WO2009037709A2

WO2009037709A2 - Magnet and gradient coil configurations for self-contained mri probes

Info

Publication number: WO2009037709A2
Application number: PCT/IL2008/001259
Authority: WO
Inventors: Aharon Blank; Yuval Zur; Eric Tammam; Amir Harel
Original assignee: Topspin Medical (Israel) Ltd.; Technion Research & Development Foundation Ltd.
Priority date: 2007-09-20
Filing date: 2008-09-21
Publication date: 2009-03-26
Also published as: WO2009037711A3; WO2009037710A3; WO2009037709A3; WO2009037711A2; WO2009037710A2

Abstract

A magnet assembly for a self-contained MRI probe having a longitudinal axis in a z-direction, the probe adapted to image an imaging region located relative to the probe, in terms of a Cartesian x-y-z coordinate system, at least partly in a positive x direction, the magnet assembly comprising magnet portions which produce a static magnetic field in the imaging region, including at least: a) a first portion located at y < 0, with magnetization having a positive y component and a negative x component; and b) a second portion, located at y > 0, with magnetization having a positive y component and a positive x component.

Description

MRI PROBE MAGNET AND COIL CONFIGURATIONS

RELATED APPICATIONS

The present application claims benefit under 35 USC 119(e) from US provisional patent applications 60/960,212, and 60/960,213, both filed on September 20, 2007. This application is related to two PCT applications filed on even date, the first titled "MRI Probe," having attorney docket number 44842, and the second titled "Image Reconstruction Methods for MRI Probes," having attorney docket number 38140. The disclosures of all of these applications are incorporated herein by reference. This application is also related to the following applications: U.S. Patent

Application No. 10/968,853 filed October 18, 2004, entitled "Magnet and Coil Configurations for MRI probes" and published as US2006/0084861; U.S. Patent Application No. 10/597,325 filed July 20, 2006 and entitled "MRI Probe for Prostate Imaging"; PCT Patent Application No. PCT/IL2005/000074 filed January 20, 2005, entitled "MRI Probe for Prostate Imaging" and published as WO 2005/067392; and U.S. Provisional Patent Application No. 60/537,030 filed January 20, 2004 and entitled "MRI Probe for Prostate Imaging". The disclosures of the above applications are each fully incoφorated herein by reference.

FIELD AND BACKGROUND OF THE INVENTION

The present invention, in some embodiments thereof, relates to self-contained MRI probes and their magnetic field configurations, and, more particularly, but not exclusively, to probes used inside the body for medical imaging.

Conventional medical MRI systems have a field of view with very uniform static magnetic field in the bore of a large magnet, and use gradient coils that produce highly linear gradient fields in three dimensions, for gradient encoding of images. Conventional MRI systems sometimes use RF receiver probes that can be used inside blood vessels, the rectum (for prostate imaging) and other body cavities, in order to improve signal to noise ratio (SNR) when imaging regions adjacent to the probe, as described, for example, by US patent 5,699,801 to Atalar, and in US patent 5,476,095 to Schnall et al. In order to reduce the cost of MRI systems, while allowing imaging of small regions inside the body with high resolution and SNR, fully self-contained MRI probes have been proposed. In these "inside out" MRI devices, inspired by NMR probes used for well- logging, the field of view is outside the probe. US 5,572,132 to Pulyer et al describes a self- contained MRI probe, for use in blood vessels or other body cavities, with permanent magnets configured to produce a static magnetic field with a saddle point, outside the probe. The imaging region of the probe is a small region around the saddle point, where the magnetic field is locally very uniform. Gradient coils produce field gradients in the imaging region, in the radial, azimuthal, and longitudinal directions, for gradient encoding of images.

US patent 5,304,930 to Crowley et al describes a larger self-contained "inside-out" MRI device, designed to be used outside the body, to image a part of the body. In this device, the static -magnetic field is not nearly uniform, but has a substantial gradient in one direction, for example the x-direction, which is used for gradient encoding of images. This field gradient is required to be very linear, and the surfaces of constant field are required to be very flat, in the y-z plane. In order to satisfy those requirements, the magnetic field is produced by a set of six permanent magnets, with hinges between them, and the angles of orientation of the magnets are adjusted to produce a magnetic field of the required characteristics. In one magnet configuration described by Crowley et al, two of the adjacent magnets, separated in the y-direction, are tilted toward each other, so that one of them, located at y < 0, is magnetized with a positive x-component, pointing toward the imaging plane, and a positive y-component, while the other magnet, located at y > 0, is magnetized with a negative x-component, pointing away from the imaging plane, and also with a positive y-component. US patent 6,704,594 to Blank et al describes a self-contained intravascular MRI probe, with a substantial static radial magnetic field gradient. The probe produces images by recording data from one voxel at a time. Voxels at different radial distances are recorded by changing the central frequency of the RF pulses used to excite nuclei. Voxels at different azimuthal angles are recorded by turning the probe on its longitudinal axis, and voxels at different longitudinal positions are recorded by moving the probe along its longitudinal axis.

R. J. Haken, in Chapter 4 of his doctoral thesis, titled "Entwicklung und Anwendung mobiler NMR-Sonden," published in 2001 by the Rheinisch-Westfalischen Technischen Hochschule in Aachen, Germany, describes a self-contained NMR probe 17 mm in diameter, said to be suitable for use in the esophagus or intestines, as well as in industrial applications. The probe is axisymmetric, and does not use gradient coils, but has a radial static magnetic field gradient, and can use different central frequencies of the RF pulses to resolve images radially. This publication is representative of many other papers published by the "NMR-MOUSE" group, particularly by Haken' s thesis advisor, B. Blϋmich, describing "inside out" NMR probes, generally for industrial or external medical use, and too large to fit inside the body. Some of these probes do use gradient coils, for example the probe described in US patent 6,489,767 to Prado et al.

The contents of all of the above documents are incorporated by reference as if fully set forth herein.

SUMMARY OF THE INVENTION

An aspect of some embodiments of the invention concerns magnet and gradient coil configurations for use in self-contained MRI probes that lead to improved uniformity of field gradients within each slice in an imaging region, with little reduction in magnetic field strength in the imaging region.

There is thus provided, according to an exemplary embodiment of the invention, a magnet assembly for a self-contained MRI probe having a longitudinal axis in a z-direction, the probe adapted to image an imaging region located relative to the probe, in terms of a Cartesian x-y-z coordinate system, at least partly in a positive x direction, the magnet assembly comprising magnet portions which produce a static magnetic field in the imaging region, including at least: a) a first portion located at y < 0, with magnetization having a positive y component and a negative x component; and b) a second portion, located at y > 0, with magnetization having a positive y component and a positive x component. Optionally, the portions of the magnet assembly also include: a) a third portion located at y < 0, further from the imaging region than the first portion, with magnetization having a positive y component and a positive x component; and b) a fourth portion located at y > 0, further from the imaging region than the second portion, with magnetization having a positive y component and a negative x component.

Optionally, the portions of the magnet assembly are grouped into three or more parts located at different longitudinal locations, including at least: a) two end parts, located respectively at z > 0 and z < 0, each end part comprising a first portion and a second portion as specified in claim 1 ; and b) a center part, located between the two end parts longitudinally, the center part comprising_a third portion located at y < 0, with magnetization having a positive y component and a positive x component, and a fourth portion located at y > 0, with magnetization having a positive y component and a negative x component.

Optionally, the center portion also comprises a first portion as specified above, closer to the imaging region than the third portion, and a second portion as specified in above, closer to the imaging region than the fourth portion.

Optionally, the first portion of the center part is smaller than the third portion, and the second portion of the center part is smaller than the fourth portion.

Optionally, the direction of magnetization is within 40 degrees of the positive y direction for most of the magnet volume.

In an embodiment of the invention, each portion is part of a separate magnet associated with that portion, and each magnet is substantially uniformly magnetized. Optionally, the magnet assembly also comprises at least one intermediate magnet not associated with any of the portions, located between two magnets associated with two of the portions, the direction of magnetization of the intermediate magnet being intermediate between the directions of magnetization of the two magnets associated with the two portions. There is further provided, in accordance with an exemplary embodiment of the invention, a self-contained MRI probe comprising a magnet assembly according to an exemplary embodiment of the invention, and a set of one or more phi-gradient coils, phi being an azimuthal coordinate, adapted to produce a phi-gradient field with a component parallel to the static magnetic field, for gradient encoding in the imaging region.

Optionally, the phi-gradient coil set is adapted to produce a component of the phi- gradient field parallel to the static magnetic field, as a function of phi for constant z and constant static magnetic field somewhere in the imaging region, that differs, at at least one value of phi, by at least 20% of the maximum value of the function of phi, from a best linear fit.

Optionally, the phi-gradient coil set is adapted to produce a component of the phi- gradient field parallel to the static magnetic field, as a function of azimuthal coordinate phi for constant z and constant static magnetic field somewhere in the. imaging region, that does not differ, for any value of phi, by more than 20% of the maximum value of the function of phi, from a best linear fit.

Optionally, the probe comprises a set of one or more z-gradient coils, adapted to produce a z-gradient field with a component parallel to the static magnetic field, for gradient encoding in the imaging region.

Optionally, the probe has size, shape and mechanical properties adapted for rectal use, with at least part of the prostate in the imaging region.

Optionally, the probe also includes an ultrasound imaging array suitable for imaging the prostate, the ultrasound array, magnets and RF transmitting and receiving elements being mounted on the probe at known relative positions and orientations to each other, such that images generated by the ultrasound array can be aligned with MRI images generated by the probe.

Alternatively, the probe is small enough for intravascular use, and sufficiently sensitive, when used inside an artery, to distinguish plaque from healthy tissue in the artery wall. Optionally, the probe has a longitudinal axis in the z direction, having a total length in the z direction greater than its total diameter in the x and y directions.

There is further provided, in accordance with an exemplary embodiment of the invention, a self-contained MRI probe adapted to image an imaging region, the probe comprising a magnet assembly which produces a static magnetic field in the imaging region, the probe having size, shape and mechanical properties adapted for positioning in the rectum such that at least part of the surface of the prostate intersects the imaging region, with a surface of constant strength of the static magnetic field being curved to more closely match said part of the surface of the prostate than any plane would. Optionally, the probe has a longitudinal axis in a z-direction, the imaging region located relative to the probe, in terms of a Cartesian x-y-z coordinate system, at least partly in a positive x direction, the magnet assembly comprising magnet portions which produce a static magnetic field in the imaging region, including at least: a) a first portion located at y < 0, with magnetization having a positive y component and a positive x component; and b) a second portion, located at y > 0, with magnetization having a positive y component and a negative x component.

There is further provided, in accordance with an exemplary embodiment of the invention, a self-contained MRI probe having a longitudinal axis in a z-direction, the probe adapted to image an imaging region located relative to the probe, in terms of a Cartesian x- y-z coordinate system, at least partly in a positive x direction, comprising: a) a magnet assembly that produces a magnetic field with a predominant y component at least somewhere in the imaging region; and b) one or more z-gradient coils, adapted to produce a z-gradient field with a component parallel to the static magnetic field for gradient encoding in the imaging region, wherein at least a portion of one z-gradient coil, defined as the y- z portion, is oriented in a direction closer to the y-z plane than to the x-y or x-z plane. Optionally, at least a portion of one z-gradient coil, defined as the x-z portion, is oriented in a direction closer to the x-z plane than to the y-z or x-y plane.

Optionally, the y-z and x-z portions are part of a single non-planar coil. Alternatively, the y-z and x-z portions are parts of separate substantially planar coils.

BRIEF DESCRIPTION OF THE DRAWINGS

Some embodiments of the invention are herein described, by way of example only, with reference to the accompanying drawings. With specific reference now to the drawings in detail, it is stressed that the particulars shown are by way of example and for purposes of illustrative discussion of embodiments of the invention. In this regard, the description taken with the drawings makes apparent to those skilled in the art how embodiments of the invention may be practiced. In the drawings:

FIG's. IA- IL schematically show magnet configurations for self-contained MRI probes, according to different exemplary embodiments of the invention;

FIG. 2 schematically shows a self-contained MRI probe using the magnet configuration of Fig. IF;

FIG. 3 schematically shows details of the magnets in the probe of Fig. 2; FIG's. 4A and 4B schematically show sets of z-gradient coils according to two embodiments of the invention, the first set similar to the one used in the probe in Fig. 2;

FIG. 4C schematically shows the magnet field lines of the field created by the z- gradient coils of Fig. 4A, for a particular ratio of currents in the coils;

FIG. 5 schematically shows a plot of static magnet field strength vs. distance from the probe along the x-axis, for the probe shown in Fig. 2; FIG. 6 schematically shows contours of static magnetic field strength in the x-y plane, with superimposed magnets, phi-gradient coils, and RF coil, for the probe shown in Fig. 2;

FIG. 7 schematically shows contours of the static magnetic field strength in the x-z plane, relative to the probe, for the probe shown in Fig. 2; FIG. 8 schematically shows a plot the RF field strength vs. distance from the probe along the x-axis, for the probe shown in Fig. 2;

FIG. 9 schematically shows a contour plot of RF field strength in the x-y plane, for the probe shown in Fig. 2; FIG. 10 schematically shows a contour plot of RF field strength in the x-z plane, for the probe shown in Fig. 2;

FIG. 11 schematically shows a circuit used to drive the RF coil, and an equivalent circuit as seen by the RF power supply, for the probe shown in Fig. 2;

FIG. 12 schematically shows a plot of the strength of the z-gradient field, as a function of longitudinal position for surfaces of different static magnetic field strength, for the probe shown in Fig. 2;

FIG. 13 schematically shows a plot of the strength of the phi-gradient field, as a function of distance from the x-axis along a field line, for surfaces of different static magnetic field strength, for the probe shown in Fig. 2; FIG. 14 schematically shows a circuit used for driving gradient coils in an MRI probe, according to an exemplary embodiment of the invention;

FIG's. 15A and 15B schematically show exemplary MRI pulse sequences that can be used with the probe shown in Fig. 2; and

FIG. 16 schematically shows a block diagram of an MRI imaging system, according to an exemplary embodiment of the invention.

DESCRIPTION OF EMBODIMENTS OF THE INVENTION

An aspect of some embodiments of the invention concerns a magnet assembly for a self-contained MRI probe, with a longitudinal axis in the z direction, using a Cartesian x-y- z coordinate system, and an imaging region that includes the positive x direction. The magnet assembly is not uniformly magnetized, but includes a portion at y < 0 with positive y component and negative x component of magnetization, and a portion at y > 0 with positive y component and positive x component of magnetization.

The y components of magnetization, optionally the dominant components and optionally uniform throughout the magnet assembly, produce a dipole component of a static magnetic field in the imaging region which falls off relatively slowly with distance from the magnet assembly, allowing relatively high image quality to be obtained relatively far from the probe. An example of the fall-off in field strength with distance from the surface of a 30 mm diameter NdFeB magnet assembly is given below in Fig. 5. The x components of magnetization, which vary in direction in different parts of the magnet assembly, produce higher multiple components of magnet field in the imaging region, which fall off more quickly with distance from the magnet assembly, but which "trim" the magnetic field in the imaging region, giving it certain desired properties, such as a field strength that is more uniform in z and/or azimuthal angle φ at a given distance r from the z-axis. To the extent that the magnetization direction is not too far from the y-direction throughout the magnet assembly, for example, an angle θ less than 40 degrees or 30 degrees or 20 degrees, the dipole component of the magnetic field is hardly reduced from what it would be if the magnet assembly were uniformly magnetized in the y-direction, being reduced approximately by a factor of cosθ, and the magnetic field strength is hardly reduced by the "trimming", especially in the furthest parts of the imaging region where the dipole component of the magnetic field may dominate. However, the x components of the magnetic field can produce local "trimming" changes in the magnet field, near the probe, that are on the order of tanθ times the dipole component of the field, which may be substantial.

This result is potentially advantageous compared to the results of an alternative method of "trimming" a magnetic field to produce desired properties in the field profile, in which a uniformly magnetized magnet has its shape adjusted in order to modify the magnetic field. Particularly in medical imaging probes that are used inside the body, there is limited space available, and it may be advantageous to maximize the magnetic field strength throughout the imaging region, in order to get adequate signal to noise ratio (SNR) within the limited acquisition time that patients will tolerate. In these circumstances, all of the available volume may be filled with a uniformly magnetized magnet to maximize the magnetic field strength in the imaging region. A substantial change in the shape of the magnet, in order to produce desired properties in the magnetic field profile, while keeping the magnet within the same envelope of available space, may result in a substantial reduction in the dipole component of the magnetic field, and a corresponding reduction in the average magnetic field in the imaging region, especially in the parts furthest from the magnet. This may be true if part of the magnet volume is replaced by soft magnetic material such as iron, as well as if it is replaced by non-magnetic material such as air.

Optionally, in addition to the two portions of the magnet assembly described above, at y < 0 and y >0, the magnet assembly has other portions at various locations and with various directions of magnetization. As will be described below in reference to Figs. ID, IE, and IF, these portions produce various effects in the magnetic field profile in the imaging region, which may have beneficial effects on image quality and on ease of image reconstruction. Optionally, the different portions of the magnet assembly comprise separate magnets, each with substantially uniform direction of magnetization, which are optionally bonded together to form the magnet assembly, or held together by other mechanical means, with enough force to overcome the large magnetic forces that may tend to align the magnets in a different way. Alternatively, at least two portions of the magnet assembly are portions of a single magnet, magnetized with non-uniform direction of magnetization.

Optionally, for at least half of the magnet assembly, or for at least 75% of it, or for at least 90% of it, the direction of magnetization is within a relatively small angle of one direction (for example, the positive y direction), for example within 40 degrees, or within 30 degrees, or within 20 degrees. Having most of the magnet assembly magnetized within a relatively small angle of the same direction ensures that the magnetic field is predominantly a dipole field, with a relatively slow fall off in field strength with distance from the magnet assembly, and with the field strength almost as high as it would be if the magnet were uniformly magnetized in that one direction. If the magnet assembly is magnetized in directions that differ by 30 degrees from the y-direction, for example, then the dipole component of the magnetic field will only be reduced to cos(30°) or 86.6% of the strength it would have if the magnet assembly were uniformly magnetized in the y-direction, but the x component of magnetization, for example, would be 50% as great as the y component of magnetization, and could make a substantial difference in shaping the field profile.

Optionally, the magnet assembly is part of an MRI probe with gradient coils, for example z-gradient coils and/or phi-gradient coils, phi being an azimuthal angle in the x-y plane. The gradient coils provide phase encoding in the z direction and/or in the phi direction (approximately the y direction) in the imaging region. Optionally, the phi (or z) gradient coils produce a phi (or z) gradient field that is substantially nonlinear, as a function of phi (or z), for a given z (or phi) and for a given field strength. For example, the gradient field differs by more than 10% of its maximum value, or more than 20%, or more than 35%, or more than 50%, from a best linear fit as a function of phi (or z), for at least some fixed values of z (or phi) and field strength. This nonlinearity may result in non-uniformity of the spatial resolution of the image in the y-direction or the z-direction, for example much higher resolution near the center of the imaging region than at the edges, and/or in the need for more complicated techniques for image reconstruction, as described in the co-filed patent application titled "Image Reconstruction Methods for MRI Probes." A potential advantage to producing static magnetic fields in which the field strength is more constant as a function of z and phi for a given r, using the "trimming" method described above, is that for such magnetic field profiles it may be easier to produce gradient fields that are nearly linear functions of the gradient variable (z or phi) for a given field strength. Optionally, the phi (or z) gradient field differs by less than 10%, or 20%, or 35%, or 50%, of its maximum value, from the best linear fit. As described in the co-filed application "Image Reconstruction Methods for MRI Probes," a probe with a substantially nonlinear gradient field may not be able to use the highly efficient FFT algorithm for reconstructing images, but may have to use slower computational methods, and/or have lower image quality.

Optionally, the self-contained MRI probe is a rectal probe used to image the prostate. Optionally, the rectal probe is dual mode probe which can do ultrasound imaging as well as MRI, as described in the co-filed application, "MRI Probe."

An aspect of some embodiments of the invention concerns a self-contained MRI rectal probe for prostate imaging, in which a magnet assembly produces a static magnetic field profile, in an imaging region that intersects the prostate, where a surface of constant field strength approximately matches the near surface of the prostate, for example better than a planar surface would. Here "matches better" is defined, for example, as having a smaller root-mean-square difference in location, within the imaging region. In contrast to the case described above, where the constant field surface is approximately a constant r surface, and hence convex with respect to the probe, this constant field surface is concave with respect to the probe. Techniques similar to those described above for trimming the field by adjusting the direction of magnetization of different portions of the magnet assembly, can also be used in this case, to produce a desired concave shape of the surfaces of constant field strength. For example, a magnet assembly with a portion at y < 0 with positive y component and positive x component of magnetization, and a portion at y > 0 with positive y component and negative x component of magnetization, may be used to produce a magnetic field with concave curvature of the constant field surfaces, in an imaging region in the positive x-direction. An aspect of some embodiments of the invention concerns a self-contained MRI probe, with a longitudinal axis in the z-direction, an imaging region in the x-direction, and a magnet assembly that produces a static magnetic field predominantly in the y-direction in the imaging region, with a set of one or more z-gradient coils, at least a portion of which are oriented predominantly in the y-z plane. Optionally, another portion of the coils is oriented predominantly in the x-z plane. Optionally, both the y-z portion and the x-z portion are portions of a single non-planar coil. Alternatively, the y-z portion and x-z portion are portions of separate substantially planar coils. These arrangements of the z-gradient coils may provide a more efficient use of the limited space generally available in self-contained MRl probes, especially medical probes used inside the body

Exemplary Basis for Magnet Design

This section will describe some of the configurations of non-uniform magnetization direction that will produce specific trimming effects in a dipole magnetic field. These results may also apply qualitatively to uniformly magnetized magnets of shapes that do not produce purely dipole fields, and such shapes can generally be expressed as a superposition of dipoles in any case. Although the novelty of particular configurations does not depend on the motivation for them, this section may help to explain the reasons for some of these configurations.

Fig. IA schematically shows a magnet assembly 100 consisting of only one magnet 102, magnetized uniformly in the y-direction, with an RF coil 104 on the side of the +x- direction, representing the direction of the imaging region. The magnet produces a static magnetic field that is substantially a dipole field. In the imaging region the field is roughly in the -y-direction, and falls off with distance r at a rate that could be between 1/r and 1/r², depending on the ratio of length to diameter of the magnet. The field strength in the imaging region is about as large as it could be, for a given magnet material, and for this envelope of available magnet volume. At a given r, the field strength is somewhat higher in the +y or -y directions, than in the +x or -x directions, again depending on the ratio of length to diameter, and falls off with z going toward the ends of the magnet from the center, for a given r and azimuthal angle φ, so the surfaces of constant field strength are not very close to surfaces of constant r. Assembly 100 is similar, for example, to the magnet assemblies used as MRI sensors in some of the embodiments of the invention shown in the related published patent application US2006/0084861, titled "Magnet and Coil Configurations for MRI Probes," cited above.

Figs. I B, 1C, and ID show modified magnet assemblies, with direction of magnetization that varies from place to place, which produce magnetic fields that have different dependence on φ than the field produced by magnet assembly 100. In Fig. IB, magnet assembly 106 has a portion 108 at y < 0, which has a positive y component and negative x component of magnetization, and a portion 1 10 at y > 0, which has a positive y component and positive x component of magnetization. If the y components of magnetization are somewhat greater than the x components, and if the magnitude of the magnetization in each portion of assembly 106 is the same as in assembly 100, being determined by the magnet material, then the magnetic field produced by assembly 106 will also be predominantly a dipole field, with average field strength, and fall off in field strength with r, almost the same as in the dipole field of assembly 100. But opposing x components of magnetization in portions 108 and 1 10 of assembly 106 will also produce a substantial quadrupole field, which will add to the dipole field, increasing the field strength, in the +x direction, and subtract from the dipole field, decreasing the field strength, in the - x direction. If the imaging region is confined to a moderately narrow range of angles around the x-axis in the +x-direction, then the field in the imaging region of assembly 106 will be stronger than the field in the imaging region of assembly 100, especially close to the magnet. Fig. 1C shows a magnet assembly 112 where the situation is reversed, with a portion 114 at y < 0 having positive y and positive x components and magnetization, and a portion 116 at y > 0 having positive y and negative x components of magnetization. In the imaging region of assembly 112, the field strength is somewhat lower than in the imaging region of assembly 100, especially close to the magnet.

Fig. ID shows a magnet assembly 1 18 with four portions. In portion 120, at y < 0 and x > 0, the magnetization has positive y component and negative x component, and in portion 122, at y > 0 and x > 0, the magnetization has positive y component and positive x component, as in assembly 106. There is also a portion 124, at y < 0 and x < 0, with positive y component and positive x component, and a portion 126, at y > 0 and x < 0, with positive y component and negative x component. The y components of magnetization produce a dipole field not much weaker than that of assembly 100, if the y components are somewhat greater than the x components. The x components of magnetization produce an octupole field, which falls off much faster with r than the dipole field, but which adds to the dipole field, raising the field strength, in the +x and -x directions, and subtracts from the dipole fields, lowering the field strength, in the +y and -y directions. If the angles of magnetization are chosen properly in assembly 1 18, then this azimuthai dependence of field strength of the octupole field will nearly cancel the azimuthai dependence of field strength of the dipole field, at least for some range of r, resulting in a field strength that is nearly independent of azimuthai angle φ. For example, if the magnet assembly has length about twice its diameter, and if the directions of magnetization are about 25 degrees from the y- direction, then the field strength will be nearly independent of φ for parts of the imaging region closest to the magnet assembly.

Fig. IE shows a magnet assembly 128, comprising three parts, a center part 130, around z = 0, and end parts 132 and 134 at z > 0 and z < 0 respectively. Parts 130, 132 and 134 are schematically shown blown apart in Fig. IE, so that they can be seen more clearly, but in fact they are joined together, either as separate magnets or as parts of one magnet with non-uniform direction of magnetization. The imaging region is in the +x-direction, but RF coil 104 is not shown, for clarity. Center part 130 has portions 114 and 116, resembling assembly 112, and consequently has decreased field strength in the +x-direction, in the imaging region. End parts 132 and 134 each have portions 108 and 110, resembling assembly 106, and consequently have increased field strength in the +x-direction, in the imaging region. The z-dependence of these field changes tends to cancel out the z-dependence of the dipole field in the imaging region, which is strongest, for a given r and φ, at z = 0, and falls off toward the ends of the assembly. For a right choice of angle of magnetization to the y-direction, and a given ratio of length to diameter of assembly 128, the z dependence of the field strength may nearly cancel out, in the imaging region over some range of r. Near the ends of the magnet, the field falls off more rapidly, in a configuration like assembly 128, than in a configuration with magnetization direction independent of z, such as assembly 106.

The optimal angles of magnetization for the different portions of assembly 128 may depend on the ratio of length to diameter of the magnet. In some embodiments of the invention, the probe is elongated in the z-direction, with greater length in the z-direction than diameter in the x and y directions, for example at least 20% greater, or at least 50% greater, or at least twice as great.

Optionally, magnet assembly 128, and other magnet assemblies described here with three slices, have four or more slices. This option applies, for example, to the magnet configurations shown in Fig. IF, and in Fig. 3, below. Optionally, the directions of magnetization, and the sizes and/or shapes of portions having different directions of magnetization in a given slice, change gradually in going from an end slice to a center slice, with intermediate slices having intermediate directions of magnetization or intermediate distributions of direction of magnetization. Although a magnet assembly having fewer slices may be easier to assemble, a magnet assembly having a greater number of slices may be better optimized with respect to a design criterion, because it has more degrees of freedom in its design. Fig. IF shows a magnet assembly 136 that combines features of assembly 118 and assembly 128, to produce a magnetic field in the imaging region that is nearly independent of both z and φ, for a given r, over some range of r in the imaging region. Assembly 136 has end parts 132 and 134 that resemble those of assembly 128, and a center part 138. Center part 138 has a pattern of magnetization that is intermediate between center part 130 in Fig. IE, and assembly 118 in Fig. ID. Center part 138 comprises four portions, 140, 142, 144, and 146, which resemble portions 120, 124, 126, and 122 of assembly 118. However, in constrast to assembly 118 where the four parts have the same size, in center part 138 parts 140 and 142, which resemble portions 114 and 1 16 in center part 130, are substantially bigger than portions 144 and 146, which have an effect opposite to that of portions 1 14 and 1 16. As a consequence, the x components of magnetization in assembly 136 create a field which is lower in the +y and -y directions than in the +x and -x directions, at least partially canceling the φ dependence of the dipole field at a given r and z. And the field created by the x components of magnetization of assembly 136 is also weaker at z = 0 than at higher and lower z, in and around the +x-direction, which is the direction of the imaging region, at least partially canceling the z dependence of the dipole field at a given r and φ. Magnetic assembly 136 is similar to the magnet assembly used in the MRI probe shown in Fig. 2, as will be described below.

Figs. IG and I H illustrate two alternative ways of manufacturing a magnet assembly such as assembly 1 12, or any of the magnet assemblies shown in Figs. IB- IF. In Fig. IG, assembly 148 comprises two separate magnets 150 and 152, each magnet corresponding to one of the two portions 1 14 and 1 16. Magnets 150 are optionally magnetized first, before assembly 148 is assembled, and then joined together, for example by bonding. Depending on the relative angles of magnetization direction and orientation of the boundaries of adjacent magnets, the magnet forces may tend to pull them together, push them apart, or twist them, but the bonding is preferably strong enough to keep the magnets in place. With modern "hard" magnetic materials, such as NdFeB, there will be little tendency for adjacent magnets to demagnetize each other.

In Fig. IH, there is a single magnet assembly 154 with non-uniform direction of magnetization, with its left and right halves corresponding to portions 114 and 116. Assembly 154 is made, for example, but taking an unmagnetized cylinder of a hard magnetic material such as NdFeB, and applying a non-uniform external field to it, strong enough to magnetize the material. For example, high field coils could be applied to assembly 154, at approximately 10 o'clock and 2 o'clock locations on the circumference of the cylinder, producing a magnetic field, and hence magnetization, of a pattern resembling that shown by the arrows on the top surface of assembly 154. The direction of magnetization of assembly 154 may vary smoothly with position, rather than changing abruptly at boundaries between magnets as in assembly 148, but, particularly if the direction of magnetization does not differ by a very large angle from one part of the magnet to another, this is not likely to have a significant effect on the magnetic field profile in the imaging region. Although manufacturing magnet assembly 154 in this way will avoid the need to cut and assemble different magnets, it may be expensive to set up a magnetization fixture with non-uniform magnetic field, which may not be something that sellers of magnets normally do, and may require a special custom job. It also may be difficult to do this for small magnets, since small coils may not be able to carry enough current without overheating.

Figs. II, IJ, and IK illustrate magnet assemblies comprising separate magnets bonded together, but with one or more extra magnets, in addition to the magnets that correspond to each portion in Fig. 1C. The extra magnets are intermediate magnets, located between the magnets that correspond to different portions in Fig. 1C, and with intermediate direction of magnetization. In Fig. I I, intermediate magnet 160 fills a parallel-slice shaped region between magnets 158 and 162, which correspond respectively to portions 1 14 and 1 16 of assembly 1 12. In Fig. U, intermediate magnet 168 fills a wedge-shaped region between magnets 166 and 170, and in Fig. IK, on the side of the assembly opposite the imaging region, and in Fig. I K, intermediate magnet 178 fills a wedge-shaped region between magnets 174 and 176, on the same side of the assembly as the imaging region. Having one or more intermediate magnets may improve the field profile in the imaging region, for the optimal design, since it provides one or more additional degrees of freedom in the design. In particular, the intermediate magnet may provide a larger y component of magnetization, in a region of the assembly where the x components of the magnets not contribute very much to the field profile in any case, thereby allowing a better tradeoff between the strength of the dipole magnetic field, and the degree of trimming of the magnetic field to reach certain goals, such as making the magnetic field independent of φ and z for a given r. Bending the direction of magnetization toward the imaging region, as in probe 1 12 in Fig. 1C, and bending the direction of magnetization away from the imaging region, as in probe 106 in Fig. IB, each has potential advantages for different applications. With probe 112, the imaging region can extend over a greater range of distance from the magnet, for a given range of resonance frequencies. The lower field gradient in probe 112 may allow the image to be acquired with either T₂ or diffusion weighting, for a probe scaled for prostate imaging, e.g. about 30 mm in diameter. The lower field gradient may also allow a lower bandwidth to be used for the same slice thickness, or a greater slice thickness for the same bandwidth, for example a bandwidth of 100 kHz, which may be advantageous for certain RF coil designs. On the other hand, the higher magnetic field found near probe 106 is also potentially advantageous for a prostate probe.

The higher magnetic field found near the probe, with probe 106, may be particularly advantageous for smaller probes, less than 5 mm in diameter, scaled for intravascular use. Higher field may be particularly useful for such small probes, because it results in higher SNR for the same voxel size and acquisition time, and SNR tends to be lower for voxels that are smaller in size, such as would be used for a small probe. For such a small probe, diffusion weighting may be dominant in any case. If an intravascular probe is located adjacent to the wall of an artery, then it may not be so important, for imaging plaque, for the imaging region to extend very far from the probe, relative to its diameter.

Fig. IL shows a magnet cross-section and field contours, in tesla, for a probe 234, which may be suitable for use as an intravascular MRl probe, and illustrates another potential advantage of having the direction of magnetization bend away from the imaging region. Probe 234 comprises two magnets 236 and 238, with direction of magnetization bending away from an imaging region 240, as in probe 106 in Fig. I B. The direction of magnetization in each of the magnets is about 30 degrees from the y direction, for example between 25 and 35 degrees, or between 15 and 45 degrees. The side of probe 234 facing the imaging region is convex, and this allows an extended area of the probe surface to be in contact, or nearly in contact, with the wall of a blood vessel, for example an artery, with the imaging region inside the wall, where the probe can detect plaque. Unlike the probes shown in Figs. 1A-1K, probe 234 has magnets that are much thinner in the x direction than in the y direction. For example, the magnets are between 0.2 and 0.25 mm thick, in the x direction, at the x-axis, and the two magnets together are between 1.25 and 1.35 mm wide in the y direction. This shape allows room for a balloon to be located between the magnets and a catheter, for example. The fact that the magnets are thinner in the x direction than in the y direction would tend to make the field on the x-axis, just outside the magnets in the imaging region, lower than if the magnets had approximately a circular cross-section, and also tends to make the gradient of the field lower in this region. Having the magnetization bend away from the imaging region tends to compensate for this, increasing both the field and the field gradient in this region, making the field more closely resemble that produced by a magnet with a circular cross-section, magnetized in the y direction.

Prostate Probe

Fig. 2 shows an exemplary self-contained MRI probe 200, designed for use as a rectal probe for imaging the prostate. Self-contained MRI probes of other sizes, and not necessarily to scale, may be used for other parts of the body, for example a probe of 2 mm or less in diameter is optionally used for intravascular MRI. Probe 200 comprises a magnet assembly 202, which is similar in design to magnet assembly 136 in Fig. IF. However, because it is difficult to see magnet assembly 202 in Fig. 2, behind other parts of the probe, the different parts of magnet assembly 202 are not labeled in Fig. 2, but are shown in Fig. 3, which shows only the magnet assembly. Optionally, the magnets in magnet assembly 202 are composed of magnetically hard permanent magnet material with high energy product, such as NdFeB, or any other such material known in the art. Such magnets have the potential advantages that they do not demagnetize easily, and that the magnetic field they produce in the imaging region will be about as strong as possible, for a given geometry of the probe and imaging region. As in Figs. 1A-1F, probe 200 has a longitudinal axis in the z-direction, an imaging region centered in the +x-direction, and magnet assembly 202 produces a magnetic field in the imaging region that is predominantly in the -y-direction, with magnets that are magnetized at angles that center around the +y-direction.

Probe 200 also comprises an RF coil 204, facing the imaging region in the +x- direction, which excites nuclei in the imaging region, and receives NMR signals from the excited nuclei. Optionally an RF shield, not shown in Fig. 2, made for example of aluminum foil, is located between the RF coil and the magnet assembly, optionally completely surrounding the magnet assembly, to prevent magneto-acoustic ringing in the magnet assembly from the RF fields. Optionally, a pair of phi-gradient coils 206, overlapping and to the sides of the RF coil, produce a phi-gradient field, optionally substantially parallel to the static magnetic field in the imaging region and anti-symmetric in φ, that provides phase encoding in the φ direction, which corresponds nearly to the y- direction in the imaging region. As used herein, "y-gradient" is synonymous with "phi- gradient" and "φ-gradient." Optionally, a set of three z-gradient coils 208 at the +z end of the probe, and a similar set of three z-gradient coils 210 at the -z end of the probe, produce a z-gradient field, optionally substantially parallel to the static magnetic field in the imaging region and anti-symmetric in z, that provides phase encoding in the z-direction.

A reference cylinder 212, 45 mm in diameter, is shown surrounding probe 200. Magnet assembly 202 is 30 mm in diameter in this design. Although phi-gradient coils 206 and RF coil 204 are shown just outside cylinder 212 in Fig. 2, with a radius of curvature of 45 mm, optionally coils 206 and/or coil 204 are located inside cylinder 212, closer to magnet assembly 202, or immediately adjacent to magnet assembly 202, with a radius of curvature 30 mm. Optionally, at least when the probe is operating, RF coil 204 and/or phi- gradient coils 206 are situated at some distance away from magnet assembly 202, whether outside or inside cylinder 212. This allows the phi-gradient coils and/or the RF coil to be closer to the prostate. Optionally, the phi-gradient coils and/or the RF coil are capable of spreading out against the rectal wall when the probe is doing MRI imaging, and to contract closer to the magnet assembly when the probe is inserted through the anus, where there is less room, into the rectum, and when the probe is removed. Means for implementing this feature are described in more detail in the related co-filed application "MRI Probe." In some embodiments of the invention, probe 200 comprises an ultrasound imaging module in addition to an MRI part of the probe. Space 214 in cylinder 212, behind magnet assembly 202, out to a radius of 30 mm, is optionally reserved for the ultrasound module. As described in the co-filed application "MRI Probe," optionally the probe is rotated, when switching between MRI and ultrasound imaging modes, bringing first one module and then the other to a part of the rectal wall adjacent to the prostate.

In some embodiments of the invention, the ultrasound module does not take up so much space, and part of region 214 may be used for additional magnet or coil volume, for the MRI part of the probe. The ultrasound module is, for example, between 25% and 35% of the volume of cylinder 212, or of a 30 mm diameter cylinder that includes magnet assembly 202, or between 35% and 45%, or less than 25%, or more than 45%. The ultrasound module is optionally shorter than the probe, and space on one or both sides of the ultrasound module in the +z or -z direction is taken up by magnet volume. Such an arrangement has the potential advantage that the greater magnet cross-section near the ends of the magnet assembly may make the magnetic field fall off more slowly with z moving away from z = 0, in the imaging region.

In these embodiments, the probe is rotated around the longitudinal axis, for example by 180 degrees, between using the MRI and ultrasound imaging modalities. Because the ultrasound module is on the opposite side of the probe from the MRI imaging region, using this volume for the ultrasound module rather than for the magnet has a relatively small effect on the magnetic field in the MRl imaging region, even if the ultrasound module takes up a relatively large fraction of the total volume of cylinder 212, of a 30 mm diameter cylinder, for example between 25% and 45% of the total volume of the cylinder. Using a relatively large volume to accommodate an ultrasound array has the potential advantage that an off-the-shelf ultrasound array could be used, not necessarily optimized to fit into a small volume, and/or that the ultrasound array could provide higher resolution, higher signal to noise ratio (SNR), and/or shorter acquisition time than a smaller ultrasound array could provide.

In another exemplary embodiment of the invention, a relatively small groove is carved into the side of the probe facing the MRI imaging region, and the groove accommodates an ultrasound array, smaller than the ultrasound array that could be accommodated in volume 214 in Fig. 2. The groove is, for example, between 5 mm and 10 mm wide, or less than 5 mm or greater than 10 mm wide, and between 3 mm and 10 mm deep, or less than 3 mm or greater than 10 mm deep. Because the groove is relatively small, removing the magnet from the volume of the groove also does not affect the magnetic field in the MRI imaging region very much. The configuration with an ultrasound module on the side of the probe facing the MRI imaging region has the potential advantage that MRI and ultrasound images could be acquired at the same time or at almost the same time, without the need to rotate the probe when changing imaging modality. In some embodiments of the invention, a similar groove, accommodating an ultrasound array of relatively small volume, is located on the side of the probe opposite the MRl imaging region, and the probe is rotated between acquiring MRI images and ultrasound images. Such a configuration has the potential advantage that the magnetic field in the MRI imaging region would be reduced even less than it would be in the configurations described above.

Fig. 3 shows magnet assembly 202 by itself. The magnets are Vacodym 722 NdFeB magnets, sold by Vacuumschmelze, in Germany. The assembly has a semi-circular cross- section 30 mm in diameter, and is 60 mm long. Like assembly 136 in Fig. IF, assembly 202 comprises a center part 202, and end parts 304, corresponding to center part 138 and end parts 132 and 134 in Fig. I F. One difference between assembly 202 and assembly 136 is that assembly 202 has a semi-circular cross-section, to allow room for the ultrasound module, while assembly 136 in Fig. IF is shown as having circular cross-section. However, the remarks made about the magnet design of assembly 136 still apply, to good approximation, to assembly 202. Center part 302 comprises four magnets 308, 310, 312, and 314, corresponding to portions 144, 146, 140 and 142 of central part 138 of assembly 136. End part 304 comprises two magnets 316 and 318, and end part 306 comprises two magnets 322 and 324, corresponding respectively to portions 108 and 1 10 in end parts 132 and 134 of assembly 136. Optionally, the direction of magnetization of the different magnets is about 25 degrees from the +y-direction, in the x-y plane, either toward the +x or -x direction, or at another angle, not necessarily the same angle for all of the magnets, between 20 and 30 degrees from the +y-direction, or between 15 and 35 degrees from the +y-direction. Optionally there is a groove 320 between magnets 316 and 318, for one of z- gradient coils 208, and a groove 326 between magnets 322 and 324, for one of z-gradient coils 210.

Alternative Designs for Z-Gradient Coil Set

Fig. 4A shows z-gradient coil set 208. The coil set comprises coils 402 and 404, substantially oriented in the y-z plane, and coil 406, substantially oriented in the x-z plane. The current in coils 402 and 404 runs in opposing directions. Near the z-axis of probe 202, all three coils 402, 404, and 406 have adjacent vertical elements, and the current in coils 402 and 404 goes in a direction opposing the current in coil 406, in these adjacent vertical elements. With this relation between the currents in the coils, the three coils all produce a z- gradient field which is predominantly in the y-direction in the imaging region, adding or subtracting from the static magnetic field, and which is symmetric with respect to the x-z plane. In some embodiments of the invention, only coil 406 is present, or only coils 402 and 404 are present, and in the latter case there need not be a groove 320 in magnet assembly 202 in Fig. 3. Having all three coils may provide higher field strength for the z- gradient field, and better linearity of the z-gradient field with z, and better independence of the z-gradient field from φ, for the whole imaging region, for a surface of a given static magnetic field strength. Partly this is because it is possible to optimize the ratio of current in coil 406 to current in coil 404, to achieve a high ratio of z-gradient field to z-gradient coil power, and to optimize the linearity of the z-gradient field with z, and to minimize its dependence on φ, in the imaging region.

Because the currents in the adjacent vertical element of coil 406 tends to cancel out the current in the adjacent vertical element of coils 402 and 404, it is possible to replace coil set 208 with an alternative coil set 408, shown in Fig. 4B. In coil set 408, there are two coils 410 and 412, each coil bent at 90 degrees so that half of it is oriented in the x-z plane and half of it is oriented in the y-z plane. If the two coils have equal currents in the directions shown, then this would produce the same z-gradient field as coil set 208 would produce, if the current in coil 406 were twice the current in each of coils 402 and 404. Other combinations of coils will also produce this z-gradient field, for example coils 410 and 404 with equal currents, or coils 402 and 412 with equal currents. These coil sets have the potential advantage over coil set 208, that two coils do not have opposing currents in the adjacent vertical elements, so that ohmic losses can be lower for the same z-gradient field. But coil set 208 has the potential advantage that the relative currents in coil 406 on the one hand, and coils 402 and 404 on the other hand, can be controlled independently, potentially allowing better optimization of the z-gradient field.

Fig. 4C schematically shows a z-gradient field 414 produced by coil set 208, for a particular ratio of current in coil 406 to current in coil 402. Keeping the current in coil 402 equal to the current in coil 404 keeps the z-gradient field symmetric about the x-z plane. Coil set 408 would produce a similar field. Flux lines 416 are oriented approximately along the y-axis in imaging region 418. This is approximately the same direction as the static magnetic field in the imaging region, so the two fields add up or subtract.

The options described for coil set 208 apply as well to coil set 210, at the other end of the probe. Optionally, coil set 210 and coil set 208, or any alternative coil sets used instead, are mirror symmetric from each other about the x-y plane, and have corresponding currents going in opposite directions, so that the z-gradient field is anti-symmetric in z.

Static Magnetic Field Profile Figs. 5, 6, and 7 show plots of the static magnetic field produced by magnet assembly 202, particularly in the imaging region. Fig. 5 shows a plot 500 of static magnetic field strength vs. distance from surface of magnet assembly 202, along the x-axis in the +x direction, which is the center of the imaging region. The solid line shows the field calculated using magnetic finite element software, and the dots show experimental data from measurements of the field using a Hall probe. The plot shows the static magnetic field strength from a distance of 5 mm to a distance of 20 mm from the edge of the probe, which is a little broader than the imaging region. For the RF coil and phi-gradient coils pushed against the rectal wall, this imaging region, which also extends from y = +25 mm to -25 mm, covers most of the peripheral zone of the prostate, where most cancers occur and where needle biopsies are performed. The field varies from 0.21 tesla down to 0.045 tesla in this region, and the imaging region is considered to range from 0.20 tesla down to 0.048 tesla, a factor of 4.2. As will be described below, the RF coil is capable of transmitting power and receiving signals over the corresponding range of nuclear magnetic resonance frequencies, ranging from 8.5 MHz down to 2.0 MHz. Alternatively, the imaging region extends over a different range of the magnetic field. For example, the maximum field is more than 0.2 tesla, or between 0.2 and 0.15 tesla, or between 0.15 and 0.1 tesla, for example 0.13 tesla, or less than 0.10 tesla. The minimum field is, for example, more than 0.07 tesla, or between 0.07 and 0.05 tesla, or between 0.05 tesla and 0.03 tesla, for example 0.04 tesla, or less than 0.03 tesla. Optionally, the imaging region extends over a different range of y, for example from +15 mm to -15 mm, or from +20 mm to -20 mm, or from +30 mm to -30 mm, or from +35 mm to -35 mm, or over a greater, smaller, or intermediate range of values of y. not necessarily symmetric about y = 0, and not necessarily the same range of y for each imaging slice. For example, for more distant imaging slices at lower resonance frequencies, where the phi-gradient fields are nearly linear over a greater range of y, the imaging region optionally extends over a greater range of y. Similarly, the imaging region may extend over various ranges of z, including any of the ranges listed above for y, and the range need not be symmetric around z = 0, and may be different for different imaging slices, for example a greater range for z for lower resonance frequencies. The relatively high field gradient shown in Fig. 5 allows diffusion weighted contrast to be used, if desired, as described below in the description of Fig. 15A. Diffusion weighted contrast may be useful for distinguishing cancerous from normal tissue.

Fig. 6 shows a contour plot 600 of the contours of static magnetic field strength in the x-y plane, at z = 0. The contours shown in plot 600 range from 0.045 tesla, at contour 602, up to 0.21 tesla, at contour 604. The contours are spaced at intervals of 0.015 tesla. Magnets 308, 310, 312, and 314, gradient coils 206, RF coil 204, and reference cylinder 212 are shown superimposed on plot 600. It should be noted that the contours lines, which represent surfaces of constant field strength, very closely follow surfaces of constant r, as they were designed to do, in the imaging region, which extends below RF coil 204 in Fig. 6. Fig. 7 shows a contour plot 700 of the contours of static magnetic field strength in the imaging region in the x-z plane, at y = 0, over the full 60 mm length of the magnet assembly in z. Side views of upper z-gradient coils 402 and 406, lower z-gradient coils 210, one of the phi-gradient coils 206, RF coil 204, and magnets 308, 312, 316 and 322, are shown to the left of the imaging region. The contours shown in plot 700 extend from contour 702 corresponding to 0.045 tesla, furthest from the magnet assembly, to contour 704 corresponding to 0.225 tesla, closest to the magnet assembly, at intervals of 0.015 tesla. It should be noted that the contours, which represent surfaces of constant field strength, correspond closely to surfaces of constant r, independent of z, as they are designed to do.

RF Coils and Fields

Probe 200 uses separate RF transmitting and receiving coils, packaged close together iaRF coil 204. Optionally, the same RF coil is used for transmitting and receiving. A potential advantage of using different transmitting and receiving coils is that the two coils can be tuned separately, and by different methods. It has been found by the inventors that, at least for this probe design, different methods of tuning are best used for the transmitting and receiving coils, as will be described below. Specifically, varactors have been found useful for tuning receiving coils, but may not be as useful for tuning high power RF transmitting coils, because high voltage can change their capacitance. The RF coil design used in probe 200 has the following potential advantageous characteristics: 1) Instantaneous RF power transmission of up to 2 kW, without arcing; 2) RF magnetic field substantially perpendicular to static magnetic field throughout the imaging region; 3) Fall off in transmitted RF magnetic field of about a factor of 4 over the imaging region, and fairly independent of z and φ; 4) Q of about 40, matching the bandwidth of the power supply; 5) Transmitting and receiving coils both electronically tunable from 2 to 8.5 MHz.

To achieve a fall-off in RF magnetic field of only a factor of 4 over the imaging region, ranging from 5 to 20 mm from the edge of the probe, RF coil 204 extends over 36 mm in z, and over 28 mm in φ, is curved with a radius of 22.5 mm, comparable to a typical radius of curvature of the inner surface of the rectal wall, and is located just outside the probe edge, centered on the x-axis, facing the imaging region. Fig. 8 shows a plot 8 of the RF field amplitude Bl as a function of distance from the probe edge, for 1 amp of current in the transmitting coil. Bl falls from about 0.96 x 10^"3 tesla at 5 mm from the probe edge, to 0.24 x 10^"3 tesla at 20 mm from the probe edge. Fig. 9 shows a contour plot 1900 of the RF field amplitude in the x-y plane at z = 0. The contours are labeled in units of 10^"3 tesla. Magnet assembly 202 is shown superimposed on the contour plot, as well as cross-sections 1902 of the transmission coil, where Bl peaks, and cross-sections 1904 of the receiving coil, where Bl falls locally to zero because of skin effects. Fig. 10 shows a similar contour plot 1000 for the x-z plane at y = 0. Over most of the imaging region, Bl is fairly independent of z and φ for a fixed r, which also corresponds closely to a fixed static magnetic field strength.

To achieve a relatively high Q, and the ability to transmit relatively high power without arcing, the RF transmitting coil uses four sub-sections in series, interspersed by tuning capacitors to make an LC circuit. This idea, described by B. Cook, and I. J. Lowe, "A Large-Inductance, High-Frequency, High-Q, Series-Tuned Coil for NMR," Journal of Magnetic Resonance 1982;49(2):346-349, has been used in RF coils for conventional MRI, because it reduces arcing, and reduces stray capacitance, thereby allowing higher Q. The sub-sections each consist of 8 turns of 0.2 mm diameter Litz wire (with 30 parallel strands each 0.02 mm in diameter). Each subsection has an inductance of about 19 μH, including mutual inductance with other sub-sections, and the total inductance of the transmitting coil is 77 μH. Since the coil has an AC resistance of about 50 ohms, Q is about 40 at 4 MHz. Tuning is accomplished by changing the capacitance of the tuning capacitors. Optionally, coarse tuning is done by switching a series of fixed capacitors of different values in and out of the circuit, for example using mechanical relays. The fixed capacitors interspersed between the four sub-sections of the coil range from values of 335, 326, 316 and 326 pF, respectively, for operation at about 2 MHz, to values of 37, 36, 35, and 36 pF, respectively, for operation at about 5 MHz. Optionally, fine tuning is done using varactors. However, it was found by the inventors that varactors did not keep a constant capacitance when they were used with high RF voltages. In order to avoid having to switch too many different sets of capacitors in and out of the circuit, the Q of the circuit is optionally deliberately spoiled, to allow the circuit to be driven at higher bandwidth by the RF amplifier. Spoiling of Q is done, for example, by using an autotransformer, as shown in Fig. 1 1. Circuit 1100 has an RF power supply 1102 in series with a resistor 1 104 of 50 ohms, representing the output impedance of the power supply. An autotransformer 1106, with one side of the primary and secondary coils grounded together, has a turn ratio of 1:4, with the larger number of turns on the side of the coil. The coil is modeled by a resistor 1108 with resistance R, in series with a capactor 1110 with capacitance C, and an inductor 11 12 with inductance L As seen by the RF power supply in equivalent circuit 1114, R, L, and C are all divided by 16, the square of the turn ratio. Since R is about 50 ohms, the equivalent circuit has Q reduced by a factor of 16 to about 2.5, and the full range of driving frequencies from 2 to 8.5 MHz can be covered with only 2 to 4 different sets of capacitors. The RF power supply can run up to 10 amps of peak current through the circuit, corresponding to 2 kW power. At 2 MHz, corresponding to the most distance part of the imaging region, the bandwidth is about 50 kHz, and at 8.5 MHz, corresponding to the closest part of the imaging region, the bandwidth is about 250 kHz.

The receiver coil optionally has windings similar to those of the transmitting coil, as well as capacitors interspersed between the sub-sections, with different sets of capacitors switched in and out of the circuit by mechanical relays, for coarse tuning. But unlike the transmitting coil, the receiving coil, which operates at very low power, uses varactors for fine tuning, so there is no need to spoil Q, which remains about 40. The receiving coil used capacitors of 299, 144, 140, and 144 pF respectively, and a varactor with varying capacitance of - 750-130 pF, for frequencies between 2 and 4 MHz, and capacitors of 74, 36, 35, and 36 pF respectively, and a varactor with varying capacitance of ~ 750-60 pF, at frequencies between 4 and 6 MHz. In order to avoid coupling between the transmitting and receiving coils, which would induce an RF voltage across the varactor during transmission of an RF pulse, as well as saturating the low noise amplifier of the receiver causing a long dead time, the receiving coil is in series with two PIN diode switches, one at each end. The switches are used to disconnect the receiving coil during transmission. In some embodiments of the invention, active toroidal protectors are used to disconnect the receiving coil from the low noise amplifier during transmission, in addition to or instead of PIN diodes, as described, for example, in WO 2006/043275.

In some embodiments of the invention, the receiving and transmitting RF coils, or a single RF coil that does both, is divided into two halves, each half on a different side of the x-z plane. In this configuration, if current flows in the same direction (clockwise or counter-clockwise) in both halves at a given RF phase, then the RF fields will look very similar to the fields from a single coil, except close to the x-z plane and close to the antennas. This configuration has the potential advantage that the two halves could be hinged, allowing them to open up once the probe is in position in the rectum, while each half could be relatively rigid. But the RF coil configuration shown in Fig. 2 has the potential advantage that there are no parts of the coila where current is flowing in the z- direction near the x-z plane, which would dissipate heat while contributing very little to the RF field in the imaging region.

Gradient Fields and Coils

The six z-gradient coils 208 and 210 are optionally each a rectangular coil made of 0.35mm diameter copper magnet wire, with 40 windings around a 13.6x7x2 mm rectangular core. Optionally, coils 208 have their centers at z = +25 mm, and coils 210 have their centers at z = -25 mm. Optionally, coils 208 are connected in series, coils 210 are connected in series, and coils 208 are then connected in parallel, anti-symmetrically, to coils 210, and have, for example, total inductance of 22.6 μH and resistance of 0.8 Ω.. In this arrangement, the NI through each of the six coils is equal in magnitude. Other arrangements are optionally made if different NI is to be used in different coils within each set 208 and 210, for example driving different coils separately at different voltages, or driving two or more coils in series but having different numbers of turns in different coils. Optionally, to make the z-gradient field anti-symmetric with respect to z and symmetric with respect to y, the four coils in the y-z plane all have the same NI, and the two coils in the x-z plane both have the same NI. The predominant component of the z-gradient fields, like the static magnetic fields, is generally the φ component (or y component) in the imaging region, so the z-gradient fields add to or subtract from the static magnetic fields in the imaging region.

Fig. 12 shows a plot 1200 of the component of the z-gradient field that is parallel to the static magnetic field, as a function of z, in the x-z plane, for four different contours of constant static magnetic field. Curves 1202, 1204, 1206 and 1208 are respectively for the constant fields resonant at 2 MHz (0.048 tesla), 3.5 MHz (0.084 tesla), 5 MHz (0.12 tesla), and 6.5 MHz (0.156 tesla). As may be seen in Fig. 7, these contours of constant static magnetic field are nearly lines of constant r, except near the ends of the imaging region in z, where the contours start to curve slightly toward the probe. The z-gradient fields shown in Fig. 12 are fairly linear functions of z, and are monotonic everywhere in the imaging region, so they may be used for unambiguous phase-encoding. Away from φ = 0° (the x-z plane), the z-gradient fields are fairly independent of φ, except near the edges of the imaging region, close to the probe, where they can be as much as 50% lower than at φ = 0°.

The two phi-gradient coils 206 are optionally located with their centers at ±65° from the x-axis, at a radius of 45 mm, when they are deployed. Optionally, they are located closer to the probe when it is inserted, and expand away from the probe, closer to the rectal wall, when the probe is in use. Possible mechanisms for accomplishing this are described in the co-filed application "MRI Probe." Each coil is optionally made of 0.7mm diameter copper wire, with 28 windings (2 layers of 14 windings each) around a 32x 16x2 mm rectangular core. Each coil has a height of approximately 42 mm in the z direction and 26 mm in the φ direction, and is about 1.4 mm thick radially. In practice, these dimensions may be a little bigger if the windings are not perfectly packed. If the two coils are connected in parallel, they have a total inductance of 16.2 μH and resistance of 0.15 Ω. Whether they are wired in parallel or in series, the two coils are wired so that the phi- gradient field they produce has a φ component that is anti-symmetric in φ. Like the static magnetic field, the predominant component of the phi-gradient field in the imaging region is the φ or y component, so that the phi-gradient field adds to or subtracts from the strength of the static magnetic field. In some embodiments of the invention, the φ-gradient coils do not extend as far in the z-direction as coils 206 do in Fig. 2, and/or the RF coils do not extend as far in the z- direction as coil 204 does in Fig. 2. For example, the coils extend only 12 mm in the z- direction, or only 16 mm, or 28 mm, or 40 mm. A potential advantage of having the coils extend a shorter distance in the z-direction is that the imaging region will be more localized in z, at least close to the probe, and the image may be resolved in z by moving the probe to different locations in z and repeating the imaging procedure, rather than by using z-gradient coils. However, it has been found that the imaging region extends further in z, moving further away from the probe, even if the coils do not extend very far in z, so using such short coils does not provide as much resolution in z, over much of the imaging region further away from the coil. Using φ-gradient and RF coils that extend further in z, as shown in Fig. 2, or even further, has the potential advantage that an extended range of z can be imagined simultaneously, even close to the probe. Such coils also have the potential advantage that the φ-gradient fields do not decrease as much with increasing distance from the probe, and the RF fields and receiving sensitivity do not decrease as much with increasing distance from the probe, as they would if the coils did not extend as far in z.

Fig. 13 shows a plot of the component of the phi-gradient field that is parallel to the static magnetic field, as a function of distance along different contours of constant static magnet field strength in the x-y plane. Curves 1302, 1304, 1306 and 1308 are respectively for the constant fields resonant at 2 MHz (0.048 tesla), 3.5 MHz (0.084 tesla), 5 MHz (0.12 tesla), and 6.5 MHz (0.156 tesla). These contours are approximately arcs of constant r, as may be seen in Fig. 6, so the distance along the contour is nearly a linear function of φ, for any given contour. The phi-gradient fields are moderately linear functions of distance along the contours, especially close to the x-axis, though not as linear as the z-gradient fields plotted in Fig. 12. The phi-gradient fields are at least monotonic functions of distance along the contours everywhere in the imaging region, so provide unambiguous phase encoding for MRI imaging. Away from the x-y plane, at higher and lower values of z, the phi- gradient field falls off somewhat, and is as much as 28% weaker near the limits of the imaging region in z. The co-filed patent application titled "Image Reconstruction Methods for MRI Probes," describes methods of image reconstruction that take into account nonlinearities in the gradient fields as a function of position on the surfaces of constant static magnetic field. These methods may be suitable for use with probe 200. Table 1 shows NMR characteristics of the slices excited by the RF transmission coil and sensed by the RF receiving coil described above, at 2, 3.5, 5, and 6.5 MHz, using the phi-gradient and z-gradient coils described above, as calculated with NMR simulation software. The field gradients shown are for 50 amps of current in the form of a 30 microsecond long half-sin gradient pulse, which is optionally the highest total current used for driving the phi-gradient coils and for driving the z-gradient coils. The SNR shown is for a single shot, with a noise figure of 3 dB, and can be reduced by averaging over multiple shots. Optionally this is done at least for frequencies below 5 MHz, to bring the SNR up to adequate levels, for example at least 3 or 4. The RF transmission coil current is for the nominally 180° refocusing pulses. The distance is the distance of the center of the excited volume from the edge of the probe.

Table 1. NMR characteristics of slices excited at different frequencies

Although the voxel size is up to several mm in the r, z and φ directions, the resolution may be sufficient for detecting prostate tumors of 5 mm diameter or larger, for example, even in the part of the imaging region most distant from the probe. The relatively large voxel size may make it easier to implement image reconstruction methods, such as some of those described in the co-filed application, "Image Reconstruction Methods for MRI Probes," which become very computationally intensive with a large number of voxels. It should be understood that, in some embodiments of the invention, methods other than gradient encoding are used for resolving an image in the r, φ, and z directions. Particularly when the probe is very small, for example an intravascular probe, use of gradient coils may be difficult, because the currents required to produce the gradient fields would produce too much ohmic heating, and/or because larger voxels are needed, relative to the probe diameter, to provide adequate SNR in reasonable acquisition times, and/or because there is not enough room for gradient coils. In these cases, resolution of the image in φ is optionally provided by rotating the probe, resolution in z is optionally provided by moving the probe longitudinally, and resolution in r is optionally provided only by using different central frequencies for the transmitted RF pulses. Another method of providing resolution in φ or z is to have a plurality of RF coils centered at different values of z or different azimuthal positions, and to record data separately from each coil. These methods are described, for intravascular MRI probes, in the patents and published applications cited above. In all these cases, deconvolution is optionally used to improve the resolution of the image. Improved resolution is also optionally obtained, possibly at some cost in SNR, by using MRI pulse sequences for which the signal is sensitive to flip angle, so that the signal comes only from a limited spatial region where the RF field is fairly uniform.

Driving Circuit for Gradient Coils Fig. 14 shows a circuit diagram for a circuit 1400 that is optionally used to drive the gradient coils, according to an exemplary embodiment of the invention. A dc voltage source 1402 with voltage V is connected to a capacitor 1406 with capacitance C, by closing a first switch 1404, while a second switch 1408 remains open, thus charging capacitor 1406 to voltage V. A MOFSET high voltage switch, for example with a transition time on the order of 10 nanoseconds, may be suitable to use for switches 1404 and 1408. When capacitor 1406 has essentially reached voltage V, first switch 1404 is opened, and second switch 1408 is quickly closed. Capacitor 1406 then forms an LC circuit with gradient coil or coils 1410, which may be modeled as an inductance L in series with a resistance R, and this RLC circuit is isolated from voltage source 1402. The current I across gradient coil 1410 would then oscillate sinusoidally in time at the frequency (LC)^"l/2, assuming that R is much less than (L/C)^1/2 so that the LC circuit is not strongly damped, starting with a current of zero when second switch 1408 is closed. After half of the oscillation period, the current across coil 1410 reaches zero again. Second switch 1408 is then quickly opened, and the current across coil 1410 remains zero. Circuit 1400 thus produces a current across the gradient coil that is a sine wave for half of a wave period, as a function to time, and is zero before and after this "half-sine" interval T_hS.

The capacitance C, and the inductance L (which depends on the number of turns N and on the gradient coil dimensions), are optionally chosen so that a desired duration of the gradient coil current, T_hs, is equal to π*J^~LC , and so that the impedance V/I has a convenient value, for example matching the impedance of an available power source, given

by , when Nl, the number of turns N times the current I, is

sufficient to produce a desired value of the field gradient in the imaging region. Some considerations that may enter into the choice of T_hs and the field gradient will be described below, in the description of Figs. 15A-B.

MRI Pulse Sequences

Fig. 15A schematically shows a pulse sequence 900 used for MRI imaging, according to an exemplary embodiment of the invention, based on a modified form of a CPMG pulse sequence. An excitation RF pulse 902 with a nominally 90 degree flip angle is applied at the beginning, to tip the spins of the resonant nuclei from the z direction to a direction in the x-y plane, say the x-direction, where they start to spread in azimuth. RF pulse 902 is followed by a gradient pulse 904 of duration T_hS, which provides gradient encoding of the nuclei, in the y and/or the z direction. This is followed by a refocusing RF pulse 906 with a nominally 180 degree flip angle, at an interval τ after excitation RF pulse 902. The refocusing pulse produces an echo 908, centered around a time τ after the center of RF pulse 906. A long train of further refocusing RF pulses 910, all with nominally 180 degree flip angle, follows refocusing pulse 906, at intervals of 2τ, continuing for a time T, producing a long train of echoes 912, each centered half-way between two refocusing pulses. A pulse train may, for example, have between a hundred and two thousand refocusing pulses and echoes, or a smaller or larger number.

All of the RF pulses are centered at a frequency fi, and have a bandwidth Af₁, which may be approximately equal to the inverse of the pulse length. The pulses excite nuclei located in a slice where the magnetic resonance frequency is within a bandwidth Δfj of frequency fi. For example, if the RF pulse length is 10 microseconds, then the bandwidth may be approximately 100 kHz. In some embodiments of the invention, however, shaped or composite pulses are used, for example a "chirped" pulse with RF frequency changing during the pulse, and in this case the bandwidth may be substantially greater than the inverse of the pulse length. The interval τ is optionally a few times greater than the RF pulse length, for example 40 or 50 microseconds, and T_hs is optionally only slightly less than τ. Successive echoes 912 decrease in amplitude on a time scale T_c, which may be the NMR coherence time T₂ or the diffusion time T_d, whichever is shorter. The interval τ can be chosen to produce T_c that is either dominated by diffusion or T₂. For sufficiently short x, the dominant relaxation mechanism may be T₂, while for longer τ, diffusion may be the dominant relaxation mechanism. Images with T₂ weighting and images with diffusion weighting may reveal different information about the tissues.

After a time T, a new pulse train optionally starts, with the RF pulses centered at a different magnetic resonance frequency f₂, with bandwidth optionally not overlapping the bandwidth of the first pulse train. This pulse train excites nuclei located in a different slice, which were not excited by the first pulse train. Like the first pulse train, the second pulse train starts with a tipping RF pulse 914 of nominally 90 degree flip angle, a gradient pulse 916, and a train of refocusing RF pulses 918, of nominally 180 degree flip angle, resulting in a train of echoes 920. Optionally, this pulse train is followed by one or more additional pulse trains with RF pulses centered at different resonance frequencies, until the entire range of magnetic fields within the imaging region has been covered, for example from 0.2 tesla to 0.05 tesla in the case of the imaging region for probe 200 in Fig. 2.

Optionally, a new set 922 of pulse trains, with the same resonance frequencies as the first set of pulse trains, is initiated after a time interval T_R. T_R may be comparable to the NMR recovery time T₁, or several times greater than Ti, depending on whether or not T₍ weighting is desired in the image. Optionally, in set 922 the y-gradient pulse and/or z- gradient pulse has a different amplitude than in the original set of pulse trains, and set 922 is followed, at further intervals T_R, by additional sets of pulse trains, each with a different amplitude of the y-gradient pulse and/or z-gradient pulse. The data from each of these sets of pulse trains optionally corresponds to a different k_y and/or k_z, which may be used to reconstruct a 1-D image in y or z, or a 2-D image in y and z, for each slice. Alternatively, a more sophisticated reconstruction algorithm, which takes into account the nonlinearity of the magnetic field as a function of y and/or z, is used to reconstruct the image in y and/or z, as described in the co-filed patent application titled "Image Reconstruction Methods for MRI Probes."

In a conventional CPMG pulse sequence, a gradient pulse is applied after each refocusing RF pulse, before the echo, and Ά gradient pulse of the opposite sign is applied after the echo, before the next refocusing RF pulse, so that there is no net gradient encoding at the time of each refocusing Rf pulse. However, this procedure may be impractical in the case of a small MRI probe, with relatively small, high resistance gradient coils that have limited ability to dissipate heat in the short intervals between refocusing RF pulses. Optionally, the gradient pulses are applied only once near the beginning of each pulse train. The entire pulse sequence is then repeated, using a different phase difference between the tipping RF pulse and the refocusing RF pulses. For example, initially the refocusing pulses are 90 degrees out of phase with the tipping pulses, but when the entire pulse sequence is repeated, the refocusing pulses are in phase with the tipping pulses. As described in the co- filed patent application titled, "Image Reconstruction Methods for MRI Probes," the data from the two pulse sequences can be combined to produce an image. A similar method is described in US 5,493,225 to Crowley, the disclosure of which is incorporated herein by reference.

Fig. 15B shows an alternative to the pulse trains shown in Fig. 15A, according to an exemplary embodiment of the invention. Instead of putting the gradient pulse within the short time τ between the excitation (nominally 90°) RF pulse and the first refocusing (nominally 180°) RF pulse, there is a longer interval ti between an excitation RF pulse 926, with a nominally 90 degree flip angle, and a first refocusing RF pulse 930, with a nominally 180 degree flip angle. RF pulse 930 produces an echo 934 at a further time interval T₁ following RF pulse 930. Optionally, there is a gradient pulse 928 between tipping RF pulse 926 and refocusing RF pulse 930, and there is a gradient pulse 932, with gradient in a different direction, between refocusing RF pulse 930 and echo 934. For example, gradient pulse 928 is a y-gradient pulse, and gradient pulse 932 is a z-gradient pulse, or vice versa. Alternatively, only one of these gradient pulses is present, for example if imaging is only being done in x and y, or in x and z. Having the y-gradient and z-gradient pulses at different times, and using fast switches to open the circuit for one gradient pulse when the other gradient pulse is being applied, has the potential advantage that it may not be necessary to take into account mutual inductance between the y-gradient coils and the z- gradient coils, when driving them. Spreading each gradient pulse over a longer time has the potential advantages that it may be possible to use simpler electronics, and that lower current can be used to produce the same amount of gradient encoding, and less heat is dissipated. For the same gradient coils and the same amount of heat dissipation, greater gradient encoding can be achieved, resulting in greater resolution in y and/or z.

Following echo 934, there is a second refocusing RF pulse 936 at a time interval τ later, which produces a second echo 938 at a further time interval τ. There follow further refocusing pulses 940, 944, ..., separated by time intervals 2τ, with echoes 942, ..., between them, extending for a time T. This part of the pulse train is the same as the CPMG pulse trains shown in Fig. 15A, but with echo 934 playing the role of initial excitation RF pulse 902.

Time interval t| is, for example, between 100 and 200 microseconds, or between 200 and 300 microseconds, or longer than 300 or shorter than 100, while time interval τ in Fig. 15A or in Fig. 15B, is, for example, between 25 and 40 microseconds, or between 40 and 75 microseconds, or longer than 75 or shorter than 25. Several considerations may influence the choice of values for these time intervals, and the numbers need not be the same for all slices. For example, the duration T_h5 of the gradient pulse is typically chosen to be slightly shorter than τi, and the shorter X₁ is, the more expensive and higher voltage the power supply may have to be to drive the current in the gradient coils. For a given power supply driving the gradient coils, making τi and T_hs longer may allow a higher gradient phase shift to be produced, resulting in greater resolution in y or z, with the same or lower heat dissipation from the gradient coils. But if Th_S is too long, then the resolution may be limited by other system properties, for example SNR, and there may be no advantage to making T_hS longer than this.

As another example of how the time intervals may be chosen, τ is typically chosen to be a few times greater than the duration of the refocusing RF pulses. This choice may maximize the number of echoes, and the SNR, for a given RF bandwidth and slice width. A greater RF bandwidth may require the use of an RF power supply, and/or an RF receiver amplifier, and/or electronics for isolating the RF receiver during transmission of RF pulses, that is expensive or difficult to obtain. A smaller RF bandwidth, on the other hand, results in thinner slices, and a longer time to acquire data for all the slices in a given imaging region.

The choice of τ may also depend on a desired balance between T₂ weighting and diffusion weighting, which depends on τ.

Optionally, the RF bandwidth is smallest, for example 100 kHz or 50 kHz, for the most distant part of the imaging region from the probe, for example at x = 40 mm. At this distance, limitations of the RF power supply and/or heating of the RF coils may limit the RF field that can be achieved, and a longer RF pulse time may be needed to obtain a flip angle of 180 degrees. Closer to the probe, where the RF field can be higher, the bandwidth is optionally greater, for example 2 times, 3 times, or 5 or more times greater at the closest part of the imaging region than at the most distant part of the imaging region. Optionally, the resulting slice thickness varies less with the position of the slice than the RF bandwidth, or does not vary at all with the position of the slice, because the static magnetic field gradient is lower, further from the probe. For example, all slices are about 0.1 mm thick, or about 0.2 mm thick, or about 0.5 mm thick, or about 1 mm thick, or about 2 mm thick. It should be noted that it is not necessary for the flip angle of the refocusing pulses, for example, to be exactly 180 degrees, or even close to 180 degrees. CPMG is fairly insensitive to flip angle θ of the refocusing pulses, with the amplitude of the echoes being proportional to sin³θ, where θ is the tipping angle of the nominally 90° excitation pulse, and it is assumed that the tipping angle of the nominally 180° refocusing pulse is twice that of the excitation pulse; see A. L. Bloom, Phys, Rev, 98 1105 (1955), the disclosure of which is incorporated herein by reference. Optionally, particularly in the slices more distant from the probe, where the RF field amplitude may be limited to relatively low values, the flip angle of the refocusing pulses is substantially less than 180 degrees, for example less than 150 degrees, or less than 120 degrees, or less than 90 degrees, or less than 60 degrees, and the SNR is correspondingly reduced.

Optionally, the reconstructed images, or the imaging data, for two or more adjacent slices are combined to form a single thicker slice, in order to increase the SNR. Optionally, a running average of slices is used to produce an image with increased SNR. These procedures may be particularly useful for the parts of the imaging region that are more distant from the probe, where the SNR is lower because the static magnetic field is lower, the RF receiving sensitivity is lower, and the refocusing flip angle may be lower if the lower transmitted RF field is not compensated for by longer RF pulse time.

Optionally, the image in a given slice is further resolved in x, by using gradient encoding due to the gradient of the static magnetic field across the slice thickness. This may be particularly useful for thicker slices and for slices closer to the probe, where SNR is greater.

Imaging Hardware Control System Fig. 16 shows a block diagram 1600 of an entire imaging hardware control system, according to an exemplary embodiment of the invention. The "home made" spectrometer is based on a RadioProcessor PC card (by SpinCore, USA) that generates the RF pulses and receives the NMR signal. This card also generates TTL control signals for the power amplifier and low noise amplifier gates. Another PC card (PulseBlaster by SpinCore), which is synchronized to the RadioProcessor card generates the rest of the TTL control signals, such as signals to the mechanical relay switches and to the gradient drivers unit. The third and final PC card that completes the control system is a Digital to Analog converter (PCI-6733 by National Instruments), that generates analog voltage to determine the magnitude of the gradient pulses and the varactor diode tuning voltage. After RF pulses are generated by the RadioProcessor card they are amplified by a

2kW power amplifier (made by Tomco, in Australia). The high power pulses then go into the probe card in which the tuning capacitors and the mechanical relay switches are located, in direct connection to the transmitting RF coil (labeled "Tx coil" in block diagram 1600). The NMR signal is picked up by the receiving RF coil (labeled "Rx coil"), which is disconnected during transmission by a PIN switch, not shown, that sits on the probe card. The Rx coil is tuned to the right frequency by a combination of fixed capacitors, selectable through a mechanical switch, and a varactor diode (all located on the probe card). The high voltage needed to tune the varactor diode is generated by the varactor control HV unit, a special unit. The NMR signal is then amplified by the FEP unit and returns to the computer for further processing. The gradient drivers can generate current of up to 50A for short durations of - 30 μs. This is achieved by a modification of the system described in M. S. Conradi, A. N. Garroway , D. G. Cory, and J. B. Miller, "Generation of Short, Intense Gradient Pulses," Journal of Magnetic Resonance l991;94(2):370-375. Here we charge a capacitor of ~5μF to voltages of up to 420V. The shape of the current produced by such system is a half sin.

As used herein the term "about" refers to ± 10 % .

The terms "comprises", "comprising", "includes", "including", "having" and their conjugates mean "including but not limited to". This term encompasses the terms "consisting of and "consisting essentially of.

The phrase "consisting essentially of means that the composition or method may include additional ingredients and/or steps, but only if the additional ingredients and/or steps do not materially alter the basic and novel characteristics of the claimed composition or method. As used herein, the singular form "a", "an" and "the" include plural references unless the context clearly dictates otherwise. For example, the term "a compound" or "at least one compound" may include a plurality of compounds, including mixtures thereof.

Throughout this application, various embodiments of this invention may be presented in a range format. It should be understood that the description in range format is merely for convenience and brevity and should not be construed as an inflexible limitation on the scope of the invention. Accordingly, the description of a range should be considered to have specifically disclosed all the possible subranges as well as individual numerical values within that range. For example, description of a range such as from 1 to 6 should be considered to have specifically disclosed subranges such as from 1 to 3, from 1 to 4, from 1 to 5, from 2 to 4, from 2 to 6, from 3 to 6 etc., as well as individual numbers within that range, for example, 1, 2, 3, 4, 5, and 6. This applies regardless of the breadth of the range.

Whenever a numerical range is indicated herein, it is meant to include any cited numeral (fractional or integral) within the indicated range. The phrases "ranging/ranges between" a first indicate number and a second indicate number and "ranging/ranges from" a first indicate number "to" a second indicate number are used herein interchangeably and are meant to include the first and second indicated numbers and all the fractional and integral numerals therebetween.

It is appreciated that certain features of the invention, which are, for clarity, described in the context of separate embodiments, may also be provided in combination in a single embodiment. Conversely, various features of the invention, which are, for brevity, described in the context of a single embodiment, may also be provided separately or in any suitable subcombination or as suitable in any other described embodiment of the invention.

Certain features described in the context of various embodiments are not to be considered essential features of those embodiments, unless the embodiment is inoperative without those elements.

Although the invention has been described in conjunction with specific embodiments thereof, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications and variations that fall within the spirit and broad scope of the appended claims.

AU publications, patents and patent applications mentioned in this specification are herein incorporated in their entirety by reference into the specification, to the same extent as if each individual publication, patent or patent application was specifically and individually indicated to be incorporated herein by reference. In addition, citation or identification of any reference in this application shall not be construed as an admission that such reference is available as prior art to the present invention. To the extent that section headings are used, they should not be construed as necessarily limiting.

Claims

WHAT IS CLAIMED IS:

1. A magnet assembly for a self-contained MRI probe having a longitudinal axis in a z-direction, the probe adapted to image an imaging region located relative to the probe, in terms of a Cartesian x-y-z coordinate system, at least partly in a positive x direction, the magnet assembly comprising magnet portions which produce a static magnetic field in the imaging region, including at least: a) a first portion located at y < 0, with magnetization having a positive y component and a negative x component; and b) a second portion, located at y > 0, with magnetization having a positive y component and a positive x component.

2. A magnet assembly according to claim I , wherein the portions of the magnet assembly also include: a) a third portion located at y < 0, further from the imaging region than the first portion, with magnetization having a positive y component and a positive x component; and b) a fourth portion located at y > 0, further from the imaging region than the second portion, with magnetization having a positive y component and a negative x component.

3. A magnet assembly according to claim 1, wherein the portions of the magnet assembly are grouped into three or more parts located at different longitudinal locations, including at least: a) two end parts, located respectively at z > 0 and z < 0, each end part comprising a first portion and a second portion as specified in claim 1 ; and b) a center part, located between the two end parts longitudinally, the center part comprising a third portion located at y < 0, with magnetization having a positive y component and a positive x component, and a fourth portion located at y > 0, with magnetization having a positive y component and a negative x component.

4. A magnet assembly according to claim 3, wherein the center portion also comprises a first portion as specified in claim 1, closer to the imaging region than the third portion, and a second portion as specified in claim 1, closer to the imaging region than the fourth portion.

5. A magnet assembly according to claim 4, wherein the first portion of the center part is smaller than the third portion, and the second portion of the center part is smaller than the fourth portion.

6. A magnet assembly according to any of the preceding claims, wherein the direction of magnetization is within 40 degrees of the positive y direction for most of the magnet volume.

7. A magnet assembly according to claim 1, wherein each portion is part of a separate magnet associated with that portion, and each magnet is substantially uniformly magnetized.

8. A magnet assembly according to claim 6, also comprising at least one intermediate magnet not associated with any of the portions, located between two magnets associated with two of the portions, the direction of magnetization of the intermediate magnet being intermediate between the directions of magnetization of the two magnets associated with the two portions.

9. A self-contained MRI probe comprising a magnet assembly according to any of the preceding claims, and a set of one or more phi-gradient coils, phi being an azimuthal coordinate, adapted to produce a phi-gradient field with a component parallel to the static magnetic field, for gradient encoding in the imaging region.

10. A probe according to claim 9, wherein the phi-gradient coil set is adapted to produce a component of the phi-gradient field parallel to the static magnetic field, as a function of phi for constant z and constant static magnetic field somewhere in the imaging region, that differs, at at least one value of phi, by at least 20% of the maximum value of the function of phi, from a best linear fit.

11. A probe according to claim 9, wherein the phi-gradient coil set is adapted to produce a component of the phi-gradient field parallel to the static magnetic field, as a function of azimuthal coordinate phi for constant z and constant static magnetic field somewhere in the imaging region, that does not differ, for any value of phi, by more than 20% of the maximum value of the function of phi. from a best linear fit.

12. A self-contained MRl probe comprising a magnet assembly according to any of the claims 1 -8, or probe according to any of claims 9-1 1 , comprising a set of one or more z- gradient coils, adapted to produce a z-gradient field with a component parallel to the static magnetic field, for gradient encoding in the imaging region.

13. A self-contained MRI probe comprising a magnet assembly according to any of claims 1-8, or a probe according to any of claims 9-12, having size, shape and mechanical properties adapted for rectal use, with at least part of the prostate in the imaging region.

14. A probe according to claim 13, also including an ultrasound imaging array suitable for imaging the prostate, the ultrasound array, magnets and RF transmitting and receiving elements being mounted on the probe at known relative positions and orientations to each other, such that images generated by the ultrasound array can be aligned with MRI images generated by the probe.

15. A self-contained MRI probe comprising a magnet assembly according to any of claims 1-8, or a probe according to any of claims 9-12, small enough for intravascular use, and sufficiently sensitive, when used inside an artery, to distinguish plaque from healthy tissue in the artery wall.

16. A magnet assembly according to any of claims 1-8, with a longitudinal axis in the z direction, having a total length in the z direction greater than its total diameter in the x and y directions.

17. A self-contained MRI probe adapted to image an imaging region, the probe comprising a magnet assembly which produces a static magnetic field in the imaging region, the probe having size, shape and mechanical properties adapted for positioning in the rectum such that at least part of the surface of the prostate intersects the imaging region, with a surface of constant strength of the static magnetic field being curved to more closely match said part of the surface of the prostate than any plane would.

18. A probe according to claim 17, having a longitudinal axis in a z-direction, the imaging region located relative to the probe, in terms of a Cartesian x-y-z coordinate system, at least partly in a positive x direction, the magnet assembly comprising magnet portions which produce a static magnetic field in the imaging region, including at least: a) a first portion located at y < 0, with magnetization having a positive y component and a positive x component; and b) a second portion, located at y > 0, with magnetization having a positive y component and a negative x component.

19. A self-contained MRI probe having a longitudinal axis in a z-direction, the probe adapted to image an imaging region located relative to the probe, in terms of a Cartesian x- y-z coordinate system, at least partly in a positive x direction, comprising: a) a magnet assembly that produces a magnetic field with a predominant y component at least somewhere in the imaging region; and b) one or more z-gradient coils, adapted to produce a z-gradient field with a component parallel to the static magnetic field for gradient encoding in the imaging region, wherein at least a portion of one z-gradient coil, defined as the y- z portion, is oriented in a direction closer to the y-z plane than to the x-y or x-z plane.

20. A probe according to claim 19, wherein at least a portion of one z-gradient coil, defined as the x-z portion, is oriented in a direction closer to the x-z plane than to the y-z or x-y plane.

21. A probe according to claim 20, wherein the y-z and x-z portions are part of a single non-planar coil.

22. A probe according to claim 20, wherein the y-z and x-z portions are parts of separate substantially planar coils.