Natural language processing for medical imaging

Abstract

A natural language processing framework for medical imaging. The framework can first receive a user request comprising natural language text. One or more imaging parameters can be generated using a trained large language model in response to the user request. A medical imaging task can be initiated based on the one or more imaging parameters. A response can then be generated based on a result of the medical imaging task.

Description

Technical Field

[0001] This disclosure relates generally to data processing, and more particularly to a natural language processing framework for medical imaging. Background Technology

[0002] Today, operators of medical imaging systems need to learn how to designate workflows for specific patient scans based on the clinical needs identified by physicians. To achieve this, operators require training with specific versions of medical imaging systems from specific vendors. Despite this training, operators often use medical imaging systems suboptimally, and in some cases, incorrectly. Furthermore, in multi-vendor hospitals, operators must be trained and proficient in all different types of medical imaging equipment, even if they rarely use specific equipment or workflows. In certain situations, rare applications or very specific workflows need to be entered into the medical imaging system so that it can perform scans.

[0003] Today, technicians typically input information using a graphical user interface (GUI) or a hardware user interface. For certain functions, these interfaces can be either fully manual, semi-automated, or fully automated. Medical imaging software often has numerous workflows, including unique workflows for specific tasks. Users can change parameters depending on specific requirements. Preset configurations for the acquisition and reconstruction process can be provided based on medical needs. This setting can be expanded into very detailed spreadsheet-like rows, allowing for multiple changes. Summary of the Invention

[0004] This paper describes a framework for natural language processing in medical imaging. The framework can first receive a user request consisting of natural language text. In response to the user request, one or more imaging parameters can be generated using a trained large language model. A medical imaging task can be initiated based on the one or more imaging parameters. A response can then be generated based on the results of the medical imaging task. Attached Figure Description

[0005] A more comprehensive understanding of this disclosure and its many accompanying aspects will be more readily obtained when considered in conjunction with the accompanying drawings and by referring to the following detailed description.

[0006] Figure 1 A block diagram illustrating an exemplary system is shown; Figure 2 This illustrates an exemplary method for natural language processing; Figure 3 This illustrates an exemplary user request that can be provided to a large language model (LLM); and Figure 4 This shows another exemplary user request provided to the LLM. Detailed Implementation

[0007] In the following description, numerous specific details, such as examples of specific components, devices, methods, etc., are set forth to provide a thorough understanding of the implementation of this framework. However, it will be apparent to those skilled in the art that these specific details need not be used to practice the implementation of this framework. In other instances, well-known materials or methods are not described in detail to avoid unnecessarily obscuring the implementation of this framework. While various modifications and alternatives to this framework may exist, specific embodiments thereof are shown by way of example in the accompanying drawings and will be described in detail herein. However, it should be understood that there is no intention to limit the invention to the specific forms disclosed, but rather, it is intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Furthermore, for ease of understanding, certain method steps are depicted as separate steps; however, these separately depicted steps should not be construed as necessarily depending on the order in which they are performed. Individuals having male, female, or other gender identities are also included in this terminology, independent of the use of grammatical terms. Unless otherwise stated, it should be understood from the following discussion that terms such as “segmentation,” “generation,” “registration,” “determination,” “alignment,” “positioning,” “processing,” “computation,” “selection,” “estimation,” “detection,” and “tracking” can refer to actions and processes of a computer system or similar electronic computing device that manipulate and convert data represented as physical (e.g., electronic) quantities in the registers and memories of the computer system into other data similarly represented as physical quantities in the computer system's memory or registers or other such information storage, transmission, or display devices. Embodiments of the methods described herein can be implemented using computer software. If written in a programming language conforming to recognized standards, sequences of instructions designed to implement these methods can be compiled to execute on various hardware platforms and interface with various operating systems. Furthermore, the implementation of this framework is not described with reference to any particular programming language. It should be understood that various programming languages ​​can be used.

[0008] For the sake of brevity, an image or image portion (e.g., a region of interest (ROI) in an image) corresponding to an object (e.g., tissue, organ, tumor, etc. of a subject, such as a patient) can be referred to as the image or image portion of that object (e.g., ROI), or an image or image portion including that object (e.g., ROI), or simply the object itself. For example, an ROI corresponding to an image of the lungs or heart can be described as: the ROI includes the lungs or heart. As another example, an image of the chest or an image including the chest can be referred to as a chest image, or simply chest. For the sake of brevity, the processing (e.g., extraction, segmentation) of a portion of an image corresponding to an object can be described as: the object is processed. For example, extracting an image portion corresponding to the lungs from the rest of the image can be described as: the lungs are extracted.

[0009] Currently, users need to understand how to translate physician requirements into setup parameters for medical imaging systems used for image formation and reconstruction. Furthermore, users must continuously monitor all relevant associations and their recommendations. This framework uses a Large Language Model (LLM), which acts as a translation layer between the human user and the medical imaging system to advantageously eliminate any guesswork and / or human error that may arise during setup and reconstruction.

[0010] LLM can convert the protocol of a medical imaging system from one vendor (or version) to a protocol compatible with a medical imaging system from another vendor (or version). LLM can be improved for medical tasks within the scope of medical imaging (e.g., SPECT) based on specific external information. LLM can map user requests for medical imaging (e.g., “Evaluate reconstructed patients for MPI to confirm postoperative success”) to image formation and reconstruction hyperparameters that achieve optimal reconstruction. These and other exemplary advantages and features will be described in more detail below.

[0011] Figure 1 A block diagram of an exemplary system 100 is illustrated. System 100 includes a computer system 101 for implementing the framework described herein. In some implementations, computer system 101 operates as a standalone device. In other implementations, computer system 101 may be connected (e.g., using a network) to other machines, such as medical imaging equipment 102 and workstation 103. In a networked deployment, computer system 101 may operate as a server (e.g., in a server-client user network environment), a client user machine in a server-client user network environment, or a peer machine in a peer-to-peer (or distributed) network environment.

[0012] In one implementation, computer system 101 includes a processor device or central processing unit (CPU) 104 coupled via an input / output interface 121 to one or more non-transitory computer-readable media 105 (e.g., computer storage or memory devices), a display device 108 (e.g., a monitor), and various input devices 110 (e.g., a mouse, touchpad, or keyboard). Computer system 101 may also include supporting circuitry such as caches, power supplies, clock circuitry, and communication buses. Various other peripheral devices, such as additional data storage devices and printing devices, may also be connected to computer system 101.

[0013] This technology can be implemented in various forms of hardware, software, firmware, dedicated processors, or combinations thereof, either as part of micro-instruction code, as part of an application or software product, or a combination thereof, which is executed via an operating system. In some implementations, the technology described herein is implemented as computer-readable program code tangibly embodied in one or more non-transitory computer-readable media 105. In particular, this technology can be implemented by processing module 106 and large language model 107. Non-transitory computer-readable media 105 may include random access memory (RAM), read-only memory (ROM), floppy disk, flash memory, and other types of memory, or combinations thereof. The computer-readable program code is executed by processor device 104 to process data provided by, for example, medical imaging device 102. Thus, computer system 101 is a general-purpose computer system, which becomes a purpose-specific computer system when the computer-readable program code is executed. The computer-readable program code is not intended to be limited to any particular programming language and its implementation. It should be understood that the teachings of the disclosure contained herein can be implemented using various programming languages ​​and their encodings. The same or different computer-readable media 105 may be used to store databases, including but not limited to image datasets, knowledge bases, individual subject data, electronic medical records (EMRs), subject diagnostic reports (or documents), or combinations thereof.

[0014] Medical imaging device 102 acquires medical image data 132. This medical image data 132 can be processed by processing module 106. Medical imaging device 102 can be a radiographic scanner and / or suitable peripheral devices (e.g., a keyboard and display device) for acquiring, collecting, and / or storing this medical image data 132. Medical imaging device 102 can acquire medical image data 132 from a subject or patient using techniques such as high-resolution computed tomography (HRCT), magnetic resonance (MR) imaging, computed tomography (CT), spiral CT, X-ray, angiography, positron emission tomography (PET), fluoroscopy, ultrasound, single-photon emission computed tomography (SPECT), or combinations thereof. Other types of imaging modalities are also useful. Medical imaging device 102 can be controlled using medical imaging software.

[0015] Medical image data 132 is data representing a medical image (or, in some examples, more than one medical image). For example, medical image data 132 may include an array or list of pixel or voxel values. When processed by suitable image viewing software, medical image data 132 results in the rendering of the medical image (or multiple medical images) it represents. An image file containing medical image data 132 may also include one or more attributes and attribute values. A specific example of an image file is a Medical Digital Imaging and Communication (DICOM) file. One or more attribute values ​​are separate from and distinct from medical image data 132, and instead include a text string indicating the content of medical image data 132. In some examples, such attribute values ​​may be referred to as metadata of the image file. In some examples, the storage attributes and attribute values ​​portion of the image file may be referred to as the image file header, and the attributes and attribute values ​​may be referred to as the image file header data.

[0016] Workstation 103 may include a computer and suitable peripherals, such as a keyboard and display device, and may operate in conjunction with the overall system 100. For example, workstation 103 may communicate with medical imaging equipment 102, such that medical image data 132 may be presented or displayed on workstation 103. Workstation 103 may communicate directly with computer system 101 to display processed data and / or output results 144. Workstation 103 may include a graphical user interface to receive user input via input devices such as a keyboard, mouse, touchscreen, voice or video recognition interface, etc., to manipulate the visualization and / or processing of data.

[0017] It should also be understood that, since some of the system components and method steps depicted in the accompanying drawings can be implemented in software, the actual connections between system components (or processing steps) may vary depending on how this framework is programmed. Given the teachings provided herein, those skilled in the art will be able to conceive of these and similar implementations or configurations of this framework.

[0018] Figure 2 An exemplary method 200 for natural language processing is illustrated. It should be understood that the steps of method 200 may be implemented in the order shown or a different order. Additional, different, or fewer steps may also be provided. Furthermore, method 200 may be implemented in… Figure 1 The system can be implemented through 100 different systems or combinations thereof.

[0019] At 204, processing module 106 receives a trained Large Language Model (LLM) 107. LLM 107 is an artificial intelligence model designed to predict the next lexical (or word) in a sequence of lexical (or word) terms based on previous lexical (or word) sequences provided to the model (i.e., context or cues). In some implementations, LLM 107 is a specialized model trained with industry-specific knowledge and capable of understanding technical language and concepts. LLM 107 can be trained on datasets of medical imaging documents, allowing it to learn the relationship between user requests and imaging parameters (e.g., image formation and / or reconstruction parameters). LLM 107 can be trained using, for example, instruction manuals, use cases based on specific instruction sets, or imaging protocols of the medical imaging device 102 for different vendors or versions.

[0020] In some implementations, LLM 107 is refined for medical tasks within the scope of medical imaging (e.g., SPECT) based on specific external information. Such external information used by LLM 107 can include, but is not limited to, publications, journal articles, recommendations or data accumulated from past examinations, workflow guidelines from associations (e.g., the National Correct Coding Initiative, or NCCI), imaging guidelines (e.g., the Society for Nuclear Medicine and Molecular Imaging, or SNMMI), or combinations thereof. For example, a typical journal article could be Hage et al.'s… Quality metrics for single-photon emission computed tomography myocardial perfusion imaging: an ASNC informationstatement , J NuclCardiol. 2023 Apr; 30(2): 864-907”, which is incorporated herein by reference.

[0021] The LLM 107 utilizes deep learning techniques, which can be applied to various tasks, including language translation, question answering, and text and / or image generation. The LLM 107 may include one or more artificial neural networks (e.g., transformers) trained using self-supervised and / or semi-supervised learning. The LLM 107 can retain long-term memory of its previous context. Bias towards sensitive attributes (e.g., age, gender, race) may be inherent in the LLM 107. This bias can be monitored and mitigated to prevent the generation of harmful or imperfect responses. One or more programmable rules (or safety barriers) can be implemented to identify and / or prevent bias in one or more imaging parameters generated by the LLM 107. Such rules can be programmed, for example, to validate the output of the LLM 107, take corrective actions when validation fails, and enforce a specific structure, type, and / or quality in the output of the LLM 107.

[0022] return Figure 2 At 206, processing module 106 receives user requests that include natural language. Natural language generally refers to any language that arises naturally in human society. Natural language can be one or more world languages, such as English, Chinese, Korean, Japanese, Spanish, French, German, or combinations thereof. In some implementations, processing module 106 is able to map user requests in different languages ​​into appropriate responses.

[0023] User requests can be provided by a user through a user interface displayed, for example, on workstation 103. User requests may include natural language text describing the imaging task the user wishes the medical imaging device 102 to perform. Imaging tasks may include, for example, image reconstruction using specific algorithms or parameters, or image acquisition using specific parameters. User requests may also include non-text media files, such as image, video, or audio files.

[0024] At 208, processing module 106, in response to a user request, generates one or more imaging parameters using LLM 107. These imaging parameters define an imaging task in a protocol compatible with medical imaging device 102 to trigger or initiate that task. Imaging parameters may include, for example, image forming parameters and image reconstruction parameters. Image forming parameters may be related to staging, planning, treatment, image acquisition, and / or follow-up in medical imaging. Image forming parameters may include, but are not limited to, modality, field of view, number of views, acquisition time, patient positioning instructions, guidance during scanning, exit instructions after scanning, and post-processing instructions (e.g., warnings and safety instructions for both patients and staff). Image reconstruction parameters may be related to the processing of acquired image data. Image reconstruction parameters may include, but are not limited to, region of interest, image resolution, pixel size, reconstruction algorithm parameters, correction parameters, or combinations thereof. Other types of parameters are also useful.

[0025] In some implementations, the LLM 107 translates user requests defined in a protocol compatible with one vendor's (or version's) medical imaging equipment into imaging parameters defined in a protocol compatible with another vendor's (or version's) medical imaging equipment. This allows for protocol conversion between different vendors' (or versions') of medical imaging equipment. In some implementations, the LLM 107 adjusts imaging parameters for workflow based on external information. For example, an association or society (e.g., SNMMI) may recommend increasing or decreasing acquisition time, while prior information from patient records may prompt the LLM 107 to increase acquisition time. As another example, the LLM 107 can analyze image data and decide to change the pixel size if there are insufficient counts.

[0026] Figure 2 At 210, processing module 106 initiates a medical imaging task based on imaging parameters. The medical imaging task can be any task related to image formation and / or image reconstruction. In some implementations, processing module 106 provides imaging parameters to medical imaging device 102 to initiate the medical imaging task and generate the desired output. For example, processing module 106 can communicate with medical imaging device 102 to perform tasks associated with staging, planning, treatment, and / or follow-up of medical imaging (e.g., PET and / or CT image acquisition, patient positioning). In some implementations, imaging parameters can be presented (e.g., displayed) as instructions to the patient or the operator of medical imaging device 102. For example, positioning instructions can be provided to the patient before medical imaging. As another example, guidance instructions can be provided to the operator during scanning. Other instructions, such as post-scan exit instructions and post-processing instructions (e.g., warnings and safety instructions for both patients and staff), can also be provided.

[0027] At 212, processing module 106 generates a response based on the results of the medical imaging task. This response may describe the results generated by the medical imaging task, such as the time spent performing the task (e.g., image reconstruction), the reconstructed image, the file location of the reconstructed image, etc. The response may include natural language text and / or other non-text media files (e.g., images, videos, audio). The response may be displayed, for example, on workstation 103.

[0028] Figure 3 An exemplary user request 302 that can be provided to the large language model 107 is shown. User request 302 can be provided in different human languages. As shown, user request 302 includes natural language text describing the imaging task “calculate cardiac reconstruction for patient KNZ001”. In response to the user request 4302, LLM 107 generates imaging parameters 306 (e.g., reconstruction(patient='KNZ001', modality='heart')), which can be provided to medical imaging device 102 to generate the desired output. It should be understood that other tasks can also be requested.

[0029] Figure 4 Another exemplary user request 402 provided to LLM 107 is shown. User request 402 includes natural language text describing the imaging task “Assessing reconstructed patients for MPI to confirm postoperative success.” Additionally, a workflow document 404 (e.g., American College of Nuclear Cardiology, or ASNC guidelines) can be provided to improve LLM 107. It should be understood that different countries, regions, and / or healthcare facilities may require specific adjustments to the workflow. Workflow document 404 can be a “local” workflow guide that is fed to LLM 107 to adapt the workflow as needed. In response to the user request 402, LLM 107 generates a detailed set of imaging parameters 406 (e.g., “Acquire data using LEHR, using 180 configuration and NCO, then reconstruct data with 48 CGAS updates, using ITC if the quality score exceeds a threshold”). Imaging parameters 406 can be provided to medical imaging device 102 to initiate image acquisition and reconstruction to generate the desired output.

[0030] The following is a list of non-limiting illustrative embodiments disclosed herein: Illustrative Example 1. A data processing system includes: a non-transitory memory device for storing computer-readable program code; and a processor device communicating with the non-transitory memory device, the processor device being capable of working with the computer-readable program code to perform steps including receiving a user request comprising natural language text, responding to the user request by generating one or more imaging parameters using a large language model, initiating a medical imaging task based on the one or more imaging parameters, and generating a response based on the result of the medical imaging task.

[0031] Illustrative Example 2. The data processing system of Illustrative Example 1, wherein the large language model is trained using a dataset of medical imaging documents.

[0032] Illustrative Example 3. A data processing system according to any of the embodiments of illustrative Examples 1-2, wherein the processor device is capable of working with the computer-readable program code to prevent damage caused by the one or more imaging parameters by using one or more programmable rules to identify or prevent bias in a large language model.

[0033] Illustrative Example 4. A data processing system according to any one of the illustrative examples 1-3, wherein natural language text includes English, Chinese, Korean, Japanese, Spanish, French, German, or combinations thereof.

[0034] Illustrative Example 5. A data processing system according to any one of the illustrative Examples 1-4, wherein the processor device is capable of working with the computer-readable program code to generate one or more imaging parameters by generating one or more image forming parameters.

[0035] Illustrative Example 6. A data processing system according to Illustrative Example 5, wherein the processor device is capable of working with the computer-readable program code to generate one or more image forming parameters by generating modalities, fields of view, acquisition time, number of views, patient positioning instructions, guidance instructions during scanning, exit instructions after scanning, post-processing instructions, or combinations thereof.

[0036] Illustrative Example 7. A data processing system according to any one of the illustrative Examples 1-6, wherein the processor device is capable of working with the computer-readable program code to generate one or more imaging parameters by generating one or more image reconstruction parameters.

[0037] Illustrative Example 8. A data processing system according to Illustrative Example 7, wherein the processor device is capable of working with the computer-readable program code to generate one or more image reconstruction parameters by generating a region of interest, image resolution, pixel size, reconstruction algorithm parameters, correction parameters, or a combination thereof.

[0038] Illustrative Example 9. A data processing method includes: receiving a user request comprising natural language text; generating one or more imaging parameters using a large language model in response to the user request; initiating a medical imaging task based on the one or more imaging parameters; and generating a response based on the result of the medical imaging task.

[0039] Illustrative Example 10. The data processing method according to Illustrative Example 9 further includes using a dataset of medical imaging documents to train the large language model.

[0040] Illustrative Example 11. A data processing method according to Illustrative Example 10, wherein medical imaging documentation includes instruction manuals, use cases based on a specific set of instructions, or imaging protocols for different vendors or versions.

[0041] Illustrative Example 12. The data processing method according to any one of the illustrative Examples 9-11 further includes improving the large language model with external information, including publications, journal articles, suggestions or data accumulated from past inspections, workflow guidelines from associations, imaging guidelines, or combinations thereof.

[0042] Illustrative Example 13. A data processing method according to any one of the illustrative Examples 9-12, wherein generating the one or more imaging parameters includes generating one or more image forming parameters.

[0043] Illustrative Example 14. The data processing method according to Illustrative Example 13, wherein generating the one or more image forming parameters includes a generation modality, field of view, acquisition time, number of views, patient positioning instructions, guidance instructions during scanning, exit instructions after scanning, post-processing instructions, or combinations thereof.

[0044] Illustrative Example 15. A data processing method according to any one of the illustrative Examples 9-14, wherein generating the one or more imaging parameters includes generating one or more image reconstruction parameters.

[0045] Illustrative Example 16. According to the data processing method of Illustrative Example 15, generating the one or more image reconstruction parameters includes generating a region of interest, image resolution, pixel size, reconstruction algorithm parameters, correction parameters, or a combination thereof.

[0046] Illustrative Example 17. A data processing method according to any one of the illustrative examples 9-16, wherein a user request is defined in a first protocol compatible with a first medical imaging device from a first supplier or version, and imaging parameters are defined in a second protocol compatible with a second medical imaging device from a second supplier or version.

[0047] Illustrative Example 18. A data processing method according to any one of the illustrative examples 9-17, wherein initiating the medical imaging task includes providing the one or more imaging parameters to a medical imaging device to initiate the medical imaging task.

[0048] Illustrative Example 19. A data processing method according to any of the illustrative Examples 9-18, wherein initiating the medical imaging task includes presenting one or more imaging parameters as instructions to a patient or an operator of a medical imaging device.

[0049] Illustrative Example 20. One or more non-transitory computer-readable media embodying a machine-executable program of instructions to perform operations including: receiving a user request comprising natural language text; generating one or more imaging parameters using a large language model in response to the user request; and generating a response based on the results of a medical imaging task initiated based on the one or more imaging parameters.

[0050] While this framework has been described in detail with reference to exemplary embodiments, those skilled in the art will understand that various modifications and substitutions can be made therein without departing from the spirit and scope of the invention as set forth in the appended claims. For example, within the scope of this disclosure and the appended claims, elements and / or features of different exemplary embodiments may be combined with and / or substituted for each other.

Claims

1. A data processing system, comprising: Non-transitory memory devices used for storing computer-readable program code; and A processor device communicating with the non-transitory memory device, the processor device being capable of working with the computer-readable program code to perform steps, the steps including Receive user requests that include natural language text. In response to the user request, one or more imaging parameters are generated using a large language model. A medical imaging task is initiated based on one or more of the imaging parameters, and A response is generated based on the results of the medical imaging task.

2. The data processing system according to claim 1, wherein the large language model is trained using a dataset of medical imaging documents.

3. The data processing system of claim 1, wherein the processor device is capable of working with the computer-readable program code to prevent damage caused by the one or more imaging parameters by using one or more programmable rules to identify or prevent bias in the large language model.

4. The data processing system according to claim 1, wherein the natural language text includes English, Chinese, Korean, Japanese, Spanish, French, German, or a combination thereof.

5. The data processing system of claim 1, wherein the processor device is capable of working with the computer-readable program code to generate one or more imaging parameters by generating one or more image forming parameters.

6. The data processing system of claim 5, wherein the processor device is capable of working with the computer-readable program code to generate one or more image forming parameters by generating modalities, fields of view, acquisition time, number of views, patient positioning instructions, guidance instructions during scanning, exit instructions after scanning, post-processing instructions, or combinations thereof.

7. The data processing system of claim 1, wherein the processor device is capable of working with the computer-readable program code to generate one or more imaging parameters by generating one or more image reconstruction parameters.

8. The data processing system of claim 7, wherein the processor device is capable of working with the computer-readable program code to generate one or more image reconstruction parameters by generating a region of interest, image resolution, pixel size, reconstruction algorithm parameters, correction parameters, or a combination thereof.

9. A data processing method, comprising: Receive user requests, including natural language text; In response to the user request, one or more imaging parameters are generated using a large language model; A medical imaging task is initiated based on one or more of the imaging parameters; and A response is generated based on the results of the medical imaging task.

10. The data processing method of claim 9 further includes using a dataset of medical imaging documents to train the large language model.

11. The data processing method of claim 10, wherein the medical imaging document includes instruction manuals, use cases based on a specific set of instructions, imaging protocols for different vendors or versions, or a combination thereof.

12. The data processing method of claim 9 further includes improving the large language model with external information, said external information including publications, journal articles, suggestions or data accumulated from past inspections, workflow guidelines from associations, imaging guidelines, or combinations thereof.

13. The data processing method of claim 9, wherein generating the one or more imaging parameters includes generating one or more image forming parameters.

14. The data processing method of claim 13, wherein generating the one or more image forming parameters includes generating modality, field of view, acquisition time, number of views, patient positioning instruction, guidance during scanning, exit instruction after scanning, post-processing instruction, or a combination thereof.

15. The data processing method of claim 9, wherein generating the one or more imaging parameters includes generating one or more image reconstruction parameters.

16. The data processing method according to claim 15, wherein generating the one or more image reconstruction parameters includes generating a region of interest, image resolution, pixel size, reconstruction algorithm parameters, correction parameters, or a combination thereof.

17. The data processing method of claim 9, wherein the user request is defined in a first protocol compatible with a first supplier or version of a first medical imaging device, and the imaging parameters are defined in a second protocol compatible with a second supplier or version of a second medical imaging device.

18. The data processing method of claim 9, wherein initiating the medical imaging task includes providing the one or more imaging parameters to a medical imaging device to initiate the medical imaging task.

19. The data processing method of claim 9, wherein initiating the medical imaging task includes presenting the one or more imaging parameters as instructions to the patient or the operator of the medical imaging device.

20. One or more non-transitory computer-readable media embodying a machine-executable program of instructions to perform operations, said operations including: Receive user requests, including natural language text; In response to the user request, one or more imaging parameters are generated using a large language model; and A response is generated based on the results of a medical imaging task, which is initiated based on one or more imaging parameters.

Images