Method for assisting a user in executing an examination workflow of a magnetic resonance examination

Abstract

The invention relates to a method for assisting a user in performing a workflow of a magnetic resonance examination on a patient, comprising the following method steps: - Input of a request by the user, whereby the input is formulated in text form or as voice input, - Determining output information corresponding to the query using a Large Language Model (LLM) and providing the output information, - Output of the output information in text form or as speech output.

Description

[0001] The present invention relates to a method for assisting a user in performing a workflow for a magnetic resonance imaging (MRI) examination on a patient. The present invention further relates to a computing unit and a system comprising a computing unit and a magnetic resonance imaging (MRI) device, wherein the computing unit is configured to implement the method for assisting a user in performing a workflow for a magnetic resonance imaging (MRI) examination on a patient. The present invention also includes a corresponding computer program and a data carrier containing the computer program.

[0002] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included.

[0003] To clarify a specific medical and / or diagnostic question, a magnetic resonance imaging (MRI) scan is often performed on a patient. Numerous options exist for tailoring the MRI scan to the medical and / or diagnostic question. However, this requires extensive expertise from the medical personnel operating the MRI scanner. Even minor changes in the settings can lead to errors in sequence parameter settings. For example, a parameter change can transform a sequence from proton density contrast to a T2-weighted sequence, although the sequence name does not reflect this and may therefore not be apparent to an inexperienced user. Another example is artifacts that can arise from setting an excessively high PAT factor.

[0004] The present invention is based in particular on the objective of assisting a user in the design and / or preparation of a magnetic resonance device. This objective is achieved by the features of the independent claims. Advantageous embodiments are described in the dependent claims.

[0005] The invention relates to a computer-implemented method for supporting a user in performing a workflow of a magnetic resonance examination on a patient, comprising the following method steps: - Input of a request by the user, whereby the input is formulated in text form or as voice input, - Determining output information corresponding to the query using a Large Language Model (LLM) and providing the output information, and - Output of the output information in text form or as speech output.

[0006] The user is preferably a medical operator supervising the magnetic resonance imaging (MRI) examination, such as a physician or an MRI technician. The user, particularly the medical operator, not only supervises the execution of the MRI examination but also prepares it. Preparation includes selecting a measurement program and setting the necessary parameters, especially sequence parameters, to suit the clinical question. This preparation, specifically the selection of a measurement program and the setting of the necessary parameters, can also be carried out well in advance of the actual MRI examination.

[0007] The selection of the measurement program is preferably carried out by means of a selection module of a computing unit and / or a control unit. The computing unit and / or control unit can be integrated into the magnetic resonance device or connected to it via an interface. The computing unit and / or control unit can also have its own user interface for interaction with the user, in particular medical personnel, or it can be connected via the interface unit to a user interface of the magnetic resonance device or to a user interface of other units, such as a mobile device. The selection module preferably includes selection software which, when implemented using a processor of the computing unit and / or control unit, provides a user interface for selecting a measurement program.

[0008] The method according to the invention can be performed both during the preparation of the magnetic resonance examination and during the execution of the magnetic resonance examination. In particular, a user, especially medical personnel, can access the method according to the invention for support at any time during the preparation of the magnetic resonance examination and / or during the execution of the magnetic resonance examination.

[0009] The workflow of a magnetic resonance imaging (MRI) scan comprises both a measurement workflow, which involves performing the MRI scan, and a preparation workflow, which involves preparing the patient for the MRI scan. The preparation workflow includes patient preparation, patient registration, and selection of a measurement program. Patient preparation can involve positioning the patient on the MRI table, applying or positioning local radiofrequency coils on the patient, particularly on the area to be examined, or attaching additional equipment such as an IV infusion unit, an ECG unit, etc.Patient registration may include the input of patient-relevant data that could be used to select a measurement program and / or individual parameter settings. This patient data may include the patient's height, age, medical history, the area of ​​the patient to be examined, and / or other patient data deemed relevant by a specialist. The selection of the measurement program is preferably based on the clinical and / or diagnostic question to be addressed by the magnetic resonance imaging (MRI) examination. Additionally, the selection of the measurement program may also depend on the area of ​​the patient to be examined. The selected measurement program preferably comprises a sequence of different measurement sequences that are executed sequentially in a defined order.This preparation may also include selecting or setting individual parameters for different measurement sequences.

[0010] The measurement workflow for magnetic resonance imaging (MRI) involves executing the measurement program, particularly the sequential execution of the various measurement sequences. This may also require user interaction, especially from medical personnel. For example, a contrast agent infusion may need to be initiated by medical personnel at a specific point in the measurement sequence for a contrast agent measurement. Furthermore, medical personnel may need to assess the image quality of acquired data to determine if it is sufficient for diagnosis. If the image quality is insufficient, the medical personnel must decide whether to abort and restart the current measurement step or to use an alternative measurement strategy for the remainder of the MRI, and so on.

[0011] During all these activities in the preparation workflow and the measurement workflow, medical operating personnel can enter a request for support or if questions or problems arise.

[0012] The output information corresponding to the query is determined by a processing unit that includes a retrieval module with a Large Language Model (LLM). The processing unit containing the retrieval module with the LLM can be the same processing unit that includes the selection module. Alternatively, the processing unit containing the retrieval module with the LLM can be separate from the processing unit containing the selection module. Furthermore, the processing unit containing the retrieval module can also be integrated into the magnetic resonance device (MRI). Or, alternatively, the processing unit containing the retrieval module can be separate from the MRI and connected to the MRI via an interface. For example, the processing unit can also be integrated into a cloud environment.

[0013] The LLM (also known as the large language model) can be trained to analyze user input, particularly user queries, and generate context-sensitive responses. For example, the LLM can analyze queries based on keywords and generate corresponding output information. Alternatively or additionally, the LLM can analyze the query within the context of a current workflow status. For instance, the discovery module can query the current workflow status and provide it to the LLM for determining the output information. Furthermore, when analyzing the query and determining the output information, the LLM can consider previous queries on the same topic, the output information generated from those queries, and / or an evaluation of the output information from previous magnetic resonance imaging (MRI) scans performed by medical personnel.Such queries and their corresponding output information can be stored in a memory unit of the investigation module. The investigation module then provides the determined output information for output. In particular, the output information is provided to an output unit of the user interface.

[0014] The LLM can represent a computational linguistic probability model that has learned statistical word and sentence sequence relationships from a large number of text documents through a computationally intensive training process.

[0015] The input of the query and the output of the information are handled via a user interface, which includes an input unit and an output unit. The input unit can include a keyboard and / or a touchscreen for entering the query in text form. It can also include a microphone for inputting the query as a voice message. The output unit can include a monitor and / or a display for displaying the output information in text form.

[0016] Furthermore, the output unit can also include speakers for outputting a voice message. Information in text form, in this context, is understood to mean that the information is in written and / or legible form.

[0017] The query can also be entered in natural language. In particular, the query can be made without precise knowledge of magnetic resonance imaging (MRI) terminology. For example, one can inquire about the causes of organ outline duplication without using keywords like "aliasing" or "parallel imaging." This results in a much higher success rate in resolving application problems. In particular, even inexperienced medical personnel can easily receive support in this way, even without prior knowledge of the exact terminology.

[0018] For example, a communication window can be displayed on a screen throughout the entire workflow of the magnetic resonance examination to allow for the input of a request. In particular, such a communication window can be displayed above a user interface or as a pop-up window. This communication window preferably includes an input area for the user to enter the request. For example, the request can be entered via text form and / or voice input can be activated. Furthermore, the output information can also be displayed via the communication window. The user can also be notified that the output information is available as a voice message, so that the user can retrieve the output information if needed.

[0019] The invention provides advantageous support for a user, particularly medical personnel, during a magnetic resonance imaging (MRI) examination. For example, medical personnel can be informed, through targeted questioning, about the effects of individual configuration options and the constraints that must be observed.

[0020] In an advantageous further development of the method according to the invention, it can be provided that at least one additional piece of information is provided which is taken into account when determining the output information using the LLM, wherein this includes at least one additional piece of information: - information from the current magnetic resonance examination, - information from at least one previous magnetic resonance imaging (MRI) scan, - a hardware feature of the magnetic resonance device, - a software feature of the magnetic resonance device, and / or - information from the patient.

[0021] The information from the current magnetic resonance imaging (MRI) scan can include current parameter settings for sequence parameters. Furthermore, the information can also include the current workflow status. For example, this can provide context from which the request from the medical operator can be derived, thereby reducing the selection of output information.

[0022] Information from a previous magnetic resonance imaging (MRI) scan can include information from an application in one or more previous MRI scans conducted in a situation identical or similar to the query context. For example, when querying parameter settings, settings from previous MRI scans can be taken into account. Preferably, information from previous MRI scans with the same or similar measurement program is considered.

[0023] The hardware properties of the magnetic resonance device can include, for example, information regarding the high-frequency coils used or intended to be used for magnetic resonance imaging. This includes, for instance, whether the correct high-frequency coils are used for the selected measurement program and whether they are correctly positioned and / or connected. Furthermore, the hardware properties of the magnetic resonance device can also include information about the magnetic field strength and / or the maximum available gradient field strength, etc.

[0024] The software properties of the magnetic resonance device can include information regarding available software licenses. For example, the response to a request to adjust sequence parameters may depend on available measurement options, which may be linked to a software license.

[0025] The patient information may include the patient's weight and / or age and / or height and / or medical history and / or implant information and / or other information and / or patient characteristics that the specialist considers relevant.

[0026] Preferably, the at least one additional piece of information is automatically provided to the LLM by the processing unit, in particular the acquisition module, in order to generate output information based on the current request together with the at least one additional piece of information. The additional information includes existing data and / or data stored and / or stored for the execution of the magnetic resonance examination, which the LLM can access to analyze the request and determine the output information.

[0027] This can increase the likelihood of providing answer information that can contribute to a solution to a current question and / or problem and / or bring about an improvement in a current situation.

[0028] In an advantageous embodiment of the method according to the invention, the output information may include a link, for example, to a defined text passage in stored documentation. When the user activates the link via a user interface, a connection to the stored documentation is established, and the defined text passage is displayed to the user. For example, the documentation may include user documentation detailing the individual steps for setting and / or executing individual measurement steps of a measurement program. Activation can be achieved by clicking with a computer mouse and / or a finger of the medical operator on a display, particularly a touch display, and / or by voice input.By linking to a defined text passage in a documentation, the provided response information to the user, especially to medical operating personnel, can be additionally substantiated and / or supported, thereby creating and / or generating a feeling of security for the medical operating personnel in a critical situation.

[0029] In an advantageous embodiment of the method according to the invention, the output information can include at least one additional indication of the current settings of the magnetic resonance examination. For example, when medical personnel request a sequence parameter setting, they can be shown what the ideal or optimal sequence parameter setting would be and, conversely, what the current sequence parameter setting is. Similarly, the medical personnel can be shown connections between individual sequence parameters if they have a specific setting for one sequence parameter, but this setting is blocked by the value of another sequence parameter.

[0030] In an advantageous embodiment of the method according to the invention, it can be provided that in a further process step, a further input from the user can be entered in response to the output information. In particular, the further input can comprise user feedback, wherein the feedback is provided to the LLM for evaluation of the determined and provided output information. In this way, user feedback regarding the output information generated by the LLM can advantageously be made available to the LLM. The feedback can, for example, include information on how well the output information helped the user with their current specific problem and contributed to solving the problem.In this way, the LLM can be continuously developed and retrained to determine the best possible output information for a user's request. For example, for requests where multiple possible output options are available, the LLM can rank the output information most relevant to solving the current problem or answering the request. This also allows the LLM to dynamically respond to changes, discarding existing output information, such as suggested solutions, or adding new output information.

[0031] For example, output information containing help that did not contribute to solving the user's specific problem and / or only marginally improved the current situation can be negatively evaluated and will no longer be provided as a first solution approach or as the first output information in the event of a subsequent request with the same problem. Conversely, output information containing help that proved helpful and contributed to solving the current problem can be positively evaluated and will again be provided as a first solution approach, and especially as the first output information, in the event of a subsequent request with the same problem.

[0032] In an advantageous embodiment of the method according to the invention, it can be provided that further output information is generated and displayed based on the additional input. For example, if the initial output information is negative, a new output information can be determined and provided. This makes it possible to provide comprehensive support to a user, particularly medical personnel, even if the initial output information with its first proposed solution did not contribute to solving the current problem or clarifying the query. For example, in the case of duplicated structures, if the minimum possible PAT factor has already been used, the additional input can inquire about other measures that could bring about an improvement.

[0033] In an advantageous embodiment of the method according to the invention, the output information may include contact information for a human contact person. This allows the user to contact an expert to resolve their query if no adequate solution has yet been provided. Furthermore, it may also be possible that, in addition to providing and outputting the contact information, a contact with the expert is also established directly.

[0034] Furthermore, the invention relates to a computing unit designed to support a user in performing a workflow of a magnetic resonance examination on a patient, wherein the computing unit comprises: - an interface module configured to connect to a user interface, wherein the user interface has an input unit for a user to input a request, wherein the input is formulated in text form or as speech input, - an investigation module with an LLM trained to determine output information corresponding to the query, and - a provisioning module that is configured to provide the output information to an output unit.

[0035] The advantages of the computing unit according to the invention essentially correspond to the advantages of the method according to the invention for supporting a user in executing a workflow for a magnetic resonance examination on a patient, which are described in detail beforehand. The claims for the computing unit can be improved by features that are described or claimed in connection with the method, and vice versa. In this case, the functional features of the method are embodied by structural units of the computing unit, and vice versa.

[0036] The interface module preferably includes an interface for connecting to a user interface. The user interface includes an input unit for the user to enter a request. The user can formulate the request in text form or as speech output. For this purpose, the input unit has a microphone and / or a keyboard and / or a touchscreen, etc., for entering the request.

[0037] The deployment module preferably has an interface, wherein the interface is configured for a data connection with a user interface. The user interface includes an output unit configured for text or speech output. The output unit preferably includes a display and / or a speaker.

[0038] The provisioning module and the interface module can also be designed as a single unit if an input unit for a user to enter the request and the output unit for outputting the provided output information are encompassed by a common user interface.

[0039] The processing unit is preferably designed to execute the procedure described above for supporting a user in performing a workflow for a magnetic resonance imaging (MRI) examination on a patient. For example, the interface can always display a communication window for inputting the request and for outputting the output information on a display of the measurement program's user interface.

[0040] The computing unit according to the invention has the advantage that it can provide beneficial support to a user, in particular to medical personnel, during a magnetic resonance examination. For example, the medical personnel can be informed, through targeted inquiries, about the effects of individual configuration options and which constraints must be observed.

[0041] Furthermore, the invention relates to a system comprising a magnetic resonance device and a computing unit, wherein the computing unit is configured as described above, and wherein the system is configured to perform a method mentioned above for supporting a user in performing a workflow of a magnetic resonance examination on a patient.

[0042] The advantages of the system according to the invention essentially correspond to the advantages of the method according to the invention for supporting a user in performing a workflow for a magnetic resonance examination on a patient, which are described in detail below. The claims for the system can be improved by features that are described or claimed in connection with the method, and vice versa. In this case, the functional features of the method are embodied by structural units of the system, and vice versa.

[0043] In an advantageous embodiment of the system according to the invention, the magnetic resonance device may have a user interface that is connected to an interface module and / or a provisioning module of the computing unit. Preferably, the user interface comprises an input unit and an output unit and is connected to both the interface module and the provisioning module to enable user input of a query and output of the information generated by the LLM.

[0044] According to another aspect, a computer program product is provided with program elements that cause a computing unit to execute the steps of the above-mentioned procedure to assist a user in performing a workflow of a magnetic resonance examination on a patient when the program elements are loaded into a memory of the computing unit.

[0045] According to yet another aspect, a computer-readable medium is provided on which program elements are stored that can be read and executed by a computing unit to carry out steps of the above-mentioned procedure to assist a user in performing a workflow of a magnetic resonance examination on a patient when the program elements are executed by the computing unit.

[0046] Further advantages, features and details of the invention will become apparent from the exemplary embodiment described below and from the drawings.

[0047] They show: Fig. 1 a method according to the invention for supporting a user in performing a workflow of a magnetic resonance examination on a patient, Fig. 2 a computing unit according to the invention and Fig. 3 a system according to the invention comprising a magnetic resonance device and a computing unit.

[0048] In Fig. Figure 1 shows a flowchart of a computer-implemented procedure to assist a user in performing a workflow of a magnetic resonance examination on a patient 202.

[0049] The process is executed using a computing unit 100, which is located in Fig. 2 is shown in more detail. The computing unit 100 has an interface module 101, a determination module 102 and a provisioning module 103 for this purpose.

[0050] The interface module 101 is configured to connect to a user interface 104, the user interface having an input unit 105 for the user to input a request. The user's request is formulated in text form or as speech output. For this purpose, the input unit 105 has a microphone and / or a keyboard for entering the request.

[0051] The investigation module 102 includes an LLM, whereby the LLM is trained to determine output information corresponding to the query.

[0052] The provisioning module 103 is configured to provide the output information determined by the detection module 102, in particular the LLM, to an output unit 106 of the user interface 104. The output unit 106 preferably includes a display and / or a loudspeaker for this purpose.

[0053] In the present embodiment, the provisioning module 103 is formed integrally with the interface module 101. The user interface 104 can be encompassed by a magnetic resonance device 201. Alternatively, the user interface 104 can also be formed separately from the magnetic resonance device 201.

[0054] In the first step of the process (10), the user enters a query at input unit 105. The input can be in text form or as speech input. The query can also be entered in natural language, meaning the user doesn't need to know the exact terminology. For example, a query could ask about the causes of organ outline duplication without using keywords like "aliasing" or "parallel imaging."

[0055] A user can enter a request during a preparation workflow for a magnetic resonance imaging (MRI) scan or during a measurement workflow. The preparation workflow includes preparing the MRI scan, for example, positioning the patient and / or selecting a measurement program, etc. The measurement workflow, on the other hand, preferably involves executing the measurement program.

[0056] In a further, second process step 11, output information corresponding to the request is determined using the LLM of the computing unit 100. The LLM can also access further information to determine the output information. This further information can include information from the current magnetic resonance examination and / or information from at least one previous magnetic resonance examination and / or a hardware property of the magnetic resonance device and / or a software property of the magnetic resonance device and / or information from the patient 202.

[0057] The information from the current magnetic resonance imaging (MRI) scan can include current parameter settings for sequence parameters. Furthermore, the information can also include the current workflow status. For example, it can provide context from which the request from the medical operator can be derived.

[0058] Information from a previous magnetic resonance imaging (MRI) scan can include application data from one or more previous MRI scans conducted in a situation identical or similar to the query context. For example, when querying parameter settings, settings from previous MRI scans can be considered. Preferably, information from previous MRI scans with the same or similar measurement program is taken into account. This data from previous MRI scans can be stored in a storage unit and / or a database, with the computing unit 100, in particular the LLM, having access rights to the storage unit and / or the database. An embedded database on the system itself, especially on the computing unit 100, or in a cloud environment could be used for this purpose.Information from a previous magnetic resonance examination can also be used to consider an identical or similar query regarding a problem from a previous magnetic resonance examination, and in determining the output information, feedback on the output information from a previous magnetic resonance examination can be taken into account.

[0059] The hardware properties of the magnetic resonance device 201 can, for example, include information regarding the high-frequency coils used or intended to be used for the magnetic resonance examination. Furthermore, the hardware properties of the magnetic resonance device can also include information regarding a magnetic field strength and / or information regarding a maximum available gradient field strength, etc.

[0060] The software properties of the magnetic resonance device 201 can include information regarding available software licenses. For example, response information to a request to set sequence parameters may depend on available measurement options, which may be linked to a software license.

[0061] The patient 202 information may include the patient 202's weight and / or age and / or height and / or medical history and / or implant information and / or other information and / or characteristics of the patient 202 that the specialist may deem useful.

[0062] Preferably, the at least one additional piece of information is automatically provided to the LLM by the computing unit 100, in particular by the determination module 102, in order to generate output information based on the current query together with the at least one additional piece of information. Preferably, the determination module 102 accesses the additional information via the interface module 101.

[0063] After the output information has been determined using the LLM, in this second process step 11 the output information is made available for output using the provisioning module 102. The output information is then provided to an output unit 106 of the user interface 104.

[0064] In a further, third process step 12, the output information is displayed at the output unit 106 of the user interface 104. The output at the output unit 106 is in text form on a display of the output unit 106 or as speech output via the loudspeaker of the output unit 106.

[0065] The output information can include a link, for example, to a defined text passage in a saved document. When the user activates the link via the user interface 104, a connection to the saved document is established, and the defined text passage is displayed to the user on the output unit 106. Activation can occur by clicking with a computer mouse and / or a finger on a display and / or by voice input. The output of the defined text passage can also be displayed via the output unit 106, specifically the display or the speakers.

[0066] In addition to a connection element, the output information can also include additional information about the current settings of the magnetic resonance imaging (MRI) scan. These current settings can include parameter settings that can be adjusted by the medical operator when selecting the measurement program.

[0067] In a further, fourth process step 13, the user can enter additional input in response to the output information. This input is also entered via the input unit 105 user interface 104 in text form or as voice input. This additional input can include user feedback, which is sent to the LLM for evaluation of the generated output information. Furthermore, the additional input can include a query regarding the output information, for example, if the user has questions about the application and / or execution of the output information.

[0068] In a further, fifth process step 14, additional output information is generated using the LLM based on the further input and displayed to the user via the output unit 106 user interface 104. This additional output information can include alternative assistance to the first output information and / or additional information to the first output information. Furthermore, the additional output information can also include contact information for a human contact person, in particular an expert.

[0069] Depending on the specific characteristics of the procedure, the fourth process step 13 and the fifth process step 14 may only be available optionally to support a user in carrying out a workflow of a magnetic resonance examination on a patient 202.

[0070] In the Fig. 3 is a system 200 with a magnetic resonance device 201 and the one made of Fig.The magnetic resonance device 201 is schematically represented by the two known computing units 100. It comprises a magnet unit 203 with a base magnet 204, a gradient coil unit 205, and a high-frequency antenna unit 206. The magnetic resonance device 201 also has a patient acquisition area 207 for acquiring the patient 202 for a magnetic resonance examination. In the present embodiment, the patient acquisition area 207 is cylindrical and is surrounded in a cylindrical shape by the magnet unit 203 in one circumferential direction. However, a different configuration of the patient acquisition area 207 is conceivable.

[0071] For positioning the patient 202, in particular an area of ​​the patient 202 to be examined, within the patient reception area 207, the magnetic resonance device 201 has a patient positioning device 208. The patient positioning device 208 has a base unit 209 and a patient table 210 that is movable relative to the base unit 209. The patient table 210 is designed to be movable within the patient reception area 207 for positioning the patient 202, in particular the area of ​​the patient 202 to be examined. In particular, the patient table 210 is mounted to be movable in the longitudinal direction of the patient reception area 207 and / or in the z-direction.

[0072] The base magnet 204 of the magnet unit 203 is configured to generate a strong and, in particular, constant base magnetic field 211. The base magnet 204 can be configured, for example, as a superconducting base magnet or as a permanent magnet. The gradient coil unit 205 of the magnet unit 203 is configured to generate magnetic field gradients used for spatial encoding during imaging. The gradient coil unit 205 is controlled by a gradient control unit 212 of the magnetic resonance device 201. The high-frequency antenna unit 206 of the magnet unit 203 is configured to excite a polarization that arises in the base magnetic field 211 generated by the base magnet 204.The high-frequency antenna unit 206 is controlled by a high-frequency antenna control unit 213 of the magnetic resonance device 201 and transmits high-frequency magnetic resonance sequences into the patient acquisition area 207 of the magnetic resonance device 201.

[0073] The magnetic resonance device 201 includes a system control unit 214 for controlling the base magnet 204, the gradient control unit 212, and the high-frequency antenna control unit 213. The system control unit 214 centrally controls the magnetic resonance device 201, for example, by performing a predetermined imaging gradient echo sequence. The system control unit 214 also includes an evaluation unit (not shown) for evaluating medical image data acquired during the magnetic resonance examination.

[0074] Furthermore, the magnetic resonance device 201 includes the user interface 104, which is connected to the system control unit 214. In addition, the user interface 104 of the magnetic resonance device 201 is connected to the computing unit 100, specifically the interface module 101 and the provisioning module 103 of the computing unit 100, for inputting the request and outputting the output information.

[0075] The illustrated magnetic resonance device 201 can, of course, include further components that magnetic resonance devices 201 typically possess. Furthermore, the general operating principle of a magnetic resonance device 201 is known to those skilled in the art, so a detailed description of the further components is omitted.

[0076] 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] Computer-implemented method for assisting a user in performing a workflow of a magnetic resonance examination on a patient, comprising the following procedure steps: - Input of a request by the user, whereby the input is formulated in text form or as voice input, - Determining output information corresponding to the query using a Large Language Model (LLM) and providing the output information, - Output of the output information in text form or as speech output. [2] Method according to claim 1, characterized by , that at least one additional piece of information is provided which is taken into account when determining the output information using the LLM, wherein the at least one additional piece of information includes: - information from the current magnetic resonance examination, - information from at least one previous magnetic resonance imaging (MRI) scan, - a hardware feature of the magnetic resonance device, - a software feature of the magnetic resonance device, and / or - information from the patient. [3] Method according to claim 2, characterized by that the patient's information includes weight and / or height and / or disease history and / or implant information and / or other characteristics of the patient. [4] Method according to any one of the preceding claims, characterized by , that the output information includes a link element to a defined text passage of a stored documentation, wherein when the link element is activated by a user via a user interface, a connection to the stored documentation is established and the defined text passage is output to the user. [5] Method according to any one of the preceding claims, characterized by that the output information includes at least one additional indication of the current settings of the current magnetic resonance examination. [6] Method according to any one of the preceding claims, characterized by , that in a further process step, further user input can be entered in response to the output information. [7] Method according to claim 6, characterized by , that the further input includes user feedback information, whereby the feedback information is fed to the LLM for evaluation of the determined output information. [8] Method according to one of claims 6 or 7, characterized by , that further output information is created and displayed based on the additional input. [9] Method according to any one of the preceding claims, characterized bythat the further output information includes contact information for a human contact person. [10] A computing unit designed to assist a user in performing a workflow of a magnetic resonance examination on a patient, the computing unit comprising: - an interface module configured to connect to a user interface, wherein the user interface has an input unit for the input of a request by the user, wherein the input is formulated in text form or as speech input, - an investigation module with an LLM that is trained to determine output information corresponding to the query, and - a provisioning module that is configured to provide the output information to an output unit. [11] System comprising a magnetic resonance device and a computing unit configured according to claim 10, wherein the system is configured to perform a method for assisting a user in performing a workflow of a magnetic resonance examination on a patient according to any one of claims 1 to 9. [12] System according to claim 11, characterized by that the magnetic resonance device has a user interface that is connected to an interface module and / or a provisioning module of the computing unit. [13] Computer program product comprising program elements which cause a computing unit to execute the steps of the method to assist a user in performing a workflow of a magnetic resonance examination on a patient according to any one of claims 1 to 9 when the program elements are loaded into a memory of the computing unit. [14] Computer-readable data carrier on which program elements are stored which can be read and executed by a computing unit to perform steps of the method to assist a user in performing a workflow of a magnetic resonance examination on a patient according to any one of claims 1 to 9 when the program elements are executed by the computing unit.

Images