Method for generating a list of examination steps and method for adapting an examination protocol during a medical imaging examination

Abstract

Method for generating a list of examination steps, wherein each examination step is assigned an instruction for adapting an examination protocol during a medical imaging examination using a computed tomography device, the method comprising the following steps: - Providing a set of training data sets, each training data set having a training instruction and an examination parameter set, the examination parameter set relating to an examination step associated with the training instruction and / or an examination result associated with the training instruction, - Generating the list of examination steps based on the set of training data sets and a machine learning algorithm, wherein each examination step of the list of examination steps is assigned an instruction for adapting an examination protocol during a medical imaging examination using the computed tomography device, wherein at least one training data set of the set of training data sets is provided by performing the following steps during an examination process: - Recording a deviation from a selected examination protocol, which was manually initiated by a user after an examination step has been carried out, as a training instruction, - Recording a set of examination parameters relating to the examination step carried out and / or an examination result of the examination step carried out.

Description

[0001] The invention relates to a method for generating a list of examination steps and a method for adapting an examination protocol during a medical imaging examination. The invention further relates to a data processing unit, a computer program, a computer-readable medium, and a medical imaging device.

[0002] Reproducibility plays a key role in quantitative findings based on radiological data, especially image data. Quantitative findings can, for example, concern the volume of a tumor, the density of a tissue, a calcium score, or similar. Comparison with corresponding values ​​from a previous examination is a common task in radiology. Post-processing applications that can derive such quantitative information from the initially reconstructed image data generally react sensitively to changes in relevant image properties, such as spatial and / or temporal resolution, contrast, or similar. Depending on the post-processing application, the data to be used as the source for quantitative evaluation must generally meet certain requirements to ensure the results are meaningful and suitable for reproducibility.

[0003] The use of the correct combination of image data and post-processing applications to derive quantitative conclusions is typically based on knowledge described in the documentation, for example, for the post-processing application. It is typically left to the user to decide whether data that does not meet the requirements in terms of meaningfulness and / or reproducibility is used for evaluation in a specific post-processing application.

[0004] DE 10 2014 107 445 A1 discloses a method for setting a parameter value for at least one adjustable device parameter of an optical observation device.

[0005] The invention aims to enable an improved performance of a medical imaging examination.

[0006] Each of the subject matters of the independent claims solves this problem. Further advantageous aspects of the invention are considered in the dependent claims.

[0007] The invention relates to a method for generating a list of examination steps according to claim 1.

[0008] Furthermore, a method for reviewing a medical image with respect to processing the medical image by means of an image processing module is disclosed, the method comprising the following steps: - Providing at least one input request of the image processing module, which relates to at least one image parameter of the medical image, - Providing at least one image parameter of the medical image, - Check whether at least one image parameter meets at least one input requirement.

[0009] In particular, depending on the result of the review, the following step can be performed: - Outputting an indication concerning a quality of an image processing result of the processing of the medical image by means of the image processing module.

[0010] In particular, depending on the result of the review, one of the following steps may be performed: - Processing the medical image using the image processing module, - Rejecting an image processing order that concerns the processing of the medical image using the image processing module.

[0011] In particular, depending on the result of the review, the following step can be performed: - Generating an image generation order for another medical image based on the at least one input request.

[0012] In particular, the method may further comprise the following steps: - Generating another medical image based on the image generation order for the another medical image, - Providing at least one image parameter of the further medical image, - Check whether at least one image parameter of the further medical image satisfies at least one input requirement.

[0013] In particular, the method may further comprise the following steps: - generating an image generation order for the medical image based on the at least one input request, - Generating the medical image based on the medical image generation job.

[0014] In particular, the at least one input requirement can be defined independently of specific properties of a medical imaging device with which the medical image and / or the further medical image is generated.

[0015] In particular, the at least one image parameter can be selected from the image parameter group consisting of a spatial resolution, a temporal resolution, an edge behavior, a contrast, a convolution kernel, a layer thickness, a rotation time, a pitch, a tube voltage, a filter property, a modulation transfer function, and combinations thereof.

[0016] The invention further relates to a method for adapting an examination protocol during a medical imaging examination using a computer tomography device, - wherein a list of examination steps is provided by generating it according to the method according to the invention for generating a list of examination steps, - where each examination step in the list of examination steps is assigned an instruction, - wherein the examination protocol comprises a start examination step from the list of examination steps, which is executed at the start of the medical imaging examination, - wherein the examination protocol is adapted after each examination step carried out by means of the instruction associated with the examination step carried out by carrying out one of the following steps based on the instruction: - Termination of the medical imaging examination regardless of the result of the performed examination step, - Termination of the medical imaging examination depending on the result of the performed examination step, - Executing a further examination step from the list of examination steps specified in the instruction, regardless of the result of the executed examination step, - Execute one of several further examination steps from the list of examination steps specified in the instruction, depending on the result of the examination step executed.

[0017] In particular, depending on a result of the performed investigation step, the following step can be performed: - Generating an image generation order for a further medical imaging examination based on the instruction and / or the result of the performed examination step.

[0018] In particular, the method may further comprise the following step: - Determining the result of the performed examination step based on a user input and / or based on an automatic evaluation of data recorded during the performed examination step.

[0019] The invention relates to a method for generating a list of examination steps, wherein each examination step is assigned an instruction for adapting an examination protocol during a medical imaging examination using a computer tomography device, wherein the method comprises the following steps: - Providing a set of training data sets, each training data set having a training instruction and an examination parameter set, the examination parameter set relating to an examination step associated with the training instruction and / or an examination result associated with the training instruction, - Generating the list of examination steps based on the set of training data sets and a machine learning algorithm, wherein each examination step of the list of examination steps is assigned an instruction for adapting an examination protocol during a medical imaging examination.

[0020] According to the invention, a training data set of the set of training data sets is provided by performing the following steps during an examination process: - Recording a deviation from a selected examination protocol, which was manually initiated by a user after an examination step has been carried out, as a training instruction, - Recording a set of examination parameters relating to the examination step carried out and / or an examination result of the examination step carried out.

[0021] This further discloses a method for training a decision support system for an examination using a medical imaging device, the method comprising the following steps: - Providing a set of training data sets, each training data set comprising a training examination order and an examination information, the examination information relating to a medical imaging examination which was carried out based on the training examination order, - Training the decision support system based on the set of training data sets and a machine learning algorithm such that the trained decision support system can generate an examination recommendation for the examination using a medical imaging device based on an examination order.

[0022] According to one embodiment, - that the training investigation order contains a training instruction, - that the examination information comprises an examination parameter set, wherein the examination parameter set relates to an examination step which is assigned to the training instruction and / or an examination result which is assigned to the training instruction, - that the examination recommendation comprises a list of examination steps which is generated based on the set of training data sets and the machine learning algorithm, wherein each examination step of the list of examination steps is assigned an instruction for adapting an examination protocol during a medical imaging examination.

[0023] According to one embodiment, - that the training examination order includes an image generation order, - that the examination information contains acquisition parameters and / or reconstruction parameters with which the image generation task was carried out during the medical imaging examination.

[0024] According to one embodiment, - that the training examination order includes an image processing order, - that the examination information contains image processing parameters with which the image processing task was carried out during the medical imaging examination.

[0025] The invention further relates to a data processing unit which is designed to carry out a method according to the invention.

[0026] The invention further relates to a computer program which can be loaded into a memory device of a data processing system, having program sections for carrying out all steps of a method according to the invention when the computer program is executed by the data processing system.

[0027] The invention further relates to a computer-readable medium on which program sections that can be read and executed by a data processing system are stored in order to carry out all steps of a method according to the invention when the program sections are executed by the data processing system.

[0028] Hereby further a decision support system is disclosed which is trained based on a method according to one of the aspects disclosed in this application.

[0029] The invention further relates to a medical imaging device comprising a data processing unit according to the invention.

[0030] Furthermore, a use of a decision support system trained on the basis of a method according to one of the aspects disclosed in this application is hereby disclosed in a medical imaging device and / or for generating an examination recommendation for an examination by means of a medical imaging device based on an examination order.

[0031] With knowledge of the operation of a post-processing algorithm, necessary and / or sufficient properties of input data, in particular of the medical image, can be specified, which indicate the suitability of the input data for evaluation using the post-processing algorithm. In a trained post-processing algorithm, these properties can be specified, in particular, with knowledge of the properties of the training data and under the assumption that the result desired by the trained algorithm functions optimally with the training data. In the context of this application, the terms post-processing module and image processing module are used synonymously.

[0032] These properties may, for example, relate to a spatial resolution in one or more spatial directions, an edge behavior, in particular an overshoot, a temporal resolution, in particular a minimum temporal resolution, a material contrast or the like. The properties can, in particular, correlate with acquisition parameters and / or reconstruction parameters. Examples of acquisition parameters and / or reconstruction parameters include, in particular, a convolution kernel, a slice thickness, a rotation time, a pitch, a reconstructed data segment, a tube voltage, or a filter. In particular, the spatial resolution correlates with the convolution kernel and / or the slice thickness; the edge behavior correlates with the convolution kernel; the temporal resolution correlates with the rotation time, the pitch, and / or the reconstructed data segment; and the material contrast correlates with the tube voltage and / or the filter.

[0033] In particular, a selection of the above-mentioned parameters can be used to characterize the medical image. Furthermore, additional parameters can be added without departing from the scope of the invention, as long as it is defined by the claims. In particular, the at least one input request can be used to define intervals or a list of values ​​for the at least one image parameter in addition to fixed values. The medical image can, in particular, include a header and / or a footer.

[0034] The image processing module can, in particular, be an image processing algorithm and / or comprise an image processing algorithm. The image processing module can, for example, be configured to segment a structure and / or perform centerline extraction in the medical image. The image processing module can, in particular, be integrated into a post-processing application. The post-processing application can, for example, be configured to detect pulmonary nodules and / or to identify a stenosis.

[0035] Preferably, the input requirements are stored in a standardized form, in particular independent of specific imaging devices and / or manufacturers. For example, instead of the convolution kernel and slice thickness, modulation transfer functions can be specified for one, several, or all spatial directions. These modulation transfer functions can be realized, in particular, by using a specific convolution kernel and setting a specific slice thickness. This makes it possible, for example, to describe the spatial resolution and edge behavior in a device-independent manner.

[0036] The medical imaging device is a computed tomography device. According to one embodiment of the invention, the medical imaging device comprises an acquisition unit configured to acquire the acquisition data. In particular, the acquisition unit may comprise a radiation source and a radiation detector.

[0037] One embodiment of the invention provides that the radiation source is configured to emit and / or excite radiation, in particular electromagnetic radiation, and / or that the radiation detector is configured to detect the radiation, in particular electromagnetic radiation. The radiation can, for example, travel from the radiation source to a region to be imaged and / or, after interacting with the region to be imaged, to the radiation detector. When the radiation interacts with the area to be imaged, it is modified and thus becomes a carrier of information concerning the area to be imaged. When the radiation interacts with the detector, this information is recorded in the form of acquisition data.

[0038] In particular, in a computed tomography device, the acquisition data can be projection data, the acquisition unit can be a projection data acquisition unit, the radiation source can be an X-ray source, and the radiation detector can be an X-ray detector. The X-ray detector can, in particular, be a quantum-counting and / or energy-resolving X-ray detector.

[0039] The data processing unit and / or one or more components of the data processing unit may be formed by a data processing system. The decision support system and / or one or more components of the decision support system may be formed by a data processing system.

[0040] The data processing system can, for example, have one or more components in the form of hardware and / or one or more components in the form of software. The data processing system can, for example, be at least partially formed by a cloud computing system. The data processing system can, for example, be and / or have a cloud computing system, a computer network, a computer, a tablet computer, a smartphone or the like, or combinations thereof. The hardware can, for example, interact with software and / or be configurable by means of software. The software can, for example, be executed by means of the hardware. The hardware can, for example, be a memory system, an FPGA (field-programmable gate array) system, an ASIC (application-specific integrated circuit) system, a microcontroller system, a processor system, and combinations thereof.The processor system may, for example, comprise a microprocessor and / or a plurality of cooperating microprocessors.

[0041] In particular, a component of the data processing unit according to one of the aspects disclosed in this application, which is configured to perform a given step of a method according to one of the aspects disclosed in this application, can be implemented in the form of hardware that is configured to perform the given step and / or that is configured to execute a computer-readable instruction such that the hardware is configurable to perform the given step using the computer-readable instruction. In particular, the system can have a memory area, for example in the form of a computer-readable medium, in which computer-readable instructions, for example in the form of a computer program, are stored.

[0042] Data transfer between components of the data processing system can, for example, be carried out using a suitable data transfer interface. The data transfer interface for data transfer to and / or from a component of the data processing system can be implemented at least partially in software and / or at least partially in hardware. The data transfer interface can, for example, be designed to store data in and / or load data from an area of ​​the storage system, with one or more components of the data processing system being able to access this area of ​​the storage system.

[0043] The computer program can be loaded into the memory system of the data processing system and executed by the processor system of the data processing system. The data processing system can, for example, be designed by means of the computer program such that the data processing system can carry out the steps of a method according to one of the embodiments disclosed in this application when the computer program is executed by the data processing system.

[0044] For example, the computer program product according to one of the embodiments disclosed in this application and / or the computer program according to one of the embodiments disclosed in this application may be stored on the computer-readable medium. The computer-readable medium can be, for example, a memory stick, a hard disk, or another data storage medium, which can be detachably connected to the data processing system or permanently integrated into the data processing system. The computer-readable medium can, for example, form a portion of the storage system of the data processing system. In the context of this application, the terms "protocol" and "investigation protocol" are used synonymously.

[0045] Within the scope of the invention, features described in relation to different embodiments of the invention and / or different claim categories (method, use, device, system, arrangement, etc.) can be combined to form further embodiments of the invention. For example, a claim relating to a device can also be further developed with features described or claimed in connection with a method. Functional features of a method can be implemented by appropriately designed physical components. In addition to the embodiments of the invention expressly described in this application, a wide variety of further embodiments of the invention are conceivable, which the person skilled in the art can arrive at without departing from the scope of the invention, insofar as it is defined by the claims.

[0046] The use of the indefinite articles "ein" or "eine" does not preclude the feature in question from being present multiple times. The use of the term "aufeinander" (to have) does not preclude the concepts linked by the term "aufeinander" (to have) from being identical. For example, the medical imaging device comprises the medical imaging device. The use of the term "einheit" does not preclude the object to which the term "einheit" refers from having multiple components that are spatially separated from one another.

[0047] In the context of the present application, the expression "based on" can be understood in particular in the sense of the expression "using." In particular, a formulation according to which a first feature is generated (alternatively: determined, determined, etc.) based on a second feature does not exclude the possibility that the first feature can be generated (alternatively: determined, determined, etc.) based on a third feature.

[0048] Selected embodiments are explained below with reference to the accompanying figures. The representations in the figures are schematic, highly simplified, and not necessarily to scale.

[0049] They show: Fig. 1 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image by means of an image processing module according to an embodiment, Fig. 2 shows a schematic representation of a data processing unit for reviewing a medical image with respect to processing the medical image by means of an image processing module according to a further embodiment, Fig. 3 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image by means of an image processing module according to a further embodiment, Fig. 4 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image by means of an image processing module according to another embodiment, Fig. 5 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image by means of an image processing module according to another embodiment, Fig. 6 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image by means of an image processing module according to another embodiment, Fig. 7 shows a flowchart for a method for adapting an examination protocol during a medical imaging examination, Fig. 8 is a flowchart for a method for training a decision support system for an examination using a medical imaging device, and Fig. 9 is a schematic diagram of a medical imaging device according to an embodiment of the invention.

[0050] Fig. 1 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image by means of an image processing module according to an embodiment, the method comprising the following steps: - Providing PQ at least one input request of the image processing module, which concerns at least one image parameter of the medical image, - Providing PX of at least one image parameter of the medical image, - Check CP whether at least one image parameter meets at least one input requirement.

[0051] Fig. 2 a schematic representation of a data processing unit 35 for reviewing a medical image with respect to processing of the medical image by means of an image processing module according to a further embodiment, comprising: - an input request provision unit PQ-M for providing PQ at least one input request of the image processing module, which relates to at least one image parameter of the medical image, - an image parameter provision unit PX-M for providing PM of at least one image parameter of the medical image, - a checking unit CP-M for checking CP whether the at least one image parameter satisfies the at least one input requirement.

[0052] Fig. Figure 3 shows a flowchart for a method for reviewing a medical image with respect to processing the medical image using an image processing module according to one embodiment. Step S marks the start of the method sequence.

[0053] A parameter set P algoassigned to a post-processing module L, which describes the input requirements, and / or stored in this post-processing module. The image parameters realized in the medical image, P re- al , in the sense of the above parameter model, the ID in which the medical image is transmitted is stored in a data structure. A header or footer can be used for this, for example. P algo is represented by the reference symbol Q. P real is represented by the reference symbol X. P real may include acquisition parameters XA for an acquisition DA and / or reconstruction parameters XR for a reconstruction IR.

[0054] Then, in step CP, it is checked whether P real with P algo This can be done in particular by checking parameter by parameter whether the realized value is in the range determined by P algospecified value range. The value range can be specified, for example, in the form of a value, an interval, and / or a list. P real and P algo can in principle be viewed as sets of values ​​for certain parameters. Checking whether at least one image parameter satisfies at least one input requirement would thus be equivalent to checking whether P real a subset of P algo is, i.e. whether P real ⊆ P algo applies.

[0055] If this requirement is met, the medical image is processed by the image processing module L. This path is marked with Y. Otherwise, as in Fig. 3, the post-processing order is rejected J1 and / or, as in Fig. 4, a warning J2 is issued to the user that the results generated with the post-processing algorithm may be limited in quality. This path is marked N. Steps E1 and E2 each mark the end of the procedure.

[0056] In the Fig. In the embodiment shown in Figure 5, if the at least one image parameter does not satisfy the at least one input requirement, an image generation order is generated which generates a further medical image taking into account the at least one input requirement P algo Based on the image generation order, a further medical image can be generated by means of a medical imaging device, which meets the requirement P real ⊆ P algo fulfilled.

[0057] In step DX, based on P real and / or P algoand / or a result of the check CP, image parameters, in particular acquisition parameters and / or reconstruction parameters, are generated for generating the further medical image. These parameters differ in particular from the initial parameters that were provided for generating the medical image in step INI. The generation of the further medical image can in particular comprise a reconstruction of the further medical image based on raw data that has already been recorded. For example, the same raw data can be used that were already used for reconstructing the medical image that does not meet the at least one input requirement. In particular, reconstruction parameters that are adapted to the at least one input requirement can be used.

[0058] The generation of the further medical image can be achieved, in particular, if the reconstruction alone meets the requirement P real ⊆ P algo cannot be met, may involve acquiring raw data using a medical imaging device. In particular, acquisition parameters adapted to the at least one input requirement may be used. The further medical image may be reconstructed based on the raw data thus acquired.

[0059] The described solution thus enables automatic control of an acquisition and / or reconstruction, which is adapted for processing a resulting medical image using a post-processing application. The inclusion of a new acquisition of data is particularly useful for modalities where it does not result in additional patient exposure, for example, through ionizing radiation.

[0060] If the condition P real ⊆ P algo cannot be fulfilled based on the further medical image, the post-processing request can be rejected and / or a notice can be issued to the user that the results generated with the post-processing algorithm may be limited in quality.

[0061] In the Fig. In the embodiment shown in Figure 6, the medical image on which the first verification step is based is already generated based on an image generation job that was generated based on the at least one input request.

[0062] In particular, based on a planned post-processing step, an image generation job can be created specifying the input requirements P algo to generate a medical image and output to a medical imaging device.

[0063] In particular, based on the image generation task, a medical image can be generated using a medical imaging device such that the at least one image parameter of the further medical image satisfies the at least one input requirement. The generation of the medical image can, in particular, comprise a reconstruction and / or an acquisition.

[0064] In particular, based on the image generation order for the medical image, a reconstruction of already acquired raw data can be carried out with reconstruction parameters that are adapted to the at least one input requirement. The aim is thus to ensure that the requirement P real ⊆ P algo is fulfilled or at least approximately fulfilled. In particular, if the reconstruction alone satisfies the requirement P real ⊆ P algo If the requirement for the medical image cannot be met, an acquisition can be performed based on the image generation order for the medical image using acquisition parameters that are adapted to the at least one input requirement. The raw data thus acquired can be reconstructed, in particular, using reconstruction parameters that are adapted to the at least one input requirement.

[0065] For example, based on a reconstruction with reconstruction parameters that are adapted to the at least one input requirement, it can first be checked whether the at least one input requirement for the medical image is met. If the at least one input requirement for the medical image is not met, an image generation job for another medical image can be generated, which concerns both an acquisition and a reconstruction. If the requirement P real ⊆ P algo is met, the post-processing step can be executed. Otherwise, the post-processing job is rejected and / or a warning is issued to the user that the results generated with the post-processing algorithm may be limited in quality.

[0066] The described solution enables, in particular, verification of whether a medical image is indeed suitable as a basis for post-processing. This can lead to improved quality, particularly in terms of accuracy and / or reproducibility, of the results derived from post-processing. Furthermore, caution can be advised when interpreting the results.

[0067] With the help of the image generation order, the best possible quality can be achieved retrospectively or in advance using post-processing algorithms in the chain of acquisition, reconstruction, and derivation of results. The execution of the process steps can be automated, in particular. This reduces the effort required by the user, or at least does not increase it.

[0068] During radiological examinations, diagnostic findings may arise that require individual modification or extension of the examination protocol.

[0069] In particular, if the investigation is intended to exclude several suspicious factors, the investigation protocol is typically designed in such a way that the maximum chain of investigation steps that could be necessary to exclude a suspicious factor can be processed.

[0070] A linear sequence of examination steps can be used, corresponding to a superset of the necessary examination steps. These examination steps can include, for example, scans under different physiological conditions, particularly native scans, scans with contrast medium, scans at different phases, or similar.

[0071] After each examination step, depending on the respective finding, a decision is made as to whether the protocol can be terminated prematurely, a step skipped, or the next examination step in the sequence can be continued. This is typically done manually based on the respective institution's indication-based rules and / or the user's general evidence base.

[0072] In the case of an incidental finding, a change to the test and / or an additional test may be necessary. Regardless of whether this is performed immediately afterward or at a later date, a new test order must usually be created manually.

[0073] The solution described makes it possible, in particular, to simplify and / or automate these dynamic adjustments.

[0074] In particular, a basic protocol can be defined that contains the list of examination steps. Such a basic protocol can also be understood, for example, as a superset of the maximum scope of the examination that can be planned using the examination steps. The examination steps can, for example, be taken from a conventional examination protocol.

[0075] The assignment of instructions to examination steps can be done manually during implementation. Alternatively, or additionally, the list of examination steps can be machine-learned in conjunction with the instructions from medical imaging examinations using conventional examination protocols. For example, the user can be asked for the reason if a deviation from the linear sequence of examination steps specified in the conventional protocol occurs during the medical imaging examination.

[0076] Such a deviation may, in particular, involve skipping an examination step in the linear sequence, prematurely aborting the medical imaging examination, or similar. The examination parameter set of the training data set may, in particular, include the reason for the deviation. According to one embodiment of the invention, the order of the examination steps in the list of examination steps is irrelevant, with the exception of the start examination step. With the exception of the start examination step, whether an examination step is executed does not depend on the relative position of the examination step in the list. Rather, it depends on whether a previously executed examination step is assigned an instruction that refers to the examination step to be executed afterwards.

[0077] According to one embodiment of the invention, at least one examination step is assigned an instruction which, without any further condition regarding the result of the examination step, points to the next examination step to be performed or terminates the entire examination. According to one embodiment of the invention, at least one examination step is assigned an instruction which, depending on a result of the examination step relating to the question to be answered by the examination, points to one of several examination steps or terminates the entire examination sequence.

[0078] According to one embodiment of the invention, an instruction is assigned to at least one examination step, which, depending on a particularly unexpected finding, defines a further indication for a subsequent examination and / or selects a further examination protocol for a clinical question related to the finding. The instruction can, in particular, comprise the generation of an examination order (requested procedure / modality worklist) in a HIS / RIS (Hospital Information System / Radiology Information System).

[0079] According to one embodiment of the invention, the result of the performed examination step is determined based on an interaction with the user, for example, a user of a medical imaging device. The interaction can, in particular, comprise a question to be answered by a user and / or a list of options from which the user can select the answer. According to one embodiment of the invention, the result of the performed examination step is determined based on an automatic analysis of data generated during the performed examination step. The data can, for example, be cross-sectional images, post-processing results, or the like. In particular, an examination protocol for coronary heart disease can be adapted based on an automatically calculated calcium score during a medical imaging examination.

[0080] Fig. Figure 7 shows a schematic representation of an examination protocol for ruling out bleeding, infarction, and / or tumors in the brain. At step 71, a list of examination steps T1, T2, and T3 is provided.

[0081] In step 72, an instruction is assigned to each examination step in the list of examination steps. In step 73, start examination step 7S is defined, which indicates, for example, that examination step T1 should be started.

[0082] Examination step T1 may, in particular, involve an examination of soft tissue without contrast agent. Examination step T2 may, in particular, involve an examination of soft tissue with contrast agent, for example, with delayed contrast agent administration. Examination step T3 may, in particular, involve an examination of a tumor volume with contrast agent, for example, with delayed contrast agent administration.

[0083] In step DT1, it can be determined, in particular, whether a hemorrhage or a heart attack is present. If yes (Y), the examination continues with examination step T2 according to instruction FT1. If no (N), the examination is terminated according to instruction ET1. In step DT2, it can be determined, in particular, whether a tumor is present. If yes (Y), the examination continues with examination step T3 according to instruction FT2. If no (N), the examination is terminated according to instruction ET2. After the examination step T3 has been executed, the examination is terminated according to the instruction ET3. Fig. Figure 8 shows a flowchart for a method for training a decision support system for an examination using a medical imaging device, the method comprising the following steps: - Providing 81 a set of training data sets, each training data set comprising a training examination order and an examination information item, the examination information relating to a medical imaging examination which was carried out based on the training examination order, - Training 82 the decision support system based on the set of training data sets and a machine learning algorithm such that by means of the trained decision support system a particularly optimal examination recommendation for the examination by means of a medical imaging device can be generated based on an examination order.

[0084] Fig.Figure 9 shows a schematic representation of a medical imaging device 1 according to an embodiment of the invention. Without limiting the general inventive concept, a computed tomography device is shown as an example for the medical imaging device 1. The medical imaging device 1 has the gantry 20, the tunnel-shaped opening 9, the patient support device 10, and the control device 30. The gantry 20 has the stationary support frame 21 and the rotor 24.

[0085] The patient 13 can be inserted into the tunnel-shaped opening 9. The acquisition area 4 is located in the tunnel-shaped opening 9. In the acquisition area 4, a region of the patient 13 to be imaged can be positioned such that the radiation 27 can reach the region to be imaged from the radiation source 26 and, after interacting with the region to be imaged, can reach the radiation detector 28.

[0086] The patient support device 10 comprises the support base 11 and the support plate 12 for supporting the patient 13. The support plate 12 is arranged on the support base 11 so that it can be moved relative to the support base 11 such that the support plate 12 can be inserted into the acquisition area 4 in a longitudinal direction of the support plate 12, in particular along the system axis AR.

[0087] The medical imaging device 1 is configured to acquire acquisition data based on electromagnetic radiation 27. The medical imaging device 1 has an acquisition unit. The acquisition unit is a projection data acquisition unit with the radiation source 26, e.g., an X-ray source, and the detector 28, e.g., an X-ray detector, in particular an energy-resolving X-ray detector.

[0088] The radiation source 26 is arranged on the rotor 24 and is configured to emit radiation 27, e.g., X-rays, with radiation quanta 27. The detector 28 is arranged on the rotor 24 and configured to detect the radiation quanta 27. The radiation quanta 27 can travel from the radiation source 26 to the region of the patient 13 to be imaged and, after interacting with the region to be imaged, impinge on the detector 28. In this way, acquisition data of the region to be imaged can be acquired in the form of projection data using the acquisition unit.

[0089] The control device 30 is configured to receive the acquisition data acquired by the acquisition unit. The control device 30 is configured to control the medical imaging device 1. The control device 30 comprises the data processing unit 35, the decision support system 37, the computer-readable medium 32, and the processor system 36. The control device 30, in particular the data processing unit 35, is formed by a data processing system comprising a computer. The control device 30 comprises the image reconstruction device 34. By means of the image reconstruction device 34, a medical image data set can be reconstructed based on the acquisition data.

[0090] The medical imaging device 1 has an input device 38 and an output device 39, each of which is connected to the control device 30. The input device 38 is designed to input control information, e.g., image reconstruction parameters, examination parameters, or the like. The output device 39 is designed, in particular, to output control information, images, and / or acoustic signals.

[0091] The solution according to the invention makes it possible, in particular, to dispense with the manual application of the finding-dependent rules, which would have to be carried out each time the protocol is used, and instead to implement an automated workflow.

[0092] This increases the reliability with which the finding-dependent rules are adhered to and executed. This enables a simplified and structured examination process. In particular, the ability to generate new examination orders for incidental findings can simplify administration and billing in such cases. Particularly when managing a large fleet of radiological equipment, these dynamic rules can be centrally managed as part of the examination protocols and distributed to the individual devices. Separate maintenance and distribution of case-specific rules is thus obsolete. An examination step can, in particular, include an acquisition and / or a reconstruction and / or image processing.

[0093] In the context of this application, a machine learning algorithm is understood to mean, in particular, an algorithm designed for machine learning. A machine learning algorithm can be implemented, for example, using decision trees, mathematical functions, and / or general programming languages. The machine learning algorithm can be designed, for example, for supervised learning and / or unsupervised learning. The machine learning algorithm can be designed, for example, for deep learning and / or reinforcement learning and / or marginal space learning.In supervised learning in particular, a class of functions can be used which is based, for example, on decision trees, a random forest, a logistic regression, a support vector machine, an artificial neural network, a kernel method, Bayesian classifiers or similar, or combinations thereof.

[0094] Possible implementations of the machine learning algorithm may, for example, use artificial intelligence. Alternatively or in addition to the first machine learning algorithm and / or the second machine learning algorithm, one or more rule-based algorithms may be used. Calculations, in particular when determining the classification system based on the set of training datasets and a machine learning algorithm, may, for example, be performed using a processor system. The processor system may, for example, comprise one or more graphics processors.

[0095] In particular, data relating, for example, to a medical image, a protocol, or a training data set can be provided by loading the data, for example, from an area of ​​a storage system, and / or generating it, for example, using a medical imaging device. In particular, one step or several steps or all steps of the method according to the invention can be carried out automatically and / or by means of a component of a data processing unit, wherein the component is formed, for example, by a processor system. In particular, the medical imaging examination can be an examination using a medical imaging device and / or can be carried out using a medical imaging device.

[0096] The set of training datasets can be provided, in particular, by collecting a large number of examination orders along with the actual examinations. A decision support system can learn from this information, particularly by recognizing patterns. For example, it can be recognized that most clinics also perform coronal reconstructions during a lung scan. This allows it to be quickly identified that the user's medical imaging examinations deviate from the majority of users performing similar medical imaging examinations.

[0097] By collecting, evaluating, analyzing, and appropriately using examination data from other clinics, users can be informed about the relevant options. For this purpose, data is collected and evaluated according to clinical indications, demonstrating the optimal examination method for the respective device and the available technical capabilities.

[0098] In particular, the decision support system can generate an examination recommendation regarding the acquisition of imaging data, for example, whether a spiral or a sequence is preferable or whether additional scans, in particular an additional late phase, could enable a better diagnosis. In particular, the decision support system can generate an examination recommendation regarding the reconstruction of a medical image, for example, whether reconstruction with a different nucleus would be more suitable. In particular, the decision support system can generate an examination recommendation regarding image processing, in particular which algorithms would be preferable.

[0099] By providing and analyzing data from multiple experts, even less experienced operators can achieve optimal examination results with the appropriate use of the decision support system. The capabilities of existing medical devices can be perfectly adapted to the indications and thus fully utilized. Both inexperienced operators and experts benefit from this, as even less frequent examinations can be performed optimally. Likewise, doctors in remote areas of the world would have the opportunity to share this knowledge. This enables an improved examination for the patient. In modalities involving radiation, the dose can also be significantly reduced in this way. Diagnosis can thus be made faster and more accurately, especially since the images can be created specifically for a given problem.

Claims

[1] Method for generating a list of examination steps, each examination step being assigned an instruction for adapting an examination protocol during a medical imaging examination using a computer tomography device, the method comprising the following steps: - Providing a set of training data sets, each training data set having a training instruction and an examination parameter set, the examination parameter set relating to an examination step associated with the training instruction and / or an examination result associated with the training instruction, - Generating the list of examination steps based on the set of training data sets and a machine learning algorithm, wherein each examination step of the list of examination steps is assigned an instruction for adapting an examination protocol during a medical imaging examination using the computed tomography device, wherein at least one training data set of the set of training data sets is provided by performing the following steps during an examination process: - Recording a deviation from a selected examination protocol, which was manually initiated by a user after an examination step has been carried out, as a training instruction, - Recording a set of examination parameters relating to the examination step carried out and / or an examination result of the examination step carried out. [2] Method for adapting an examination protocol during a medical imaging examination using a computer tomography device, - wherein a list of examination steps is provided (71) by being generated according to a method according to claim 1, - where each examination step in the list of examination steps is assigned an instruction, - wherein the examination protocol comprises a start examination step from the list of examination steps, which is carried out at the start of the medical imaging examination using the computed tomography device, - wherein the examination protocol is adapted after each examination step carried out by means of the instruction associated with the examination step carried out by executing one of the following steps based on the instruction: - Termination of the medical imaging examination using the computer tomography device regardless of the result of the examination step carried out, - Termination of the medical imaging examination using the computer tomography device depending on the result of the examination step carried out, - Executing a further examination step from the list of examination steps specified in the instruction, regardless of the result of the executed examination step, - Execute one of several further examination steps from the list of examination steps specified in the instruction, depending on the result of the examination step executed. [3] Method according to claim 2, wherein depending on a result of the performed examination step, the following step is carried out: - Generating an image generation order for a further medical imaging examination using the computed tomography device based on the instruction and / or the result of the performed examination step. [4] A method according to any one of claims 2 to 3, wherein the method further comprises the following step: - Determining the result of the performed examination step based on a user input and / or based on an automatic evaluation of data recorded during the performed examination step. [5] Data processing unit (35) which is designed to carry out a method according to one of claims 1 to 4. [6] A computer program loadable into a memory device of a data processing system, comprising program sections for carrying out all steps of the method according to any one of claims 1 to 4 when the computer program is executed by the data processing system. [7] Computer-readable medium on which program sections are stored which can be read and executed by a data processing system in order to carry out all the steps of the method according to one of claims 1 to 4 when the program sections are executed by the data processing system. [8] Medical imaging device (1) comprising a data processing unit (35) according to claim 5, wherein the medical imaging device is a computed tomography device.

Images