Computer-implemented method for operating a medical imaging device and imaging device

Abstract

The invention relates to a computer-implemented method for operating a medical imaging device, wherein for evaluating an image data set of a computed tomography recorded with the imaging device, in particular by means of a multi-energy computed tomography, from which by means of image processing a processing data set reflecting different image data contents can be determined, the following steps are provided: receiving at least one user information which describes a desired display form of the image data set; providing at least one request information which describes an evaluation algorithm to be used, in particular at least one input data requested in accordance with the user information; determining at least one first processing data set corresponding to the display form in accordance with the user information and at least one second processing data set which can be used as input data for the respective evaluation algorithm in accordance with the request information; applying at least one evaluation algorithm to the respective second processing data set in order to determine evaluation information; and outputting the first processing data set and the evaluation information to a display device.

Description

TECHNICAL FIELD

[0001] The invention relates to a computer-implemented method for operating a medical imaging device, wherein an evaluation of an image data set, in particular a computed tomography, recorded with the imaging device, in particular by means of a multi-energy computed tomography, is carried out, from which by means of image processing a processing data set can be determined which reflects different image data contents. Furthermore, the invention relates to an imaging device, in particular a computed tomography device, a computer program and an electronically readable data carrier. BACKGROUND

[0002] Medical imaging is the basis of modern medical diagnostics. By means of increasingly advanced imaging methods and imaging techniques a large amount of information about the interior of the human body can be obtained, which, however, is not always directly visible by observing a processing data set derived from an image data set in a display form. Therefore, software means, in particular evaluation algorithms for evaluating the data set, play an increasingly greater role in daily clinical practice. In addition to conventionally programmed evaluation algorithms or corresponding computer programs, the use of artificial intelligence is increasingly proposed, so that evaluation algorithms are known which comprise a training function, for example a neural network. Evaluation algorithms, in particular those using artificial intelligence, are used to solve clinical problems, for example to track blood vessels in the human body, for which the evaluation algorithm is trained with respect to input data or developed for input data, in particular sufficiently unambiguous display of the image contents to be evaluated, so that later reliable result information (output data) can be provided. This means that many evaluation algorithms depend on the data pool which has been used for the training or development of the evaluation algorithm and only provide reliable results on the basis of similar input data. In clinical practice, it is often also the case for conventional imaging that the image data used for training or development of the evaluation algorithm are sufficiently similar to the image data of the field of application, so that the evaluation algorithm trained on the data pool can be used for any image of the field of application. This is due to the fact that the recording mode sets the final image characteristics within certain limits, which cannot be changed afterwards.

[0003] However, advanced imaging techniques enable a variety of processing variants in order to emphasize or specifically display diverse image content. In addition to the fact that using different filters, specifically different filter kernels, can produce significantly different image impressions in computed tomography data sets, there is also a particularly great diversity in so-called spectral X-ray imaging, also referred to as multi-energy X-ray imaging or multi-energy imaging. For example, in the special case of multi-energy computed tomography, two different X-ray spectra will be produced at the radiation source side of the recording device in order to obtain two different partial data sets. So-called photon-counting X-ray detectors ("photon-counting CT") are also increasingly being accepted in modern computed tomography devices. Here it can be said that individual incidence events of X-ray quanta on the X-ray detector can be counted and evaluated in terms of their energy, so that events are, for example, classified into energy histograms, so that partial data sets also arise for different energies and / or energy intervals. In multi-energy computed tomography, it is possible that image properties can also fundamentally change after the recording of the image data set is complete, so that the processed data produced are no longer comparable / sufficiently similar. This applies, for example, to the calculation of mono- energy images, basis material decomposition, i.e. the calculation of the equivalent densities of two attenuating materials, and material differentiation, in which different materials are highlighted or suppressed differently in the image. Therefore, in the case of multi-energy computed tomography, it is not possible to sufficiently train evaluation algorithms with a generalized data pool within a reasonable cost category.

[0004] This in turn leads to the result that if the recording of the image data set is used for a specific examination purpose, the user of the imaging device designed as a computed tomography device is severely restricted in the choice of the display form. Instead of the display form that the user would most like to see, he is restricted in the display form that is suitable as input data for the evaluation algorithm. This in turn leads to the result that the user either cannot use the full range of possibilities of spectral X-ray imaging or has to do without the support of the evaluation algorithm. In any case, the user has to make compromises, for example in order to be able to make a diagnosis, and cannot exploit the full potential of the imaging and diagnostic tool.

[0005] From this, in one example, the calculation of an unrolled spine is less successful in the case of a calculated mono-energy image of 40 keV, but only has sufficient reliability at higher energies, specifically at more than 70 keV, even though the user can clearly identify the spine in both cases. In another example, so-called VNC images (virtual non-contrast images) are helpful to the user for certain questions, for example the degree of calcification, but, for example, a blood vessel display cannot be calculated on the VNC image.

[0006] In this connection, it has already been proposed in the prior art to train the evaluation algorithms individually and repeatedly for different display forms in order to provide the use of the evaluation algorithms for as many display forms as possible. However, this can only be carried out very restrictively, since on the one hand this means a separate development for each display form, while on the other hand, particularly in the case of multi-capable computed tomography, there are a number of display forms which cannot be processed meaningfully by the evaluation algorithms but which have a clinical benefit, such as the above-mentioned blood vessel display without contrast agent, which would result in the algorithm no longer finding blood vessels. SUMMARY

[0007] It is therefore an object of the present application to propose an evaluation method which is comparatively improved and which in particular allows synergistic effects.

[0008] This object is achieved by the method, the imaging device, the computer program and the electronically readable data carrier according to the application.

[0009] According to the application, a computer-implemented method for operating a medical imaging device is proposed for evaluating an image data set recorded with the imaging device, in particular by means of multi-capable computed tomography, in particular computed tomography, from which image data set processing data sets reflecting different image data contents are determinable by means of image processing, the method comprising the following steps:

[0010] receiving at least one user information which describes a desired display form of the image data set,

[0011] providing at least one request information which describes at least one input data requested for the evaluation algorithm to be used, in particular in accordance with the user information,

[0012] determining at least one first processing data set corresponding to the display form in accordance with the user information and at least one second processing data set which is usable as input data for the respective evaluation algorithm in accordance with the request information,

[0013] applying the at least one evaluation algorithm to the respective second processing data set in order to determine evaluation information, and

[0014] outputting the first processing data set and the evaluation information to a display device.

[0015] In this connection, in an embodiment, the recording of the image data set itself can also constitute a part of the method, since, for example, it is known that, already before the examination, i.e. the recording of the image data, the desired display form is specified not only at the input device of the imaging device, but also that an evaluation algorithm or the required evaluation information is already provided in accordance with the examination purpose. This applies in particular after prior art design concepts, in which first an evaluation algorithm or the desired evaluation information to be used is selected, which then limits the available display forms, as described above. The present application can now particularly advantageously allow a selection from the entire available pool of display forms and evaluation algorithms, since, independently of the selected display form, a further second processing data set is determined in the background, which makes it possible to use the respective evaluation algorithm. The user information can in particular be received from the input device of the imaging device, for example via an interface of a control device for carrying out the method, so that it advantageously already contains information about the evaluation algorithm to be used and / or the desired evaluation information which makes it possible to determine and provide the appropriate request information, in addition to the information about the desired display form, if necessary and further processing parameters of the first processing data set. The request information of different evaluation algorithms to be used can be provided, for example, in a respective database and / or look-up table.

[0016] Both the user information and the request information can comprise processing parameters which describe how the respective processing data set is to be derived from the image data set. Thus, for example, it can be provided that the user information comprises at least one preparation output processing parameter, in particular a slice thickness and / or a rotation increment. Furthermore, the user information can comprise image content parameters which describe the desired image content, in particular contrast information, in particular also as processing parameters. As described above, it can be considered that the user information is received from a user interface, in particular an input device, of the imaging device.

[0017] The imaging device is particularly advantageously a computed tomography device, since, in particular in the field of computed tomography, a plurality of recording techniques exist, in which processing data sets of a display form can be derived from the generated image data sets, which highlight or display different image content. However, in principle, it can also be considered within the scope of the present application to apply the method described here to other types of imaging devices, for example two-dimensional imaging X-ray devices or magnetic resonance devices. In the case of magnetic resonance devices, for example, image data sets can be considered which, for example, display different contrasts in partial data sets and / or from which processing data sets can be derived which are related to different contrasts, for example. One example in the field of magnetic resonance imaging is the so-called Dixon technique, in which partial data sets are recorded at different echo times in order to be able to distinguish between spin types which have a chemical shift between them, for example spins of protons bound in fat and spins of protons bound in water.

[0018] By means of the method according to the application, the user can directly at the imaging device, preferably by an input made there, prepare to obtain all the information he needs, wherein as the image data set becomes available, the user will also correspondingly obtain the outputted information (without further necessary interaction) for example at a display device in the immediate vicinity of the recording component. In principle, in addition to the originally image recording device with the recording component, the imaging device can also have an evaluation workstation with a computing device for the diagnosis and / or an archive system with a computing device, wherein the evaluation step can also be made there. Of course, the evaluation method according to the application can in principle also be used for other computing devices in which an image data set is present, however it has a particular advantage in the context of the imaging device itself.

[0019] Finally, the application proposes a combined evaluation system, wherein the display form of the recorded image data set displayed to the user is decoupled from the processing data set used for the calculation of the at least one evaluation algorithm. The results of the individual sub-steps, i.e. the determination of the first processing data set and the result information of the at least one evaluation algorithm, are finally summarized, so that the applicant will obtain the evaluation algorithm results, in particular in the form of evaluation information, in addition to the desired display form as output.

[0020] In this method, the user information obtained by the user input about the desired display form is used to determine a first processing data set having this display form, which can also be referred to as a first auxiliary body. Furthermore, for each evaluation algorithm to be used, processing parameters, i.e. request information, suitable for the display form of the respective evaluation algorithm are determined, for example queried, wherein with this request information at least one second processing data set, in particular as a second auxiliary body, is determined, which is provided as input data to the respective evaluation algorithm in order to obtain the result information of the evaluation algorithm, which can then be used as or for the determination of the evaluation information.

[0021] Thus, an algorithm-independent infrastructure is advantageously used for a generalized solution of the problem of optimal evaluation and information provision, instead of adjusting each evaluation algorithm and each display form separately. The user input is decoupled from the algorithmic calculation of the first processing data set, i.e. the user image, which in particular makes it possible to apply the result of the at least one evaluation algorithm to a processing data set, i.e. the first processing data set, which is not suitable for the evaluation algorithm, which will be discussed in more detail below.

[0022] In comparison to an extensive training of the evaluation algorithms, for example, the method according to the application makes it possible to improve the quality of the results. In comparison to an adaptation of the evaluation algorithms, the effort is significantly reduced. Furthermore, by means of the application, consistent evaluation algorithm results are available for the initial diagnosis and the progress check for the evaluation of different display forms, in other words also a standardization is achieved. The solution described here is scalable to an arbitrary number of evaluation algorithms. This is different from previously known prior art solutions, in which the intersection of the display forms becomes smaller and smaller when a plurality of evaluation algorithms is used, which in the extreme case leads to the fact that there is no longer a possibility of implementation, since no display form is suitable as input data for all evaluation algorithms.

[0023] Furthermore, in the application of the application, the user no longer needs to know the internal structure of the evaluation algorithms in order to select a suitable display form, since this is decoupled from the display form desired by the user at least for the application of the evaluation algorithms, and the individual results are only merged downstream. Here, it should also be noted at this point that in principle it is also possible to provide specific evaluation information without selecting a specific display form for this or without having to be selected by the user. The decision of the user on the display form, i.e. the visualization, does not mean any limitation or restriction on the diagnostic tools used.

[0024] In summary, therefore, the application provides a simplified, less limited operation, a higher flexibility and an improved evaluation and display of the results in terms of quality and quantity for imaging devices.

[0025] Here, a basic aspect of the application is the fact that the first processed data set is not suitable for obtaining meaningful results from it by means of the at least one evaluation algorithm. In other words, the first processed data set in particular does not meet the requirements of the requested information. The image content necessary for obtaining a suitable evaluation result by means of the evaluation algorithm is at least not to a sufficient extent available. However, according to the application, evaluation information can be determined and placed in the context of the first processed data set, in particular of its display form.

[0026] Although the evaluation algorithms, which do not use artificial intelligence, are basically developed for specific image content, so that specific input data can be required, the application can be particularly advantageously applied in the case in which at least one of the at least one evaluation algorithm comprises a training function, which has been or will be trained by using a corresponding second processed data set. The training function can also be referred to as an artificial intelligence algorithm, which in general maps cognitive functions associated with the human brain. By means of a training based on training data, the training function is able to adapt to new environments, which must be detected and extrapolated.

[0027] Thus, in general, the parameters of the training function can be adapted by training (machine learning). In particular, supervised learning, semi-supervised learning, unsupervised learning, reinforcement learning and / or self-adaptive learning can be used. Furthermore, also representation learning (another expression is "feature learning") can be used. The parameters of the training function can in particular be adapted iteratively by various training steps.

[0028] The training function can comprise, for example, a neural network, a support vector machine (SVM), a decision tree and / or a Bayesian network. The training function can be based on k-means clustering, Q-learning, a genetic algorithm and / or a rule of assignment. The neural network can in particular be a deep neural network, a convolutional neural network (CNN) or a deep CNN. Furthermore, the neural network can be an adversarial network, a deep adversarial network and / or a generative adversarial network (GAN).

[0029] By using the application, it is sufficient to use a specific type of processing data set as training data to train the respective evaluation algorithm, which is also provided as second processing data set in practice according to the request information. In this way, an extremely robust, reliable evaluation algorithm can be created, since it is not necessary to compensate for differences in the display form.

[0030] In principle, the image data sets can be processed in different ways. Thus, for example, it can be computer tomography data sets to which filters with different filter kernels are applied. Such filters are used, for example, to reconstruct a three-dimensional processing data set from two-dimensional projection images of the image data set and can contribute to improving the highlighting of specific anatomical structures and / or are particularly suitable for specific body parts. Here, with different filter kernels completely different image results can be achieved, resulting in completely different display forms. For example, the image content of a filter kernel produced specifically to highlight blood vessels within the torso differs significantly from the image content of a filter kernel for the torso in relation to the remaining anatomical structures. Thus, there are already significant differences at this level, which justify the use of the application.

[0031] However, it is particularly preferred in the context of the application that the image data set is a multi-energy data set, i.e. a multi-energy computed tomography image data set, and comprises different sub-data sets assigned to different energies and / or energy intervals, wherein in the context of the processing a combined data set is determined as a processed data set, in particular by a pixel-wise calculation of the sub-data sets. As mentioned in the outset, multi-energy tomography is an energy-resolving computed tomography imaging which is mainly used in the medical field and is also referred to as spectral imaging. In this imaging technique, the energy dependence of the X-ray attenuation is used to obtain additional information which can be used in various ways in the context of the evaluation. One well-known and frequently used example for evaluating multi-energy data sets is the so-called material decomposition, in which the proportion of individual materials or material classes within the image data set can be determined, so that for example material images or different combined images can be generated. Contrast agents can also be used in the context of examinations with multi-energy computed tomography, which can comprise a substance which exhibits completely different energy-dependent attenuation properties than at least most of the anatomical structures, so that it can be distinguished and extracted by evaluation of the multi-energy data set. A contrast agent which is frequently used for examinations of the vascular system is iodine.

[0032] In this regard, there are a large number of possible embodiments in which the application can be used.

[0033] For example, if a user wants to evaluate the viability of a tumor in the liver, he will be interested in the uptake of contrast agent, in particular iodine, in the liver, but it is also important that the liver can be localized. However, an evaluation algorithm which segments the liver cannot work on the iodine image, as desired by the user in this regard, as a first processed data set. It is therefore possible to determine a second processed data set from the requested information of the evaluation algorithm and to provide it to the evaluation algorithm, which clearly shows the anatomical structures and for example also highlights the organ, in particular the liver, wherein the localization information or segmentation information about the liver can be particularly advantageously displayed directly as evaluation information together with the first processed data set.

[0034] In another example, in order to be able to identify calcifications, the user can request a VNC image (virtual non-contrast image) as a display form, so that a corresponding first processed data set is determined. However, this does not show the contrast agent, for example iodine, so that an evaluation algorithm which aims to derive vascular tree properties cannot reasonably work on it, so that here too a suitable second processed data set, for example an iodine image, can be generated in the background on the basis of the corresponding requested information, and the evaluation information generated can be merged with the first processed data set into a result to be displayed. Another important problem when using iodine contrast agents is that it is sometimes more difficult to distinguish between blood vessels filled with iodine contrast agent and bone, so that an evaluation algorithm can also involve the detection and localization of bone material, which can then be removed from the display and / or can be abstracted from the evaluation by another evaluation algorithm.

[0035] Thus, more generally, if the image data set recorded as a pluripotent data set shows at least one anatomical region of the patient having blood vessels with an iodine contrast agent, it can be provided that at least one evaluation algorithm involves the detection and localization of bone material and / or at least one evaluation algorithm involves the segmentation and / or localization of anatomical structures, in particular at least one blood vessel and / or a tumor and / or an organ.

[0036] The pluripotent data set can be recorded by using different X-ray spectra and / or by using a photon counting X-ray detector. In the case of the use of an X-ray tube, different X-ray spectra can be produced, for example, by using different tube voltages, wherein the respective X-ray spectrum has a maximum at a different X-ray energy. However, in the context of the present application, the use of a photon counting X-ray detector is preferred. Photon counting is a technique in which individual photons are counted by using a corresponding X-ray detector, which can also be referred to as a "single photon detector" (SPD). The energy of each detected photon can also be determined, so that, for example, event inputs can be made into an energy histogram comprising "bins" defined by energy intervals. From this, sub-data sets of a particular energy and / or a particular energy interval can then be derived.

[0037] In a particularly advantageous refinement of the application it can be provided that, if the requirements of the requested information cannot be completely fulfilled by using the image data set, a processing data set is determined as a second processing data set, which processing data set can be determined from the image data set and is similar to the data set required in accordance with the requested information, in particular contains as many of the at least one image contents defined by the requested information as possible. Depending on the recording parameters used for recording the image data set, for example the X-ray spectrum used in the case of photon counting or the selection thereof in the case of a plurality of X-ray spectrum measurements, it can be taken into account that a particular processing data set can not be derivable from the image data set because of a lack of information, for example cannot be measured in the relevant energy range. In this respect, this refinement of the application provides that it is first checked whether the display form in accordance with the requested information, i.e. the target input data, can be calculated from the recorded image data set at all. If this is not the case, the most similar processing data set can be determined, which can be calculated using the image data set. For example, if a mono-energetic image at 70 keV is required as input data for the evaluation algorithm, which cannot be determined from the image data set, a mono-energetic image at 50 keV can be used as the most similar second processing data set which can be determined. It can be specified here in particular that the requested information is provided alternatively describing the processing data set which can be determined, so that it is obtained as a necessary hint which input data can alternatively be used, while the result information of the evaluation algorithm has sufficient reliability. However, additionally or alternatively, it can also be taken into account that the processing parameters for determining the similar processing data set are determined by means of an assignment rule describing the degree of similarity. Such an assignment rule can be, for example, a database and / or a look-up table, in which processing data sets and / or requested information are assigned similar processing data sets, in particular their processing parameters, and / or in which processing parameters can be assigned an allowed deviation for determining a processing data set which can be, for example, part of the requested information.

[0038] A particularly advantageous design of the application is obtained when the evaluation information is evaluated and / or output together with the first processing data set, in particular when the display result to be output is determined from the evaluation information and the first processing data set as the display data set. In this embodiment, the results of the individual steps, i.e. in particular the first processing data set and the evaluation information, are also combined in terms of processing technology, wherein, for example, the fact can be utilized that both relate to the same image data set and in particular even exist in the same coordinate system without the need for registration. Thus, in particular, the evaluation information can be associated with the first processing data set and the display form for which the evaluation information would not have been suitable at all. By this means, significant benefits in terms of information and knowledge about the anatomical region displayed in the image data set are achieved, in other words the quality and quantity of the evaluation result is improved. Furthermore, it is possible to visualize the evaluation information in an intuitively easily understandable display form, i.e. the first processing data set, desired by the user. On the other hand, the quality of the first processing data set can also be optimized by using the evaluation information, in particular in terms of the examination goal and the improved user evaluation.

[0039] In particular, it can be provided, for example, that at least one position-dependent part of the evaluation information is output in the first processing data set in particular by superimposition in full positional correspondence. In this context, it can be provided, for example, that if the evaluation information comprises segmentation results and / or other localization results, the segmented and / or localized anatomical structures or other structures can be visualized in full positional correspondence in the first processing data set, even if the structures are not visible there. Thereby, for example, the blood vessel distribution can be displayed in a VNC image and / or other anatomical components, for example organs, can be displayed in a pure contrast agent image. However, information other than segmentation and localization can also be displayed in full positional correspondence, for example the size of an anatomical structure or other structure measured at a particular point, for example the vessel diameter, the tumor size, etc. In this case, for example, a section can be displayed whose vicinity the respective value can be seen. Obviously, many different possibilities exist.

[0040] However, it is also conceivable in a particularly advantageous embodiment of the application that at least a part of the evaluation information is used to modify the first processed data set before the output. This means that the display form can be optimized by further processing steps with the aid of the evaluation information, either outside or within the user preset. For example, a specific design proposal in this context provides that the evaluation information used for the modification relates to image portions to be removed in the first processed data set, in particular material classes to be removed. The image portions are then removed accordingly for the modification. In this regard, one example is bone material, which can be confused by the user, for example in the examination of the vascular system, with calcification and / or contrast agent, so that a corresponding evaluation algorithm has already been proposed to mark individual pixels as containing bone, which of course can also be transferred to the first processed data set on the basis of the same underlying image data set. It is also possible in other words to remove the image portions to be removed, for example bone material, for the modification, for which known methods can be used, in particular interpolation and / or extrapolation and / or other ways of calculating the information to be removed. In this way, the first processed data set will eliminate unwanted image content, so that it can be interpreted significantly better and focused on the relevant desired information. Of course, other possibilities can also be considered in principle to use the evaluation information to improve the first processed data set. For example, if the visualization or recognition of a specific organ with a known location is involved, edge sharpening or the like can be carried out locally.

[0041] However, it is also conceivable in a particularly advantageous embodiment of the application that at least a part of the evaluation information is used to modify the first processed data set before the output. This means that the display form can be optimized by further processing steps with the aid of the evaluation information, either outside or within the user preset. For example, a specific design proposal in this context provides that the evaluation information used for the modification relates to image portions to be removed in the first processed data set, in particular material classes to be removed. The image portions are then removed accordingly for the modification. In this regard, one example is bone material, which can be confused by the user, for example in the examination of the vascular system, with calcification and / or contrast agent, so that a corresponding evaluation algorithm has already been proposed to mark individual pixels as containing bone, which of course can also be transferred to the first processed data set on the basis of the same underlying image data set. It is also possible in other words to remove the image portions to be removed, for example bone material, for the modification, for which known methods can be used, in particular interpolation and / or extrapolation and / or other ways of calculating the information to be removed. In this way, the first processed data set will eliminate unwanted image content, so that it can be interpreted significantly better and focused on the relevant desired information. Of course, other possibilities can also be considered in principle to use the evaluation information to improve the first processed data set. For example, if the visualization or recognition of a specific organ with a known location is involved, edge sharpening or the like can be carried out locally.

[0042] Although in the scope of the present description primarily particularly suitable examples of application of the multi- purpose computer tomography are discussed, one advantage of the present application is that it can be extended arbitrarily to very large application fields, since various evaluation algorithms and display forms can be used for various image data sets. For example, if an evaluation algorithm should be added as a further possibility, this can be done without problems without modification or reprogramming.

[0043] A large number of further specific possibilities also exist for the first processed data set or the display result, in particular with respect to the desired display form. Thereby, for example, a display form can be desired which allows a kind of walkthrough through hollow organs or the like, so that for example cross-sectional images are created in a plane which is perpendicular to the local orientation direction of the hollow organs or the general structure. For example, in this respect it is known to "unfold" the distribution of a blood vessel, for example the aorta, wherein in this display particularly advantageous evaluation information, for example the radius, flow information or the like, can also be integrated.

[0044] In addition to the method, the present application also relates to an imaging device, in the present case a computer tomography device, which has a control device which is designed to carry out the method according to the present application. All statements relating to the method according to the present application can similarly apply to the imaging device according to the present application, so that thereby also the advantages already mentioned can be obtained.

[0045] The imaging device comprises at least one recording assembly, which for example can consist of an X-ray emitter and an X-ray detector, in particular a photon counting X-ray detector, wherein in a computer tomography device the recording assembly can be rotatably supported for example in a gantry around a patient. In general, the control device comprises at least one processor and at least one storage device, wherein by means of the processor functional units for carrying out the method according to the present application can be realized.

[0046] In the storage device of the control device for example evaluation algorithms, image data sets and / or request information assigned to the evaluation algorithms can be stored. The control device can advantageously comprise an interface for receiving user information, which can also describe the evaluation algorithms to be used. The control device can for example in particular comprise:

[0047] a providing unit for providing request information,

[0048] a first determination unit for determining a first processed data set,

[0049] a second determination unit for determining a second processed data set,

[0050] an evaluation unit for applying at least one evaluation algorithm to the respective second processed data set,

[0051] - a merging unit for merging the first processing data set and the evaluation information into a display result (display data set), and

[0052] - an output unit for outputting the display result on a display device.

[0053] Of course, other functional units can also be considered for other design variants of the application, and the control device can in particular also have a recording unit for recording the control image data set.

[0054] A computer program according to the application can for example be directly loadable into the memory of a computing device, in particular of a control device, of an imaging device, and have program means for executing the steps of the method according to the application when the computer program is run on the computing device. The computer program can be stored on an electronically readable data carrier according to the application, and thus comprise control information stored thereon, which comprises at least one computer program according to the application and is designed in such a way that, when the data carrier is used in a computing device, in particular a control device, of an imaging device, the computing device is caused to execute the steps of the method according to the application. The electronically readable data carrier can be a non-transitory data carrier, for example a CD-ROM. BRIEF DESCRIPTION OF DRAWINGS

[0055] Further advantages and details of the application result from the embodiments explained below and by means of the drawings. Therein:

[0056] Figure 1 A general flow chart of an embodiment of the method according to the application is shown,

[0057] Figure 2 A diagrammatic illustration of a specific embodiment of the method according to the application is shown,

[0058] Figure 3 An imaging device according to the application is shown, and

[0059] Figure 4 A functional structure of a control device of an imaging device is shown. DETAILED DESCRIPTION

[0060] Figure 1 A general flow chart of an embodiment of the method according to the application is shown. Here, the method focuses on the evaluation of an image data set of a computed tomography, here a multi-energy computed tomography, but can also include the recording thereof and thus the control of the entire examination process at an imaging device, here a computed tomography device.

[0061] In the present embodiment, at the beginning of the examination process at the imaging device, in step S1 an examination is prepared. This preparation also includes that a user makes settings at a user interface of the imaging device, which also comprises an input device, regarding the recording of an image data set and a subsequent evaluation. Here, these user pre-settings also include user information which is used in the method explained here, which is received by the user interface in step S1. The user information describes the display form which the user desires for the image data set to be recorded, for example whether the user wants to see mono-energetic images at a certain X-ray energy, whether the user wants to see VNC images or certain material images. Here, various processing parameters regarding the display form can be set, for example layer thickness, angular step size in a rotatable image, filters to be used, etc. More complex display forms can also be considered, for example a view of a hollow organ, in which a cross-sectional image perpendicular to the extension course is calculated.

[0062] In addition to the display form desired by the user, the user information can also describe which evaluation information is desired, directly or indirectly, for example in the context of a certain evaluation package. The evaluation information desired or meaningful can also result from the examination goal, sometimes also directly from the selected display form. Thereby, different examination goals can for example be assigned a certain display form and a certain useful evaluation information. For example, if the user is interested in the blood flow in a lesioned tissue using a contrast agent which is transported in the blood, for example an iodine contrast agent, the display form can relate to the contrast agent, however wherein the position of the lesioned tissue or at least the relevant organ represents useful evaluation information in order to make the display easier to understand on the one hand and to carry out further evaluation steps, for example to also quantify the blood flow in the region of interest, etc. For example, if a calcification of a blood vessel is involved, the display of the blood vessel and, if necessary, further evaluation information regarding the blood vessel, for example its radius or flow, is useful.

[0063] The significant thing about the two examples described above is that here the display form is very useful for the examination goal, but is not suitable for determining the mentioned meaningful evaluation information therefrom. Here, the method shown provides a remedy, as will be shown in the following, wherein in the present case the desired display form and the evaluation information are determined completely automatically by the imaging device based on the user information alone and are combined in a coordinated display result. In the present case, this is all done compactly at the imaging device and by the imaging device, the image data set on the basis of which the evaluation is based is initially recorded in step S2, as described above, which can but does not have to form part of the method.

[0064] The image data sets recorded in step S2 are multi-energy data sets, so that the multi-energy data sets have sub-data sets for different energies or energy intervals of the X-ray radiation. Here, not only different X-ray spectra on the X-ray emitter side can be used, for example different tube voltages for an X-ray tube, but also photon-counting X-ray detectors can be used, which are preferred because in particular a fixed X-ray spectrum can be used and if necessary a more fine resolution of the energies is possible.

[0065] The automatic evaluation of the image data sets in steps S3 to S8 is carried out by the control device of the imaging device without further user interaction. As a basis, in step S3, in addition to the already existing user information which is intended for each evaluation algorithm for determining the evaluation information, request information is also provided. The request information describes which input data are required by the respective evaluation algorithm and can also include suitable processing parameters for processing the image data sets in order to obtain these input data. The request information can for example be assigned to each evaluation algorithm and already exist in the storage means of the control device of the imaging device, however additionally or alternatively can also be called up in some other way. The same applies to the evaluation algorithms themselves. It is preferred here in any case that the evaluation algorithms to be used can also be derived at least partially from the user information, wherein for example the above-mentioned assignment of meaningful evaluation information to the examination object can be used. The display form can also be assigned useful evaluation information. Other information can be used, for example the recorded body region. It is also conceivable that a specific examination protocol, for example "multi-energy tomography of the chest - vascular system", is assigned evaluation algorithms and evaluation information which are in principle advantageous.

[0066] The processing data sets are then determined independently of and separately from one another in steps S4 and S5, on the one hand in step S4 a first processing data set is determined from the user information which reflects the desired display form, and on the other hand in step S5 at least one second processing data set is determined which is suitable as input data for at least one of the evaluation algorithms. It is noted at this point here that at least a part of the evaluation algorithms available at the imaging device can comprise a training function, i.e. an artificial intelligence. These evaluation algorithms, more precisely the training functions of the evaluation algorithms, are also trained with the respective input data, i.e. the second processing data sets derived from the training image data sets. Thereby, the desired result information for the evaluation information will be provided in a robust and reliable manner if suitable input data are provided.

[0067] The respectively determined second processing data set can then in step S6 itself correspond to a specific display form, i.e. the image content required for the display evaluation algorithm, which second processing data set is provided as input data to the respective evaluation algorithm which determines the respective result information from which the evaluation information is then derived.

[0068] At this point it is also noted here that it can be considered that due to the recording parameters the image dataset does not suit the requested information for fully satisfying at least one of the at least one evaluation algorithm. It can then be advantageously alternatively generated a similar processed dataset providing at least the same image content and provided as input data. The requested information itself can already contain the respective alternative solution. However it can also be considered to use an assignment rule, for example a database or a lookup table, in which the requested information or a second processed dataset satisfying the requested information is assigned an alternative solution, for example by providing corresponding processing parameters. The processing parameters can also be directly assigned an interval or an allowed deviation. Nonetheless, in this way the respective evaluation algorithm can still be used.

[0069] In step S6 the result information can also be further associated to the desired evaluation information in part if the result information does not directly represent the evaluation information. This means that it can be considered to use a plurality of evaluation algorithms with different result information in order to determine the actually intended evaluation information therefrom.

[0070] The results of steps S4 and S6 are then combined in step S7 to obtain a display dataset as a display result. In this regard it can be considered on the one hand that the extended information is supplemented by for example merging the at least one result information and the processed data of the first processed dataset of step S4 in an evaluation manner, for example in the sense of local image information in the processed data of the first processed dataset.

[0071] However, it is provided in this context in any case that the evaluation information is evaluated together with the first processed dataset in step S7. For this at least a part of the evaluation information is used to modify the first processed dataset before the output, for example for removing unwanted image components or the like. However, at the same time the display result is also generated in step S7 by this joint evaluation, in which the evaluation information is at least partially visualized within the first processed dataset.

[0072] For this it is particularly advantageous to output the position-dependent part of the evaluation information in the first processed dataset in full positional correspondence, in which for example an overlay can be made. It can be considered for example to display the borders of detected or segmented anatomical structures or other structures or even quantitative data thereon in full positional correspondence, thereby including them into the display result.

[0073] The first processed dataset and the evaluation information are then output together in the form of the display result in step S8. Here it is of course also considered to additionally output the position-independent, in particular general part of the evaluation information, for example adjacent to the image output of the first processed dataset.

[0074] In addition to being output to a display device, the display results can also be stored, particularly along with the image dataset and / or the first processing dataset and / or evaluation information.

[0075] Figure 2 The schematic diagram illustrates the following based on Figure 1 An exemplary application of the method is provided. Here, the user information 1 input for the examination process may describe, for example, that at least a recorded VNC image of the aorta is required as the display form, which needs to be reconstructed using a filter specifically adapted for the reproduction of the vascular system. A thin-slice dataset for display needs to be prepared, and it needs to be able to rotate in 10° increments. In addition, the radius region of the aorta should be analyzed based on the user information.

[0076] Furthermore, it is known in the control device that during such a recording process (here, multi-potential chest recording), the image dataset may contain unwanted skeletal portions that should be removed.

[0077] Accordingly, three processing datasets are determined. For the purpose of display, the first processing dataset 2 is a VNC image with a thin layer using the desired filter.

[0078] The first processing dataset 2 will be completely unsuitable for determining the radius of the aorta because the contrast agent (in this case, iodine contrast agent) will be invisible in the VNC image, making the vessel almost unidentifiable automatically. However, in the current situation, at least aortic segmentation is necessary, for which the corresponding evaluation algorithm 3 should be used, whose input data, as described by the corresponding request information 22, should include a monoenergetic 70keV image with a thin layer and a specific filter during reconstruction. The corresponding determination unit in the control device precisely determines such a second processing dataset 4.

[0079] Regarding the removal of bone material, another evaluation algorithm 5 should be used, which provides pixels containing bone as result information, and here also as evaluation information. Its input data requires a monoenergetic 55keV image with a thin layer and a specific filter element, which is also provided as another second processing dataset 6.

[0080] In a merging step (S7) indicated here by an arrow 8, the skeletal parts in the first processed data set 2 are now calculated using the evaluation information of the evaluation algorithm 5. At the same time, the aorta segmented as position-dependent evaluation information by the evaluation algorithm 3 is marked by superimposition in the first processed data set 2. The aorta radius determined as part of the evaluation can likewise be displayed in different locations in full position correspondence, at which the user can place it directly in the background of the VNC image in the calcifications to be viewed. Thereby a display result 9 is produced, which is then output onto a display device, namely a VNC image with a thin layer and which can be rotated in steps of 10°, which is reconstructed by using the required filter kernel, which additionally contains information about the aorta and its radius in full position correspondence, although this information cannot be obtained from the first processed data set 2. Furthermore, no interfering skeletal parts are seen any more.

[0081] It should be noted that, of course, other evaluation information can also be determined by using the processed data of the first processed data set 2, in particular after removal of the skeletal parts, for example calcifications can be detected purposefully and can also be quantified in part by using the bounding box determined from the segmentation of the aorta, etc.

[0082] Figure 3 A schematic diagram of an imaging device 10, here a computed tomography device, according to the application is shown. It is known in principle that the imaging device 10 has a gantry 11 comprising a patient accommodation 12 into which a patient 13 can be moved by means of a patient couch 14 in such a way that the anatomical region to be recorded lies in the visible range of an accommodation assembly rotatably supported in the gantry 11. Here, the recording assembly comprises an X-ray emitter 15 and an X-ray detector 16, here a photon counting X-ray detector 16.

[0083] The operation of the imaging device 10 is controlled by a control device 17, which is also designed to carry out the method according to the application.

[0084] Furthermore, the imaging device 10 also has a user interface 18, which here comprises an input device 19 and a display device 20.

[0085] Figure 4 The functional structure of the control device 17, which can carry out the method according to the application, is shown in more detail. The control device 17 firstly comprises a storage 21, in which, in addition to the evaluation algorithms 3, 5 and their assigned request information 22, the assignment rules can also be stored, if applicable. Furthermore, the storage 21 is suitable for storing user information obtained from the user interface 18 via an interface 23, just as image data sets recorded by a recording unit 24, in particular according to step S2.

[0086] The processing data sets 2, 4, 6, the display results 9, the evaluation information, etc. can also be stored in the storage device 21 at least for the checking process or the evaluation thereof.

[0087] For carrying out the method according to the application, the control device 17 has first of all a provision unit 25 for providing, in accordance with step S3, the respective required request information 22, which is selected here. The first processing data set 2 is determined in a first determination unit 26, and the at least one second processing data set 4, 6 is calculated in a further second determination unit 27, wherein the processing data sets 2, 4, 6 are determined from the image data sets, respectively. The first determination unit 26 thus carries out step S4, while the second determination unit 27 carries out step S5.

[0088] In accordance with step S6, at least one evaluation algorithm 3, 5 is implemented in an evaluation unit 28 by using the respective second processing data set 4, 6 as input data, in order to obtain the result information / evaluation information. The result information / evaluation information is forwarded together with the first processing data set 2 of the first determination unit 26 to a merging unit 29, which determines the display results 9, see also step S7. An output unit 30 controls the display device 20 to output the display results 9.

[0089] Although the application has been illustrated and described in detail by the preferred embodiments, the application is not restricted to the disclosed examples and other variations can be derived therefrom by a person of skill in the art without departing from the scope of the present application.

Claims

1. A computer-implemented method for operating a medical imaging device (10), wherein for evaluating an image data set of a computed tomography recorded with the imaging device (10), processing data sets (2, 4, 6) reflecting different image data contents are determinable from the image data set by image processing, the following steps being provided: - receiving at least one user information (1) describing a desired display form of the image data set, - providing at least one request information (22) describing at least one input data requested by an evaluation algorithm (3, 5) to be used, - determining at least one first processing data set (2) corresponding to the display form from the user information (1) and at least one second processing data set (4, 6) usable as input data for the respective evaluation algorithm (3, 5) from the request information (22), - if the request of the request information (22) cannot be completely fulfilled by using the image data set, determining a processing data set which is determinable from the image data set and which is similar to the data set required according to the request information (22) as a second processing data set (4, 6), - applying the at least one evaluation algorithm (3, 5) to the respective second processing data set (4, 6) to determine evaluation information, and - outputting the first processing data set (2) and the evaluation information to a display device (20). At least one of the at least one evaluation algorithm (3, 5) comprises a training function which has been or will be trained by using the respective second processing data set (4, 6) and / or the first processing data set (2) does not fulfill the request of the at least one request information (22). The image data set is a poly-energetic data set and comprises different sub-data sets assigned to different energies and / or energy intervals. The poly-energetic data set is recorded by using different X-ray spectra and / or by using a photon-counting X-ray detector. The determined processing data set as a second processing data set (4, 6) contains as much as possible of at least one image content defined by the request information (22). A request information (22) is provided which describes an alternative determined processing data set (4, 6), and / or processing parameters for determining a similar processing data set (4, 6) are determined according to an assignment rule describing a similarity rate. The evaluation information and the first processing data set (2) are evaluated and / or output together.

2. The method of claim 1, wherein, At least one position-dependent part of the evaluation information is output in the first processing data set (2) in full correspondence with the position.

3. The method according to claim 1 or 2, characterized in that, At least one part of the evaluation information is used to modify the first processing data set (2) before output.

4. The method of claim 3, wherein, The evaluation information used for the modification relates to image parts to be removed in the first processing data set (2).

5. The method according to claim 1 or 2, characterized in that, ​ 6. The method of claim 5, wherein, ​ 7. The method of claim 1, wherein, ​ 8. The method of claim 7, wherein, ​ 9. The method according to claim 7 or 8, characterized in that, ​ 10. The method of claim 9, wherein, ​ 11. The method of claim 1 or 2, wherein, In the case of the use of a plurality of evaluation algorithms (3, 5), the result information of at least two evaluation algorithms (3, 5) with respect to at least one evaluation information of the at least one evaluation information is evaluated together and / or at least one evaluation information is determined by using the processing data of the first processing data set (2) and from the result information of at least one of the at least one evaluation algorithm (3, 5).

12. The method of claim 1 or 2, wherein, The image data set recorded as a multipurpose data set shows at least one anatomical region of the patient having blood vessels with an iodine contrast agent, wherein at least one evaluation algorithm (3, 5) involves the detection and localization of bone material and / or at least one evaluation algorithm (3, 5) involves the segmentation and / or localization of anatomical structures.

13. The method of claim 1, wherein, The image data set is recorded by means of a multipurpose computed tomography.

14. The method of claim 1, wherein, The at least one input data is requested in accordance with the user information (1).

15. The method of claim 3, wherein, In the processing, a combined data set is determined as a processing data set (2, 4, 6) by means of a pixel-by-pixel calculation of the sub data sets.

16. The method of claim 7, wherein, At least one position-related part of the evaluation information is outputted in the first processing data set (2) by means of superimposition in such a way that it corresponds to the position.

17. The method of claim 9, wherein, The evaluation information for the modification involves a material class to be removed in the first processing data set (2).

18. The method of claim 12, wherein, At least one evaluation algorithm (3, 5) involves the segmentation and / or localization of blood vessels and / or tumors.

19. An imaging device (10) designed as a computed tomography device, having a control device (17) designed for carrying out the method according to any one of claims 1 to 18.

20. A computer program product having a computer program which, when running on a control device (17) of an imaging device (10), carries out the steps of the method according to any one of claims 1 to 18.

21. An electronically readable data carrier on which a computer program is stored which, when running on a control device (17) of an imaging device (10), carries out the steps of the method according to any one of claims 1 to 18.

Images