Magn Reson Med Sci, Vol. 6, No. 1, pp. 29–42, 2007

REVIEW

New Horizons in MR Technology: RF Coil Designs and Trends

Hiroyuki FUJITA1,2,3*

Departments of 1Physics, Case Western Reserve University, and 2Radiology, University Hospitals of Cleveland, 10900 Euclid Avenue, Cleveland, Ohio 44106-7079, USA 3School of Information Technology and Electrical Engineering, The University of Queensland, Brisbane Qld 4072, Australia (Received January 9, 2007; Accepted February 14, 2007)

Parallel imaging techniques have developed very rapidly, and realization of their full potential has required the design of magnetic resonance (MR) scanners with ever-increasing numbers of receiver channels (32 to 128). In particular, 1.5- and 3-Tesla fast MR imaging applications are now used in everyday clinical practice. Both strengths require maximum achievable signal-to-noise ratio (SNR) and multi-detector array coil optimization within the framework of the parallel imaging scheme for more advanced and faster clinical MR scanning. Preampliˆers are key components in the detector array coils and serve many functions beyond mere signal ampliˆcation. One critical function is to aid in the decoupling of individual coils, which is essential for optimal SNR and the performance of parallel imaging. To support a large number of detector array coils for parallel imaging, preampliˆers must be physically very small so that they may be tightly packed together to form an optimized detector array. The author herein reviews the state-of-the-art work reported by those skilled in the art to consider the rationale for determining how many channels are enough and how fast we can go. The paper explores the important and fundamental principles of RF array coils for MR imaging and reviews cutting-edge array coils, including those for transmit-SENSE or parallel transmission applications. The future of radiofrequency (RF) coil technology is also considered.

Keywords: phased array RF coils, preampliˆer decoupling, parallel imaging, parallel transmission, wireless RF coils

today's advanced and sophisticated clinical appli- Introduction cations. Modern radiofrequency (RF) array coil design is In 1978, in his historic paper, Hoult detailed and complex, and its understanding requires broad discussed the nuclear magnetic resonance (NMR) knowledge of magnetic resonance (MR) physics receiver;1 in 1990, Roemer and associates proposed and electronics. Today's coils are very intricate and the phased-array coil technology that underlies the delicate. For example, as many as 400 to 1500 design of most of today's RF receive coils;2 and in individual componentsWparts are used in the con- 1991, Hayes's group discussed the application of struction of 8-channel head coils and 16-channel that technology to volume imaging.3 Original and brain-spine combo coils, both of which are typical general phased-array radar technology4,5 and array detector array coils. Every component must be coil theory and application6 are reviewed in detail correctly integrated; if even one component is elsewhere. broken or damaged, the coil does not function A small surface coil is well known to yield higher optimally. The author attempts to explain the SNR nearer the surfaces of the coil than does a critical components that constitute an RF array volume coil, such as the birdcage coil,7 but the coil, especially emphasizing principles and methods B1-sensitive region of a small surface coil is (much) for decoupling array coils, which is essential to smaller than that of a volume coil. It is remarkable that in an approach utilizing phased-array coils, the *Corresponding author, Phone: ＋1-216-368-3894, Fax:＋1- high signal-to-noise ratio (SNR) associated with a 216-368-4671, E-mail: hiroyuki.fujita＠case.edu small surface coil can be achieved and maintained

29 30 H. Fujita

Fig. 1. The mechanism of electromagnetic coupling between the patient and the magnetic resonance (MR) radiofrequency (RF) coils, due to the stray capacitances present, is illustrated. Baluns are added to block the common mode current and allow the diŠerential mode current to ‰ow. over an extended ˆeld of view (FOV) by using a set author uses state-of-the-art examples to address the of RF coils. This is possible because of the develop- questions of what constitutes a su‹cient number of ment of various array coil designs for imaging channels and how coil elements in an array are applications of interest and of techniques to decou- decoupled and to consider how array detectors can ple the mutual inductances and thereby elimi- be optimized. nate cross-talk among the elements of the array coils. Adjacent coil elements are often partially Signal-to-Noise Ratio overlapped to cancel the mutual inductance be- tween the elements. The next nearest and other Signal-to-noise ratio (SNR), one of the most coil elements can be mutually decoupled using a important parameters to be optimized in MR low-input impedance (typically less than 5 Ohms) imaging applications, was detailed by Hoult and preampliˆer decoupling circuit. Other decoupling Richards.14 Here, it is essential to understand how techniques and methods are brie‰y discussed later; SNR relates to the coil-related parameters. The detailed discussions of low noise ampliˆers may be signal is written as found elsewhere.8 S1g3B 2 B xy(r)Eq.(1), In conjunction with a coil matchingWdecoupling 0 1 circuit, the low-input impedance preampliˆer where g is the nuclear gyromagnetic ratio, B0 is the xy eŠectively eliminates current ‰ow in the coil ele- static magnetic ˆeld, and B1 (r)istheRFmagnetic ment loop, thereby eliminating the magnetic ˆelds ˆeld produced by a coil with unit current 1A. In xy induced in neighboring coil elements (shown later). designing a coil, B1 (r) must be optimized, which Baluns and considerations pertaining to cable requires the appropriate choices of coil size over a routing are also important factors in array coil target FOV and of distance from the FOV. Size design. Figure 1 illustrates how cables couple with choice depends generally on the number of availa- the patient and why baluns need to be implement- ble receiver channels. ed; details are discussed elsewhere.9 On the other hand, NMRWMR imaging noise is RF coil technologies are driven by the rapid thermal noise, and the noise generated from the advances in development of MR scanners with coil is given by greater numbers of receiver channels as well as in N＝ 4kTDfR Eq. (2), development of parallel imaging techniques;10–12 the 96-channel head array coil is one such example.13 where k is the Boltzmann constant, T is the temper-

Whatever the application at any ˆeld strength, ature in K, Df is the bandwidth, and R＝RC＋RS. maximum achievable SNR is key. Because SNR is RC isthecoilresistanceandRS thesampleenergy destined to be degraded or lowered from the loss, i.e., the equivalent series resistance resulting starting point, i.e., the signal source, to the end from the induced eddy current losses in the conduc- point of the receiver chain, minimizing its loss in tive sample. Combining Eqs. (1) and (2), SNR can the process is of primary concern. In this paper, the be expressed using the coil-related parameters as components of an array coil are explained, and issues requiring special attention are noted. The

Magnetic Resonance in Medical Sciences RF Coil Designs and Trends 31

Table 1. Relationship between dB expression and common factor

Factor 1W100 1W10 1W51W31W223456 7 10100 dB －20 －10 －7 －4.8 －3 3 4.8 6 7 7.8 8.45 10 20

S B xy(r) 1 1 Eq. (3). N RC＋RS

xy To optimize SNR, B1 (r) can be maximized by having the coil closer to the sample, and RS can be minimized by matching coil size to the target FOV. In fact, Wright and Wald compared an N-element Fig. 2. Depiction of a preampliˆer noise model. V array to a single coil with respect to the resultant denotes the signal; NS, the noise generated by the resistance, r, of the input signal source; and N ,the SNR while overall coil dimension remained un- P noise generated inside the preampliˆer. changed.6 They showed, for example, that when comparing a single coil with arrays measuring 8×8, 4×4, and 2×2, improvement in SNR disappears where.1,16 Figure 2 depicts a simple model for at a depth equal to the diameter of the array. preampliˆer noise. V denotes the signal; NS,the However, for distances in between, SNR curves noise generated by the resistance, r, of the input vary according to the number of array (i.e., coil) signal source; and NP, the noise generated inside elements. If each coil is made too small, the coil the preampliˆer. Using these quantities, NF is resistance loss dominates the sample noise. This deˆned as implies that optimization is required for a given N 2 ＋N 2 speciˆc application, e.g., target depth and available NF＝10 log P S where N ＝ 4kTDfr 10 Ø N 2 » S number of receiver channels. R is minimized by S C Eq. (4). making the unloaded coil Q high. It is well known that at low-B0 ˆeld systems (Ã0.3T), coil resistance The industry standard for preampliˆer NF is less is dominant or at least comparable to sample noise, than 0.5 dB, and the ˆrst NF and gain are well but at 1.5T and above, sample noise dominates known to have the most signiˆcant impact in the

(i.e., RS:RC ). Furthermore, it is useful to be aware entire electronics circuit, which is often cascaded. that if each of 2 coils yields the same relative Kucera, Lott, Horowitz and Hill demonstrate this sensitivity in free space and is dominated by sample in their respective chapters on the design of low- noise, the two have the same absolute sensitivity.15 noise ampliˆers.8,16 Because noise is generated in (The detailed discussion of the coil unloaded Q, any passive elements that dissipate power, includ- loaded Q, and sensitivity in this reference should be ing cables, the design of most detector array coils useful to the coil engineer.) integrates preampliˆers so as to minimize undesira- ble noise. The multi-detector array is a key compo- nent for parallel imaging and comprises many (32 Low Noise Preampliˆer to 128) detectors. As stated, preampliˆers not only The characteristic parameters of preampliˆers, amplify signals; they are key features of these such as the noise ˆgure (NF), aŠect SNR perfor- detectors that assist in the decoupling of individual mance. Thus, a preampliˆer is a key component in detectors, and decoupling is critical to optimizing an RF coil and plays a critical role in the design of a parallel imaging. The very small size of the pream- detector array coil. A brief summary of the pream- pliˆers allows them to be tightly packed to form an pliˆer's function follows. The electromotive force optimized array. or induced voltage (i.e., signal) in a coil is very small, typically on the order of a few mV. A pream- ASingleCoilModel pliˆer increases this small signal to a few mV, e.g., 30 dB (i.e., 1000 times) greater. In practice, gain To understand the mechanism underlying the usually refers to power gain; thus, a 20-dB gain in function of detector array coils, we start with a power corresponds to a 10-dB gain in voltage. schematic representation of a single coil that (Quick conversion factors are shown in Table 1.) includes a simpliˆed preampliˆer decoupling circuit The NF, one parameter for measuring the per- (Fig. 3[a]). This coil circuit is a building block of an formance of the preampliˆer, is detailed else- array coil. L1 represents the coil inductance and

Vol. 6 No. 1, 2007 32 H. Fujita

R1, the coil resistance (typically around 0.5 Ohms in air and 5 Ohms when placed on a phantom). C1 and C2 are the respective tuning and matching capacitances. When looked at from the side of the preampliˆer (i.e., that side of the circuit), a part of

L2, a matching inductor, functions to match the coil impedance to 50 Ohms together with C2.Atthe same time, L2 achieves a parallel resonant circuit formed with C2, in particular, when the input impedance of the preampliˆer, rpreamp,becomes small (e.g., 0 to 2 Ohms). When looked at from the side of the coil, impedance is inˆnity, so the circuit has high impedance and thus is equivalently open. It is extremely important to note that because current cannot ‰ow in an open coil, no magnetic ˆeld can be induced through non-zero mutual inductance in neighboring coil elements. This principle underlies the elimination of cross-talk among coil elements. This decoupling mechanism is explained in more detail later, but it is worth mentioning that the original signal source changes from the current source (without preampliˆer decoupling) to the voltage source (with preampliˆer decoupling) that contains the necessary informa- tion without losing the integrity of the original signal information.

Preampliˆer Decoupling We now consider a system of 2 coupled coils and the underlying mechanism of their decoupling. In Fig. 3(b), Coils 1 and 2 are coupled through mutual inductance, M12. Coil 1(2) is represented by its self- inductance, L1(2); resistance, R1(2); and capacitance

C1(2).Vsignal indicates the signal induced in Coil 1, and Vout indicates the output voltage. When the AC current, I1, ‰ows in Coil 1, the induced AC current,

I2, will ‰ow in Coil 2 through the non-zero mutual coupling between the 2 coils. Thus, assuming harmonic time dependence (i.e., e－ivt ), the output voltage is expressed as 1 Vout＝Vsignal＋ØR1＋iØvL1－ »»I1＋ivM12 I2 vC1 Eq. (5). This equation is essential to understand the need for the coil to be resonated at the target Larmor Fig. 3. (a) A single-coil model with a low-input im- frequency and to comprehend the mechanism of pedance preampliˆer. (b) A system of 2 coils coupled preampliˆer decoupling. The second term in the through non-zero mutual inductance. (c)Asystemof2 right side in Eq. (5) associated with I1 may be coils coupled through non-zero mutual inductance. A realized as the noise related to Coil 1, and the third low-input impedance preampliˆer is added to Coil 2 for term associated with I2 is recognized as the noise preampliˆer decoupling. (d) The matching inductor L2 resulting from the coupling between Coils 1 and 2. (L2A＋L2B ) plays several critical functions in preampliˆ- ThenoiserelatedtoCoil1canbeminimizedby er decoupling. Each quarter-l circuit has its characteris- tuning and matching Coil 1 for resonance, i.e., tic impedance for necessary impedance matching.

Magnetic Resonance in Medical Sciences RF Coil Designs and Trends 33 the imaginary part of the term associated with I1 vanishes, leaving the intrinsic real resistance, R1. Decoupling requires that the third term in the right side in Eq. (5), ivM12 I2, i.e., the noise through the mutual coupling, be zero. To satisfy this condition, we can arrive at 2 cases: either M12＝0orI2＝0.

When M12＝0, Coils 1 and 2 are overlapped to nullify the mutual inductance between the two. Capacitive decoupling may be employed here to cancel the mutual inductance.17 The principle behind the capacitive decoupling is that the non- zero mutual impedance is modeled as an induc- tance, and the inductance is canceled by the addi- tion of capacitance (i.e., the 2 coils are connected by a capacitor). I2＝0 corresponds to the case of so-called preampliˆer decoupling. Now Coil 2 is integrated with a preampliˆer decoupling and matching circuit (Fig. 3[c]). Here,

L2 is a matching inductor, and rpreamp represents the input impedance of the preampliˆer; as rpreamp becomes 0, the inductor, L2,andthecapacitor,C2, can be chosen to form a parallel resonant circuit at the target MR frequency, yielding a high impedance (i.e., inˆnity, theoretically speaking). Coil 2 then becomes an open circuit, and there is no Fig. 4. (a) Schematic representation of a detector current ‰ow, i.e., I2＝0. At this stage, even if the array coil (8-channel). (b) Depiction of B1 sensitivity mutual coupling between the 2 coils is not zero, proˆle in an 8-channel array coil in arbitrary units. Coils 1 and 2 are decoupled using a low-input impedance preampliˆer. This is the art of preampli- ˆer decoupling. Furthermore, it is emphasized that the matching inductor, L2, denoted as L2A and L2B in Fig. 3(d), plays several crucial functions that may not appear obvious. L2A functions to match the coil impedance to 50 Ohms together with CM when looked at from the side of the preampliˆer

(IIIªII). At the same time, L2A (and L2B and C2, which are adjusted to be ``short,'' i.e., the series resonance) achieves a parallel resonant circuit Fig. 5. The structure of a micro-strip line coil. To formed with CM, in particular, when the input achieve necessary impedance matching, the thick- impedance of the preampliˆer, rpreamp (approximate- ness, h, of the dielectric medium and the width, w,of ly RL2B＋RC2), becomes small (e.g., 0 to 2 Ohms). the conducting strip are adjusted such that the

When looked at from the side of the coil (IªII), characteristic impedance Z0＝50 V. the impedance is inˆnity, thereby having a high- impedance or open circuit. In other words, the decoupling circuit considered here and shown in Z 2 Fig. 3(d) (sandwiched by 2 dashed lines) is equiva- Z ＝ 0 Eq. (6), in Z lently a mismatched lW4 (quarter wavelength cable) Load circuit; one end is open and the other end is short. where Z0＝characteristic impedance. In turn, Z0＝

It is instructive and meaningful to view the circuit Zin ZLoad. In other words, by appropriately choos- (regions denoted as II and III) such that each region ing the characteristic impedance of the matching may be represented as a corresponding lW4 circuit section, a quarter-wave section can be used to with its characteristic impedance, Z0.Atransmis- match any 2 impedances, such as in sections II and sion line that is an odd multiple of quarter III of Fig. 3(d). The transformations in impedance wavelengths long that terminates in an impedance fromItoIItoIIIandinreverseareexecuted

ZLoad presents an input impedance of correctly and consistently because each region has

Vol. 6 No. 1, 2007 34 H. Fujita its associated and appropriate characteristic im- decoupling); solenoidal array employing anti- pedance. With this quarter-wave matching section turn loop;18 capacitor decoupling to cancel induc- mechanism, we turn to the preampliˆer. tance;17,19 decoupling network after coils to mani- A preampliˆer is noise matched; for a given ˆeld- pulate the electric ˆeld induced by the mutual in- eŠect transistor (FET), there is a signal-source ductive coupling;20 and shield or transmission line impedance that optimizes or minimizes the noise designs, which are often employed for very high ˆgure. The optimum NF can be achieved when the ˆeld applications (Fig. 5) because most electromag- impedance of the receive coil matches the source netic ˆelds are contained between the coil element impedance of the FET. Because the FET im- and the shield.21,22 In these shielded or transmission pedance is relatively high (:50 Ohms), an line designs, for imaging, electromagnetic-ˆeld impedance transformer comprising L2B and C2 leakage is small, and because the leaked ˆelds are together forms a lW4 circuit in section III of only a small portion of all ˆelds, the isolation Fig. 3(d) and transforms 50 Ohms to the high input between coils remains adequate even if leaked impedance of the FET for the lowest NF. An electromagnetic ˆelds couple with each other. optimal impedance for the lowest noise ˆgure of typical FET is on the order of a few hundred V. Various Parallel Imaging Methods (SMASH Thus, L and C transform the coil's 50 Ohms to 2B 2 versus SENSE) the high impedance of FET for the lowest NF. As an additional note, the input impedance of the With the advent of parallel imaging tech- preampliˆer in Fig. 3(d) is approximated as the sum niques,10–12,23 array coil design plays a major role of resistances associated with L2B and C2. However, in achieving desired parallel-imaging perfor- since the Q of the capacitor is typically much higher mance.24–33 However, with respect to coil design, than that of the inductor, the input impedance is there is a major diŠerence between simultaneous dominated by the contribution of the resistance acquisition of spatial harmonics (SMASH), a associated with the inductor. ``k-space'' reconstruction approach, and sensitivity With an understanding of this preampliˆer encoding (SENSE), an ``x-space'' (i.e., image decoupling method, it is straightforward to extend domain) reconstruction approach. In SMASH, the array concept to 8-, 16-, and even larger num- spatial harmonics are achieved by manipulating the bers of channels. Figure 4(a) shows an 8-channel B1 sensitivity of each coil in the array to replace array coil in which each coil has its own B1 sensitivi- some k-space data lines that would otherwise have ty proˆle (Fig. 4[b]). Because it is a surface coil, been collected by the application of conventional which yields a high SNR over only a small region, gradient phase encoding. This requires that the an overall high SNR over the target FOV or region directions of the array and desired phase encoding of interest (ROI; Fig. 4[b]) can be generated if each coincide. In clinical practice, the complex shapes channel SNR is combined by a sum-of-squares and curvature of the body often make it di‹cult to method,2 for example. As discussed, taking into construct a SMASH array coil whose directionW consideration the appropriate size of each coil with orientation coincides with that of the desired phase SNR optimization, the number of coil channels is encoding, at the same time generating the spatial determined by the number of available receiver harmonics required in that direction. In contrast, channels and the target FOV or ROI. Many refer- SENSEislessrestrictiveaboutplacementofeach ences discuss diŠerent array coil designs for various channel coil in conjunction with a chosen phase imaging applications at diŠerent ˆeld strengths. encoding direction. However, SENSE requires that

Although a detector array coil using quadrature onebeabletotellwhichB1 sensitivity belongs to pairs (for each channel) is possible, as the number whichcoilsothatthealiasedimagethatiscreated of available receiver channels increases, the current unfolds and the desired acceleration factor can be trend is to use a set of various (size and shape) loop achieved. The acceleration factor is the inverse of coils. the reduction factor in imaging time. The measure

of ease to distinguish B1 sensitivity is given by a so-called g-factor, and for a given imaging Other Array Coil Designs application, coil design requires that the g-factor be This paper focuses on detector array coil designs optimized.34 Furthermore, the acceleration factor that employ preampliˆer decoupling because these depends upon the number of receiver channels; are the most popular and widely accepted designs at e.g., to achieve an acceleration factor of 3 in x- 1.5 and 3 Tesla. Other decoupling techniques direction, there must be at least 3 individual coils in include: overlap;2 quadrature (intrinsic isolationW that direction. Additionally, Ohliger and associates

Magnetic Resonance in Medical Sciences RF Coil Designs and Trends 35

Fig. 6. (a) Prototype coil of 3T, 128-channel cardiac array coil. (Courtesy Lawrence L. Wald, Massachusetts General Hospital). (b) A 7T base shielded birdcage TWR brain coil measuring 18 cm long with a diameter of 27 cm. The shield protects against radiation loss at very high frequencies. (Courtesy John L. Patrick, Philips Medical Systems). (c) A 7T, 24-channel whole-brain array. Twenty-four elements are arranged in 8 radially gapped columns (30z gap between elements). The array has anterior columns (2 elements) and posterior columns (4 elements). The image is obtained with MPRAGE: repetition timeW echo time (TRWTE) 5W11 ms; 512×512; 2-mm slice; 0.5×0.5×2mm3 voxels. (Courtesy Patrick J. Ledden, Nova Medical, and JeŠ H. Duyn, National Institutes of Health [NIH], Bethesda, MD). (d) A 7T, 16-channel volume strip array with 16-way TWR interface. (Courtesy Ray F. Lee, New York University).

attempted to establish an upper bound on the tion factor can be achieved as B0 becomes higher spatial encoding capabilities of array coils in (due to shorter wavelength); and tissue conductivity parallel imaging.35 A few highlights from their aŠects parallel imaging performance. The achieva- work include that: a limit for acceleration is 5× for ble SNR with ˆnite coil arrays is discussed in.36 1D undersampling of k-space; a net acceleration These insightful considerations should be helpful to factor of 15× is the limit for 2D; a higher accelera- coil designersWengineers.

Vol. 6 No. 1, 2007 36 H. Fujita

Fig. 7. Illustration of the 3T QD birdcage transmit and 8-channel receive hybrid coil and 3T knee images obtained with the coil. (left: 3D SPGR, a＝39;right:3DSPGR,a＝189)(Im- ages Courtesy Brian Rutt, Robarts Research Institute)

Table 2. The diŠerences between parallel imaging and transmit-SENSE

Parallel Imaging Transmit-SENSE (parallel transmission of multidimensional RF pulses)

What do we know? Aliased images, Full ˆeld-of-view excitation, coil sensitivities coil sensitivities

What do we want to know? Full ˆeld-of-view Aliased excitations images

Can we use preampliˆer decoupling? Yes Noªother decoupling techniques needed

State-of-the-Art Examples array coils are available at these diŠerent static magnetic-ˆeld strengths. Figure 6(d) shows an One-and-a-half-Tesla MR imaging systems are example of shieldWtransmission-line type designs, the gold standard in many clinical applications, and referred to as volume-strip arrays.38 some systems are now available with 96 RF receiver Nonetheless, the essential questions remain: channels.13 3T systems have proven their advan- what are the clinical applications; how many tages in brain and musculoskeletal imaging, and channelsWcoilsdowehaveavailableanduse;how rapid development in body applications are in do we decouple the coils; and how do we optimize progress. The 3T market is being recognized as a the SNR? growing high-end clinical market, and some 3T Another important consideration is the speciˆc systems are now equipped with 128 RF receiver absorption rate, SAR, which is the quantity charac- channels (Fig. 6[a]). On the other hand, a few 7T terizing the energy deposited into a unit mass:39 systems are available worldwide that are targeting SAR1v2B 2a2 Eq. (7), the study of brain function; functional brain imag- 0 1 ing requires a ˆner spatial scale, with resolutions at where v0＝Larmor angular frequency, B1＝trans- a sub-millimeter scale. Researchers are hoping that mit RF magnetic ˆeld amplitude, and a＝coil di- such brain analysis enables improved treatment of ameter (or patient size). As the static magnetic ˆeld neurodegenerative diseases, such as Alzheimer's increases from 1.5 to 3T, the resultant SAR at 3T is disease. Because of the available intrinsic high SNR 4timesthatat1.5T.Thus,athigherˆeldstrengths, at 7T (Fig. 6[b] and [c]),37 images based on 30 or SAR becomes the issue for particular sequences more physicochemical parameters are used to exa- that require a large ‰ip angle andWor many slices to mine tissue morphology, blood ‰ow, metabolism, be collected within the repetition time (TR). In such and chemistry in vivo. Many diŠerent RF detector cases, a hybrid coil can be advantageous. A hybrid

Magnetic Resonance in Medical Sciences RF Coil Designs and Trends 37 coil consists of a localized dedicated transmitter coil whose size is just optimized for a target ROIW FOV and an independent set of receiver coils as shown in Fig. 7.40 Because the transmitter coil is much smaller than a typical whole body transmitter coil, the localized transmitter requires less input power to generate the necessary B1 or a much larger B1 can be generated that translates into a shorter RF pulse duration and potentially reduced SAR. Furthermore, Eq. (7) means that an RF coil with Fig. 8. A9.4T,400MHz half the diameter yields one fourth the SAR provid- TWR multi-element TEM ed that B1 remains unchanged. For this reason, the parallel transceiver head hybrid coil is considered an eŠective approach at coil for B1 control. (Court- higher static magnetic-ˆeld strengths. esyThomasVaughan,Uni- versity of Minnesota)

Fig. 9. (a) Illustration of direction of ‰ow and location of aliased information. (left: parallel imaging; right: Transmit-SENSE). (b) Normal excitation versus inner volume excitation. (c) Example of homogeneity correction. (Courtesy Mark A. Griswold, Case Western Reserve University)

Vol. 6 No. 1, 2007 38 H. Fujita

Fig. 10. The world's ˆrst simultaneously obtained MR-PET human brain im- ages. The patient was injected with FDG, and the PET acquisition took 20 min.

During the PET acquisitions MR images with T2 (a) and Dark Fluid contrast (b) were acquired. The images show the ˆndings: periventricular gliotic foci in the white matter mostly present in the bioccipital region, and calciˆcation-iron deposition of the basal ganglia. (Courtesy Siemens Medical Solutions, Univer- sity of Tennessee, and University of Tuebingen)

time.43 Transmit-SENSE Table 2 illustrates the diŠerences between At very high frequencies, e.g., 3T and above, parallel imaging and Transmit-SENSE. With wavelength eŠects within a dielectric object have regard to RF coil design, the major diŠerence been reported to produce non-uniformity issues, as between SENSE and Transmit-SENSE is that seen in dielectric resonance in which object dimen- SENSE can utilize preampliˆer decoupling among sion and wavelength coincide. Thus, achieving the detector array coil elements, but Transmit-

B1 transmission to yield a uniform B1 ˆeld is chal- SENSE cannot because the current is required to lenging. This is why some researchers attempt to produce a B1 transmit ˆeld. Thus, learning how to correct the B1 ˆeld by manipulating the magni- decouple multiple transmit coil elements still tudes and phases associated with each transmitter remains a challenge. Kurpad's and Vernickle's channel, known as ``mode-scanning excitation'' groups have presented examples of coil designs for 38 41 44,45 (Fig. 6[d]) and ``B1 shimming'' (Fig. 8). Transmit-SENSE.

However, this B1 shimming process is an iterative As recognized in Table 2, with respect to direc- process, and keen insight is required to control B1 tion of ‰ow and location of aliased information, spatial modulation. Transmit-SENSE,42 on the process ‰ow is opposite for parallel imaging and other hand, is an inverse approach. It is well known Transmit-SENSE. The ‰ow is depicted in Fig. 9(a). that a Fourier transform relationship exists in the Use of multi-dimensional RF pulses enables follow- reception (omitting constants for simplicity) be- ing applications such as inner volume excitation tween signal and image (i.e., object): (Fig. 9[b]; faster imaging in localized regions) and

homogeneity correction (B1 shimming;Fig.9[c]). M(…x)＝fS(k)e－i…xkdk Eq. (8). By designing RF pulses to compensate for transmis- sion inhomogeneity and susceptibility correction,

Katscher's group realized that, in transmission, which is for B0, inhomogeneity is corrected. there is a Fourier transform relationship between the excitation RF pulse shape and the 3-D excited Emerging Future RF Coil Technologies spatial proˆle: W The healthcare industry and medical technol- －i…xk Mexcite(…x)＝fB1(k)e dk Eq. (9). ogies are rapidly advancing, and each medical specialty may be supported by a large industrial In the Transmit-SENSE approach, diŠerent time- eŠort. However, because of the variances in func- varying B1 pulses in each coil of the array are used tion between specialties, it is anticipated that many to excite a multi-dimensional proˆle in a reduced diŠerent medical technologies will be integrated to

Magnetic Resonance in Medical Sciences RF Coil Designs and Trends 39

Table 3. Various imaging modalities

Imaging At what are Ionizing Modality we looking? Radiation Pros Cons

Computed Electron density Yes Fast scan time, Radiation TomographyW high spatial exposure, X-ray resolution, compromised widespread soft tissue access contrast

PET (Nuclear Injection of a Negligible Characterization Compromised Medicine*): radioactively (equivalent to of biochemical spatial resolution imaging with labeled drug that amount we functions of (blurring), about positron emitters accumulates in receive from cells, organs, and one hour to wait target organs or natural body structures for the injection pathologic environment in vivo, to go through the tissue, tracers each year) metabolic body prior to labeled with activity at a imaging positron emitters molecular level

Magnetic Proton density No High spatial Patients with Resonance (mostly) resolution, pacemakers, Imaging excellent soft claustrophobia, tissue contrast acoustic noise

SPECT Radiotracers in Negligible Simpler scanner Compromised (Nuclear the human body than PET localization Medicine*): with various capability imaging with decay energies compared with single photon PET emitters

Ultrasound Propagation of No Portability, Image quality sound wave real-time too dependent on (f＝1to15 imaging speciˆc patient MHz) into the patient (back scattering)

*Nuclear medicine involves the emission of radiation from the organ or tissue, whereas radiology involves the transmission of radiation into the tissue. realize more powerful and useful diagnostic tools. take the requirements of other modalities into The principle underlying the combination of consideration. For instance, Kuroda is working on imaging modalities, such as PET-CT,46 is that each an MR-compatible endoscope made of titanium, provides diŠerent image contrasts and gives glass, and polymers with a 1.5-cm ID for a tip RF diŠerent information. In this example, because CT coil.48 This yet unknown and emerging ˆeld will reveals the shape of cancerous tissue and PET further constrain the design of MRI RF coils. shows the activity of the cancer, PET and CT The possibility of wireless RF coils is worth together simultaneously reveal information about mentioning. As discussed earlier and elsewhere,9 tumor shape, location, and activity. Siemens cabling is a challenge in the design of RF receiver Medical Solutions (Erlangen, Germany) is working coils with many channels because of couplings on combining the strengths of MR (soft tissue among the cables49 and coil elements, space limita- contrast) and PET (molecular information),47 as tions, and weight. Obviously, a cable with galvanic reported in Fig. 10. Table 3 summarizes the various interconnections exists between an RF coil and MR imaging modalities. Successful integration requires imaging system to: send the MRI RF signal with other modalities to at least be compatible with MR, its high dynamic range (typically 150 dB with i.e., non-magnetic, and RF coil designs will have to modern systems) and bandwidth; control signals

Vol. 6 No. 1, 2007 40 H. Fujita for tuning detuning, recognizing coil ID and coil W Acknowledgments status; and supply power to the onboard compo- nents. Wireless technologies may advance safety by I am grateful to all my friends and colleagues eliminating baluns and oŠering convenience and with whom I have interacted professionally. In ease of operation for particular clinical applica- particular, I would like to thank Dr. Xiaoyu Yang tions (e.g., in the use of the anterior cardiac array and Tsinghua Zheng at Case Western Reserve coil). On the other hand, challenges include ˆnding University (CWRU) for their invaluable discussions methods to maintain the phase information of in- on preampliˆer decoupling; Professors Mark A. dividual signals.50 Thiscanbedoneintheanalog Griswold (CWRU), Ray F. Lee (New York Univer- domain by mixing the RF signals on to higher fre- sity), and Larry L. Wald (Massachusetts General quency carriers, or it can be done digitally if each Hospital[MGH])andDrs.PatrickLedden(Nova coil is equipped with its own AWD converter. Medical), JeŠ H. Duyn (National Institutes of Another issue is that the necessary wireless data Health), Brian Rutt (Robarts Research Institute), transmission can be greater than 1 GBWs for a high- J. Tommy Vaughan (University of Minnesota), end MR imaging scanner with high receiver-chan- Ioannis Panagiotelis (Siemens Medical Solutions), nel count, assuming the required bits accuracy, Juergen Kampmeier (Siemens Medical Solutions), sampling rate, maximum gradient strength, and and John L. Patrick (Philips Medical Systems) FOV. This speciˆcation is not yet available from for providing their invaluable picturesWimages; the current WLAN industry. Furthermore, the im- Professor Robert W. Brown (CWRU) for his plementation of wireless technology requires a bat- invaluable comments and discussions to improve tery to power the coil and its associated on-board the manuscript; and Matthew Finnerty for his electronics. However, it is anticipated that the assistanceingeneratingaplotofFig.4(b).Iam aforementioned challenges will be overcome be- also grateful for discussions with Drs. Johan cause of the rapid technological advancement in the Overweg (Philips Research Laboratories, Ham- industry, as evidenced by the growth of the wireless burg, Germany), Arne Reykowski (InVivo Corpo- communication industry. Thus, the challenge ration), and Ralph Oppelt (Siemens Corporate remainsastohowwecantakeadvantageofthese Research,Germany)atthe2005ISMRMMR emerging technologies. Engineering Study Group Workshop on Wireless MRI Coil Technology and with Wayne Dannels (Toshiba Research Institute USA) that guided my Summary description on the approach to desirable wireless The subject of array coil design is complex and data signal processing. Professors Steve Wright encompasses the entire reception chain and (Texas A & M University), Stuart Crozier (The emerging applications of parallel imaging and University of Queensland), Steve Conolly (Stan- transmission technologies. Although there are ford University and University of California, many coil designs for diŠerent applications at Berkeley), Dr. Hubertus Fischer (Siemens Medical various B0 ˆeld strengths, the principles underlying Solutions), and Kazuya Okamoto (Toshiba Medi- the design of detector array coils are essentially the cal Systems Corporation) are gratefully acknowl- same. In particular, preampliˆer decoupling in edged for their useful discussions over the years. view of a mismatched lW4matchingcircuitis Finally, I would like to thank Professor Katsumi important. Design of RF detector array coils must Kose (Tsukuba University), Dr. Fumiyuki Mitsu- take into consideration that: (1) SNR be optimized; mori (National Institute for Environmental Stud-

(2) for a given clinical application at a selected B0, ies), and Professor Tsuneya Watabe (Editor-in- the number of available receiver channels, desirable Chief, Magnetic Resonance in Medical Sciences) phase encoding direction, target acceleration fac- for providing me with this wonderful opportunity tor, and FOV determine the coil design of choice; to write a review article on this RF coil-related and (3) SNR (for spatial resolution) and speed of subject. acquisition (for temporal resolution, i.e., accelera- ApartofthisworkwaspresentedattheFour- tion factor) are best balanced depending upon clini- teenth Scientiˆc Meeting of the ISMRM51 and the cal application. 34th Annual Scientiˆc Meeting of the Japanese Additional integration of diŠerent imaging Society of Magnetic Resonance in Medicine held at modalities is anticipated, which will further con- the Tsukuba International Conference Hall in strain design of RF array coils intended for either 2006.52 reception-only or transmissionWreception.

Magnetic Resonance in Medical Sciences RF Coil Designs and Trends 41

array coils. Magn Reson Med 2002; 47:794–799. References 19. Jevtic J. Ladder networks for capacitive decou- 1. Hoult DI. The NMR receiver: a description and pling in phased-array coils, In: Proceedings of analysis of design. Progress in NMR spectroscopy the ninth annual meeting of the ISMRM. Glasgow, 1978; 12:41–77. 2001; 17. 2. Roemer PB, Edelstein WA, Hayes CE, Souza SP, 20. Wang J, Friedrich A. Antenna arrangement for a Mueller OM. The NMR phased array. Magn Reson magnetic resonance apparatus. U.S. Patent Med 1990; 16:192–225. 5,708,361. 3. Hayes CE, Hattes N, Roemer PB. Volume imaging 21. Lee RF, Giaquinto RO, Hardy CJ. Coupling and with MR phased arrays. Magn Reson Med 1991; decoupling theory and its application to the MRI 18:309–319. phased array. Magn Reson Med 2002; 48:203–213. 4.FennAJ,TemmeDH,DelaneyWP,Courtney 22. Lee RF, Westgate CR, Weiss RG, Newman DC, WE. The development of phased-array radar tech- Bottomley PA. Planar strip array (PSA) for MRI. nology. Lincoln Lab J 2000; 12:321–340. Magn Reson Med 2001; 45:673–683. 5. Stutzman WL, Thiele GA. Antenna theory and 23. Madore B, Pelc NJ. SMASH and SENSE: experi- design. John Wiley & Sons, 1981; 108. mental and numerical comparisons. Magn Reson 6. Wright SM, Wald LL. Theory and application of Med 2001; 45:1103–1111. array coils in MR spectroscopy. NMR Biomed 24. Herlihy DJ, Larkman DJ, Fujita H, Burl M, 1997; 10:394–410. Hajnal JV. A 4-channel head coil for SENSE 7. Hayes CE, Edelstein WA, Schenck JF, Mueller imaging, In: Proceedings of the eighth annual OM, Eash M. An e‹cient, highly homogeneous meeting of the ISMRM. Denver, 2000; 1394. radiofrequency coil for whole-body NMR imaging 25. Hajnal JV, Larkman DJ, Herlihy DJ. An array at 1.5T. J Magn Reson 1985; 63:622–628. that exploits phase for SENSE imaging, In: 8. Kucera J, Lott U. Low noise ampliˆer design, In: Proceedings of the eighth annual meeting of the Golio M, ed. The RF and microwave handbook. ISMRM. Denver, 2000; 1719. CRC Press, 2001; 548–571. 26. Fujita H, Spence DK. A novel 8-channel ``saddle- 9. Yang X, Zheng T, Fujita H. TWR switches, baluns, train'' array coil for abdominal SENSE imaging at and detuning elements in MRI RF coils, In: 1.5T, In: Proceedings of the 10th annual meeting ISMRM Fourteenth Scientiˆc Meeting Weekend of the ISMRM. Hawaii, 2002; 326. Syllabus 2006. 27. Boskamp E, Blawat L, King KF, Lorbiecki J, 10. Sodickson DK, Manning WJ. Simultaneous acqui- Tiwari AB. A 16-channel neurovascular SENSE sition of spatial harmonics (SMASH): fast imaging array, In: Proceedings of the tenth annual meeting with radiofrequency coil arrays. Magn Reson Med of the ISMRM. Hawaii, 2002; 852. 1997; 38:591–603. 28. Blawat LR, Hoppel BE, Boskamp EB, Iwadate Y. 11. Pruessmann KP, Weiger M, Scheidegger MB, 3-T breast array optimized for parallel imaging, In: Boesiger P. SENSE: sensitivity encoding for fast Proceedings of the twelfth annual meeting of the MRI. Magn Reson Med 1999; 42:952–962. ISMRM. Kyoto, 2004; 1596. 12. Griswold MA, Jakob PM, Heidemann RM, et al. 29. Okamoto K, Takagi M, Ishii M, et al. A novel Generalized autocalibrating partially parallel acqui- ``smart'' neurovascular array coil system for sitions (GRAPPA). Magn Reson Med 2002; 47: parallel imaging, In: Proceedings of the twelfth 1202–1210. annual meeting of the ISMRM. Kyoto, 2004; 1595. 13. WigginsGC,PotthastA,TriantafyllouC,etal.A 30. Ohliger MA, Greenman RL, Giaquinto R, 96-channel MRI system with 23- and 90-channel McKenzie CA, Wiggins G, Sodickson DK. Con- phase array head coil at 1.5T, In: Proceedings of centric coil arrays for parallel MRI. Magn Reson the thirteenth annual meeting of the ISMRM. Med 2005; 54:1248–1260. Miami, 2005; 671. 31. Spencer D, Akao J, Duensing R, et al. Design of a 14. Hoult DI, Richards RE. The signal to noise ratio 32-channel cardiac array for parallel imaging, In: of the nuclear magnetic resonance. J Magn Reson Proceedings of the thirteenth annual meeting of 1976; 24:71–85. theISMRM.Miami,2005;911. 15. Froncisz W, Jesmanowicz A, Kneeland JB, Hyde 32. Lanz T, Kellman, Nittka M, Greiser A, Griswold JS. Counter rotating current local coils for high- M. A 32-channel cardiac array optimized for resolution magnetic resonance imaging. Magn parallel imaging, In: Proceedings of the fourteenth Reson Med 1986; 3:590–603. annual meeting of the ISMRM. Seattle, 2006; 16. Horowitz P, Hill W. In: Chapter 7, The art of 2578. electronics. Second edition. Cambridge University 33. Reykowski A, Hemmerlein M, Fischer H. 16-chan- Press, 1989; 434–435. nel headWneck matrix coils for 3 Tesla, In: Pro- 17. Li BK. The design and analysis of high frequency ceedings of the thirteenth annual meeting of phased array coils for MRI. Ph.D. Thesis. The theISMRM.Miami,2005;908. University of Queensland, 2006; 84–89. 34. Weiger M, Pruessmann KP, Leussler C, 18. Su S, Zou MX, Murphy-Boesch J. Solenoidal Roschmann P, Boesiger P. Speciˆc coil design for

Vol. 6 No. 1, 2007 42 H. Fujita

SENSE: a six-element cardiac array. Magn Reson 43. Zhu Y. Parallel excitation with an array of trans- Med 2001; 45:495–504. mit coils. Magn Reson Med 2004; 51:775–784. 35. Ohliger MA, Grant AK, Sodickson DK. Ultimate 44. Kurpad KN, Boskapm EB, Balanis SM. A parallel intrinsic signal-to-noise ratio for parallel MRI: transmit volume coil with independent control of electromagnetic ˆeld considerations. Magn Reson currents on the array elements, In: Proceedings of Med 2003; 50: 1018–1030. the thirteenth annual meeting of the ISMRM. 36. Wiesinger F, De Zanche N, Pruessmann KP. Miami, 2005; 16. Approaching ultimate SNR with ˆnite coil arrays, 45. Vernickle P, Roschmann P, Findeklee C, et al. An In: Proceedings of the thirteenth annual meeting of 8-channel transmitWreceive body coil for 3T, In: theISMRM.Miami,2005;672. Proceedings of the foureenth annual meeting of 37. Ledden PJ, Mareyam A, Wang S, van Gelderen P, the ISMRM. Seattle, 2006; 123. Duyn JH. 24-channel receive-only array for brain 46. Arnulf Oppelt, ed. Imaging systems for medical imaging at 7T, In: Proceedings of the fourteenth diagnostics. Erlangen: Publicis Corporate Publish- annual meeting of the ISMRM. Seattle, 2006; 422. ing, 2005; 830–835. 38. Lee RF, Brown R, Mizsei G, et al. Implementation 47. Fischer H, Ladebeck R, Schmand MJ. Challenges of mode-scanning excitation method with a 16- and technical solutions of a combined MR PET channel transmitWreceive volume strip array at 7T, scanner. Jpn J Magn Reson Med 2006; 26 (Suppl): In: Proceedings of the fourteenth annual meeting 111. of the ISMRM. Seattle, 2006; 125. 48. Kuroda K. MR compatible endoscope. Nihon 39. Haacke EM, Brown RW, Thompson MR, Keizai Shinbun 2005; 18 November. Venkatesan R. Magnetic resonance imaging: physi- 49. Koste GP, Nielsen MC, Tolliver TR, Frey RL, cal principles and sequence design. New York: Watkins RD. Optical MR receive coil array inter- Wiley-Liss, 1999; 858–859. connect, In: Proceedings of the thirteenth annual 40. Yang X, Fujita H, Zheng T, Richmond K, Colwell meeting of the ISMRM. Miami, 2005; 411. C. A 3-T QD transmitW8-channel receive hybrid 50. Scott G, Yu K. Wireless transponders for RF coils: knee coil, In: Proceedings of the thirteenth annual systems issues, In: Proceedings of the thirteenth meeting of the ISMRM. Miami, 2005; 907. annual meeting of the ISMRM. Miami, 2005; 330. 41. Vaughan J, DelaBarre L, Snyder C, et al. 9.4-T hu- 51. Fujita H, Yang X, Zheng T. Array coil types and man imaging: preliminary results, In: Proceedings design principles, In: ISMRM Fourteenth Scientiˆc of the fourteenth annual meeting of the ISMRM. Meeting Weekend Syllabus 2006. Seattle, 2006; 529. 52. Fujita H. New horizons in MR technology: from a 42. Katscher U, Boernert P, Leussler C, van den Brink perspective of RF coil designs and trends. Jpn J JS.TransmitSENSE.MagnResonMed2003; Magn Reson Med 2006; 26 (Suppl): 111. 49:144–150.

Magnetic Resonance in Medical Sciences