Device and method for real-time three-dimensional distortion correction of magnetic resonance imaging

Abstract

The present invention relates to a magnetic resonance tomography device and a method for correcting images in real time. A first layer and a second layer are acquired by three-dimensional encoding, wherein the two layers are positioned at an angle, preferably at an angle of 90 degrees, relative to each other, and the device or trajectory to be imaged is located in the intersection of the two layers. Preferably, two-dimensional images are alternately generated from the two layers and displayed to the user.

Description

[0001] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included. Technical Field

[0002] The present invention relates to a magnetic resonance tomography device and a method for real-time tracking of a medical device within the scope of interventional magnetic resonance imaging. Background Art

[0003] In minimally invasive medical procedures, medical devices such as catheters and / or interventional needles are inserted into a patient's body, and the procedure is usually monitored by images. Image monitoring enables recording of images in which the medical device is made visible relative to its anatomical environment. Although X-ray imaging has traditionally been used for image monitoring of minimally invasive medical interventions, especially fluoroscopy, the use of magnetic resonance devices, i.e., magnetic resonance imaging (MR imaging), has also been proposed for image monitoring. This is generally referred to as interventional magnetic resonance imaging.

[0004] Here, most medical devices are metallic or made of other materials that do not allow direct imaging of protons in water or fat in the body using magnetic resonance imaging. Instead, indirect methods are used for imaging, in which imaging is carried out using indirect effects or artificial marks. However, for real-time tracking, a short repetition period is required, which only provides poor image quality. In addition, geometric distortion is also common in magnetic resonance tomography, making it more difficult to track thin devices compared to geometrically accurate imaging by X-rays. Summary of the Invention

[0005] Therefore, the technical problem to be solved by the present invention is to improve the tracking of medical devices during intervention.

[0006] The above technical problem is solved by a magnetic resonance tomography device according to the present invention and a method for operating a magnetic resonance tomography device according to the present invention.

[0007] The magnetic resonance tomography device according to the present invention is configured to image a medical device on a trajectory in real time. The magnetic resonance tomography device has a controller that coordinates the activities required for image recording using the magnetic resonance tomography device. Here, the controller can also be divided into sub-units, for example, it can be divided into a gradient and high-frequency controller for real-time control of the process, an operation interface for interacting with the operator, and an image reconstruction computer with high computing power.

[0008] The controller is designed to control a magnetic resonance tomography device such that a first layer is acquired by three-dimensional encoding. Herein, a three-dimensional volume as described below is regarded as a layer, which is also referred to as a "Slab" in English. This three-dimensional volume also has a thickness in the sense of magnetic resonance imaging, that is, this three-dimensional volume is not only composed of a single two-dimensional arrangement formed by the voxels of magnetic resonance imaging in a plane, but also composed of at least two voxel layers overlapping vertically, preferably more voxel layers. Therefore, although there is image distortion, the medical device can still be fully acquired.

[0009] Herein, the controller is designed such that, during imaging, it selects the position of the first layer according to the trajectory and selects the layer thickness according to the pre-determined distortion of the magnetic resonance tomography device, or sets parameters for acquiring the first layer, that is, sets the position of the first layer according to the trajectory and sets the layer thickness according to the pre-determined distortion of the magnetic resonance tomography device, so that the trajectory or the medical device is fully acquired through the first layer. In this case, the position includes not only the location of the layer but also the orientation of the layer. For example, the pre-given trajectory can be intervened through the coordinates determined in the planning. However, it is also conceivable that, for example, the trajectory is determined in advance by acquiring the position and / or orientation of the medical device by means of imaging. With the knowledge of the position of the trajectory and the distortion of the imaging performed by the magnetic resonance tomography device, the controller can determine the volume where the trajectory is located considering the distortion, such as a cube composed of voxels. Herein, preferably, the layer thickness is designed such that the layer only includes the trajectory, that is, there is no such voxel layer or only a few such voxel layers that are not penetrated by the trajectory or have no voxels belonging to the trajectory, in order to accelerate the three-dimensional encoding.

[0010] In addition, the controller is designed to acquire a second layer by three-dimensional encoding in the described manner, wherein, during imaging, the controller selects the position of the second layer according to the trajectory and determines the layer thickness according to the pre-determined distortion of the magnetic resonance tomography device, so that the trajectory is fully acquired through the second layer.

[0011] In addition, the controller is designed to select the positions of the first layer and the second layer such that the first layer and the second layer enclose an angle greater than 20 degrees, 40 degrees or 60 degrees and less than or equal to 90 degrees, and the trajectory of the medical device and / or the medical device is located in the intersection or cross volume of the first layer and the second layer. For example, it is conceivable that the controller rotates the coordinates of the first layer around the trajectory by this angle to keep the position of the second layer.

[0012] The corresponding layer also provides an image of another environment of the trajectory in a direction extending transversely or in a plane. Here, it is preferred that the extension of the layer corresponds accordingly to the orientation of the two-dimensional image, which is subsequently displayed to the user for navigation, so that this other environment can also be displayed for orientation. By displaying the images from the two layers, preferably in the case of a combination of the image data of the two layers, a high repetition rate can be advantageously achieved for the trajectory while displaying the other environment.

[0013] Finally, the controller is designed to generate a sequence of two-dimensional images of the medical device in a predetermined plane. The trajectory, and thus the medical device, lies in this plane. Images are generated by the first layer and the second layer, or by repeating the generation of images by a plurality of first layers and second layers. Preferably, the predetermined plane is parallel to the first layer or the second layer, but can also enclose a small angle with the first layer or the second layer in order to be able to reproduce as much of the environment of the trajectory as possible. Preferably, the controller also generates another image and displays this other image, the display plane of which is parallel to the other layer or encloses a small angle with the other layer, so that the trajectory can be displayed from different angles or the medical device can be tracked.

[0014] Furthermore, the controller is designed to finally output the sequence of two-dimensional images to the user, for example on a display screen. Here, the output is carried out as soon as possible after the generation of the images in order to ensure a real-time or almost real-time display of the medical device or the trajectory. In particular, the first layer is recorded, two-dimensional images are generated in this plane, and preferably its output is carried out before or during the recording of the second layer and the generation of the two-dimensional images.

[0015] In a preferred embodiment, two-dimensional images are correspondingly generated and displayed alternately from the recordings of the first layer and the second layer in order to generate the trajectory from different viewing directions or angles that are substantially perpendicular to the respective layer. Here, it is preferred to update or display the region of the image corresponding to the intersection or intersection volume of the two layers correspondingly using the acquisitions of the first layer and the second layer.

[0016] The recordings of the two layers aligned at an angle to each other advantageously enable the deviation of the movement of the medical device from the trajectory to be rapidly acquired in space using the trajectory within the layer and to be displayed to the user together with the other environment. In particular, compared to a magnetic resonance tomography device that alternately records layers arranged at an angle to each other, the image repetition rate for the trajectory in the intersection of the layers is doubled in order to track the movement in three dimensions, but then two-dimensional images are correspondingly only generated by layers that are substantially oriented parallel to the display plane. That is to say, it is preferred to also update the region of the intersection of the two layers with the trajectory using each acquisition of one of the two layers in an image oriented parallel to the other layer, so that for this region, the image repetition rate is doubled.

[0017] The method according to the invention shares the advantages of the device according to the invention.

[0018] Other advantageous embodiments are given in the following description.

[0019] In a possible embodiment of a magnetic resonance tomography device according to the invention, the controller of the magnetic resonance tomography device is designed to correct a predetermined distortion of the first layer and / or the second layer. The distortion of the image acquisition of the magnetic resonance tomography device can be caused dynamically by the inhomogeneity of the static magnetic field, the gradient field, and by eddy currents, and is predetermined by the magnetic resonance tomography device in this regard. Here, the predetermined distortion can be determined in a type-specific manner for the magnetic resonance tomography device by calculation during research and development or test measurements. It is also conceivable to determine the predetermined distortion individually for each device by means of a calibration measurement during installation or before image acquisition. For the determined distortion that is a linear mapping, for example, a distortion defined by a matrix, there is a corresponding inverse function, which can be determined from the predetermined distortion using linear algebraic methods in order to correct the acquired first and / or second layer so that the trajectory or the medical device is geometrically correctly displayed, especially in the case of a straight or linear trajectory, and the trajectory is reproduced as part of a straight line.

[0020] Advantageously, by means of the correction, the geometric structure of the acquired three-dimensional data, and thus also the geometric structure of the two-dimensional image generated therefrom, is corrected, so that a distortion-free display of the medical device and the trajectory is provided to the user, and thus the intuitive navigation of the device is made easier.

[0021] In a conceivable embodiment of a magnetic resonance tomography device according to the invention, the magnetic resonance tomography device is also designed to average the first layer and / or the second layer after the correction. Averaging in the sense of the present invention should be understood as improving the signal-to-noise ratio (SNR) of the image acquisition by averaging over a plurality of adjacent voxels in a three-dimensional magnetic resonance recording. In the simplest case, and in the original meaning of the term, averaging is binning, i.e., by summing the values, adjacent voxels are combined into a virtual larger voxel. However, the spatial resolution is reduced at the same time. However, with the reduction in resolution, especially the result of the correction is significantly worse. However, other forms of averaging are also conceivable, such as averaging using weighted summation.

[0022] Therefore, in a possible embodiment of a magnetic resonance tomography device according to the invention, the controller is also designed to perform the correction at a resolution higher than that of the subsequent two-dimensional image.

[0023] Thus, in an advantageous manner, in a magnetic resonance tomography device according to the invention, correction is carried out before the resolution is reduced due to averaging or sorting, so that the orientation of the medical device can be displayed more accurately. Averaging is then carried out in order to improve the SNR of the corrected recordings. Since averaging itself is not an invertible mapping or function, the order is decisive and changes the result. In the case where the magnetic resonance data has a higher resolution, the magnetic resonance data can be corrected with better results. Then, the deteriorated SNR achieved at a higher resolution with the same measurement time can be improved again by subsequent averaging.

[0024] In a conceivable embodiment of a magnetic resonance tomography device according to the invention, the controller is also designed to acquire a third layer containing the medical device using three-dimensional encoding. The third layer is preferably acquired with a higher resolution and / or a longer integration time than the first layer and / or the second layer, so that in particular the environment of the medical device can be acquired and displayed more accurately. Here, the controller is designed to generate an image based on the third layer when generating a sequence of two-dimensional images. Here, preferably, the recording of the third layer is used to reproduce the environment of the trajectory with a higher resolution and / or a higher SNR in the two-dimensional image. The environment of the trajectory hardly changes during the operation of the medical device, so that it is conceivable to use the data of the recording of the third layer for the reproduction of the environment in the two-dimensional image, in particular also in a plurality of consecutive images, without repeating or updating the recording of the third layer between them. It is also conceivable to segment the orientation of each organ or blood vessel in the first and / or second layer here and to correspondingly correct the orientation in the recording in the third layer for the two-dimensional image.

[0025] In an advantageous manner, the combination of the image acquisition of the environment with high image quality but low repetition rate and the rapid recording of the first and second layers with the medical device enables real-time tracking of the device with high resolution of the surrounding organs.

[0026] In a possible embodiment of a magnetic resonance tomography device according to the invention, the trajectories acquired using the first and second layers are curved trajectories, or the medical device is designed to be substantially non-linear but curved. As already described with regard to the distortion caused by the image acquisition of the magnetic resonance tomography device, it is also conceivable that by appropriately selecting the first and second layers, not only the linear device distorted due to magnetic resonance imaging is completely acquired, but also the curved device or the curved trajectory is completely acquired in the case where the geometry or orientation of the trajectory or the medical device is known. For this purpose, a cube defining the first and / or second layer is selected such that the cube completely encloses the curved device or the curved trajectory taking into account the distortion of the magnetic resonance tomography device.

[0027] Thus, the magnetic resonance tomography device according to the invention and the method according to the invention are also capable of tracking a curved medical device or a corresponding device on a curved trajectory in an advantageous manner in real time.

[0028] In a conceivable embodiment of the magnetic resonance tomography device according to the invention, the controller is designed to linearize a curved trajectory when generating a two-dimensional image. Here, it is conceivable that the linearization corrects or linearizes the distortion caused by the magnetic resonance tomography device and / or the geometric curvature of the trajectory or the medical device. Here, the correction can be carried out in a plane or also in space. For example, it is conceivable to carry out the linearization in a plane perpendicular to the plane of the two-dimensional image.

[0029] The magnetic resonance tomography device according to the invention and the method according to the invention also advantageously enable a curved medical device or a flexible medical device on a curved trajectory to be quickly and fully displayed to the user in a two-dimensional image and improve the navigation of the medical device. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The features, characteristics and advantages of the invention described above and the ways in which they are implemented will become clearer and easier to understand in conjunction with the following description of the embodiments, which will be described in more detail in conjunction with the accompanying drawings. In the drawings:

[0031] Figure 1 A schematic view of a magnetic resonance tomography device according to the invention with a device according to the invention is shown;

[0032] Figure 2 A schematic view of a layer acquired by the magnetic resonance tomography device according to the invention in the method according to the invention is shown;

[0033] Figure 3 A schematic view of a layer acquired by the magnetic resonance tomography device according to the invention in the method according to the invention is shown;

[0034] Figure 4 A schematic view of the relative arrangement of two layers acquired by the magnetic resonance tomography device according to the invention in the method according to the invention is shown;

[0035] Figure 5 A schematic flow chart of the method according to the invention is shown. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0036] Figure 1The figure shows a schematic view of an embodiment of a magnetic resonance tomography device 1 according to the present invention, and the magnetic resonance tomography device 1 is also used to perform the method according to the present invention. The magnet unit 10 has a field magnet 11 that generates a static magnetic field B0 for aligning the nuclear spins of a sample or a patient 100 in the recording area. The recording area is characterized by an extremely uniform static magnetic field B0, where the uniformity particularly relates to the magnetic field strength or magnitude. The recording area is almost spherical and is arranged in the patient channel 16, and the patient channel 16 extends through the magnet unit 10 along the longitudinal direction 2. The patient couch 30 can be moved by a moving unit 36 in the patient channel 16. The field magnet 11 is usually a superconducting magnet, which can provide a magnetic field with a magnetic flux density of up to 3 T, and even higher in the case of the latest devices. However, for lower magnetic field strengths, permanent magnets or electromagnets with normal conducting coils can also be used.

[0037] In addition, the magnet unit 10 has gradient coils 12, which are designed to superimpose a magnetic field that is variable in time and space on the magnetic field B0 in three spatial directions in order to spatially distinguish the imaging area in the acquired examination volume. The gradient coils 12 are usually coils made of normal conducting wires, which can generate fields orthogonal to each other in the examination volume.

[0038] The magnet unit 10 also has a body coil 14, which is designed to radiate high-frequency signals fed through signal lines into the examination volume, receive resonance signals emitted by the patient 100, and output them through signal lines.

[0039] The control unit 20 supplies different signals to the magnet unit 10 for the gradient coils 12 and the body coil 14, and evaluates the received signals.

[0040] Therefore, the control unit 20 has a gradient controller 21, which is designed to supply a variable current to the gradient coils 12 through a feeder line, and the variable current provides a desired gradient field in the examination volume in a temporally coordinated manner.

[0041] In addition, the control unit 20 has a high-frequency unit 22, which is designed to generate high-frequency pulses with a pre-given time curve, amplitude, and spectral power distribution for exciting the nuclear spin magnetic resonance in the patient 100. Here, a pulse power in the kilowatt range can be achieved. The excitation signal can be radiated into the patient 100 through the body coil 14, or also through a local emission antenna.

[0042] The controller 23 communicates with the gradient controller 21 and the high-frequency unit 22 through a signal bus 25.

[0043] A medical device 60, such as a biopsy needle, is positioned on a patient 100 in a patient channel 16 or inserted into the patient 100 within the intervention range, and is acquired by means of a magnetic resonance tomography device according to the invention and displayed to the user on an output device such as a display screen 24.

[0044] Both the B0 static magnetic field of the field magnet 11 and the gradient field of the gradient coils 12 have inhomogeneities or non-linearities, which lead to spatial distortions during imaging.

[0045] In Figure 2 a first layer 71 of image acquisition according to the invention with a written track 101 is schematically shown. In Figure 2 a true undistorted geometry is shown, i.e., a medical device 60, such as a biopsy needle, which is substantially a partially straight line geometrically, is also shown here as a partially straight line without bends. In contrast, the first layer 71 of the image acquisition performed by the magnetic resonance tomography device 1 is not a cube with straight edges due to the distortion of magnetic resonance imaging, but is curved or distorted.

[0046] The distortion experienced by the imaging performed by the magnetic resonance tomography device 1 is basically determined by the magnetic resonance tomography device 1 and the sequences used. Therefore, through calibration measurements, for each magnetic resonance tomography device 1 of the same type of device series, and with higher precision for individual devices and pre-determined sequences, the distortion can be determined. This can be done, for example, by imaging a phantom with a grid composed of magnetically active samples using the magnetic resonance tomography device 1 and the corresponding sequence. It is also conceivable to estimate the distortion through simulation.

[0047] The determined distortion provides a mapping of each image point of the examination object in the true geometry to the corresponding voxel in the three-dimensional imaging. Corresponding inverse mappings can be determined using linear algebra methods.

[0048] Therefore, for a given track 101, the positions of the individual points of the track 101 in the distorted voxel space can be determined. Thus, a cube can also be determined in this voxel space, which completely encloses or writes the distorted track 101 while minimizing the number of voxels to be acquired within a limited range. In most cases, for example, the voxels associated with the start and end points of the track in this voxel space can be selected as the diagonally opposite corner points of the cube.

[0049] Conversely, in Figure 3A perspective view of a magnetic resonance tomography device is shown. The first layer 71 is defined by a cube having a plurality of parallel layers 75, 76, 77 composed of voxels in parallel planes. Conversely, the trajectory 101, which is actually straight in this example, appears curved here.

[0050] Here, it can also be seen that when the voxels must be small enough to display the trajectory accurately enough for the navigation of medical devices, due to the curvature, it is not possible to completely acquire the trajectory 101 only using the voxels of the layers 75, 76, 77. Therefore, the present invention also proposes to perform three-dimensional acquisition of the first layer 71 using the voxels of the three layers 75, 76, 77 of the exemplary number here. However, by restricting the cube such that the cube exactly completely encloses the trajectory, the number of voxel values to be acquired can be reduced, thereby increasing the repetition rate with the same SNR.

[0051] Preferably, each layer 75, 76, 77 of the layers 71, 72 is acquired three-dimensionally by phase encoding. For this purpose, the entire layer is excited, and the entire extension of the layer is completely acquired in the FoV using three-dimensional phase encoding to avoid folding of the voxels outside the cube. Although this increases the number of voxels and thus prolongs the acquisition time in the phase encoding, this disadvantage is offset by the much faster excitation of the layers 71, 72 compared to the selective excitation of the cube.

[0052] The coordinate system shown is basically a Cartesian coordinate system, which is only exemplary here even if distorted in the voxel space and is used for better description. Since the user can operate intuitively during the navigation process, it is preferably used for the acquisition of the layer and thus for the resulting imaging. However, the present invention can also be correspondingly applied to other coordinate systems.

[0053] In Figure 4 the first layer 71 and the second layer 72 are shown. For clarity, the individual layers 75, 76, 77 of the respective layers are not shown separately. The first layer 71 and the second layer 72 have an intersection volume 78, and the intersection volume 78 is located inside the two layers 71, 72. The first layer 71 and the second layer 72 are arranged at an angle relative to each other. In other words, the normal vectors of the flat surfaces, i.e., the normal vectors that are not the edge surfaces, enclose this angle. This angle is preferably 90 degrees, that is, the first layer 71 is arranged perpendicular to the second layer 72. Other angles greater than 30 degrees or 60 degrees are also conceivable, so that the layers 71, 72 have significantly different orientations relative to the trajectory 101 accordingly.

[0054] The trajectory 101 is written into the intersection volume 78, so that the trajectory 101 is correspondingly acquired together not only when the first layer 71 is acquired, but also when the second layer 72 is acquired. In this way, the controller 23 can update the position of the medical device 60 along the trajectory 101 in the image at twice the frequency from two different angles parallel to the layers 71, 72.

[0055] In Figure 5 FIG. shows a schematic flow chart of an embodiment of the method according to the invention. The method is performed on a magnetic resonance tomography apparatus 1 according to the invention.

[0056] In step S20, the magnetic resonance tomography apparatus 1 acquires the first layer 71 by means of three-dimensional encoding under the control of the controller 23. Here, the first layer 71 has a plurality of sub-layers 75, 76, 77, and the voxels 79 are arranged in these sub-layers 75, 76, 77. Here, the controller 23 selects the position, i.e., the location and orientation, and the layer thickness of the first layer 71 during imaging according to the trajectory 101 and the predetermined distortion of the magnetic resonance tomography apparatus 1, so that the trajectory 101 is completely acquired by the first layer 71. For this, reference is also made to the descriptions regarding Figure 2 and Figure 3 hereof.

[0057] It is also possible for the controller 23 to predetermine the predetermined distortion in advance by means of a calibration measurement in step S10, for example using a phantom having a grid, the grid being composed of magnetically active samples at predetermined positions.

[0058] In step S30, the magnetic resonance tomography apparatus 1 acquires the second layer 72 by means of three-dimensional encoding under the control of the controller 23. Similarly according to the trajectory 101 and the predetermined distortion of the magnetic resonance tomography apparatus 1, the orientation and the layer thickness of the second layer 72 are selected, so that the trajectory 101 is completely acquired by the second layer 72.

[0059] As shown in Figure 4 and described with respect thereto, the controller 23 selects the orientation of the first layer 71 and the second layer 72 relative to each other such that the first layer 71 and the second layer 72 enclose an angle greater than 30 degrees or 60 degrees, preferably equal to 90 degrees, and the trajectory 101 of the medical device 60 and / or the medical device 60 is located in the intersection volume 78 of the first layer 71 and the second layer 72.

[0060] In step S70, the controller 23 generates a two-dimensional image sequence of the medical device 60 in a predetermined plane. For example, it is conceivable to reconstruct the voxels 79 passing through the intersection volume 78 along this plane or a sectional view of the projection onto this plane.

[0061] Here, the image sequence has an image generated by the first layer 71 and an image generated by the second layer 72. Here, the display plane of the corresponding image is substantially parallel to the layer from which the image is generated. Here, it is preferable to alternately acquire the first layer 71 in step S20 with the second layer 72 in step S30. It is also preferable to directly generate a sectional view according to the magnetic resonance data of the corresponding layer 71, 72 acquired, respectively, before or during the acquisition of the corresponding other layer 71, 72. Thus, images can be generated as close as possible in time. Preferably, by updating the part of the image of the display cross volume 78 according to the data or voxels of the first layer 71 and the second layer 72 acquired, the display of the medical technical device in the two images can be updated at twice the repetition rate.

[0062] Then, in step S80, the controller 23 outputs a two-dimensional image to the user correspondingly on the display screen 24, for example. Here, the output is performed as close as possible in time to the acquisition and generation of the images of the corresponding layers 71, 72. The step sequence preferably is to acquire the second layer 72 in step S30, generate a two-dimensional image according to the second layer 72 in step S70, and before or during the direct output in step S80, acquire the first layer 71 in S20, generate an image of the first layer 71 in S70, and directly output the image of the first layer in S80. Preferably, within a predetermined duration, within the duration of a control signal of a foot switch, for example, or within the duration of an intervention, the sequence of acquisition in S20, S30, generation of the two-dimensional image in S70, and output in S80 is repeated.

[0063] In a possible embodiment of the method according to the invention, in step S40, the controller geometrically corrects the magnetic resonance data of the first layer 71 or the second layer 72 acquired. The magnetic resonance tomography device 1 has geometric distortions in the acquired images of the object. These distortions may be due to the inhomogeneity of the B0 static magnetic field, the nonlinearity of the gradient field, or due to eddy currents generated by the gradient.

[0064] The distortions are specific or predetermined for the corresponding magnetic resonance tomography device and the sequence used. For example, calibration scans can be used to determine the distortions, in which a phantom with a grid composed of magnetically active samples at predetermined positions is acquired. The deviation of the determined positions from the positions predefined by the grid results in the distortions, which can be given in the form of a matrix. To perform the correction, the inverse function in the form of an inverse matrix can be determined therefrom.

[0065] For the correction in step S40, for example, the image points or voxels in the position space can be determined by image reconstruction in the k-space based on the magnetic resonance data of the first layer 71 or the second layer 72 collected, for example, based on three-dimensional Fourier transform. However, their coordinates undergo distortions, which are caused by the magnetic resonance tomography device 1 and can be corrected by applying the inverse matrix to the coordinates.

[0066] In order to better meet the real-time requirements when tracking the medical device 60, image acquisition must be performed at a high rate. Therefore, only very little time is available for signal integration, and thus the SNR is low. This applies in particular when a large number of points are to be acquired in the k-space to achieve a high position resolution. However, the SNR can be improved by averaging over multiple image points or voxels or by binning.

[0067] However, when the underlying voxels in the image space are large, there are relatively large errors in the correction of the geometry. Therefore, in one embodiment of the method according to the invention, it is proposed to first perform the correction on the data in the image space with the highest possible resolution, even if these data are noisy, in order to achieve better spatial correction with a higher resolution, even if the data do not initially provide a better display of the medical device 60 for the observer. According to this embodiment, averaging is applied after the geometrically corrected voxels or image points in step S50 to improve the SNR. Here, the spatial density of the image points can remain unchanged, but binning can also be envisaged, by which the spatial resolution is reduced.

[0068] In one embodiment of the method according to the invention, it can also be envisaged that the trajectory 101 and / or the medical device 60 itself is geometrically curved, i.e., not part of a straight line. From the perspective of the magnetic resonance tomography device 1 or its controller 23, the only difference between the bending of the trajectory caused by image distortion and the geometric bending of the trajectory 101 or the medical device is that this characteristic is not determined by the magnetic resonance tomography device 1 and the sequence. If the bending is known, the bending can be corrected in a separate step as already described for the correction in step S40. For the linearization of the bending, for example, a linear mapping in matrix form can be envisaged. Here, it can also be envisaged to jointly perform the correction and the linearization in step S40, for example, in such a way that in step S40, the inverse matrix described for S40 is multiplied by the matrix for linearization and then applied to the coordinates of the image points or voxels in the magnetic resonance acquisition.

[0069] The bending of the medical device 60 or the trajectory 101 is not known in advance to the controller 23, and in particular, the bending of the medical device 60 or the trajectory 101 is not constant for the corresponding magnetic resonance tomography device 1 and the sequence.

[0070] In order to determine the thicknesses and orientations of the first layer 71 and the second layer 72 in steps S20 and S30, for example, the trajectory can be pre-given from the outside, for example, it can be known through treatment planning and the user input of the controller 23 or electronic data transmission.

[0071] It is also conceivable that the controller 23 determines the orientation of the medical device in such a way that the controller 23, for example, performs a scan and acquires the position and orientation of the medical device 60 by segmentation. It is also conceivable to use the images of the first layer in S20 and / or the second layer in S30 for adjusting the layer thickness and / or the orientation for subsequent repetitions of steps S20 and / or S30.

[0072] In a possible embodiment, in step S60, the controller 23 records a third layer in a three-dimensional acquisition using the magnetic resonance tomography device 1. The third layer is acquired in a so-called through-plane Aufnahme, in other words, the trajectory 101 or the medical device 60 penetrates the third layer in the direction of the layer thickness, that is, substantially in the direction of the plane normal of the layer. For example, the third layer can be arranged orthogonally to the first layer 71 and to the second layer. Thus, the third layer can advantageously provide a three-dimensional image with a higher resolution of the environment of the trajectory 101. The third layer can be acquired with a low repetition rate because the change in another environment of the trajectory 101 is small and is not compulsorily required for real-time tracking of the medical device 60.

[0073] Preferably, in step S70 when generating a two-dimensional image, this is carried out based on the third layer, in particular for another environment of the trajectory 101. For example, organs and blood vessels can be segmented based on the image data of the first layer 71 and the second layer 72, and the corresponding image data can be faded in correspondingly based on the image of the third layer.

[0074] Although the present invention has been further illustrated and described in detail by the preferred embodiments, the present invention is not limited to the disclosed examples, and those skilled in the art can derive other variants therefrom without departing from the protection scope of the present invention.

Claims

1. A magnetic resonance tomography device for imaging a medical device on a trajectory (101) in real time, wherein, The magnetic resonance tomography device (1) has a controller (23), wherein the controller (23) is designed to acquire a first layer (71) by means of three-dimensional encoding, wherein the controller selects the position of the first layer (71) according to the trajectory during imaging and selects the layer thickness according to a predetermined distortion of the magnetic resonance tomography device (1), such that the trajectory (101) is completely acquired by the first layer (71); the controller (23) is designed to acquire a second layer (72) by means of three-dimensional encoding, wherein the controller (23) selects the position of the second layer (72) according to the trajectory (101) during imaging and selects the layer thickness according to a predetermined distortion of the magnetic resonance tomography device (1), such that the trajectory (101) is completely acquired by the second layer (72); wherein the controller (23) is designed to select the positions of the first layer (71) and the second layer (72) such that the first layer (71) and the second layer (72) enclose an angle greater than 20 degrees, and the trajectory (101) of the medical device (60) and / or the medical device (60) is located in the intersection volume (78) of the first layer (71) and the second layer (72); the controller (23) is designed to generate a sequence of two-dimensional images of the medical device (60) in a predetermined plane, wherein the sequence has images generated by the first layer (71) and images generated by the second layer (72); and the controller (23) is designed to output the sequence of two-dimensional images to the user.

2. The magnetic resonance tomography device according to claim 1, wherein, The controller (23) of the magnetic resonance tomography device (1) is designed to correct a predetermined distortion of the first layer (71) and / or the second layer (72).

3. The magnetic resonance tomography device according to claim 2, wherein, The controller (23) is further designed to average the first layer (71) and / or the second layer (72) after the correction.

4. The magnetic resonance tomography device according to claim 3, wherein, The controller (23) is further designed to perform the correction at a resolution higher than the resolution of the two-dimensional images.

5. A magnetic resonance tomography device according to any one of the preceding claims, wherein, The controller (23) is further designed to acquire a third layer containing the medical device by means of three-dimensional encoding, wherein the controller (23) is designed to generate an image according to the third layer when generating the sequence of two-dimensional images.

6. The magnetic resonance tomography device according to any one of the preceding claims, wherein, The trajectory (101) is a curved trajectory (101).

7. The magnetic resonance tomography device according to claim 6, wherein, The controller (23) is designed to linearize the curved trajectory (101) when generating the two-dimensional image.

8. A method for imaging a medical device (60) on a trajectory (101) in real time by means of a magnetic resonance tomography device (1), wherein, The magnetic resonance tomography device (1) has a controller (23), and the method has the following steps: (S20)Acquire a first layer (71) by means of three-dimensional encoding using the magnetic resonance tomography device (1), wherein the controller (23) selects the position of the first layer (71) according to the trajectory (101) during imaging and selects the layer thickness according to the predetermined distortion of the magnetic resonance tomography device (1), such that the trajectory (101) is fully acquired by the first layer (71); (S30)Acquire a second layer (72) by means of three-dimensional encoding using the magnetic resonance tomography device (1), wherein the controller (23) selects the position of the second layer (72) according to the trajectory (101) during imaging and selects the layer thickness according to the predetermined distortion of the magnetic resonance tomography device (1), such that the trajectory (101) is fully acquired by the second layer (72); Wherein the positions of the first layer (71) and the second layer (72) are selected such that the first layer (71) and the second layer (72) enclose an angle greater than 20 degrees, and the trajectory (101) of the medical device (60) and / or the medical device (60) is located in the intersection volume of the first layer (71) and the second layer (72); (S70)The controller (23) generates a sequence of two-dimensional images of the medical device (60) in a predetermined plane, wherein the sequence has images generated by the first layer (71) and images generated by the second layer (72); (S80)Output the sequence of two-dimensional images to the user.

9. The method according to claim 8, wherein The method further has a step (S40) for correcting the predetermined distortion of the first layer (71) and / or the second layer (72) by the controller (23).

10. The method according to claim 9, wherein, The method further has a step (S50) for averaging the first layer (71) and / or the second layer (72) by the controller (23) after the correction.

11. The method according to claim 10, wherein, The correction step (S40) is performed at a resolution higher than the resolution of the two-dimensional images.

12. The method according to any one of claims 8 to 11, wherein, The method further has a step (S60) for acquiring a third layer containing the medical device by means of three-dimensional encoding using the magnetic resonance tomography device (1), wherein in the step of generating the sequence of two-dimensional images, images are generated according to the third layer.

13. The method according to any one of claims 8 to 12, wherein The trajectory (101) is a curved trajectory (101).

14. The method according to claim 13, wherein, In the step of generating the two-dimensional images, the curved trajectory (101) is linearized.

15. A computer program product having instructions which, when executed by the magnetic resonance tomography device (1) according to any one of claims 1 to 7, cause the magnetic resonance tomography device (1) to perform the method according to any one of claims 8 to 14.

Images