, as well as a study by Herfkens et al, demonstrated the feasibility of gated MRI for defining cardiac anatomy (47,55). A 1985 study by Lanzer et al found that resolution of heart anatomy on gated MR images was adversely affected by prolonged spin-echo time delay, imaging in late diastole, image acquisition at the cardiac apex, irregular triggering, and artifacts (56). The investigators concluded, however, that ECG-gated MRI (at 0.35T) appears safe for patients, permits diagnostic resolution of images, allows image acquisition at distinct points during the cardiac cycle, and enables monitoring of patients during imaging.

In a study of determine the best method for obtaining respiratory signals suitable for respiratory gating, Ehman et al also demonstrated improved quality of gated versus nongated images of the abdomen (57). Their analysis of the different segments of the respiratory cycle found the greatest reduction of artifacts was achieved from end-expiration to the beginning of inspiration. As in other respiratory gating imaging studies, imaging time was an issue. The authors concluded that, even though respiratory gating greatly improves abdominal imaging, the inherent time delays limit its usefulness. Some investigators have introduced prescribed breathing patterns by the patient that lengthen the time between breaths to reduce imaging time without compromising image quality. More recently, Groch et al, in a written communication, described their finding have found that by reducing the number of sequence repetitions by a factor of 2, respiratory-gated images of improved quality can be acquired in approximately the same time as nongated images (58).

While MRI utilizing respiratory is associated with severe limitations from greatly increased imaging time, gated MRI of the heart is not. Higgins et al in a number of studies demonstrated that gated MRI is feasible for imaging the heart without adverse increases in imaging time (59-61). In gated MRI studies of the heart performed by Lieberman et al, acquisition time was approximately 5 minutes for each image obtained with cardiac gating only (62). Cardiac gating was accomplished by initiating the imaging sequence at the occurrence of the R wave of the ECG. According to the investigators, gated images were superior to nongated images, and images that were the product of both cardiac and respiratory gating gave better detail than did images produced by cardiac gating alone. A rubber bellows attached to the patient's chest was the respiratory gating technique used to signal end-expiration and the beginning of data acquisition. The investigators observed excellent anatomic detail of the cardiac chambers, valves, and myocardium. However, when cardiac and respiratory gating were combined, acquisition time increased to 10 minutes.

In another early study of gated MRI, Stark et al evaluated the pericardium and pericardial abnormalities with the use of gated and nongated MRI techniques (63). The 10 patients with pericardial abnormalities also were evaluated with serial chest radiographs, ultrasound, computed tomography, or angiography. Stark et al demonstrated gated MR sensitivity in detecting abnormalities to be as good as the comparison imaging modalities. The investigators concluded that gated MRI can detect and differentiate a variety of pathologic processes affecting the pericardium. According to Fletcher et al, gated MRI also has proved useful in diagnosing several types of congenital cardiac abnormalities (64). Of the 19 malformation cases studied by Fletcher et al, all but three were detected by MRI. Additionally, no false positive MRI diagnoses were made. These early positive experiences in small groups of subjects resulted in more extensive studies to evaluate gated MRI of the heart and abdomen as well as studies of the brain and spine.

In 1985 Higgins et al reported finding gated MRI a practical cardiac imaging modality for the evaluation of a variety of central cardiovascular abnormalities (65). From a study of 172 subjects, the investigators determined that image quality had sufficient depiction of internal cardiac anatomy to permit the diagnosis of abnormalities in more than 90 percent of the cases. They reported that the gated MR images provided a visually more precise demonstration of abnormalities, and because of the large field of view, a clearer depiction of the extent of pathologic processes. Although in most cases the diagnosis had already been established by another imaging technique, the investigators concluded that gated MRI was an effective technique for cardiac diagnosis. In 1986 Didier et al, in a large series of patients with congenital heart disease, demonstrated the clinical effectiveness of gated MRI for detecting congenital abnormalities of the heart and great vessels (66). Gated MR images in 72 patients with a variety of congenital anomalies of the heart and great vessels were graded and corroborated by angiography or two-dimensional echocardiography. According to the authors, gated studies that were considered to be diagnostic were obtained in 96 percent of the cases. In a study of patients with congenital heart disease (atrial-septal defects), Diethelm et al reported that gated MRI had high sensitivity and specificity in the diagnosis and location of atrial-level shunts (67).

Other studies of gated cardiac imaging have shown the technical capacity of MR scanning for imaging of cardiac and pericardiac tumors (68,69). , myocardial ischemia and infarction (70,71). , flow measurements (72,73). , and parameters of heart function (left ventricular wall thickening and ejection fraction) (74-76).

More recent studies also have shown the usefulness of gated MR in aortic imaging. A study Pernes et al of 30 patients with suspected or known chronic aortic dissection found cardiac-gated MRI to be an accurate method for evaluation and followup (77). More recently, Kersting-Sommerhoff et al have also demonstrated that gated MRI is a highly effective modality for the evaluation of the thoracic aorta (78). When there was experience with flow artifacts on the part of the radiologis, MRI was fond to be highly sensitive and specific in the diagnosis of aortic dissection. Unlike angiography and computed tomagraphy, gated MRI does not require catheterization or contrast medium.

A review of MR application to the cardiovascular system was presented by Pohost and Canby in 1987 (79). They reported that gated high-resolution MR images depict both normal and abnormal cardiovascular structures with excellent detail. According to the authors, recent comparative studies have corroborated the accuracy of gated MRI to quantify cardiac chamber volume and myocardial mass. Pohost and Canby concluded that the excellent resolution of gated MR images, and their inherent contrast, sensitivity to blood flow, dimensional nature, and other characteristics make this procedure suitable for imaging the cardiovascular system. They also found that the relatively high cost and lengthy image acquisition times limit the clinical utility of MRI when compared with other effective and less expensive noninvasive modalities, such as echocardiography and radionuclide procedure. However, the noninvasive nature of MRI may extend its clinical utility in certain areas.

According to Higgins the techniques used in MRI of the heart depend on the primary goal of the study (80). Because the ECG-gated spin-echo technique provides images with high signal-to-noise ratios, it has the technical capacity to evaluate anatomic abnormalities. The review by Higgins has shown gated MRI to be effective for evaluating a variety of cardiovascular diseases, including cardiomyopathy, pericardial disease, intracardiac thrombus, and thoracic aortic disease.

Because cardiac gating was shown to reduce blurring and motion artifacts for pulsating structures such as the heart and great vessels, investigators attempted as utilize cardiac gating to improve imaging of other structures affected by the degradation of cardiac motion. Westcott et al evaluated the effect of cardiac gating on the diagnostic quality of MR images of the hilum and mediastinum by comparing gated with nongated scans in 20 patients (48). An analysis of these comparative scans showed reduced blurring and motion artifacts on the gated images. However, the advantages of cardiac gating were not uniform and varied with the organs under consideration, their location within the chest, and the presence and extent of disease. In other studies investigators have found that cardiac gating significantly increases the resolution of the mediastinal great vessels, the pulmonary hila, and mediastinum (47,62,81). In 1987 Mark et al studied gated and nongated MR scans of the chest in five normal subjects and 20 patients with chest disease (82). In 17 of the 20 patients, the gated images provided improved definition of hilar and mediastinal structures. Because of the improvement in image quality, the investigators recommended the use of ECG-gated sequences for evaluation of the hilar and mediastinal pathology at and below the level of the aortic arch. According to Webb, cardiac gating is not necessary in all MR scans of the chest (83). Respiratory gating has also been used to facilitate the identification of structures in the mediastinum and hilum (52,53). According to Groch et al, respiratory and gating alone was observed to improve image quality in the thorax (58).

In 1987 Enzmann et al described the use of gating of the cerebrospinal fluid to improve MR images of the spine and brain (84,85). In one study, nongated and gated MR images of normal and abnormal spinal cord conditions were compared in 17 patients (84). Gating was accomplished using a peripheral pulse. According to the investigators the gated studies proved superior in signal-to-noise ratio, object contrast, and resolving power. They concluded that peripheral pulse (cardiac) gating effectively eliminates or reduces the signal loss and ghost images caused by the oscillatory pulsation of the cerebrospinal fluid. Similar results were reported by Ezmann et al in the study comparing nongated and gated images of the brain (85). In both normal and abnormal brain tissue, the investigators found gated images superior to nongated images in object contrast and resolving power; the images were equivalent in signal-to-noise ratio. According to the investigators cardiac gating in MRI studies of the brain diminishes the ghost effects caused by pulsating cerebrospinal fluid and flowing blood within the carotid and vertebrobasilar systems. In the presence of cerebrospinal fluid ghost artifacts, Hahn et al recommend cardiac gating and rotation of the phase-encoding gradient by 90 degrees in order to direct the artifact away from the region of interest (86). Quencer et al, while recommending cardiac gating to image the heart, recommend a motion-artifact suppression technique (MAST) for imaging regions distant from the heart, such as the brain (87). They recommend motion-compensating gradients as an alternative to cardiac gating to suppress motion artifacts due to flowing blood and cerebrospinal fluid motion.

While reports of cardiac-gated MRI studies of the heart and other structures have shown continued progress, reports of respiratory-gated MRI studies of the chest and upper abdomen have not. Although the approach to respiratory gating is similar to the approach for cardiac gating, the increased imaging time remains problematic. In 1986 Lewis et al demonstrated the success of respiratory gating in reducing spatial blurring and ghosts of MR images of the chest and abdomen (88). They reported, however, that respiratory gating, which limits data acquisition to end-expiration, increased imaging time two to three times. According to the investigators, the long data acquisition time limits the use of this technique to sequences, with short repetition times of less than 1,000 msec. Because repetition times typically used in abdominal imaging are usually longer than 1,000 msec, the investigators developed another method, respiratory triggering, for the removal of respiratory artifacts in According to Pattany et al, abdominal imaging would benefit substantially from an artifact-suppressed MR technique that was not associated with increased imaging time (89). A review of MR application to the acute abdomen was presented by Shaff et al in 1988 (90). They reported that the development of abdominal MRI has ben impeded by motion artifacts secondary to both respiratory and peristaltic motion. To reduce peristaltic activity they recommended the use of intravenous glucagon. The investigators found that images degrated by motion artifacts were being corrected mainly by fast scanning techniques.

Similar findings were reported by Kressel in his review of the technical approaches employed to improve image quality in the liver (91). Although respiratory gating had been advocated as a methodology for improving upper abdominal image quality, Kressel found that the length of time required to complete a scan precluded its use on a routine basis.

Discussion

The focus of the gating section of this report has been on the value of gating in enhancing the quality of MR images of structures that are not stationary. The published literature indicates, as it did with surface coils, that the development and application of gating techniques have been important aspects of advancing the capabilities of MRI. The value of gating to the technical capacity of MR imaging of the hear and other structures has been well documented (34). MRI subject to motion produces images with artifacts, distortions, and signal loss. The image artifacts consist of ghost replicas of moving anatomic structures that become superimposed on other structures. Movement of tissue protons generating the radiofrequency signal within the targeted area during the course of the imaging procedure distorts the location of the signal and the image that is reconstructed. Gating techniques have been used to eliminate movement artifacts that cause degradation of the radiofrequency signal and associated blurring of the image produced from the signals. To understand the importance of gating techniques it is necessary to understand that MR images are built up from information that is acquired over a relatively long period of time. A lengthy series of pulses of radiofrequency energy is applied to the tissue to be imaged. After each pulse or group of pulses, the tissue emits the radiofrequency signal, which is detected and used to form the image. Repetition rate is the time interval that elapses between successive pulses. Since the image is constructed from several scans varying both in time and location, motion distorts the image reconstruction.

Early cardiac MRI showed that the continuous motion of the heart during the lengthy imaging process caused a severe degradation of the image. A review of recent imaging studies has demonstrated that cardiovascular imaging improves with physiologic gating of pulse sequence and data acquisition to fixed segments of the cardiac cycle. Cardiac imaging studies by Lanzer et al, Herfkens et al, and other investigators have demonstrated improved image quality with gating techniques as compared with those images obtained without gating (47,55,62). Electrocardiographically gated MRI studies of the heart were capable of depicting cardiac anatomy and pathoanatomy with excellent resolution (55,65). The recent cardiovascular MRI review by Reed and Soulen and the report on MRI of the cardiovascular system from the American Medical Association's Council on Scientific Affairs support the findings of other studies of gated cardiac imaging that have demonstrated the technical capacity of MR scanning for evaluating ischemic heart disease, cardiomyopathy, pericardial diseases, intracardiac and paracardiac masses, thoracic-aortic diseases, and congenital heart disease (92,93). In general, respiratory gating is not required for cardiac imaging (94). Moreover, images may not be improved enough to justify the additional scan time (62). However, Peshock believes that the majority of cardiac studies will require gating (95).

Important recent advances in cardiac MRI have resulted from the development of fast (rapid) scanning techniques (94,96). Images of the heart with a new fast-scan MRI technique termed GRASS (gradient recalled acquisition in the steady state) has been introduced. GRASS and other fast scanning techniques enable the acquisition of images with pulse repetition times as short as 20 to 30 msec (97). With fast scanning techniques the number of imaging pulses (time frames) per cardiac cycle is greatly enhanced. When used to image the heart, these techniques permit 20 to as many as 40 images per cardiac cycle (98). According to Higgins and other investigators, the introduction of fast imaging techniques to cardiac MRI has provided the temporal resolution necessary to attain precise evaluations of cardiovascular function in a short imaging time (96,98). In contrast to gated spin-echo imaging, the pulse repetition in fast scanning techniques is independent of the ECG signal.

Although cardiac gating is used primarily in cardiovascular imaging, it is also valuable in reducing motion artifacts and improving image quality in MRI of the chest, brain, and spine. Cardiac gating is considered the most important method for reducing cardiovascular motion and flow-related ghost artifacts in the mediastinum (99). Studies by Wescott et al and Mark et al demonstrated that ECG-gated MR images of the chest are better than nongated images (48,82). Moreover, in these studies cardiac gating did not increase the imaging time. Mark et al compared gated and nongated MRI studies and showed that the definition of the mediastinum and hila was improved, and the signal-to-noise ratio was increased when ECG-gated imaging was performed (82).

In addition cardiac gating techniques have been used successfully in MRI studies of the brain and spine to minimize cerebrospinal fluid flow artifacts (84,85). In studies comparing gated and nongated images of the spine and brain, Enzmann et al demonstrated that peripheral pulse gating effectively eliminates or reduces the signal loss and ghost images caused by the oscillatory pulsation of the cerebrospinal fluid. In these types of studies data acquisition can be synchronized to the cardiac cycle by either ECG triggering of the excitation pulse or triggering to some peripheral pulsation. ECG triggering is recommended by Peshock because of problems with variability of the timing between cardiac contraction and the appearance of the pulse in the periphery (95). According to Hahn et al, correction of cerebrospinal fluid ghost artifacts also can be achieved by 90 degrees rotation of the phase-encoding gradient to direct the artifact away from the region of interest (86). Other MRI studies of the brain showed that with MAST, artifacts on brain images could be reduced or eliminated and image quality improved without cardiac gating (87).

The purpose of gating is to coordinate signal acquisition with the heartbeat (ECG activated) and/or respiratory movement (postexpiratory pause) so that each repetition of the scanning pulse sequence is applied at the identical phase in the cardiac or respiratory cycle. With gating, physiologic motion can be accounted for effectively by limiting useful signal acquisition to a constant period of the physiologic cycle. In the studies by Schultz, Runge, Ehman, and their colleagues, limiting data acquisition to a constant period of the respiratory cycle improved the quality of gated compared with nongated images of the chest and abdomen (52,53,57). However, since the gating process uses only a small part of the cardiac or respiratory cycle and the respiration rate is much lower than the cardiac rate, imaging with the respiratory gating process is significantly lengthened and throughput lowered (44). It has been concluded that even though respiratory gating greatly improves thoracic and abdominal imaging, the inherent time delays limit its usefulness (57,88). The improved image quality is not considered by some investigators to be significant enough to justify the additional scanning time and the inconvenience of respiratory monitoring (100,101). Because motion artifacts degrade images to such a degree, other investigators argue that the additional complexity of a monitor is often warranted (102). Improvements in scan time with respiratory gating have been made by using a relatively wide data acceptance "window" around end-expiration and by reducing the number of sequence repetitions. Nevertheless, clinical adoption of respiratory motion compensation techniques that require monitoring of motion (gating) has been slow, due in part to the cumbersome process of positioning the pneumatic belt, extended examination time, and lack of feedback during scanning to determine whether the system is functioning properly (2).

Motion artifacts constitute significant obstacles for MRI of the chest and abdomen, where respiratory as well as cardiovascular and peristaltic motion occur (102). Some investigators consider respiration, the dominant source of motion artifacts, as a major obstacles to further improvement of the quality of MR images of the chest and upper abdomen (57). Therefore, in addition to gating, other effective methods have emerged to remove respiratory artifacts. Some of the new motion compensation techniques require respiratory monitoring while others do not. Developments in motion compensation techniques include respiratory-ordered phase-encoding, short T1 inversion recovery, signal averaging, pseudogating, gradient moment nulling, fast imaging, and echo planar imaging (2,102,103).

Respiratory ordered phase-encoding methods (ROPE COPE, EXORCIST) employ a respiratory monitor such as a bellows to monitor the rate and phase of respiration. By ordering the phase-encoding steps according to the respiratory cycle, motion artifacts on the images are less discrete (90). Although these techniques still require the inconvenience of respiratory monitoring, they do not prolong imaging time (2). These techniques have been employed in combination with long TR/TE scanning at higher field strength systems to improve image quality. Recently introduced, these methods are considered difficult to implement (102).

Short T1 inversion recovery (STIR) is a special inversion recovery technique that can be used to nullify the high signal intensity of fat (102). Rendering this tissue low in signal eliminates ghost artifacts from the movement of fat. STIR sequences selectively decrease ghosts from fat but do not affect ghosts from other moving tissues (101). The sequences are considered easy to implement and do not require monitoring of the patient's breathing (103). According to Stark et al, although the technique has not been widely available it shows promise for improving abdominal imaging (101).

The signal averaging method combines (averages) more than one acquisition of MR data (measurements). Total acquisition time is increased in proportion to the number of measurements. The method has been effective in suppressing all types of motion artifacts and does not require additional hardware, software, or synchronization with any particular type of motion (2). Stark et al have demonstrated that the combination of short repetition time and short echo delay (TE) with signal averaging on breath-holding is an effective method for reducing motion artifacts, improving signal-to-noise ratio, and maximizing the anatomic resolution of T1-weighted abdominal MR images (101).

Pseudogating is considered a simple motion compensation technique that does not require physiologic gating and can be applied on any standard MRI unit (104). Haacke et al have described the technique and illustrated various successful imaging schemes for reducing motion artifacts. The technique has been used to remove both respiratory and cardiac motion artifacts.

"Gradient moment nulling" is also known as the motion artifact suppression technique (MAST). This technique involves the application of additional magnetic field gradients during conventional spin-echo or gradient-echo pulse sequences to correct phase encoding errors and restore proper anatomic localization of the signal (2). According to Mitchell et al, motion artifacts limit the usefulness of MR images of the upper abdomen obtained with long repetition and long echo times. In the study by Mitchell et al, gradient moment nulling reduced motion artifacts which improved the quality of T2-weighted sequences on long TR/TE images of the upper abdomen without increasing setup or imaging time (105). The investigators found that presaturation pulses (software modification technique) applied outside the volume of interest further improved the quality of these images by virtually eliminating artifact from the aorta and decreasing the signal of its lumen. Pattany et al have demonstrated markedly improved images of the head, abdomen, chest, and spine using MAST (89).

As described elsewhere, a recent major development in motion compensation techniques has been fast (rapid) scanning techniques. These techniques (sequences) provide methods of accumulating images in much shorter total times than conventional sequences (spin-echo pulse sequence) that can be obtained during breath holding (106). Initially, decreased scan times were obtained by simple modifications of pulse sequence software. Reduction of repetition time (TR) was the most efficient way to achieve fast scanning. The recent introduction of gradient-echo pulse sequences has further decreased scan time of fast imaging techniques (FLASH, GRASS); however, this has been achieved at the expense of decreased image contrast and signal-to-noise ratio. Currently, manufacturers are developing fast scan gradient-echo techniques for numerous clinical applications. These modifications require only upgraded software and little if any additional cost to the user (2). Margulis et al and other investigators are in favor of using fast imaging techniques for removing respiratory motion artifacts (2,100).

Another rapid scanning technique samples more than one line during one data collection period. With this echo planar imaging technique a single excitation creates a train of echoes that, if suitably gradient encoded, provides all the views necessary to reconstruct the image. According to Wehrli the technique has seen limited use as a result of instrumental problems (107). More success has been achieved with hybrid imaging, a method related to the echo planar imaging technique that collects several lines of data in a single pass. Reduced scan times have been reported with these methods.

In addition to gating, numerous methods are available for suppressing motion artifacts. The effectiveness of a particular method depends on the anatomic site being imaged as well as the diagnostic question and the specific imaging protocol employed. Under the appropriate circumstances, any method can be effective (102). Moreover, the methods are not mutually exclusive. There are advantages to combining various methods to achieve even greater suppression of motion artifacts. The vast flexibility of MR sequences has led to the development of a wide variety of techniques for all types of imaging. However, the selection of a technique for motion artifact suppression may determine which pulse sequence is most effective for any particular MRI system. Because of the effectiveness of gating and other motion artifact suppression methods, gains in image quality and the performance (technical capacity) of MRI have been demonstrated. However, future studies are needed to compare gated MRI and other imaging techniques in all organs and for all diseases.

Consultations. NIH has informed OHTA that conventional MR imaging of the heart requires cardiac gating to obtain diagnostic quality images. It adds, however, that rapid imaging techniques currently being researched may be able to overcome this by obtaining high signal-to-noise acquisition of the total heart volume within a single cardiac cycle. According to NIH, cardiac gating also provides improved images and clinically useful information in studies of the hilar areas, mediastinum, spine, brain, and liver. NIH finds fully successful respiratory gating not yet accomplished because it lengthens the imaging process and makes patient cooperation difficult. Some investigators are using fast MR scanning techniques as well as some methods of artifact suppression to mitigate the problems associated with respiratory motion. According to NIH, investigators using respiratory gating report improvement in images of those organs that move with the diaphragm such as the kidneys, liver, and spleen. Because there are no major physiologic movements in the pelvic region that interfere with MRI, gating is not considered necessary for good visualization of the pelvic organs.

NIH has found MRI with gated techniques to be better than conventional MRI, because gated MRI improves the quality of the scans and the diagnostic capabilities of the method and may result in better patient treatment. NIH believes that MRI has become the imaging technique of choice for diseases of the nervous system, especially the demyelinating diseases, the cerebrovascular diseases, the spinal cord diseases, and basal ganglia pathology. According to NIH, the use of gating techniques does not represent any hazard for the patient.

As of September 1989 the FDA had approved the MRI device of 18 companies for marketing in general interstate commerce. Eight of the companies offer cardiac and respiratory gating. Six of these companies offer only cardiac gating.

Medical Specialty Responses. According to information provided by the American Heart Association, cardiac-gating in conjunction with MRI has become well established as an important method for improving the quality of cardiac MR images. It is considered necessary for MR studies of the heart. The information indicated that although cardiac gating does not completely eliminate motion artifacts due to cardiac contraction, it results in a marked improvement in cardiac image quality. The information from the Association indicates that cardiac gated images of the heart are effective for a variety of indications including pericardial disease, cardiac masses, and some types of congenital heart disease. Moreover, cardiac gating substantially improves MRI studies of any region of the chest including the hila, mediastinum, thoracic aorta, and pulmonary artery by reducing noise and ghost artifacts. Because motion artifacts can occur due to the pulsatile flow of blood or cerebrospinal fluid, cardiac gating may have advantages in imaging other structures such as the spine and brain. Although there are no contraindications to the use of cardiac gating and no complications have been reported, hazards such as shock and skin burns could be present.

The information provided by the American Heart Association to OHTA states that the use of respiratory gating in conjunction with MRI is far more limited because the repetition rate of the scanner must be synchronized with the respiratory cycle, which is much slower than the repetition rates normally used in The significant increase in scan time resulting from respiratory gating limits its clinical use. Other approaches have therefore been employed to reduce respiratory motion artifacts. One such nongated technique is respiratory ordered phasing encoding (ROPE), which increases scan time 10 to 20 percent while providing a variable improvement in image quality.

Comments received from the National Electrical Manufacturers Association (NEMA) stress that cardiac-gated MR images are superior to nongated images in those areas of the body where motion synchronous to the beating heart can create image artifacts and degrade image quality. According to the Association, cardiac gating is particularly useful for images of the heart, chest, brain, and spine.

NEMA states that respiratory-gated MR images are superior to nongated images in areas of the body where motion synchronous to respiration can create image artifacts and degrade image quality. It suggests that respiratory gating is useful in the chest and abdomen. NEMA believes that improvements in the clarity of an image by reducing motion with gating leads to improved diagnostic confidence for the clinician.

The American College of Cardiology has stated that cardiac gating is an effective anatomical imaging technique that provides clinically useful information about the heart, great vessels, pericardium, and intracardiac masses. Although it lengthens acquisition time, the College considers the use of cardiac gating essential for providing useful imaging in the diagnosis of aortic dissection and other abnormalities and diseases of the aorta, pulmonary arteries, pericardium, and heart. The College has pointed out that MRI permits data acquisition in any plane, making it possible to orient tomographic sections to the intrinsic axes of the heart. According to the College, respiratory gating has yet to be proven effective in cardiovascular diagnosis and adds additional time to data acquisition.

According to the American College of Radiology, gating is essential to obtain tomographic MR images of the heart in which the myocardium, valves, and chambers are distinguishable from each other. In the absence of gating, an image of the heart would be characterized by low spatial resolution and complete invisibility of any of the smaller structures. The College pointed out that the motion artifacts from transmitted pulsations around the great vessels and hila also are reduced by cardiac gating. They also recommend cardiac gating when cardiac pulsations create a series of motion artifacts through movement of cerebrospinal fluid that degrades central nervous system images.

The American Society of Colon and Rectal Surgeons has pointed out that gating techniques are not particularly useful in imaging the abdomen and pelvis, and generally they are not used clinically for this purpose. The Society believes that MRI will be useful as a secondary imaging test in many situations.

The American Association of Neurological Surgeons has indicated that cardiac gating has provided clinically useful information for MRI of the heart, abdominal organs close to the diaphragm, and some specific types of central nervous system structures where elimination of artifacts from cerebrospinal fluid pulsation is necessary for adequate image resolution. In the Association's view respiratory gating has been important in the improvement of images from abdominal structures such as the liver, which moves with the diaphragm.

Summary

Gating is a process developed to obtain high-resolution magnetic resonance images despite the presence of motion. In MRI, motion has the detrimental effect of causing artifacts and distortions to appear in images, as well as producing signal loss. When the acquisition of data in MRI is prolonged, physiologic motion associated with respiration, cardiac motion, blood flow, and peristalsis results in both blurring and motion artifacts. Motion artifacts appear as ghosts in the image. Gating techniques that synchronize the start of the MR pulse sequence to a constant time point of the cardiac or respiratory cycle have been used to eliminate movement artifacts that cause degradation of the radiofrequency signal and associated blurring of the image. Gating has become an important technique for reducing motion artifacts and improving temporal resolution in MRI.

Early cardiac MRI showed that the continuous motion of the heart during the lengthy imaging process caused a severe degradation of the image. An analysis of recent imaging studies has demonstrated that cardiovascular imaging improves with physiologic gating of pulse sequence and data acquisition to fixed segments of the cardiac cycle. Developments in cardiac-gated imaging have enabled the acquisition of high-resolution diagnostic quality cardiac images with conventional MRI. Studies of gated cardiac imaging have demonstrated the technical capacity of MR scanning for evaluating ischemic heart disease, cardiomyopathy, pericardial diseases, intracardiac and paracardiac masses, thoracic-aortic diseases, and congenital heart disease. Researchers are attempting to use fast scanning MRI techniques to produce a diagnostic-quality cardiac image without gating. Although cardiac (ECG) gating is used primarily in cardiac imaging it also has been found valuable in reducing motion artifacts in MRI of the chest, brain, spine, and liver.

MRI of the chest and abdomen is hampered by the normal periodic motion of the diaphragm associated with breathing. For reducing motion artifacts in thoracic and abdominal imaging, respiratory gating was introduced as a modification to conventional MRI. Using respiratory gating, the MR pulse sequence is synchronized in the respiratory cycle to decrease movement-induced artifacts. Various methods are available to monitor the respiratory cycle, and detecting chest wall motion with a pneumatic bellows has become a common method to monitor the respiratory cycle and generate signals for data acquisition. Gated imaging studies demonstrated that MR respiratory gating greatly improves thoracic and abdominal imaging, particularly abdominal imaging.

Since the gating process uses only a small part of the respiratory cycle, and the respiratory cycle is much slower than the repetition rates normally used in scanning, imaging time with the gating process is significantly lengthened. Because respiratory motion artifacts are a major obstacle to further improvement of the quality of MR thoracic and abdominal images, other effective methods have emerged to remove respiratory artifacts. Developments in motion compensation techniques include respiratory ordered phase-encoding, short T1 inversion recovery, signal averaging, pseudogating, gradient moment nulling, fast imaging, and echo planar imaging. The effectiveness of a particular method will depend on the anatomic site being imaged, as well as the diagnostic question, the MR system, and the specific imaging protocol employed. Under the appropriate circumstances any method can be effective. Motion compensation techniques are not mutually exclusive, and there are advantages to combining various methods to achieve even greater suppression of motion artifacts. Because respiratory gating greatly improves thoracic and abdominal imaging, improvements in scan time with respiratory gating have been attempted by using a relatively wide data acceptance "window" around end-expiration and by reducing the number of sequence repetitions. With the effectiveness of gating and other motion compensation techniques, gains in image quality and the performance (technical capacity) of MRI have been demonstrated.

Use of gating procedures in conjunction with MRI should not represent any hazard for the patient. No patient-related risks appear to be involved when gating procedures are used appropriately. The use of gating techniques may add substantial time to the imaging procedure and subsequently reduce the number of patients imaged per day, depending on the procedure, equipment, software, and indications. In addition to capital and operating costs, the estimated cost per MRI scan is strongly dependent on a given facility's annual patient throughput.

NIH has informed OHTA that MRI with gated techniques has been found in studies to be better than conventional MRI because MRI with gated techniques improves the quality of the scans and the diagnostic capabilities of the method and may result in better patient treatment. NIH finds respiratory gating not yet practical because of the lengthened imaging time. According to NIH, the use of gating techniques does not represent any hazard for the patient.

As of September 1989 the FDA had approved the MRI devices of 18 companies for marketing. Eight of the companies offer cardiac and respiratory gating. Six offer only cardiac gating.

References

1.
[no authors listed] Magnetic Resonance Imaging Health Technology Assessment Report, 1985, No. 13. Rockville, MD: National Center for Health Services Research and Health Care Technology Assessment.
2.
[no authors listed] Future improvements in magnetic resonance imaging Health Technol 1988. 2(1):12–19. [PubMed: 10286000]
3.
Jacobson HG (Ed). Magnetic resonance imaging: Prologue. (AMA Council on Scientific Affairs; Report of the Panel on Nuclear Magnetic Resonance Imaging) JAMA 1987. 258(22)3283–3285. [PubMed: 3316721]
4.
Jacobson HG (Ed). Fundamentals of magnetic resonance imaging. (AMA Council on Scientific Affairs; Report of the Panel on Nuclear Magnetic Resonance Imaging) JAMA 1987. 258(23):3417–3423. [PubMed: 3316726]
5.
Pearce RB. New surface coil designs enhance MRI performance Diagn Imag 1988. January:117–129.
6.
American Hospital Association. Guideline report: NMR, Issues for 1985 and beyond, executive summary Hospital Technology Series 4 (3 and 4). , Chicago IL: American Hospital Association, 1985.
7.
Newhouse J. Written communication June 25, 1987.
8.
Pavlicek W. Safety considerations in Stark DD, Bradley WG(Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1988. pp. 224–251.
9.
Stark DD, Bradley WG, Ferrucci JT. Advances in body imaging enhance utility of MRI Diagn Imag 1987; . December:74–82.
10.
Ackerman JJH, Grove TH, Wong GG, et al. Mapping of metabolites in whole animals by 31 PNMR using surface coils Nature 1980. 283: 167–170. [PubMed: 7350541]
11.
Bernardo ML, Cohen AJ, Lauterbur PC. Radiofrequency coil designs for nuclear magnetic resonance zeugmatographic imaging. Institute of Electrical and Electronic Engineers, . 1982
12.
Axel L. Surface coil magnetic resonance imaging J Comput Assist Tomogr 1984. 8(3):381–384. [PubMed: 6725683]
13.
Axel L, Hayes C. Surface coil magnetic resonance imaging. in Proceedings of the 2nd Symposium on Progress of NMR in Medicine. 1985. pp. pp 11–17.
14.
Hayes CE, Axel L. Noise performance of surface coils for magnetic resonance imaging at 1.5T Medical Phys 1985. 10(50):604–607. [PubMed: 4046995]
15.
Kulkarni MV, Patton JA, Price RR. Technical considerations for the use of surface coils in MRI AJR 1986. 147:373–378. [PubMed: 3487956]
16.
Schenck JF, Foster TH, Henkes JL, et al. High-field surface-coil MR imaging of localized anatomy AJNR 1985. 6:181–186. [PubMed: 3920878]
17.
Fisher MR, Barber Barker B, Amparo EG, et al. MR imaging using specialized coils Radiology 1985. 57: 443–447. [PubMed: 4048453]
18.
Roschmann P, Tischler R. Surface coil protein proton MR imaging at 2T Radiology 1986. 161:251–255. [PubMed: 3763875]
19.
Hornak JP, Szumowski J, Rubens D, et al. Breast MR imaging with loopgap resonators Radiology 1986. 161:832–834. [PubMed: 3786741]
20.
Fitzsimmons JR, Thomas RG, Mancuso AA. Communication: Proton imaging with surface coils on a 0.15T resistive system Magn Reson Med 1985; . 2:180–185. [PubMed: 3831686]
21.
Lufkin RB, Vortruba Votruba J, Reicher M, et al. Solenoid surface coils in magnetic resonance imaging AJR 1986. 146:409–412. [PubMed: 3484592]
22.
Hoover LA, Wortham DG, Lufkin RB, et al. Magnetic resonance imaging of the larynx and tongue base: Clinical applications Otalaryngol Head Neck Surg 1987. 97:245–256. [PubMed: 3118304]
23.
Kier R, Blinder RA, Herfkens RJ, et al. MR imaging with surface coils in primary hyperparathyroidism J Comput Assist Tomogr 1987. 11(5):863–868. [PubMed: 3655050]
24.
McArdle CB, Bradley Bailey BJ, Amparo EG. Surface coil magnetic resonance imaging of the normal larynx Arch Otolaryngol Head Neck Surg 1986. 112:616–622. [PubMed: 3964448]
25.
Edelman RR, Shoukimas GM, Stark DD, et al. High-resolution surface-coil imaging of lumbar disk disease AJR 1985. 144:1123–1129. [PubMed: 3873790]
26.
Holtas SL, Plewes DB, Simon JH, et al. Technical aspects on magnetic resonance imaging of the spine at 1.5 tesla Acta Radiol 28 1987. 4:375–381. [PubMed: 2958049]
27.
Edelman RR, McFarland E, Stark DD, et al. Surface coil MR imaging of abdominal viscera. Part 1: Theory, technique, and initial results Radiology 1985. 157:425–430. [PubMed: 4048450]
28.
Simeone JF, Edelman RR, Stark DD, et al. Surface coil MR imaging of abdominal viscera. Part III: The pancreas Radiology 1985. 157:437–441. [PubMed: 4048452]
29.
White ME, Edelman RR, Stark DD, et al. Surface coil MR imaging of abdominal viscera. Part II: The adrenal glands Radiology 1985. 157:431–436. [PubMed: 4048451]
30.
Bartentsz JAM, Barentsz JO, Lemmens JOA, Boskamp EB, et al. Improved MR imaging of the bladder by using a new surface coil Fortschr Rontigenstr 1986. 145(3);351–353. [PubMed: 3020646]
31.
McCarthy S, Fritzsche PJ. Male pelvis in Stark DD, Bradley WG (Eds). Magnetic Resonance Imaging, St. Louis: CV Mosby Co, 1988. pp. 1233–1235.
32.
Akins EW, Fitzsimmons JR, Mancuso AA, et al. Technical note: Double loop receiver coil for MR imaging at 0.15T J Comput Assist Tomogr 1986. 10(6):1083–1088. [PubMed: 3782558]
33.
Doornbos J, Brimbergen HAA, Grimbergen HA, Booijen PE, et al. Application of anatomically shaped surface coils in MRI at 0.15T Mag Reson Med 1986. 3:270–281. [PubMed: 3713491]
34.
Stark DD, Bradley, WG (Eds). Magnetic Resonance Imaging, St. Louis: CV Mosby Co., 1988.
35.
Kneeland JB. Instrumentation in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby Co., 1988. pp 56–65.
36.
Haughton VM. Magnetic resonance: State of the art Radiology 1988. 166:297–301. [PubMed: 3275973]
37.
Boskamp EB. Improved surface coil imaging in MR: Decoupling of the excitation and receiver coils Radiology 1985. 157:449–452. [PubMed: 4048454]
38.
El Yousef, SJ. The breast in Higgins CB, Hricak H (Eds): Magnetic Resonance Imaging of the Body, New York: Raven Press, 1987. –.
39.
Murphy WA, Gohagan JK. Breaks in: Stark DD and Bradley WG (Eds) Magnetic Resonance Imaging, St. Louis: CV Mosby Co., 1988. pp 861–863.
40.
Stark DD. Biliary system, pancreas, spleen, and alimentary tract in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby Co., 1988. pp 1060–1087.
41.
Sell JJ, Lamott Ga, Ablin DS. Magnetic resonance imaging of the pelvis in Mettler FA, Muroff LR, Kulkarni MV (Eds):, Magnetic Resonance Imaging and the Spectroscope, New York: Churchill Livingstone, 1980. pp 187–205.
42.
Rholl KS, Lee JKT, Ling D, et al. MR Imaging of the scrotum with a high-resolution surface coil Radiology 1987. 163:99–103. [PubMed: 3547497]
43.
Ehman RL. MR imaging with surface coils Radiology 1985. 157:549–550. [PubMed: 4048470]
44.
Mettler FA, Muroff, LR, Kulkarni MV. Basic principles of magnetic resonance imaging in Mettler FA, Muroff LR, Kulkarni MV (Eds): Magnetic Resonance Imaging and Spectroscopy, New York: Churchill Livingstone, 1986. pp 1–14.
45.
Hammond M. Written communication, July 14, 1987.
46.
Eckel CG, Mettler FA, Wicks JD, et al. Magnetic resonance imaging in cardiovascular disease in Mettler FA, Muroff LR, Kulkarni MV (Eds): Magnetic Resonance Imaging and Spectroscopy, New York: Churchill Livingstone, 1986. pp 1–14.
47.
Lanzer P, Botvinick EH, Schiller NB, et al. Cardiac imaging using gated magnetic resonance Radiology 1984. 150:121–127. [PubMed: 6227934]
48.
Westcott JL, Hensche Henschke CI, Berkmen Y. MR imaging of the hilum and mediastinum: Effects of cardiac gating J Comput Assist Tomogr 1985. 9(6):1073–78. [PubMed: 4056140]
49.
Runge V. Respiratory gating in magnetic resonance imaging in Mettler FA, Muroff, LR, Kulkarni MV (Eds): Magnetic Resonance Imaging and Spectroscopy, New York: Churchill Livingston, 1986.
50.
Wood ML, Henkelman RM. Suppression of respiratory motion artifacts in magnetic resonance imaging Med Phys 1986. 13(6):794–805. [PubMed: 3796475]
51.
Aisen AM, Glazer GM, Carson PL, et al. Motion artifacts in quantitative magnetic resonance imaging Mag Reson Imaging 1986; . 4:207–213. [PubMed: 3669932]
52.
Schultz CL, Alfidi RJ, Nelson AD, et al. The effect of motion on two-dimensional Fourier transformation magnetic resonance images Radiology 1984. 152:117–121. [PubMed: 6729101]
53.
Runge VM, Clanton JA, Partain CL, et al. Respiratory gating in magnetic resonance imaging at 0.5 tesla Radiology 1984. 151:521–523. [PubMed: 6709928]
54.
Higgins CB, Lanzer P, Stark D, et al. Assessment of cardiac anatomy using nuclear magnetic resonance imaging J Am Coll Cardiol 1985. 5:778–815. [PubMed: 3965536]
55.
Herfkens RJ, Higgins CB, Hricak H, et al. Nuclear magnetic resonance imaging of the cardiovascular system: Normal and pathologic findings Radiology 1983. 147:749–759. [PubMed: 6601813]
56.
Lanzer P, Barta C, Botvinick EH, et al. ECG-synchronized cardiac MR imaging: Method and evaluation Radiology 1985. 155:681–686. [PubMed: 4001369]
57.
Ehman RL, McNamara MT, Pallack M, et al. Magnetic resonance imaging with respiratory gating: Techniques and advantages AJR 1984. 143: 1175–1182. [PubMed: 6333787]
58.
Groch MW, Turner, DA, Kaplan HC. Written communication April 17, 1987.
59.
Higgins CB, Stark D, McNamara M, et al. Multiplane magnetic resonance imaging of the heart and major vessels: Studies in normal volunteers AJR 1984. 142:661–667. [PubMed: 6608219]
60.
Higgins CB, Lanzer P, Stark D, et al. Imaging by nuclear magnetic resonance in patients with chronic ischemic heart disease Circulation 1984. 69:523–531. [PubMed: 6692514]
61.
Higgins CB, Byrd BF III, Farmer DW, et al. Magnetic resonance imaging in patients with congenital heart disease Circulation 1984. 70(5):851–860. [PubMed: 6488498]
62.
Lieberman JM, Alfidi RJ, Nelson AD, et al. Gated magnetic resonance imaging of the normal and diseased heart Radiology 1984. 152:465–470. [PubMed: 6739817]
63.
Stark DD, Higgins CB, Lanzer P, et al. Magnetic resonance imaging of the pericardium: Normal and pathologic findings Radiology 1984. 150:469–474. [PubMed: 6691103]
64.
Fletcher BD. Jacobstein MD, Nelson AD, et al. Gated magnetic resonance imaging of congenital cardiac mal-formations malformations Radiology 1984. 150:137–140. [PubMed: 6689753]
65.
Higgins CB, Byrd BF, McNamara MT, et al. Magnetic resonance imaging of the heart: A review of the experience in 172 subjects Radiology 1985. 155:671–679. [PubMed: 3159039]
66.
Didier D, Higgins CB, Fisher MR, et al. Congenital heart disease: Gated MR imaging in 72 patients Radiology 1986. 158:227–235. [PubMed: 3940387]
67.
Diethelm L, Dery R, Lipton MJ, et al. Atrial-level shunts: Sensitivity and specificity of MR in diagnosis Radiology 1987. 162:181–186. [PubMed: 3786759]
68.
Amparo EG, Higgins CB, Farmer D, et al. Gated MRI of cardiac and paracardiac masses: Initial experience AJR 1984; . 143:1151–1156. [PubMed: 6333784]
69.
Go RT, O'Donnell JK, Underwood DA, et al. Comparison of gated cardiac MRI and 2D echocardiography of intracardiac neoplasms AJR 1985; . 145:–. [PubMed: 3873848]
70.
McNamara MT, Higgins CB, Schechtmann N, et al. Detection and characterization of acute myocardial infarction in man with use of gated magnetic resonance Circulation 1985. 71(4):717–724. [PubMed: 3971541]
71.
Fisher MR, McNamara MT, Higgins CB. Acute myocardial infarction: MR evaluation in 29 patients AJR 1987. 148: 247–251. [PubMed: 3492097]
72.
Smith MA. Cardiac gating in nuclear magnetic resonance imaging J Med Eng Tech 1986; . 10(6):325–328. [PubMed: 3820264]
73.
Matsuda T, Shimizu K, Sakurai T, et al. Measurement of aortic blood flow with MR imaging: Comparative study with Doppler US Radiology 1987; . 162:857–861. [PubMed: 3544040]
74.
Stratemeier EJ, Thompson R, Brady TJ, et al. Ejection fraction determination by MR imaging: Comparison with left ventricular angiography Radiology 1986. 158:775–777. [PubMed: 3945753]
75.
Van Rossum AC, Visser FC, Sprenger M, et al. Evaluation of magnetic resonance imaging for determination of left ventricular ejection fraction and comparison with angiography. Am J Cardiol . 1988:62:628–633. [PubMed: 3414556]
76.
White RD, Cassidy MM, Cheitlin MD, et al. Segmental evaluation of left ventricular wall motion after myocardial infarction: Magnetic resonance imaging versus echocardiography Am Heart J 1988. 155:166–175. [PubMed: 3336970]
77.
Pernes JM, Greinier Grenier P, Desbleds MT, et al. MR evaluation of chronic aortic dissection J Comput Assist Tomogr 1987. 11(6):975–981. [PubMed: 3680710]
78.
Kersting-Sommerhoff BA, Higgins CB, White RD, et al. Aortic dissection: Sensitivity and specificity of MR imaging Radiology 1988. 166:651–655. [PubMed: 3340758]
79.
Pohost GM, Canby RC. Nuclear magnetic resonance imaging: Current applications and future prospects Circulation 1987; . 75:88–95. [PubMed: 3024861]
80.
Higgins CB. Malcolm Hanson Memorial Lecture: MR of the heart: Anatomy, physiology, and metabolism AJR 1988. 151:239–248. [PubMed: 3293369]
81.
Mazer MJ, Carroll FE, Falke THM. Practical aspects of gated magnetic resonance imaging of the pulmonary artery J Thorac Imaging 1988. 3(3):73–84. [PubMed: 3292785]
82.
Mark AS, Winkler ML, Peltzer M, et al. Gated acquisition of MR images of the thorax: Advantages for the study of the hila and mediastinum Magn Reson Imaging 1987. 5:57–63. [PubMed: 3586873]
83.
Webb WR. Magnetic resonance imaging of the hila and mediastinum Cardiovasc Intervent Radiol 1986; . 8:306–313. [PubMed: 3084095]
84.
Enzmann DR, Rubin JB, Wright A. Use of cerebrospinal fluid gating to improve T2-weighted images. Part I. The spinal cord Radiology 1987; . 162:763–767. [PubMed: 3809491]
85.
Enzmann DR, Rubin JB, O'Donohue J, et al. Use of cerebrospinal fluid gating to improve T2-weighted images. Part II. Temporal lobes, basal ganglia, and brain stem Radiology 1987. 162:768–773. [PubMed: 3809492]
86.
Hahn FJ, Chu, WK, Coleman PE, et al. Artifacts and diagnostic pitfalls of magnetic resonance imaging: A clinical review Radiol Clin North Am 1988. 26(4):717–735. [PubMed: 3380943]
87.
Quencer RM, Hinks RS, Pattany PH, et al. Improved MR imaging of the brain by using compensating gradients to suppress motion-induced artifacts AJR 1988. 151:163–170. [PubMed: 3259801]
88.
Lewis CE, Prato FS, Drost DJ, et al. Comparison of respiratory triggering and gating techniques for the removal of respiratory artifacts in MR imaging Radiology 1986. 160:803–810. [PubMed: 3737921]
89.
Pattany PM, Phillips JJ, Chiu LC, et al. Motion artifact suppression technique (MAST) for MR imaging J Comput Assist Tomogr 1987. 11:369–337. [PubMed: 3571576]
90.
Shaff MI, Tarr RW, Partain CL, et al. Computed tomography and magnetic resonance imaging of the acute abdomen Surg Clin North Am 1988. 68(2):233–254. [PubMed: 3279544]
91.
Kressel HY. Strategies for magnetic resonance imaging of focal liver disease Radiol Clin North Am 1988. 26(3):607–615. [PubMed: 2836885]
92.
Reed JD, Soulen RL. Cardiovascular MRI: Current role in patient management Radiol Clin North Am 1988; . 26(3):589–606. [PubMed: 3287436]
93.
Jacobson HG (Ed). Magnetic resonance imaging of the cardiovascular system: Present state of the art and future potential. (AMA Council on Scientific Affairs; Report of the Panel on Nuclear Magnetic Resonance Imaging) JAMA 1988. 259(2):253–259. [PubMed: 3275813]
94.
Higgins CB. The heart: Acquired disease in Higgins CB, Hricak H (Eds): Magnetic Resonance Imaging of the Body, New York: Raven Press, 1987. pp 239–242.
95.
Peshock RM. Heart and great vessels in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1980. pp 887–920.
96.
Utz JA, Herfkens RU. Dynamic and physiologic cardiac MR in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1988. pp 921–929.
97.
Pflugfelder PW, Sechtem UP, White RD, et al. Rapid (cine) magnetic resonance imaging of the heart New Concepts Imaging 1988. 4:281–300.
98.
Higgins CB, Holt W, Plugfelder P, et al. Functional evaluation of the heart with magnetic resonance imaging Mag Reson Med 1988. 6: 121–139. [PubMed: 3367771]
99.
Hahn D. Mediastinum and lung in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1980. pp. 804–847.
100.
Margulis AR, Hricak H, Crooks L. Medical applications of nuclear magnetic resonance imaging Q Rev of Biophys 1987; . 19(3/4):221–237. [PubMed: 3310080]
101.
Stark DD, Hendrick RE, Hahn PF, et al. Motion artifact reduction with fast spin-echo imaging Radiology 1987. 164:183–191. [PubMed: 3588902]
102.
Wood ML. Thoracic and abdominal motion artifacts in Stark DD and Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1988. pp 792–803.
103.
Haacke EM, Bellon EM. Artifacts in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1988. pp 138–153.
104.
Haacke EM, Lenz GW, Nelson DA. Pseudogating Pseudo-gating: Elimination of periodic motion artifacts in magnetic resonance imaging without gating Mag Reson Med 1987. 4:162–174. [PubMed: 3561245]
105.
Mitchell DG, Vinitski S, Burk DL, et al. Motion artifact reduction in MR imaging of the abdomen: Gradient moment nulling versus respiratory sorted phase encoding Radiology 1988. 169:155–160. [PubMed: 3420252]
106.
Hendrick RE. Image contrast and noise in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1988. pp 66–83.
107.
Wehrli FW. Principle of magnetic resonance in Stark DD, Bradley WG (Eds): Magnetic Resonance Imaging, St. Louis: CV Mosby, 1988. pp 3–23.

DHHS Publication No. (PHS) 90-3458

PubReader format: click here to try

Views

  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...