DEVICE AND METHOD FOR REAL-TIME 3D DISTORTION CORRECTION OF MAGNETIC RESONANCE IMAGES

Description

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

[0002] The invention relates to a magnetic resonance imaging scanner and a method for real-time tracking of a medical instrument in the context of interventional MR imaging.

[0003] In minimally invasive medical procedures, medical instruments, such as catheters and / or surgical needles, are inserted into a patient, typically with image monitoring. This image monitoring allows for the acquisition of images that visualize the medical instrument in relation to its anatomical surroundings. While X-ray imaging, particularly fluoroscopy, has traditionally been used for image monitoring in minimally invasive medical procedures, the use of magnetic resonance imaging (MRI) has also been proposed. This is typically referred to as interventional MRI.

[0004] Most medical instruments are made of metal or other materials that cannot be directly imaged using magnetic resonance imaging (MRI) of protons in the body's water or fat. Instead, indirect imaging techniques are used, which utilize indirect effects or artifacts. However, real-time tracking requires short scan intervals, resulting in poor image quality. Furthermore, geometric distortions are common in MRI, which, unlike the geometrically accurate imaging provided by X-rays, make tracking these thin instruments difficult.

[0005] From the article Elgort DR et al., "A Review of Technical Advances in Interventional Magnetic Resonance Imaging", Academic Radiology, 2005, Vol. 12, No. 9, pages 1089-1099, a method for imaging a medical instrument in real time is known, whereby first and second mutually orthogonal layers are recorded.

[0006] It is therefore an object of the present invention to improve the tracking of a medical instrument during an intervention.

[0007] The problem is solved by a magnetic resonance tomograph according to claim 1 and a method for operating the magnetic resonance tomograph according to claim 8.

[0008] The magnetic resonance imaging (MRI) scanner according to the invention is designed to image a medical instrument along a trajectory in real time. The MRI scanner has a control unit that coordinates the activities required for image acquisition. The control unit can also be divided into subunits, for example, a gradient and high-frequency control unit for controlling processes in real time, a user interface for interaction with an operator, and a high-performance image reconstruction computer.

[0009] The control system is designed to operate the magnetic resonance imaging (MRI) scanner in such a way that a first layer is acquired using 3D encoding. A layer, also referred to as a "slab," is defined as a three-dimensional volume that also has a thickness in the sense of the MRI image. This means it does not consist of just a single two-dimensional arrangement of MRI voxels in a plane, but rather of at least two layers of voxels stacked on top of each other, preferably several layers. This allows the medical instrument to be fully captured despite image distortions.

[0010] The control system is designed to select the position of the first slice based on the trajectory and the slice thickness based on a predetermined distortion of the magnetic resonance imaging (MRI) scanner during imaging, or to adjust the parameters for acquiring the first slice so that the trajectory or the medical instrument is completely captured by the first slice. In this context, "position" encompasses both the location and orientation of the slice. The trajectory can be predetermined by the intervention itself, for example, by coordinates determined during planning. However, it is also conceivable that the trajectory was determined beforehand, for instance, by capturing the position and / or orientation of the medical instrument through imaging.Knowing the position of the trajectory and the distortions of the image caused by the magnetic resonance imaging (MRI) scanner, the controller can define a volume, for example, a cuboid of voxels, within which the trajectory lies, taking the distortion into account. Preferably, the slice thickness is designed such that the slice only encompasses the trajectory, i.e., it has no or only a few layers of voxels that are not penetrated by the trajectory or that do not contain voxels belonging to the trajectory, in order to accelerate 3D encoding.

[0011] Furthermore, the control system is designed to capture a second layer in the described manner using 3D encoding, whereby the control system selects a position of the second layer depending on the trajectory and the layer thickness depending on a predetermined distortion of the magnetic resonance tomograph during imaging, such that the trajectory is completely captured by the second layer.

[0012] Furthermore, the controller is designed to select the positions of the first and second layers such that they enclose an angle greater than 20, 40, or 60 degrees, but less than or equal to 90 degrees, and that the trajectory of the medical instrument and / or the medical instrument itself lies within an intersection or volume of the first and second layers. For example, the controller could rotate the coordinates of the first layer around the trajectory by the specified angle to obtain the position of the second layer.

[0013] Each layer also provides an image of the surrounding area of ​​the trajectory in the direction of its lateral or planar extent. Preferably, the extent of the layer corresponds to the orientation of the two-dimensional images that are subsequently displayed to the user for navigation, so that the surrounding area can also be shown for orientation. By displaying images from both layers, preferably by combining image data from both layers, a high refresh rate for the trajectory can advantageously be achieved while simultaneously displaying the surrounding area.

[0014] Finally, the controller is designed to generate a sequence of two-dimensional images of the medical instrument in a predetermined plane. The trajectory, and thus the medical device, lies within this plane. The images are generated from the first and second layers, or, in the case of repetition, from multiple first and second layers. Preferably, the predetermined plane is parallel to the first or second layer, but it can also form a small angle with the first or second layer to represent as much of the trajectory's surroundings as possible. Preferably, the controller also generates and displays a further image whose display plane is parallel to, or at a small angle to, the other layer, thus enabling the representation of the trajectory or tracking of the medical instrument from different perspectives.

[0015] Furthermore, the control system is designed to output the sequence of two-dimensional images to a user, for example, on a screen. The output occurs as immediately as possible after the image is generated to ensure a real-time or near-real-time representation of the medical instrument or trajectory. Specifically, the acquisition of a first layer, the generation of the two-dimensional image in the plane, and its output preferably occur before or while the second layer is acquired and its two-dimensional image is generated.

[0016] In a preferred embodiment, two-dimensional images are alternately generated and displayed from images of the first layer and the second layer in order to create the trajectory from different perspectives or viewing directions, essentially perpendicular to the respective layer. Preferably, the area of ​​the images corresponding to the intersection or intersection volume of the two layers is updated or displayed with each acquisition of the first and second layers.

[0017] Advantageously, the acquisition of the two slices, oriented at an angle to each other, with the trajectory within the slices, allows deviations of the medical instrument's movement from its trajectory in space to be quickly detected and displayed to the user, including the surrounding environment. In particular, the image refresh rate for the trajectory in the intersection of the slices is doubled compared to a magnetic resonance imaging (MRI) scanner that alternately acquires slices arranged at an angle to each other to track movement three-dimensionally, but then only produces the two-dimensional image from the slice that is essentially parallel to the imaging plane.The area where both layers intersect the trajectory is therefore preferably updated with each acquisition of one of the two layers in the image that is aligned parallel to the other layer, thus doubling the frame rate for this area.

[0018] The method according to the invention shares the advantages of the device according to the invention.

[0019] Further advantageous embodiments are specified in the dependent claims.

[0020] In one possible embodiment of the magnetic resonance imaging (MRI) scanner according to the invention, the scanner's control system is designed to correct the predetermined distortion of the first and / or second slice. Distortion in an image acquired by the MRI scanner can be caused by inhomogeneities in the static magnetic field, gradient fields, and dynamically by eddy currents, and is therefore predetermined by the MRI scanner. The predetermined distortion can be determined for the specific MRI scanner model through calculations during development or test measurements. It is also conceivable that the predetermined distortion is determined individually for each device during installation or prior to image acquisition through calibration measurements.For the determined distortion as a linear mapping, for example defined by a matrix, there is a corresponding inverse function that can be determined from the predetermined distortion using the means of linear algebra in order to correct the captured first and / or second layer in such a way that the trajectory or the medical device is represented geometrically correctly, in particular in the case of a straight or linear trajectory, representing it as a segment of a straight line.

[0021] In an advantageous way, rectification corrects the geometry of the captured 3D data and thus also of the resulting 2-dimensional image, so that the user is offered an undistorted representation of the medical instrument and the trajectory, thereby facilitating intuitive navigation of the instrument.

[0022] In one conceivable embodiment of the magnetic resonance imaging (MRI) scanner according to the invention, the MRI scanner is further designed to perform an averaging of the first and / or second slice after rectification. Averaging, as used in the invention, means that by averaging over several adjacent voxels of the 3D MRI scan, the signal-to-noise ratio (SNR) of the image acquisition is improved. In the simplest case and according to the original meaning of the term, averaging is binning, a process of combining adjacent voxels into larger virtual voxels by summing their values. However, this simultaneously reduces the spatial resolution. With the reduced resolution, the rectification result is significantly worse. Other forms of averaging, such as weighted summation, are also conceivable.

[0023] In one possible embodiment of the magnetic resonance tomograph according to the invention, the control system is therefore further designed to perform the rectification with a resolution greater than the resolution of the subsequent two-dimensional image.

[0024] Advantageously, in the magnetic resonance tomograph according to the invention, rectification is performed before the resolution is reduced by averaging or binning, thus enabling a more precise representation of the position of the medical instrument. Averaging is then performed subsequently to improve the signal-to-noise ratio (SNR) of the corrected image. Since averaging itself is not a reversible process, the order is crucial and affects the result. Higher resolution magnetic resonance data allows for better rectification. The reduced SNR achieved with the higher resolution at the same measurement time can then be improved by the subsequent averaging.

[0025] In a conceivable embodiment of the magnetic resonance tomograph according to the invention, the control system is further configured to acquire a third layer containing the medical instrument with a 3D encoding. Preferably, the third layer is acquired with a higher resolution and / or longer integration time than the first and / or second layer, so that, in particular, the environment of the medical instrument can be acquired and displayed more accurately. The control system is designed to generate the sequence of two-dimensional images depending on the third layer. Preferably, the acquisition of the third layer is used to reproduce the environment of the trajectory with higher resolution and / or better signal-to-noise ratio in the two-dimensional image.The environment surrounding the trajectory hardly changes during the handling of the medical instrument, so it is conceivable to use data from the third-layer acquisition for representing the environment in the 2D image, especially in several successive images, without having to repeat or update the third-layer acquisition in between. It is also conceivable to use segmentation to capture the position of individual organs or vessels in the first and / or second layer and to correct their position in the third-layer acquisition accordingly for the 2D image.

[0026] The advantageous combination of high-quality, low-refresh-rate imaging of the environment with rapid first and second layer imaging using the medical instrument allows for real-time tracking of the instrument at high resolution of surrounding organs.

[0027] In one possible embodiment of the magnetic resonance tomograph according to the invention, the trajectory captured by the first and second layers is a curved trajectory, or the medical instrument is not essentially linear, but curved. As already explained with regard to distortion caused by image acquisition by the magnetic resonance tomograph, it is also conceivable that, by appropriately selecting the first and second layers, not only is a linear instrument distorted by the magnetic resonance imaging completely captured, but, given knowledge of the geometry and position of the trajectory or the medical instrument, a curved instrument or a curved trajectory is also completely captured. For this purpose, the cuboid that bounds the first and / or second layer is selected such that it encompasses the curved instrument or the curved trajectory.the curved trajectory fully includes, taking into account the distortions of the magnetic resonance imaging scanner.

[0028] Advantageously, the magnetic resonance tomograph and the method according to the invention are able to track a curved medical instrument or a corresponding instrument on a curved trajectory in real time.

[0029] In one conceivable embodiment of the magnetic resonance imaging (MRI) scanner according to the invention, the control system is designed to linearize the curved trajectory when generating the two-dimensional image. It is conceivable that this linearization corrects or linearizes the distortion caused by the MRI scanner and / or the geometric curvature of the trajectory or the medical instrument. The correction can be performed in a plane or in space. For example, it is conceivable that the linearization takes place in a plane perpendicular to the plane of the two-dimensional image.

[0030] Advantageously, the magnetic resonance tomograph and the method according to the invention also enable curved medical instruments or bendable medical instruments on a curved trajectory to be quickly and completely displayed to a user in a 2-dimensional image and to improve navigation with the medical instrument.

[0031] The properties, features and advantages of this invention described above, as well as the manner in which they are achieved, will become clearer and more easily understood in connection with the following description of the exemplary embodiments, which are explained in more detail in conjunction with the drawings.

[0032] They show: Fig. 1 a schematic representation of a magnetic resonance tomograph according to the invention with an instrument according to the invention; Fig. 2 a schematic representation of a layer acquired by the magnetic resonance tomograph according to the invention in the method according to the invention; Fig. 3 a schematic representation of a layer acquired by the magnetic resonance tomograph according to the invention in the method according to the invention; Fig. 4 a schematic representation of a relative arrangement of two layers acquired by the magnetic resonance tomograph according to the invention in the method according to the invention; Fig. 5 a schematic flow chart of a method according to the invention.

[0033] Fig. 1 Figure 1 shows a schematic representation of an embodiment of a magnetic resonance tomograph 1 according to the invention, as it is also used to carry out the method according to the invention.

[0034] The magnet unit 10 includes a field magnet 11 that generates a static magnetic field B0 for aligning the nuclear spins of samples or the patient 100 within a recording area. The recording area is characterized by an extremely homogeneous static magnetic field B0, with the homogeneity referring in particular to the magnetic field strength or magnitude. The recording area is nearly spherical and arranged within a patient tunnel 16 that extends longitudinally 2 through the magnet unit 10. A patient bed 30 can be moved within the patient tunnel 16 by the traversing unit 36. The field magnet 11 is typically a superconducting magnet capable of generating magnetic fields with a magnetic flux density of up to 3 T, and even higher in the latest devices. However, for lower magnetic field strengths, permanent magnets or electromagnets with normal-conducting coils can also be used.

[0035] Furthermore, the magnetic unit 10 comprises gradient coils 12, which are designed to superimpose temporally and spatially variable magnetic fields in three spatial directions onto the magnetic field B0 in order to spatially differentiate the captured imaging areas within the investigation volume. The gradient coils 12 are typically coils made of normally conducting wires that can generate mutually orthogonal fields within the investigation volume.

[0036] The magnet unit 10 also has a body coil 14, which is designed to radiate a high-frequency signal supplied via a signal line into the examination volume and to receive 100 resonance signals emitted by the patient and to transmit them via a signal line.

[0037] A control unit 20 supplies the magnet unit 10 with the various signals for the gradient coils 12 and the body coil 14 and evaluates the received signals.

[0038] The control unit 20 thus has a gradient control 21 which is designed to supply the gradient coils 12 with variable currents via supply lines, which provide the desired gradient fields in the investigation volume in a time-coordinated manner.

[0039] Furthermore, the control unit 20 includes a high-frequency unit 22 designed to generate a high-frequency pulse with a predefined temporal profile, amplitude, and spectral power distribution to excite magnetic resonance of the nuclear spins in the patient 100. Pulse powers in the kilowatt range can be achieved. The excitation signals can be radiated into the patient 100 via the body coil 14 or via a local transmitting antenna.

[0040] A controller 23 communicates via a signal bus 25 with the gradient controller 21 and the high-frequency unit 22.

[0041] A medical instrument 60, for example a biopsy needle, is arranged in a patient tunnel 16 on a patient 100 or is inserted into the patient 100 during a procedure and detected by means of the magnetic resonance tomograph according to the invention and displayed to a user on an output device such as a screen 24.

[0042] Both the static B0 magnetic field of the field magnet 11 and the gradient fields of the gradient coils 12 exhibit inhomogeneities or nonlinearities, which lead to spatial distortions in the imaging.

[0043] In Fig. 2 A first layer 71 of an image acquisition system according to the invention with an inscribed trajectory 101 is schematically represented. Fig. 2The real, undistorted geometry is depicted; that is, a medical instrument 60, such as a biopsy needle, which geometrically represents essentially a straight line, is also depicted here as a straight line without curvature. The first layer 71 of an image acquisition by the magnetic resonance tomograph 1, however, is not a cuboid with straight edges due to the distortions of the magnetic resonance image, but rather curved or bent.

[0044] The distortions experienced by the magnetic resonance imaging (MRI) scanner 1 are essentially determined by the MRI scanner 1 and the sequence used. They can therefore be determined for an individual MRI scanner 1 within a series of identical devices, and with greater accuracy for a specific device and a predetermined sequence, through calibration measurements. This can be done, for example, by imaging a phantom with a grid of magnetic resonance-active samples using the MRI scanner 1 and the corresponding sequence. It is also conceivable to estimate the distortion through simulation.

[0045] The determined distortion maps individual pixels of an object under investigation in real geometry to corresponding voxels in a 3D representation. A corresponding inverse mapping can be determined using linear algebra.

[0046] For a given trajectory 101, the locations of the individual points of the trajectory 101 in the distorted voxel space can be determined. In this way, a cuboid can also be found in this voxel space that completely contains or inscribes this distorted trajectory 101, while simultaneously minimizing the number of voxels to be identified to a limited extent. In most cases, for example, the voxels that correspond to the start and end points of the trajectory in the voxel space can be chosen as diagonally opposite vertices of the cuboid.

[0047] In Fig. 3 In contrast, the view from the magnetic resonance tomography scanner is shown. The first layer 71 is defined by a cuboid that has several parallel layers 75, 76, 77 of voxels in parallel planes. The trajectory 101, which is actually straight in the example, appears curved here.

[0048] It also becomes apparent that, due to the curvature, it is not possible to fully capture the trajectory 101 with only one layer 75, 76, 77 of voxels if the voxels must be small enough to represent the trajectory accurately enough for navigation of the medical instrument. The invention therefore also proposes a 3D capture of the first layer 71 with the exemplary number of three layers 75, 76, 77 of voxels. By limiting the cuboid so that it just completely inscribes the trajectory, the number of voxel values ​​to be captured can be reduced, thus improving the repetition rate for the same signal-to-noise ratio (SNR).

[0049] The individual layers 75, 76, 77 of the layers 71, 72 are preferably acquired in 3D by phase coding. For this purpose, the entire layer is excited and its entire extent is captured in the field of view using 3D phase coding to avoid infolding of voxels outside the cuboid. While this increases the number of voxels and thus lengthens the acquisition time in the phase coding, this disadvantage is offset by the significantly faster excitation of layers 71, 72 compared to selective excitation of the cuboid.

[0050] The depicted, essentially Cartesian coordinate system, although distorted in voxel space, is used here only as an example and for the sake of clarity. Due to its intuitive usability for user navigation, it is preferably used for capturing the layers and thus the images generated from them. However, the invention can also be applied to other coordinate systems.

[0051] In Fig. 4 A first layer 71 and a second layer 72 are shown. For the sake of clarity, the individual layers 75, 76, 77 of each layer are not shown separately. The first layer 71 and the second layer 72 have an intersection volume 78, which is located inside both layers 71, 72. The first layer 71 and the second layer 72 are arranged at an angle to each other; in other words, normal vectors of the planar surfaces, i.e., not the edge surfaces, enclose this angle. The angle can preferably be 90 degrees, i.e., the first layer 71 can be 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 each have significantly different orientations with respect to the trajectory 101.

[0052] Trajectory 101 is inscribed within this section volume 78, such that trajectory 101 is captured both when acquiring the first layer 71 and when acquiring the second layer 72. In this way, the controller 23 can update the position of the medical instrument 60 along trajectory 101 in the images from two different perspectives parallel to layers 71 and 72 at twice the frequency.

[0053] In Fig. 5 Figure 1 shows a schematic flowchart of an embodiment of the method according to the invention. The method is carried out on a magnetic resonance imaging scanner 1 according to the invention.

[0054] In step S20, a first layer 71 is acquired by the magnetic resonance imaging (MRI) scanner 1 using 3D encoding under the control of the controller 23. The first layer 71 has multiple layers 75, 76, 77 in which voxels 79 are arranged. The position (i.e., location and orientation) and the layer thickness of the first layer 71 are selected by the controller 23 based on the trajectory 101 and a predetermined distortion of the MRI scanner 1 during imaging, such that the trajectory 101 is completely captured by the first layer 71. Reference is also made to the explanations regarding Fig. 2 and Fig. 3 referred.

[0055] The predetermined distortion can also be determined in advance by the controller 23 through a calibration measurement in a step S10, for example with a phantom with a grid of MR-active samples at predetermined positions.

[0056] In step S30, a second layer 72 is acquired by the magnetic resonance imaging (MRI) scanner 1 using 3D encoding under the control of the controller 23. The position of the second layer 72 and the layer thickness are also selected depending on the trajectory 101 and the predetermined distortion of the MRI scanner 1 such that the trajectory 101 is completely captured by the second layer 72.

[0057] The orientation of the first layer 71 and the second layer 72 relative to this is as shown in Fig. 4 as shown and explained, selected by the control 23 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 instrument 60 and / or the medical instrument 60 lies in an intersection volume 78 of the first layer 71 and the second layer 72.

[0058] In step S70, the controller 23 generates a sequence of two-dimensional images of the medical instrument 60 in a predetermined plane. For example, a cross-sectional image through the reconstructed voxels 79 of the section volume 78 along this plane or a projection onto this plane is conceivable.

[0059] The sequence of images includes images generated from the first layer 71 and images generated from the second layer 72. The representation plane of each image is essentially parallel to the layer from which it is generated. Preferably, the first layer 71 is acquired in step S20, alternating with the second layer S72 in step S30. Preferably, the cross-sectional image is also generated directly from the acquired magnetic resonance data of the respective layer 71, 72, before or while the other layer 71, 72 is acquired. This allows the images to be generated as quickly as possible. Preferably, the parts of the images representing the cross-sectional volume 78 are derived from the acquired data.When voxels of the first layer 71 and the second layer 72 are updated, the representations of the medical instrument in the two figures can be updated at twice the repetition rate.

[0060] In step S80, the two-dimensional images are then output by the controller 23 to a user, for example, on a screen 24. The output occurs as soon as possible after the image is acquired and generated from the respective layer 71, 72. The sequence of steps preferably involves acquiring (S20) a first layer 71, generating (S70) the image of the first layer 71, and immediately outputting (S80) the image of the first layer before or while (S30) the second layer 72 is acquired, a two-dimensional image from the second layer 72 is generated in step S70, and immediately output in step S980. Preferably, this sequence of acquiring (S20, S30), generating two-dimensional images (S70), and outputting (S80) is repeated for a predetermined duration, such as the duration of a control signal (e.g., from a foot switch), or for the duration of an intervention.

[0061] In one possible embodiment of the method according to the invention, in step S40 the acquired magnetic resonance data of the first layer 71 or the second layer 72 are geometrically rectified by the controller. The magnetic resonance tomograph 1 exhibits geometric distortions in the imaging of acquired objects. These distortions can be caused by inhomogeneities of the static B0 field, nonlinearities of the gradient fields, or by eddy currents generated by the gradients.

[0062] The distortions are characteristic or predetermined for a specific magnetic resonance imaging (MRI) scanner and the sequences used. These distortions can be determined, for example, with a calibration scan in which a phantom with a grid of MR-active samples is scanned at predetermined positions. The deviation of the determined position from the position defined by the grid yields the distortion, which can be expressed as a matrix. To correct the distortion, an inverse function in the form of an inverse matrix can be derived from this matrix.

[0063] For rectification in step S40, for example, image points or voxels in spatial space can be determined from the acquired magnetic resonance data of the first layer 71 or the second layer 72 in k-space by image reconstruction, for example based on a 3D Fourier transform. However, their coordinates are subject to the distortions caused by the magnetic resonance tomograph 1 and can be corrected by applying the inverse matrix to the coordinates.

[0064] To better meet real-time requirements when tracking medical instrument 60, image acquisition must occur at a high rate, leaving little time for signal integration and thus resulting in a low signal-to-noise ratio (SNR). This is especially true when a large number of points in k-space need to be acquired for high spatial resolution. However, the SNR can be improved by averaging or binning across multiple pixels or voxels.

[0065] However, geometric rectification is subject to a greater error when the underlying voxels in the image space are larger. Therefore, in one embodiment of the inventive method, it is proposed to first perform the rectification on data in the image space that has the highest possible resolution, even if it is noisy, in order to achieve better spatial correction with higher resolution, even if the data does not initially provide a better representation of the medical instrument 60 to a viewer. According to this embodiment, an averaging process is then applied to the geometrically rectified voxels or pixels in a step S50 to improve the signal-to-noise ratio (SNR). The spatial density of the pixels can remain unchanged, but binning, which reduces the spatial resolution, is also conceivable.

[0066] In one embodiment of the method according to the invention, it is also conceivable that the trajectory 101 and / or the medical instrument 60 itself are geometrically curved, i.e., not part of a straight line. From the perspective of the magnetic resonance imaging (MRI) scanner 1 or its control unit 23, the difference between a curvature of the trajectory due to distortion of the image and a geometric curvature of the trajectory 101 or the medical instrument 101 is merely that this property is not determined by the MRI scanner 1 and the sequence. If the curvature is known, it can be corrected in a separate step, as already explained for rectification in step S40. For example, a linear mapping in the form of a matrix is ​​conceivable for linearizing the curvature.It is also conceivable that the rectification from step S40 and the linearization are performed together in one step, for example by multiplying the inverse matrix described in S40 with the matrix for linearization and then applying it to the coordinates of the pixels or voxels from the MR acquisition in step S40.

[0067] The curvature of the medical instrument 60 or the trajectory 101 for the control 23 is not known a priori, in particular it is not a constant for the respective magnetic resonance tomograph 1 and the sequence.

[0068] In order to determine the thickness and orientation of the first layer 71 and the second layer 72 in steps S20 and S30, the trajectory can, for example, be specified externally, e.g. by a treatment plan, and be known by user input or electronic data transmission to the controller 23.

[0069] It is also conceivable that the controller 23 determines the position of the medical instrument by, for example, performing a scan and using segmentation to capture the position and orientation of the medical instrument 60. It is also conceivable to use the images of the first layer S20 and / or the second layer in S30 to adjust the layer thickness and / or position for a subsequent repetition of steps S20 and / or S30.

[0070] In one possible embodiment, in step S60, a third layer is acquired in a 3D scan by the controller 23 using the magnetic resonance imaging (MRI) scanner 1. The third layer is acquired in a so-called through-plane scan; in other words, the trajectory 101, or the medical instrument 60, penetrates the third layer in the direction of the layer's thickness, essentially in the direction of the layer's surface normal. For example, the third layer can be arranged orthogonally to the first layer 71 and to the second layer. The third layer can thus advantageously provide a higher-resolution 3D image of the trajectory 101's surroundings. The third layer can be acquired at a low repetition rate because the surroundings of the trajectory 101 change only slightly and are not essential for real-time tracking of the medical instrument 60.

[0071] Preferably, in step S70, when generating the two-dimensional image, this is carried out depending on the third layer, in particular for a further neighborhood of the trajectory 101. For example, organs or vessels could be segmented from the image data of the first layer 71 and second layer 72 and corresponding image data from the image of the third layer could be superimposed accordingly.

[0072] Although the invention has been illustrated and described in detail by the preferred embodiment, the invention is not limited by the disclosed examples and other variations can be derived by the person skilled in the art without leaving the scope of protection of the invention.

Claims

1. Magnetic resonance tomography device for mapping a medical instrument 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 capture a first slice (71) by means of 3D encoding, wherein the controller selects the position of the first slice (71) as a function of the trajectory and the slice thickness as a function of a predetermined distortion of the magnetic resonance tomography device (1) during mapping, so that the trajectory (101) is captured in full by the first slice (71); the controller (23) is designed to capture a second slice (72) by means of 3D encoding, wherein the controller (23) selects a position of the second slice (23) as a function of the trajectory (101) and the slice thickness as a function of a predetermined distortion of the magnetic resonance tomography device (1) during mapping, so that the trajectory (101) is captured in full by the second slice (72); wherein the controller (23) is designed to select the position of the first slice (71) and the second slice (72) such that the first slice (71) and the second slice (72) confine an angle greater than 20 degrees, and the trajectory (101) of the medical instrument (60) and / or the medical instrument (60) lies in an intersecting volume (78) of the first slice (71) and the second slice (72); the controller (23) is designed to generate a sequence of two-dimensional images of the medical instrument (60) in a predetermined plane, wherein the sequence contains images that were generated from the first slice (71) and images that were generated from the second slice (72); and the controller (23) is designed to output the sequence of two-dimensional images to a user.

2. Magnetic resonance tomography device according to claim 1, wherein the controller (23) of the magnetic resonance tomography device (1) is designed to perform an equalisation of the predetermined distortion of the first slice (71) and / or the second slice (72).

3. Magnetic resonance tomography device according to claim 2, wherein the controller (23) is further designed to perform averaging of the first slice (71) and / or second slice (72) after the equalisation.

4. Magnetic resonance tomography device according to claim 3, wherein the controller (23) is further designed to perform the equalisation with a resolution greater than a resolution of the two-dimensional image.

5. Magnetic resonance tomography device according to one of the preceding claims, wherein the controller (23) is further designed to capture a third slice containing the medical instrument with 3D encoding, wherein the controller (23) is designed to generate the images as a function of the third slice during generation of the sequence of two-dimensional images.

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

7. Magnetic resonance tomography device according to claim 6, wherein the controller (23) is designed to perform a linearisation of the curved trajectory (101) during generation of the two-dimensional image.

8. Method for mapping a medical instrument (60) in real time by means of a magnetic resonance tomography device (1) on a trajectory (101), wherein the magnetic resonance tomography device (1) has a controller (23) and the method contains the steps: (S20) capture of a first slice (71) by the magnetic resonance tomography device (1) by means of 3D encoding, wherein a position of the first slice (71) is selected by the controller (23) as a function of the trajectory (101) and the slice thickness as a function of a predetermined distortion of the magnetic resonance tomography device (1) during mapping, so that the trajectory (101) is captured in full by the first slice (71), (S30) capture of a second slice (72) by means of 3D encoding by the magnetic resonance tomography device (1), wherein a position of the second slice (72) is selected by the controller (23) as a function of the trajectory (101) and the slice thickness as a function of a predetermined distortion of the magnetic resonance tomography device (1) during mapping, so that the trajectory (101) is captured in full by the second slice (72), wherein the position of the first slice (71) and the second slice (72) is selected such that the first slice (71) and the second slice (72) confine an angle greater than 20 degrees, and the trajectory (101) of the medical instrument (60) and / or the medical instrument (60) lies in an intersecting volume of the first slice (71) and the second slice (72); (S70) generation of a sequence of two-dimensional images of the medical instrument (60) in a predetermined plane by the controller (23), wherein the sequence contains images that were generated from the first slice (71) and images that were generated from the second slice (72); (S80) output of the sequence of two-dimensional images to a user.

9. Method according to claim 8, wherein the method further contains the step (S40) of the performance by the controller (23) of an equalisation of the predetermined distortion of the first slice (71) and / or the second slice (72).

10. Method according to claim 9, wherein the method further contains the step (S50) of the performance by the controller (23) of averaging of the first slice (71) and / or second slice (72) after the equalisation.

11. Method according to claim 10, wherein the step (S40) of equalisation takes place with a resolution greater than a resolution of the two-dimensional image.

12. Method according to one of claims 8 to 11, wherein the method further contains the step (S60) of capturing a third slice containing the medical instrument with 3D encoding by the magnetic resonance tomography device (1), wherein in the step of the generation of a sequence of two-dimensional images the images are generated as a function of the third slice.

13. Method according to one of claims 8 to 12, wherein the trajectory (101) is a curved trajectory (101).

14. Method according to claim 13, wherein in the step of generation of the two-dimensional image a linearisation of the curved trajectory (101) takes place.

15. Computer program product with commands which during execution by a magnetic resonance tomography device (1) according to one of claims 1 to 7 cause the magnetic resonance tomography system (1) to perform a method according to one of claims 8 to 14.