EP1924867A2

EP1924867A2 - Asymmetric gradient coil for mri

Info

Publication number: EP1924867A2
Application number: EP06765955A
Authority: EP
Inventors: Gert B. J. Mulder; Robert-Paul Buitendijk
Original assignee: Koninklijke Philips Electronics NV
Current assignee: Koninklijke Philips NV
Priority date: 2005-06-30
Filing date: 2006-06-29
Publication date: 2008-05-28
Also published as: WO2007004164A3; WO2007004164A2

Abstract

A gradient coil assembly (500) for use in magnetic resonance imaging is presented. The gradient coil assembly includes a first gradient coil portion (510) and a second gradient coil portion (520) disposed from the first gradient coil portion along a z-axis. The first gradient coil portion is shielded and the second gradient coil portion is unshielded.

Description

ASYMMETRIC GRADIENT COIL FOR MRI

DESCRIPTION

The following relates to the magnetic resonance imaging (MRI) arts. It has particular applicability to a magnet resonance imaging system having an open geometry and will be described with particular reference thereto.

MRI designers have been directing efforts towards open and patient-friendly geometries of the MRI system and, in particular, the main magnet. One of the challenges is that a more open geometry can come with a tradeoff in magnet performance and/or magnet cost.

One type of MRI magnet uses a cylindrical shape, as illustrated in Figure 1. The magnetic field has a homogeneous volume, or examination region, 18, necessary for MRI imaging, which is often located in the center of the system along the z-axis of the cylinder. As shown in Figure 1, the z-axis of the cylindrical system is along the same direction as the main magnetic field. As a result, a patient lies in the cylinder at a position along the z-axis which places a region of interest of the patient within the homogeneous volume. A disadvantage is that many times the cylindrical system encloses the patient, leading to a patient unfriendly situation. Shortening the magnet length can improve the openness but it can also come with a penalty in cost price and/or magnetic field homogeneity. An open magnet geometry is shown in Figure 2. Such a magnet includes two symmetric halves, an upper half and a lower half, with the patient positioned between the halves. In this geometry, assuming the patient is lying down, the magnetic field is in the vertical direction, hence the term "vertical field magnet" for this type of magnet. Another open geometry, though relatively inefficient, is that of a so-called "projection magnet". It includes a single-sided magnet, with a homogenous volume on the z-axis, in front of the magnet, projected outside the magnet as shown in Figure 3. If such a magnet were hidden behind a wall, the homogenous volume would be floating in the room without giving the patient the impression of being enclosed by a magnet. An extremely friendly "walk-up magnet" is the result. However, such projection magnets have not achieved much success due to problems with inefficiency.

Figure 4 shows a projection magnet with a single supplementary coil 22. It leads to an asymmetric magnet design as shown in Figure 4. The efficiency and homogeneity are significantly improved over the pure projection magnet shown in Figure 3. The disadvantages of a projection magnet with a supplementary coil can be explained as follows. The supplementary coil 22 at the small pole (right side) have a current density comparable to that of the large pole (left side), or even higher. So, if the large pole includes a superconductive magnet, then the small pole also includes a superconductive magnet. Therefore, as shown in Figure 4 the main coils 20 are disposed within a first cryostat 30 and the supplementary coils 22 are disposed within a second cryostat 32. The cryostat at the small pole adds cost, reduces the openness, and reduces the gap between the poles, thereby reducing the space available to accommodate a patient. Other considerations when designing an MRI system relate to gradient coils. Shielded gradient coils are known for both cylindrical magnets as well as for open magnets. In Figures 5 and 6 shielded gradient coils for cylindrical and open systems, respectively, are shown. In both of the cases a magnetic field gradient within the imaging volume is created by a primary set of windings. In the case of a cylindrical system as shown in Figure 5, the primary set of windings are represented by a first, or inner, cylinder 50. In the case of an open system, as shown in Figure 6, the primary set of windings are represented by upper and lower inner surfaces 60, 61. The shielding is achieved by using a secondary set of windings, separated from the primary set by a certain distance, which generate a magnetic field that substantially cancels the magnetic fringe field generated by the first set of windings. In the case of the cylindrical system shown in Figure 5, the secondary set of windings are represented by a second, or outer, cylinder 52. In the case of the open system shown in Figure 6, the secondary windings are represented by upper and lower outer surfaces 62, 63.

An important design consideration for gradient coils is the stored energy in the coil. A small amount of stored energy allows fast on and off switching of the gradient resulting in a very efficient gradient coil suitable for fast MR imaging. In general, shielding the gradient coil will lead to higher stored energy and therefore a decreased efficiency. Although active shielding of the gradient coil is less efficient, it is generally necessary when there is a cryostat adjacent to the gradient coil in order to reduce eddy currents in the cryostat to an acceptable level. In another design it is possible for the windings to cross over between the inner and outer cylinders, in the case of a cylindrical system, or surfaces, in the case of an open system. These cross-overs can increases the efficiency of the gradient coils. Depending on the efficiency, without cross-overs, it may or may not be necessary to allow the windings to cross over.

In Figure 7, plane A contains the primary windings, plane B contains the secondary windings and rim C contains the cross-overs. Depending on the diameter of the gradient coils this increases the efficiency. This type of design is also used on cylindrical gradient coils in order to combine short length with good efficiency.

The aforementioned shielded gradient coils have disadvantages. In open systems, both sides of the gradient coil are actively shielded. Active shielding of a gradient coil typically requires extra space needed in comparison to an unshielded gradient coil. Active shielding of both sides is not necessarily needed for certain special open MRI systems. In configurations where active shielding is not necessary for both sides, for example because of the use of a non-superconducting side, the use of shielded gradient coils will lead to inefficient designs. The energy stored in that type of configuration is higher than needed, and space required for the gradient coil is more than necessary, and the openness is reduced.

In light of the foregoing, it is desirable to have a gradient coil which is relatively small as well as efficient. It is also desirable to have a gradient coil which is switchable between two gradient magnetic fields.

In accordance with one embodiment of the present invention, a gradient coil assembly for use in magnetic resonance imaging is provided. The gradient coil assembly includes a first gradient coil portion and a second gradient coil portion disposed from the first gradient coil portion along a z-axis. The first gradient coil portion being shielded and the second gradient coil portion being unshielded.

In accordance with another aspect of an embodiment of the present invention, a gradient coil assembly for use in magnetic resonance imaging is provided. The gradient coil includes a shielded gradient coil portion disposed in a first plane; and an unshielded gradient coil portion disposed in a second plane, the second plane being substantially parallel to the first plane and the shielded gradient portion acting in combination with the unshielded gradient coil portion to generate a first gradient magnetic field. The invention may take form in various components and arrangements of components, and in various process operations and arrangements of process operations. The drawings are only for the purpose of illustrating preferred embodiments and are not to be construed as limiting the invention.

Figure 1 shows an illustration of a cylindrical magnetic resonance imaging system.

Figure 2 shows an illustration of an open magnetic resonance imaging system.

Figure 3 shows an illustration of a projection magnet.

Figure 4 is an illustration of an asymmetric open magnet. Figure 5 shows an illustration of cylindrical shielded gradient coil.

Figure 6 is an illustration of a shielded gradient coil for an open MR system.

Figure 7 is an illustration of one side of a shielded gradient coil with cross-overs for an open MR system.

Figure 8 is an illustration of an asymmetric open system having superconducting coils at the large pole and non-superconducting coils at the small pole.

Figure 9 shows an illustration of a design strategy for designing asymmetric main magnets for an MR system.

Figure 10 shows an asymmetric open system using normal coils at the small pole in which the homogeneous volume is centered and having a patient in a sitting position whose head is being scanned.

Figure 11 shows an asymmetric open system, using normal coils at the small pole, in which the homogeneous volume is near the normal pole (right side).

Figure 12 shows an asymmetric open system, using ferro-magnetic rings (iron) at the small pole. Figure 13 shows a cross section of a coil element.

Figure 14 shows an open gradient coil which is shielded on only one side.

Figure 15 shows an exploded view of a gradient coil.

Figure 16 shows windings of a gradient coil which is shielded on one side.

Figure 17 shows a view of the windings in the gradient coil of Figure 16 showing both sets of primary windings and the secondary windings of the shielded part.

Figure 18 shows an overview of the fields generated by the gradient coils.

Figure 19 shows windings of the Baseline and Correction coils. Figure 8 shows an asymmetric main magnet system of a magnetic resonance imaging apparatus 100. The asymmetric main magnet system includes a first main magnet portion 120 and a second main magnet portion 130 disposed away from the first main magnet portion along the z-axis. In the embodiment shown, the first main magnet portion, or large pole, includes a number of coils 1-9, 13 disposed within a cryostat 125. The second main magnet portion, or small pole, includes a number of coils 10-12. However, the small pole is not disposed within a cryostat. Further, the total diameter of the small pole is substantially smaller than that of the large pole. In one embodiment, the ratio of diameters is less than 0.7.

Accordingly, the embodiment shown in Figure 8 includes a superconducting large pole in combination with a non- superconducting small pole. In other words, the two poles are not only different in size, but also different in applied technology.

In designing such a main magnet system, multiple coils on the small pole are allowed. The use of current on the small pole, however, is discouraged as much as possible. This is done by means of a cost function of the design solver. For example, the dissipation in the entire magnet system is minimized while all coils of the small pole are given a much higher resistivity (e.g. factor 100) than the coils in the large pole. This approach results in a design where most of the field is generated by the big pole, using a relatively large number of ampere-turns, while the small pole becomes a set of coils carrying relatively few ampere-turns, which can be realized with resistive coils or magnetic rings. In such asymmetric geometries the field contributions generated by the small and the big pole are manipulated. For example, during the design process, for a given target field (homogeneity and fringe field) current flow in the big pole can be encouraged, and currents on the small pole can be discouraged. By doing this, a situation where most of the field is generated by the large pole (necessitating superconductivity), while the small pole provides significantly less field (achievable with non-superconducting technology) can be achieved.

The strategy for designing such an asymmetric magnetic resonance imaging system is described in more detail as follows. First the areas at the large and small pole where current is in principle allowed to flow are set. This is illustrated in Figure 9. As shown in Figure 9, there are first 251, second 252, and third 253 areas in the large pole where current is allowed to flow. With respect to the small pole shown in this embodiment, there is one area, or fourth area 254, in which current is allowed to flow. By doing this, the dimensions of the large and small poles respectively are defined. As can also be seen in Figure 9, whether the large pole should define a recess 260 or other geometrical configuration can be established. For example, a recess can be established to accommodate a gradient coil, radio-frequency coil, or for any other reason.

The current-carrying areas are then meshed in small elements. The field targets are also defined. On the surface of the intended homogeneous volume 200 a large number of points, e.g. 50, are selected in which the magnetic field is pre-scribed with a certain tolerance, e.g. B_z = 1000000 ± 10 μT. To obtain active shielding, several fringe field points 210 in the external region around the magnet are selected and the fringe field is prescribed with certain tolerance, e.g. B = 0 ± 500 μT.

Next the necessary currents in the mesh are determined using a linear solver, which solves the currents while satisfying the target field constraints and simultaneously minimizing a cost function. A suitable cost function is for example the total dissipation in the magnet system, where sections in the normal (non-superconducting) conducting area, e.g. the small pole, are given an α times higher resistivity than those in the superconducting regions e.g. the large pole. The larger the factor α, the more current is pushed from the normal conducting side towards the superconducting side. In the examples α has been chosen to be 250, but other values are applicable.

The result is a homogeneous magnet design having currents in many small mesh elements. Due to the design approach, the current density in the normal conducting side is substantially smaller than in the superconducting side. If the current density in the normal conducting region is still too high the dimensions of this side can either be increased, or the value of α can be increased and the solver can be run again.

The next step is to group clusters of current elements with equal current direction into discrete rectangular coils that can be practically wound. During this process, dimensions, positions, and current densities of the new coils are chosen such that the target fields are preserved. In the examples a total of 13 to 15 coils is reached. Tiny coils can often be eliminated to reduce the number of coils (i.e. less complex magnet) without giving up too much of the target field. Finally, if needed, normal conducting coils can replaced by ring-shaped elements of ferro-magnetic or permanently magnetic materials. This requires shift in positions and dimensions in order to preserve the target field.

The above design strategy leads to designs as illustrated in Figures 8 and 10-12. These magnets all have the following features in common: - 1.0 T central field;

- field homogeneity of ± 10 ppm on a sphere of 20 cm diameter;

- expected net patient gap 35 cm;

- diameter of the small pole ranging from 0.45 to 0.60 m; - approximate outer dimensions of big pole arel .8 m diameter and 0.7 m width;

- 5 cm deep recess in the big pole for, for example, a gradient coil; and

- actively shielded, 0.5 mT contour smaller than RxZ=6x7 m.

It is to be understood that various alterations to the above parameters are contemplated. The figures show a variety of clinical applications possible with this type of asymmetric magnet. In most of the cases, the large pole can be hidden behind a wall, in the ceiling, or in the floor, in order to improve the perception of openness.

Figures 8, 10, and 11 show how the homogeneous volume can be chosen near one of the magnet sides or simply in the middle, the optimum choice depending on the intended clinical application. These three designs, which employ coils in the main magnet portions, are shown in more detail in Table 1. As shown in Figure 13 the values for r correspond to the radial distance from the z-axis to the individual coil element, the values for w correspond to the width of the coil element, and the values for t correspond to the thickness of the coil elements. As can be seen, the current density in the superconducting coils (100- 120 A/mm²) is much higher than the current density in the normal conducting coils (9-24 A A/mm²). Design 1:

# z[m] r[m] w [m] t[m] J[A/mm2]

1 -0.250000 0.006856 0.019833 0.035401 -114.0390

2 -0.250000 0.045603 0.025667 0.033443 114.0390

3 -0.250000 0.080445 0.027483 0.032442 -114.0390

4 -0.250000 0.132073 0.030221 0.041746 114.0390

5 -0.250000 0.200522 0.037466 0.046324 -114.0390

6 -0.250000 0.285693 0.057945 0.057945 114.0390

7 -0.170000 0.342044 0.021285 0.041325 114.0390

8 -0.170000 0.405800 0.069891 0.088924 -114.0390

9 -0.170000 0.628364 0.103344 0.206687 114.0390

10 0.300000 0.009956 0.072915 0.074729 9.1207

11 0.300000 0.106653 0.065498 0.056657 -9.1207

12 0.300000 0.178114 0.086429 0.151880 9.1207

13 -0.700000 0.674464 0.089098 0.178197 -114.0390

Design 2:

# z[m] r[m] w [m] t[m] J[A/mm2]

1 -0.325000 0.002825 0.044798 0.071180 -119.9570

2 -0.325000 0.084148 0.041736 0.057811 119.9570

3 -0.325000 0.164410 0.043147 0.061437 -119.9570

4 -0.325000 0.266472 0.058382 0.073209 119.9570

5 -0.245000 0.346340 0.028114 0.028114 119.9570

6 -0.245000 0.401760 0.074918 0.086880 -119.9570

7 -0.245000 0.630751 0.102570 0.205141 119.9570

8 0.225000 0.012281 0.020570 0.027684 23.9929

9 0.225000 0.054820 0.014331 0.026502 -23.9929

10 0.225000 0.099509 0.027073 0.034952 23.9929

11 0.225000 0.161785 0.028467 0.035194 -23.9929

12 0.225000 0.221281 0.046841 0.095310 23.9929

13 -0.775000 0.668719 0.089610 0.179220 -119.9570

Design 3:

# z[m] r[m] w [m] t[m] J[A/mm2]

1 -0.400000 0.009633 0.042581 0.085982 117.5530

2 -0.400000 0.118450 0.039519 0.070578 -117.5530

3 -0.400000 0.243919 0.058403 0.081064 117.5530

4 -0.320000 0.389294 0.066842 0.084562 -117.5530

5 -0.320000 0.629736 0.102262 0.204525 117.5530

6 0.150000 0.009038 0.011183 0.011183 23.5120

7 0.150000 0.020493 0.011026 0.011026 -23.5120

8 0.150000 0.037331 0.013008 0.013008 23.5120

9 0.150000 0.057466 0.013652 0.012056 -23.5120

10 0.150000 0.072347 0.015006 0.026391 23.5120

11 0.150000 0.102543 0.017798 0.017986 -23.5120

12 0.150000 0.131306 0.030358 0.030698 23.5120

13 0.150000 0.180185 0.033878 0.031838 -23.5120

14 0.150000 0.227283 0.048844 0.090898 23.5120

15 -0.850000 0.663642 0.090061 0.180122 -117.5530

Table 1 : Values relating to the designs of Figures 8 (Design 1), 10 (Design 2), and 11 (Design 3). Figure 12 shows a main magnet where the coil elements 9, 10 are ferro-magnetic (iron) rings rather than the normal conducting coils discussed above.

Figure 14 shows an embodiment of a gradient coil assembly 500 which leaves one side unshielded. An advantage of such an embodiment is that a reduction in stored energy in the gradient coil is facilitated. Additionally, a reduction in the amount of space needed for the gradient coil is facilitated. As shown in Figure 14, the gradient coil assembly includes a first, or baseline, gradient coil portion 510 and a second, or correction, gradient coil portion 520. In the embodiment shown, the first and second gradient coil portions are disposed apart from each other along the z-axis. In one embodiment, the first gradient coil generates most of the gradient field and acts in combination with the second gradient coil. Also as shown in Figure 14, the first gradient coil is shielded.

Turning to the embodiment of Figure 15, the gradient coil assembly is build up out of two coaxial planes (planes A) containing primary sets of winding that generate the main gradient field. At one side a secondary (shield) plane (plane B) is added coaxial to the two primary planes, separated by a distance from one of the primary planes. The embodiment shown in Figure 15 also includes a connecting rim C, on the shielded side. This rim connects the primary plane with the secondary plane, and contains the windings that cross over between planes A and B.

An example is worked out in detail as follows, an x-gradient coil with the homogeneous area slightly off centre between the primary planes, as shown in Figure 16. It is to be understood that a design of a y-gradient coil can be obtained by rotating the total x-gradient coil by 90°.

In Figure 16 it can be seen that one side of the design is shielded. Almost no field lines penetrate the area behind the shielded part of the gradient coil whereas behind the unshielded part there are many field lines visible.

The outer plane represents the conducting outer surface of a magnet cryostat. During the design a limitation is set on the amplitude of eddy currents that are induced in this extra plane, or on the fields that are generated by these eddy currents.

Figure 17 shows the discretised windings in the primary planes and the shielding plane for the example.

The homogeneous area created by the gradient coil can be placed at an arbitrary position, on the axis, in the gap between the two primary planes. It is not necessary to have the same position as in the example. A 10mT/m gradient field is created with a 360A operating current.

According to another embodiment of the invention, a gradient coil assembly with a switchable homogeneous area is provided. This gradient coil is build up out of two gradient coils described above. In this embodiment, the baseline coil carries a larger part of the current to generate the gradient field (>60%) and the correction coil generates a smaller part of the current to generate the gradient field (<40%).

In designing such a gradient coil, two gradient coil designs that make the two different homogeneous areas are used as a start. "A" and "B" current patterns of the two gradient coils are thus defined.

"Baseline" and "Correction" are used to denote the current patterns of the Baseline and Correction coil respectively.

BC and CC are obtained by adding and subtracting the current patterns as follows:

Baseline = ^(A+B) Correction = ¹A(A-B) Therefore:

Baseline + Correction = A Baseline - Correction = B

By linear combination of these current patterns a corresponding linear combination of their magnetic fields is obtained.

So when the Baseline coil and Correction coil are connected in series they create magnetic field A, and when they are in anti-series they create magnetic field B. Therefore switching from homogeneous area 1 to homogeneous area 2 by switching from a series- to anti-series connection between the Baseline- and Correction coil is enabled.

As another approach BC and CC can be designed in such a way that they generate the desired magnetic field. BC would then make the average field of A and B: %(A+B).

Because the Correction coil generates far less magnetic field than the Baseline coil it has a very low amount of stored energy, 0.074J vs. 0.843 J. Therefore by using the

Baseline and Correction coil switching between two different homogeneous areas, without much penalty in stored energy in both coils, is allowed. Figure 18 shows the magnetic fields generated by the Baseline coil, Correction coil and both coils A and B.

The windings of the primary planes of the Baseline and Correction coil are displayed in Figure 19. To generate 10mT/m one has to operate the gradient coils at 170A. The Correction coil contains far less windings and is less complex.

The invention has been described with reference to the preferred embodiments. Obviously, modifications and alterations will occur to others upon reading and understanding the preceding detailed description. It is intended that the invention be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

1. A gradient coil assembly (500) for use in magnetic resonance imaging, the gradient coil assembly comprising: a first gradient coil portion (510); and a second gradient coil portion (520) disposed from the first gradient coil portion along a z-axis, the first gradient coil portion being shielded and the second gradient coil portion being unshielded.

2. The gradient coil assembly as set forth in claim 1 wherein the first gradient coil portion contributes substantially more to a gradient field than the second gradient coil portion.

3. The gradient coil assembly as set forth in claim 2 wherein the first gradient coil portion contributes approximately sixty percent of the gradient field and the second gradient coil portion contributes approximately forty percent of the gradient field.

4. The gradient coil assembly as set forth in claim 1 wherein the gradient coil assembly is switchable between first and second magnetic fields.

5. The gradient coil assembly as set forth in claim 4 wherein the first magnetic field corresponds to the first and second gradient coil portions being connected in series and the second field of view corresponds to the first and second gradient coil portions being connected in anti-series.

6. A gradient coil assembly for use in magnetic resonance imaging, the gradient coil comprising: a shielded gradient coil portion disposed in a first plane; and an unshielded gradient coil portion disposed in a second plane, the second plane being substantially parallel to the first plane and the shielded gradient portion acting in combination with the unshielded gradient coil portion to generate a first gradient magnetic field.

7. The gradient coil assembly as set forth in claim 6 wherein the shielded gradient coil portion and the unshielded gradient coil portion act in combination to generate a second gradient magnetic field.

8. The gradient coil assembly as set forth in claim 7 wherein the first and second gradient magnetic fields correspond to series and anti-series connections, respectively, of the shielded and unshielded gradient portions.

9. The gradient coil assembly as set forth in claim 6 wherein the first and second planes are two dimensional.

10. The gradient coil assembly as set forth in claim 6 wherein the first and second planes are disposed along a z-axis from one another.