Provision of forecast data sets

Abstract

The invention relates to a computer-implemented method for providing a prediction data set of an examination object, comprising: receiving a first image data set of an examination region of the examination object; receiving at least one operating parameter of a medical object arranged at the examination region and positioning information of the medical object arranged at the examination region; generating a prediction data set by applying a trained function to input data, wherein the input data is based on the first image data set and at least one operating parameter and positioning information of the medical object, wherein at least one parameter of the trained function is based on a comparison with a first comparison image data set, wherein the first comparison image data set has changes at the examination region affected by the medical object relative to the first image data set; providing the prediction data set. The invention also relates to a computer-implemented method for providing a trained function, a training unit, a providing unit, a medical device, a computer program product and a computer-readable storage medium.

Description

Technical Field

[0001] The invention relates to a computer-implemented method for providing a prediction data set of an examination object, a computer-implemented method for providing a trained function, a training unit, a provision unit, a medical device, a computer program product and a computer-readable storage medium. Background Art

[0002] In medical practice, it is essential for a good treatment result to support medical personnel, such as doctors, particularly accurately and as comprehensively as possible. In this case, medical personnel can be supported, for example, by medical imaging of the part to be treated of the object to be examined, such as a human and / or animal patient, in treatment and diagnosis.

[0003] In particular, in medical interventions in which tissue is to be ablated by means of a medical object, in particular an ablation needle and / or a laparoscope and / or an endoscope, accurate knowledge of the ablation surface or ablation area is crucial. Here, the behavior of the ablation area usually depends to a large extent on the properties of the corresponding tissue and the medical object.

[0004] A method for simulating ablation zones based on preoperatively recorded three-dimensional imaging data of an examination object and intraoperative blood flow imaging is known from the prior art. Here, assumptions about tissue parameters are usually made based on the preoperatively recorded image data, which may have an adverse effect on the accuracy of the simulated ablation zones. Summary of the invention

[0005] Therefore, the object of the present invention is to provide a particularly accurate prediction for an examination region which may be affected by a medical object.

[0006] According to the invention, the above technical problem is solved by a computer-implemented method for providing a prediction data set of an examination object, a computer-implemented method for providing a trained function, a training unit, a providing unit, a medical device, a computer program product and a computer-readable storage medium according to the invention. The invention also provides advantageous embodiments with appropriate extensions.

[0007] The solution to the above technical problem according to the present invention is described below with respect to the claimed device and the claimed method. The features, advantages or alternative embodiments mentioned here can also be transferred to other claimed objects, and vice versa.

[0008] Furthermore, the solution to the above technical problem according to the invention is described with respect to a method and device for providing a prediction data set and with respect to a method and device for providing a trained function. Here, the features and alternative implementations of the data structures and / or functions in the method and device for providing a prediction data set can be transferred to similar data structures and / or functions in the method and device for providing a trained function. Here, in particular, similar data structures can be characterized by using the prefix "Training". Furthermore, the trained function used in the method and device for providing a prediction data set for an examination object can be adjusted and / or provided, in particular, by the method and device for providing a trained function.

[0009] In a first aspect, the invention relates to a computer-implemented method for providing a prediction data set for an examination object. It is then provided that a first image data set of an examination area of ​​the examination object is received. In addition, at least one operating parameter of a medical object arranged at the examination area of ​​the examination object and positioning information of the medical object arranged at the examination area of ​​the examination object are received. In addition, a prediction data set is generated by applying a trained function to input data. In this case, the input data is based on the first image data set and at least one operating parameter and positioning information of the medical object. In addition, at least one parameter of the trained function is based on a comparison with a first comparison image data set. In this case, the first comparison image data set has changes at the examination area that are affected by the medical object relative to the first image data set. Thereafter, the prediction data set is provided.

[0010] The first image data set may include in particular two-dimensional and / or three-dimensional medical image data of an examination region of the examination object. Advantageously, the first image data set may in particular map the examination region of the examination object before surgery. In this case, the first image data set may include, for example, a tissue parameter map. In addition, a predetermined tissue region, for example a tumor, may be segmented in the first image data set.

[0011] Furthermore, the examination region of the examination subject can include an anatomical and / or spatial region of the examination subject which may be affected by the medical object and / or a predetermined tissue region, such as a tumor.

[0012] In this case, in particular, a first image data set can be recorded and / or provided from an examination region of an examination subject by means of a medical imaging device. In this case, the examination subject can include, for example, an animal patient and / or a human patient.

[0013] For example, the medical imaging device can be designed as a medical X-ray device, in particular a C-arm X-ray device and / or a Dyna-CT, and / or a magnetic resonance device (MRT) and / or a computed tomography device (CT) and / or an ultrasound device.

[0014] Furthermore, the first image data set can advantageously include metadata. The metadata can include information about recording parameters and / or operating parameters of the medical imaging device and / or patient data.

[0015] The receiving of the first image data set may in particular comprise acquiring and / or reading out a computer-readable data memory and / or receiving from a data memory unit, such as a database. Furthermore, the first image data set may be provided by a processing unit of the medical imaging device.

[0016] The medical object may include, for example, an ablation needle and / or a laparoscope and / or an endoscope and / or an electrocautery device and / or a catheter.

[0017] Receiving at least one operating parameter of the medical object arranged at the examination region of the examination object may in particular include acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. Furthermore, the at least one operating parameter may be provided by a processing unit of the medical object.

[0018] In this case, the at least one operating parameter can advantageously include a current intensity and / or a voltage and / or a temperature and / or a magnetic field strength and / or an electric field strength and / or a frequency of the medical object.

[0019] Furthermore, the positioning information of the medical object arranged at the examination region of the examination object can advantageously include information about the orientation and / or orientation and / or position of the medical object, in particular information about the orientation and / or orientation and / or position of the medical object relative to the examination region of the examination object and / or relative to the medical imaging device used to record the first image data set. Furthermore, the orientation and / or orientation and / or position of the medical object relative to the first image data set can also be determined based on the positioning information, in particular by registration.

[0020] Furthermore, the localization information can be determined by a medical imaging device, for example a medical X-ray device and / or a magnetic resonance device and / or a computer tomography device and / or an ultrasound device and / or a camera system and / or an electromagnetic localization system. Here, for example, the localization information can be determined by segmenting the medical object in a mapping of the medical object with the aid of the medical imaging device. Furthermore, for example, the localization information can be determined by a marker structure and / or arrangement of at least one marker object on the medical object with the aid of the medical imaging device.

[0021] The receiving of the positioning information of the medical object arranged at the examination area of ​​the examination object can in particular comprise the acquisition and / or reading of a computer-readable data memory and / or the receiving from a data memory unit, for example a database. In addition, the positioning information can be provided by a processing unit of the medical object. For this purpose, the medical object can have, for example, a position sensor, in particular a gyroscope.

[0022] Furthermore, the trained function can advantageously be trained by machine learning methods. In particular, the trained function can be a neural network, in particular a convolutional neural network (CNN) or a network comprising a convolutional layer.

[0023] The trained function maps input data to output data. In this case, the output data may also depend in particular on one or more parameters of the trained function. One or more parameters of the trained function may be determined and / or adjusted by training. The determination and / or adjustment of one or more parameters of the trained function may in particular be based on a pair consisting of training input data and associated training output data, wherein the trained function for generating the training mapping data is applied to the training input data. The determination and / or adjustment may in particular be based on a comparison of the training mapping data with the training output data. In general, a trainable function, i.e. a function with one or more parameters that have not yet been adjusted, is also referred to as a trained function.

[0024] Other terms for a trained function are a trained mapping rule, a mapping rule with trained parameters, a function with trained parameters, an algorithm based on artificial intelligence, a machine learning algorithm. An example of a trained function is an artificial neural network, wherein the edge weights of the artificial neural network correspond to the parameters of the trained function. Instead of the term "neural network", the term "neural network" can also be used. In particular, the trained function can also be a deep artificial neural network (deep neural network, deep artificial neural network). Another example of a trained function is a "support vector machine", and in addition, other machine learning algorithms can also be used as trained functions.

[0025] The trained function, in particular the neural network, advantageously has an input layer and an output layer. Here, the input layer can be designed to receive input data. In addition, the output layer can be designed to provide mapping data. Here, the input layer and / or the output layer can each include a plurality of channels, in particular neurons.

[0026] Advantageously, the prediction data set is determined by applying a trained function to input data. The input data is based on the first image data set and at least one operating parameter and positioning information of the medical object. Furthermore, at least one parameter of the trained function can be based on a comparison with a first comparative image data set. The first comparative image data set has changes in the examination region affected by the medical object relative to the first image data set.

[0027] The first comparison image data set may include in particular two-dimensional and / or three-dimensional medical image data of an examination region of the examination object. Advantageously, the first comparison image data set may map the examination region of the examination object, in particular after surgery. In this case, the first comparison image data set may include, for example, a tissue parameter map. In addition, predetermined tissue regions, such as ablation regions and / or ablation lesions, may be segmented and / or compared in the first comparison image data set. In this case, the predetermined tissue region may have in particular changes that are influenced by the medical object.

[0028] In particular, at least one parameter of the trained function can be based on a comparison of a further prediction data set, in particular a training prediction data set, with the first comparison image data set. In this case, the further prediction data set can be determined in particular by applying the trained function to input data, which are based on the further first image data set and at least one further operating parameter and further positioning information of the medical object. In this case, the training prediction data set can be determined in particular as part of the proposed computer-implemented method for providing a trained function, which will be described below.

[0029] Furthermore, the prediction data set can correspond in particular to a comparative image data set, wherein predetermined tissue regions of the examination region, for example ablation regions and / or ablation lesions, can be segmented and / or compared in the prediction data set. Furthermore, the prediction data set can preferably include information about changes in the examination region, in particular predetermined tissue regions, that are affected by the medical object.

[0030] Furthermore, the provision of the forecast data set can include, in particular, storage on a computer-readable storage medium and / or display on a display unit and / or transmission to a processing unit.

[0031] As a result, a particularly precise and comprehensive prediction of changes in the examination region of the examination object that are influenced by the medical object can be provided to the medical personnel in a supportive manner.

[0032] In a further advantageous embodiment of the proposed computer-implemented method for providing a prediction data set, an elastic imaging data set of an examination region of the examination object can be received. Furthermore, the elastic imaging data set can be registered to the first image data set. Furthermore, the input data can also be based on the elastic imaging data set.

[0033] In this case, the elastic imaging data set can advantageously include information about the elasticity and / or viscosity of at least one tissue region of the examination region. The elastic imaging data set can in particular include associating elastic imaging values ​​and / or elastic imaging information with at least one tissue region, in particular contained in the examination region. The elastic imaging data set can in particular include two-dimensional and / or three-dimensional medical image data of the examination region of the examination object. Advantageously, the elastic imaging data set can in particular map the examination region of the examination object preoperatively.

[0034] In this case, the elastography data set can be recorded and / or provided from the examination region of the examination subject, in particular preoperatively, by means of a medical imaging device. For example, the medical imaging device for recording the elastography data set can be designed as a magnetic resonance device (MRT) and / or an ultrasound device.

[0035] Receiving the elastography data set may in particular comprise acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. Furthermore, the elastography data set may be provided by a processing unit of a medical imaging device for recording the elastography data set.

[0036] Advantageously, the input data of the trained function can also be based on an elastography dataset. Preferably, by taking into account the elastography dataset when generating the prediction dataset, a higher accuracy of the prediction dataset can be achieved with respect to changes in the examination region affected by the medical object.

[0037] Furthermore, the registration of the elastography data set to the first image data set can be based in particular on a predetermined anatomical structure and / or a predetermined tissue region (e.g. an ablation region and / or an ablation lesion) and / or a marker structure. In this case, the registration of the elastography data set to the first image data set can preferably include a rigid and / or non-rigid transformation of the elastography data set. In particular, the registration of the elastography data set to the first image data set can be performed image-based, for example based on anatomical and / or geometric features.

[0038] In a further advantageous embodiment of the proposed computer-implemented method for providing a prediction data set, at least one second image data set of at least a portion of the examination region of the examination object can be received. In this case, the at least one second image data set can advantageously map changes over time caused by the medical object at the examination region of the examination object. Furthermore, the input data can also be based on the at least one second image data set.

[0039] The at least one second image data set may in particular comprise two-dimensional and / or three-dimensional medical image data of at least a portion of an examination region of the examination object. Advantageously, the at least one second image data set may in particular map at least a portion of the examination region of the examination object during surgery. Furthermore, the at least one second image data set may, for example, comprise a tissue parameter map and / or a temperature map. Advantageously, the at least one second image data set may in particular map changes over time caused by the medical object at the examination region of the examination object relative to the first image data set. Furthermore, predetermined and / or changed tissue regions, for example ablation regions and / or ablation lesions, may be segmented and / or contrasted in the at least one second image data set.

[0040] In this case, the at least one second image data set can be recorded and / or provided from an examination region of the examination object, in particular by means of a medical imaging device. For example, the medical imaging device for recording the at least one second image data set can be designed as a medical X-ray device, in particular a C-arm X-ray device and / or a Dyna-CT, and / or a magnetic resonance device (MRT) and / or a computed tomography device (CT) and / or an ultrasound device.

[0041] The receiving of the at least one second image data set can in particular comprise acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. In addition, the at least one second image data set can also be provided by a processing unit of the medical imaging device for recording the at least one second image data set.

[0042] Furthermore, at least one second image data set can be registered to the first image data set. In this case, the registration of at least one second image data set to the first image data set can in particular be based on a predetermined anatomical structure and / or a predetermined tissue region (e.g. of an ablation region and / or an ablation lesion) and / or a marker structure. In this case, the registration of at least one second image data set to the first image data set can preferably include a rigid and / or non-rigid transformation of the at least one second image data set. In particular, the registration of at least one second image data set to the first image data set can be performed image-based, for example based on anatomical and / or geometrical features.

[0043] Advantageously, the input data of the trained function can also be based on at least one second image data set. By taking the second image data set into account when generating the prediction data set, a higher accuracy of the prediction data set can be achieved with respect to changes over time at the examination region affected by the medical object.

[0044] In a further advantageous embodiment of the provided computer-implemented method for providing a prediction data set, at least one second image data set can be recorded with the aid of an ultrasound device. Here, the at least one second image data set can in particular include two-dimensional and / or three-dimensional medical ultrasound image data of at least a portion of the examination area of ​​the examination object. In addition, tissue areas that may be affected by the medical object and are predetermined to be changed, such as ablation areas and / or ablation lesions and / or at least one fluid blister (Fluid blase), can be segmented and / or contrasted in the at least one second image data set. Fluid blisters can be formed, for example, due to local thermal effects in the case of thermal ablation of tissue areas with the aid of the medical object. Here, the at least one fluid blister can, for example, include a spatially delimited fluid, in particular a gas and / or a gas mixture and / or a liquid and / or a liquid mixture.

[0045] Thus, lesion formation (in particular of an ablation lesion) and / or fluid blister formation is advantageously mapped in the at least one second image data set, in particular as a change over time at the examination region of the examination object.

[0046] Advantageously, at the time of recording the second image data set, localization information of the ultrasound device can be acquired by means of a position acquisition unit, in particular an electromagnetic and / or optical sensor, thereby making it possible to achieve particularly accurate registration of the at least one second image data set to the first image data set.

[0047] In a further advantageous embodiment of the proposed computer-implemented method for providing a prediction data set, a plurality of second image data sets can be received in chronological order. Advantageously, a prediction data set can be generated in each case based on the previously received chronological second image data sets.

[0048] Advantageously, a plurality of second image data sets can be acquired by one or different medical imaging devices, thereby making it possible to achieve a particularly comprehensive mapping of changes over time caused by the medical object in the examination region of the examination object.

[0049] Furthermore, after receiving the respective second image data set, the step of generating a prediction data set by applying the trained function to the input data can be repeated (in particular iteratively). Advantageously, the input data for the trained function can be generated in each case from the second image data set received so far in chronological order. In particular, the input data for generating the prediction data set can be based on all the second image data sets received so far in chronological order. This can achieve a continuous improvement in the accuracy of the prediction data set. In particular, the time course of changes caused by the medical object in the examination region of the examination object, which have been acquired so far in chronological order by a plurality of second image data sets, can be taken into account when generating the respective prediction data set.

[0050] The receiving of the plurality of second image data sets can in particular comprise acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. Furthermore, the plurality of second image data sets can be provided by a processing unit of the medical imaging device for recording the plurality of second image data sets. In particular, all second image data sets recorded so far in chronological order can be stored in the processing unit and / or the data memory unit and / or the computer-readable data memory.

[0051] In a further advantageous embodiment of the proposed computer-implemented method for providing a prediction data set, the prediction data set can preferably include probability information and / or representation information, in particular with respect to the temporal progression, of fluid blister formation and / or lesion formation in the examination region of the examination object. In this case, the probability information can, for example, include a temporal and / or spatial progression model of fluid blister formation and / or lesion formation in the examination region of the examination object. For example, the prediction data set can respectively associate at least one probability value of fluid blister formation and / or lesion formation with at least one tissue region in the examination region. Furthermore, the prediction data set can include representation information, such as a volume model, about the representation of fluid blister formation and / or lesion formation in the examination region.

[0052] Here, the fluid blister can for example comprise a spatially bounded fluid, in particular a gas and / or a gas mixture and / or a liquid and / or a liquid mixture. In addition, the lesion and / or the fluid blister can be acquired and / or displayed, in particular, by a contrast of at least one tissue region of the examination region that is changed relative to the first image data set.

[0053] Here, the probability of fluid blister formation and / or lesion formation in the examination area can, for example, depend on the heat capacity and / or thermal conductivity and / or elasticity and / or viscosity of the tissue area. In addition, the probability of fluid blister formation and / or lesion formation in the examination area can also depend on at least one operating parameter of the medical object.

[0054] In order to improve the accuracy of the prediction data set, it is advantageous to take into account all input data of the trained function. In particular, in tissue regions of the examination region where there is no or only a small amount of fluid blister formation and / or lesion formation, it may also be advantageous to interpolate the probability information and / or the performance information. For this purpose, the prediction data set can be interpolated, for example, by applying an interpolation function. In addition, the prediction data set can also be extrapolated by applying an extrapolation function, for example a simulated extrapolation function including the probability information and / or the performance information, in particular in tissue regions of the examination region. In this case, for example, the temperature trend around the medical object can be extrapolated. In addition, the extrapolation can be based on an exponential and / or Gaussian temperature gradient around the medical object, for example. The temperature trend around the medical object can be determined based on at least one tissue parameter of the examination region, for example the thermal conductivity, in particular by means of a convolution of the spatial thermal effects from the medical object with the temperature gradient of the tissue region of the examination object.

[0055] In a further advantageous embodiment of the proposed computer-implemented method for providing a prediction data set, the prediction data set can have a validity range with respect to at least one operating parameter. Advantageously, a trained function for generating a prediction data set is trained according to the proposed computer-implemented method for providing a trained function. In this case, in particular, a definition range for a valid range of a producible prediction data set, in particular a trained function, can be predefined by at least one training operating parameter as input data, in particular training input data. For example, a current intensity interval and / or a voltage interval and / or a temperature interval and / or a magnetic field strength interval and / or an electric field strength interval and / or a frequency range of at least one training operating parameter of the respective medical object can predefine a valid range of the prediction data set.

[0056] Furthermore, the prediction data set can include information about a validity range related to at least one operating parameter. In particular, the prediction data set can be generated independently of the validity range by applying a trained function to the input data. Based on the information about the validity range contained in the prediction data set, the validity of the prediction data set with respect to the at least one operating parameter can be inferred. For example, the information about the validity range of the prediction data set can be displayed to an operator by means of a display unit.

[0057] In a further advantageous embodiment of the proposed computer-implemented method for providing a prediction dataset, the first image dataset can be recorded by means of a magnetic resonance device and / or a medical X-ray device and / or a computed tomography device. As a result, a two-dimensional and / or three-dimensional mapping of the examination region of the examination object can be achieved in the first image dataset, in particular before surgery.

[0058] In a second aspect, the invention relates to a computer-implemented method for providing a trained function. In this case, a first training image data set is received from an examination region of an examination object. In addition, at least one training operating parameter of a medical object arranged at the examination region of the examination object and training positioning information of the medical object arranged at the examination region of the examination object are received. In addition, a further training image data set is received from the examination region of the examination object. In this case, a further training image data set is recorded temporally after the first training image data set, wherein changes caused by the medical object at the examination region of the examination object occur after the recording of the first training image data set and before the recording of the further training image data set. In addition, a further comparative prediction data set is generated based on the further training image data set. In this case, the further comparative prediction data set has changes at the examination region affected by the medical object relative to the first training image data set. In addition, a training prediction data set is generated by applying the trained function to input data. In this case, the input data is based on the first training image data set and at least one training operating parameter and training positioning information of the medical object. Thereafter, at least one parameter of the trained function is adjusted based on a comparison of the further comparative prediction data set with the training prediction data set. In addition, a trained function is provided.

[0059] The first training image data set can in particular have all the properties of the first image data set, and vice versa, all the properties of the first image data set have been described in relation to the method for providing a prediction data set. In particular, the first training image data set can be the first image data set. The first training image data set can in particular include two-dimensional and / or three-dimensional medical image data of an examination region of the examination object. Advantageously, the first training image data set can in particular map the examination region of the examination object preoperatively. In this case, the first training image data set can, for example, include a tissue parameter map. Furthermore, a predetermined tissue region, for example a tumor, can be segmented in the first training image data set.

[0060] Furthermore, the examination region of the examination subject can include an anatomical and / or spatial region of the examination subject which may be affected by the medical object and / or a predetermined tissue region, such as a tumor.

[0061] In this case, in particular, the first training image data set can be recorded and / or provided from an examination region of an examination subject by means of a medical imaging device. In this case, the examination subject can include, for example, an animal patient and / or a human patient.

[0062] For example, the medical imaging device can be designed as a medical X-ray device, in particular a C-arm X-ray device and / or a Dyna-CT, and / or a magnetic resonance device (MRT) and / or a computed tomography device (CT) and / or an ultrasound device.

[0063] Furthermore, the first training image data set can advantageously include metadata. The metadata can include information about acquisition parameters and / or operating parameters of the medical imaging system used to acquire the first training image data set.

[0064] Receiving the first training image data set may in particular include acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. In addition, the first training image data set may be provided by a processing unit of a medical imaging device for recording the first training image data set.

[0065] The medical object may include, for example, an ablation needle and / or a laparoscope and / or an endoscope and / or an electrocautery device and / or a catheter.

[0066] In particular, the at least one training operating parameter can have all the characteristics of the at least one operating parameter, and vice versa, all the characteristics of the operating parameter have been described in relation to the method for providing a prediction data set. In particular, the at least one training operating parameter can be an operating parameter. The reception of the at least one training operating parameter of the medical object arranged at the examination region of the examination object can in particular comprise the acquisition and / or reading of a computer-readable data memory and / or the reception from a data memory unit, such as a database. Furthermore, the at least one training operating parameter can be provided by a processing unit of the medical object.

[0067] In this case, the at least one training operating parameter can advantageously include a current intensity and / or a voltage and / or a temperature and / or a magnetic field strength and / or an electric field strength and / or a frequency of the medical object.

[0068] The training positioning information may in particular have all the properties of the positioning information and vice versa, all the properties of the positioning information have already been described with respect to the method for providing a prediction data set. In particular, the training positioning information may be positioning information. Furthermore, the training positioning information of a medical object arranged at an examination region of the examination object may advantageously include information about an orientation and / or orientation and / or position of the medical object, in particular relative to the examination region of the examination object and / or relative to a medical imaging device for recording the first training image data set. Furthermore, the orientation and / or orientation and / or position of the medical object relative to the first training image data set may also be determined based on the training positioning information, in particular by registration.

[0069] Furthermore, the training positioning information can also be determined by a medical imaging device, such as a medical X-ray device and / or a magnetic resonance device and / or a computer tomography device and / or an ultrasound device and / or a camera system and / or an electromagnetic positioning system. Here, for example, the training positioning information can be determined by segmenting the medical object in a mapping of the medical object with the aid of the medical imaging device. Furthermore, for example, the training positioning information can be determined by a marker structure and / or arrangement of at least one marker object on the medical object with the aid of the medical imaging device.

[0070] The receiving of the training positioning information of the medical object arranged at the examination region of the examination object can in particular comprise the acquisition and / or reading of a computer-readable data memory and / or the receiving from a data memory unit, for example a database. In addition, the training positioning information can be provided by a processing unit of the medical object. For this purpose, the medical object can, for example, have a position sensor, in particular a gyroscope.

[0071] The further training image data set may in particular comprise two-dimensional and / or three-dimensional medical image data of at least a portion of the examination region of the examination object. Advantageously, the further training image data set may in particular postoperatively map at least a portion of the examination region of the examination object. In this case, the further training image data set may for example comprise a tissue parameter map. In this case, the further training image data set may advantageously be recorded temporally after the first training image data set. Furthermore, changes caused by the medical object in the examination region of the examination object may occur after the recording of the first training image data set and before the recording of the further training image data set.

[0072] The reception of the further training image data set can in particular comprise acquisition and / or reading a computer-readable data memory and / or reception from a data memory unit such as a database. In addition, the further training image data set can be provided by a processing unit of the medical imaging device for recording the further training image data set.

[0073] Generating a comparison prediction data set based on a further training image data set may in particular comprise registering the further training image data set to the first training image data set. In this case, registering the further training image data set to the first training image data set may in particular be based on predetermined anatomical structures and / or predetermined tissue regions and / or marker structures (e.g. of ablation regions and / or ablation lesions). In this case, registering the further training image data set to the first training image data set may preferably comprise a rigid and / or non-rigid transformation of the further training image data set. In particular, registering the further training image data set to the first image data set may be performed image-based, for example based on anatomical and / or geometrical features.

[0074] Furthermore, the generation of the comparative prediction data set may include segmenting and / or comparing changes caused by the medical object in the examination region of the examination object in the further training image data set. Furthermore, the generation of the comparative prediction data set may include matching and / or subtracting the further training image data set from the first training image data set.

[0075] Advantageously, this makes it possible to detect changes in the examination region of the examination object caused by the medical object.

[0076] Advantageously, the training prediction data set can be determined by applying a trained function to input data. The input data can be based on a first training image data set and at least one training operating parameter and training positioning information of the medical object. Furthermore, at least one parameter of the trained function can be based on a comparison of the prediction data set with the training prediction data set.

[0077] Furthermore, the first training image data set and / or the further training image data set and / or at least one training operating parameter and / or training positioning information can be simulated. For example, at least one training operating parameter can be simulated for different medical objects. Furthermore, the first and / or further medical training data sets can be simulated for different tissue types and / or tissue regions within the examination region of the examination object. Furthermore, further training image data sets can be simulated for different changes caused by the medical object at the examination region of the examination object. Furthermore, the training positioning information can be simulated for different medical objects and / or arrangements of medical objects and / or different examination objects.

[0078] Furthermore, the provision of the trained function may in particular include storage on a computer-readable storage medium and / or transmission to a processing unit.

[0079] With the proposed method, it is advantageously possible to provide a trained function which can be used in a method for providing a prediction data set.

[0080] In a further advantageous embodiment of the proposed computer-implemented method for providing a trained function, a training elastography data set of an examination region of the examination object can be received. Furthermore, the training elastography data set can be registered to the first training image data set. Furthermore, the input data can also be based on the training elastography data set. The training elastography data set can in particular have all the characteristics of the elastography data set, and vice versa, all the characteristics of the elastography data set have been described in relation to the method for providing a prediction data set. The training elastography data set can in particular be an elastography data set.

[0081] In this case, the training elastography data set can advantageously include information about the elasticity and / or viscosity of at least one tissue region of the examination region. The training elastography data set can in particular include associating elastography values ​​and / or elastography information with at least one tissue region, in particular, contained in the examination region. The training elastography data set can in particular include two-dimensional and / or three-dimensional medical image data of the examination region of the examination object. Advantageously, the training elastography data set can in particular map the examination region of the examination object preoperatively.

[0082] In this case, in particular before the operation, a training elastography data set is recorded and / or provided, in particular by means of a medical imaging device from an examination region of the examination object. For example, the medical imaging device for recording the training elastography data set can be designed as a magnetic resonance imaging device (MRT) and / or an ultrasound device. Furthermore, the training elastography data set can be simulated.

[0083] Receiving the training elastography data set may in particular comprise acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. Furthermore, the training elastography data set may be provided by a processing unit of the medical imaging device for recording the training elastography data set.

[0084] Advantageously, the input data of the trained function can also be based on a training elastography dataset. Preferably, by taking into account the training elastography dataset when generating the training prediction dataset, a higher accuracy of the training prediction dataset can be achieved with respect to changes in the examination region affected by the medical object.

[0085] Furthermore, the registration of the training elastography data set to the first training image data set can be based in particular on predetermined anatomical structures and / or predetermined tissue regions (e.g. of ablation regions and / or ablation lesions) and / or marker structures. In this case, the registration of the training elastography data set to the first training image data set can preferably include a rigid and / or non-rigid transformation of the training elastography data set. In particular, the registration of the training elastography data set to the first image data set can be performed, for example, image-based, for example based on anatomical and / or geometrical features.

[0086] In a further advantageous embodiment of the proposed computer-implemented method for providing a trained function, at least one second training image data set of at least a portion of the examination region of the examination object can be received. The at least one second training image data set can in particular have all the properties of the second training image data set, or vice versa, all the properties of the second training image data set have been described with respect to the method for providing a prediction data set. In particular, the at least one training image data set can be the second image data set.

[0087] At least one second training image data set may in particular include two-dimensional and / or three-dimensional medical image data of at least a portion of an examination region of the examination object. Advantageously, at least one second training image data set may map at least a portion of the examination region of the examination object, in particular intraoperatively and / or postoperatively. In addition, at least one second training image data set may, for example, include a tissue parameter map and / or a temperature map. Here, in particular relative to the first training image data set, at least one second training image data set may advantageously map changes over time caused by the medical object at the examination region of the examination object. In addition, predetermined and / or changed tissue regions, such as ablation regions and / or ablation lesions, may be segmented and / or contrasted in at least one second training image data set. In addition, at least one second training image data set may map at least a portion of an examination region of an examination object ex vivo. For this purpose, for example, a tissue sample from the examination region of the examination object may be mapped ex vivo in at least one second training image data set.

[0088] In this case, at least one second training image dataset can be recorded and / or provided from an examination region of the examination object, in particular by means of a medical imaging device. For example, the medical imaging device for recording the at least one second training image dataset can be designed as a medical X-ray device, in particular a C-arm X-ray device and / or a Dyna-CT, and / or a magnetic resonance device (MRT) and / or a computed tomography device (CT) and / or an ultrasound device.

[0089] The receiving of the at least one second training image data set can in particular comprise acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. In addition, the at least one second training image data set can be provided by a processing unit of the medical imaging device for recording the at least one second training image data set.

[0090] Furthermore, at least one second training image data set can be registered to the first training image data set. In this case, the registration of the at least one second training image data set to the first training image data set can be based in particular on a predetermined anatomical structure and / or a predetermined tissue region (e.g. of an ablation region and / or an ablation lesion) and / or a marker structure. In this case, the registration of the at least one second training image data set to the first training image data set can preferably include a rigid and / or non-rigid transformation of the at least one second training image data set.

[0091] Advantageously, the input data of the trained function can also be based on at least one second training image dataset. Preferably, by taking into account at least one second training image dataset when generating the training prediction dataset, a higher accuracy of the training prediction dataset can be achieved with respect to changes over time in the examination region affected by the medical object.

[0092] In a further advantageous embodiment of the proposed computer-implemented method for providing a trained function, the at least one second training image data set can advantageously be recorded by means of an ultrasound device.

[0093] In this case, the at least one second training image dataset may in particular comprise two-dimensional and / or three-dimensional medical ultrasound image data of at least a portion of an examination region of the examination object. Furthermore, tissue regions that may be affected by the medical object and that are predetermined to be changed, such as ablation regions and / or ablation lesions and / or at least one fluid blister, may be segmented and / or contrasted in the at least one second training image dataset.

[0094] Thus, ablation formation and / or fluid blister formation, in particular as a change over time in the examination region of the examination subject, can advantageously be mapped in the at least one second training image data set.

[0095] Advantageously, at the time of recording the at least one second training image data set, training position information of the ultrasound device can be acquired by means of a position acquisition unit, in particular an electromagnetic and / or optical sensor, thereby making it possible to achieve a particularly accurate registration of the at least one second training image data set to the first training image data set.

[0096] In a further advantageous embodiment of the proposed computer-implemented method for providing a trained function, the training prediction data set can preferably include probability information and / or performance information of fluid blister formation and / or lesion formation in the examination region of the examination object, in particular with regard to the temporal course.

[0097] Here, the probability information may include, for example, a temporal and / or spatial progression model of fluid blister formation and / or lesion formation in the examination region of the examination object. For example, the training prediction data set may associate at least one probability value of fluid blister formation and / or lesion formation with at least one tissue region in the examination region, respectively. In addition, the training prediction data set may include performance information about the performance of fluid blister formation and / or lesion formation in the examination region, such as a volume model.

[0098] Here, the fluid blister may for example comprise a spatially bounded gas and / or gas mixture and / or liquid and / or liquid mixture. In addition, the lesion and / or fluid blister may be acquired and / or displayed, in particular, by a contrast of at least one tissue region of the examination region that is changed relative to the first training image data set.

[0099] In order to improve the accuracy of the training prediction data set, it is advantageous to take into account all input data of the trained function. In particular, in tissue regions of the examination region where there is no or only a small amount of fluid blister formation and / or lesion formation, it may also be advantageous to interpolate the probability information and / or the performance information. For this purpose, the training prediction data set can be interpolated, for example, by applying an interpolation function.

[0100] In a further advantageous embodiment of the proposed computer-implemented method for providing a trained function, the training prediction data set can advantageously have a validity range with respect to at least one training operating parameter.

[0101] In this case, in particular, a validity range for a producible training prediction data set, in particular a definition range of a trained function, can be predefined by at least one training operating parameter as input data for the trained function. For example, a current intensity interval and / or a voltage interval and / or a temperature interval and / or a magnetic field intensity interval and / or an electric field intensity interval and / or a frequency range of at least one training operating parameter of the corresponding medical object can predefine the validity range of the training prediction data set.

[0102] Furthermore, the training prediction data set may include information about a validity range associated with at least one training operating parameter. In particular, the training prediction data set may be generated independently of the validity range by applying a trained function to the input data. Based on the information about the validity range contained in the training prediction data set, the validity of the training prediction data set with respect to the at least one training operating parameter may be inferred.

[0103] In a further advantageous embodiment of the proposed computer-implemented method for providing a trained function, a first training image dataset can be recorded by means of a magnetic resonance device and / or a medical X-ray device and / or a computed tomography device. As a result, a two-dimensional and / or three-dimensional mapping of an examination region of the examination object can be achieved in the first training image dataset, in particular before surgery.

[0104] In a third aspect, the present invention relates to a providing unit for providing a prediction data set. Here, the providing unit comprises a computing unit and an interface. In addition, the interface is designed to receive a first image data set of an examination area of ​​an examination object. In addition, the interface is designed to receive at least one operating parameter of a medical object arranged at the examination area and positioning information of the medical object arranged at the examination area. In addition, the computing unit is designed to generate a prediction data set by applying a trained function to input data. In this case, the input data is based on the first image data set and at least one operating parameter and positioning information of the medical object. In addition, at least one parameter of the trained function is based on a comparison with a first comparison image data set. In this case, the first comparison image data set has changes at the examination area affected by the medical object relative to the first image data set. In addition, the interface is designed to provide the prediction data set.

[0105] Such a provision unit can be designed in particular to carry out the above-described method according to the invention for providing a prediction data set and its aspects. The provision unit is designed to carry out the method and its aspects by designing the interface and the computing unit to carry out the corresponding method steps.

[0106] In a fourth aspect, the invention relates to a medical device comprising the proposed provision unit. In this case, the medical device is designed to control a medical object based on a prediction data set. Furthermore, the medical device can be designed, for example, as a medical imaging device, in particular a magnetic resonance device and / or a computer tomography device and / or an ultrasound device and / or a medical X-ray device.

[0107] The medical device may in particular comprise a display unit, such as a display and / or monitor, which is designed to display information and / or graphic representations of the medical device and / or provision unit and / or other components. In particular, the display unit may be designed to display graphic representations of the prediction data set.

[0108] The advantages of the proposed medical device correspond essentially to the advantages of the proposed computer-implemented method for providing a prediction data set. Features, advantages or alternative embodiments mentioned here can also be transferred to other claimed subjects and vice versa.

[0109] In a fifth aspect, the invention relates to a training unit for providing a trained function. Here, the training unit comprises a training interface and a training calculation unit. In addition, the training interface is designed to receive a training image data set of an examination area of ​​an examination object. In addition, the training interface is designed to receive at least one training operating parameter of a medical object arranged at the examination area of ​​the examination object and training positioning information of the medical object arranged at the examination area. In addition, the training interface is designed to receive another training image data set of the examination area of ​​the examination object. Here, the other training image data set is recorded after the first training image data set in time. In addition, changes caused by the medical object at the examination area of ​​the examination object occur after recording the first training image data set and before recording the other training image data set. In addition, the training calculation unit is also designed to generate another comparative prediction data set based on the other training image data set. Here, the other comparative prediction data set has changes affected by the medical object at the examination area relative to the first training image data set. In addition, the training calculation unit is designed to generate a training prediction data set by applying the trained function to input data. Here, the input data is based on the first training image data set and at least one training operating parameter and training positioning information of the medical object. Furthermore, the training calculation unit is designed to adjust at least one parameter of the trained function based on a comparison of the further comparison prediction data set with the training prediction data set. Furthermore, the training interface is designed to provide the trained function.

[0110] In particular, such a training unit can be designed to perform the previously described method according to the invention for providing a trained function and its aspects. The training unit is designed to perform the method and its aspects by designing the training interface and the training calculation unit to perform the corresponding method steps.

[0111] In a sixth aspect, the present invention relates to a computer program product having a computer program, which can be directly loaded into the memory of a providing system, and has a program part so that when the providing unit executes the program part, all steps or aspects of the method for providing a prediction data set are performed, and / or the computer program can be directly loaded into the training memory of a training unit, and has a program part so that when the training unit executes the program part, all steps or aspects of the method for providing a trained function are performed.

[0112] In a seventh aspect, the present invention relates to a computer-readable storage medium, on which a program part is stored that can be read and executed by a providing unit, so that when the providing unit executes the program part, all steps or aspects of the method for providing a prediction data set are executed; and / or on which a program part is stored that can be read and executed by a training unit, so that when the training unit executes the program part, all steps or aspects of the method for providing a trained function are executed.

[0113] In an eighth aspect, the present invention relates to a computer program or a computer readable storage medium comprising a trained function provided by the method for providing a trained function or an aspect thereof.

[0114] The advantage of a largely software-based implementation is that provision units and / or training units that have already been used can also be upgraded in a simple manner by means of a software update so that they can operate in the manner according to the invention. In addition to the computer program, such a computer program product may also include additional components, such as documentation and / or additional components, and hardware components, such as a hardware key (dongle, etc.) for using the software. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] Embodiments of the present invention are shown in the drawings and are described in more detail below. In different drawings, the same reference numerals are used for the same features. In the drawings:

[0116] Figures 1 to 3 A schematic diagram showing a data flow of different specific embodiments of a computer-implemented method for providing a prediction data set of an examination object;

[0117] Figure 4 A schematic flow chart of a computer-implemented method for providing a prediction data set of an examination object is shown;

[0118] Figure 5 A schematic flow chart of a computer-implemented method for providing a trained function is shown;

[0119] Figure 6 An embodiment of a providing unit is shown;

[0120] Figure 7 An embodiment of a training unit is shown;

[0121] Figure 8 A medical device for controlling a medical object based on a prediction data set is shown. DETAILED DESCRIPTION

[0122] exist Figure 1The data flow of the proposed computer-implemented method for providing a prediction data set of an examination object is schematically depicted in . In this case, a first image data set B1 of an examination region of the examination object can be received. In addition, at least one operating parameter BP of a medical object arranged in the examination region of the examination object and positioning information P of the medical object arranged in the examination region of the examination object can also be received. In addition, the prediction data set PD can be generated by applying a trained function TF to the input data ED. In this case, the trained function TF can be a neural network, in particular a convolutional neural network or a network comprising convolutional layers.

[0123] Furthermore, the prediction data set PD can be regarded in particular as mapping data AD of a trained function. In this case, the input data ED can be based on the first image data set B1 and at least one operating parameter BP and positioning information P of the medical object. Furthermore, at least one parameter of the trained function TF can be based on a comparison with a first comparison image data set. In this case, the first comparison image data set can have changes in the examination region affected by the medical object relative to the first image data set B1. Furthermore, a PROV-PD prediction data set PD can be provided.

[0124] The prediction data set PD can advantageously include probability information and / or behavior information, in particular regarding the temporal course, of fluid blister formation and / or lesion formation in the examination region of the examination object. Furthermore, the prediction data set PD can have a validity range regarding at least one operating parameter BP.

[0125] Advantageously, the first image data set B1 can be acquired by means of a magnetic resonance device and / or a medical X-ray device and / or a computer tomography device.

[0126] Figure 2 The data flow of a further embodiment of the proposed computer-implemented method for providing a prediction data set PD of an examination object is schematically shown. In this case, an elastic imaging data set E1 of an examination region of the examination object can also be received. Furthermore, the elastic imaging data set E1 can be registered R1 to the first image data set B1. Furthermore, the input data ED of the trained function TF can also be based on the elastic imaging data set E1.

[0127] Furthermore, at least one second image data set B2 of at least a part of the examination region of the examination object can also be received. In this case, the at least one second image data set B2 can map changes over time caused by the medical object at the examination region of the examination object. Furthermore, the at least one second image data set B2 can be registered R2 to the first image data set B1. Furthermore, the input data ED of the trained function TF can also be based on the at least one second image data set B2.

[0128] exist Figure 3 The data flow of a further embodiment of the method provided for providing a prediction data set PD is schematically shown in . Advantageously, a plurality of second image data sets B2 can be received in chronological order. Furthermore, the prediction data set PD can be generated, in particular iteratively, based on the second image data sets B2 received in chronological order so far.

[0129] Figure 4 A schematic flow chart of a computer-implemented method for providing a prediction dataset PD of an examination object is shown. In this case, a first image dataset B1 , at least one operating parameter BP and a positioning information P can be received in a first step REC- 1 .

[0130] Receiving can in particular include acquiring and / or reading a computer-readable data memory and / or receiving from a data memory unit, such as a database. In addition, the first image data set B1 and / or the positioning information can be provided by a processing unit of the medical imaging device. In addition, at least one operating parameter BP and / or the positioning information P can be provided by a processing unit of the medical object.

[0131] In a further step DET-PD, a prediction data set PD may be generated by applying the trained function TF to the input data ED. Thereafter, a PROV-PD prediction data set PD may be provided.

[0132] Figure 5 A schematic flow chart of an embodiment of a computer-implemented method for providing a trained function TF is shown. In this case, a first training image data set TB1 of an examination region of an examination object can be received.

[0133] Advantageously, the first training image data set TB1 can be acquired by means of a magnetic resonance device and / or a medical X-ray device and / or a computed tomography device.

[0134] In addition, at least one training operation parameter TBP and training positioning information TP of a medical object arranged at the inspection area of ​​the inspection object can also be received. In addition, another training image data set WTB of the inspection area of ​​the medical object can also be received. Here, the other training image data set WTB can be recorded in time after the first training image data set TB1, in particular after surgery. In addition, changes caused by the medical object at the inspection area of ​​the inspection object can occur after the recording of the first training image data set TB1 and before the recording of the other training image data set WTB. In addition, F comparison prediction data set VPD can be generated based on the other training image data set WTB. Here, the comparison prediction data set VPD can advantageously have changes affected by the medical object at the inspection area relative to the first training image data set TB1. In addition, the training prediction data set TPD can be generated by applying the trained function TF to the input data ED. Here, the input data ED can advantageously be based on the first training image data set TB1 and at least one training operation parameter TBP and training positioning information TP of the medical object. Here, the training prediction data set TPD can be regarded as the training mapping data TAD of the trained function TF.

[0135] The training prediction data set TPD may preferably include probability information and / or performance information, in particular about the temporal course, of fluid blister formation and / or lesion formation in the examination region of the examination object. In addition, the training prediction data set TPD may have a validity range for at least one training operating parameter TBP.

[0136] Furthermore, at least one parameter of the ADJ-TF trained function TF may be adjusted based on a comparison of the comparison prediction data set VPD with the training prediction data set TPD.Thereafter, the PROV-TF trained function TF may be provided.

[0137] Furthermore, the proposed computer-implemented method for providing a trained function TF may also comprise receiving a training elastography data set (not shown here) of an examination region of the examination object. Furthermore, the training elastography data set may be registered to the first training image data set TB1. Furthermore, the input data ED of the trained function TF may also be based on the training elastography data set.

[0138] Furthermore, at least one second training image data set (not shown here) of at least a part of the examination region of the examination object can be received. In this case, the at least one second training image data set can advantageously map changes over time caused by the medical object in the examination region of the examination object. Furthermore, the input data ED of the trained function TF can advantageously also be based on the at least one second training image data set.

[0139] Figure 6shows the providing unit PRVS, Figure 7 A training unit TRS is shown. Advantageously, the provision unit PRVS shown can be designed to carry out a computer-implemented method according to the invention for providing a prediction data set PD for an examination object. Advantageously, the training unit TRS shown can be designed to carry out the proposed computer-implemented method for providing a trained function TF. The provision unit PRVS can advantageously comprise an interface IF, a calculation unit CU and a memory unit MU. Furthermore, the training unit TRS can advantageously comprise a training interface TIF, a training calculation unit TCU and a training memory unit TMU.

[0140] In this case, the interface IF can be designed to receive a first image data set B1 of an examination region of the examination object. Furthermore, the interface IF can be designed to receive at least one operating parameter BP of a medical object arranged in the examination region of the examination object and positioning information P of the medical object arranged in the examination region of the examination object. Furthermore, the calculation unit CU can be designed to generate a DET-PD prediction data set PD by applying a trained function TF to the input data ED. Furthermore, the interface can be designed to provide a PROV-PD prediction data set PD.

[0141] Furthermore, the training interface TIF may be designed to receive a first training image data set TB1 of an examination region of an examination object. Furthermore, the training interface may be designed to receive at least one training operating parameter TBP of a medical object arranged at the examination region of the examination object and training positioning information TP of a medical object arranged at the examination region of the examination object. Furthermore, the training interface TIF may be designed to receive a further training image data set WTB. Furthermore, the training calculation unit may be designed to generate an F comparison prediction data set VPD based on the further training image data set WTB. Furthermore, the training calculation unit may be designed to generate a training prediction data set TPD by applying a trained function TF to input data ED. Furthermore, the training calculation unit may be designed to adjust at least one parameter of a trained function TF of an ADJ-TF. Furthermore, the training interface TIF may be designed to provide a trained function TF of a PROV-TF.

[0142] The provision unit PRVS and / or the training unit TRS can be, in particular, a computer, a microcontroller or an integrated circuit. Alternatively, the provision unit PRVS and / or the training unit TRS can be a real or virtual association of computers (the English term for a real association is "cluster", and the English term for a virtual association is "cloud"). The provision unit PRVS and / or the training unit TRS can also be designed as a virtual system (virtualization) executed on a real computer or on a real or virtual association of computers.

[0143] The interface IF and / or the training interface TIF can be a hardware or software interface (e.g. a PCI bus, USB or Firewire). The computing unit CU and / or the training computing unit TCU can, for example, have hardware or software elements, such as a microprocessor or a so-called FPGA (Field Programmable Gate Array). The memory unit MU and / or the training memory unit TMU can be implemented as a non-permanent system memory (random access memory, RAM for short) or as a permanent mass storage device (hard disk, USB stick, SD card, solid state disk).

[0144] The interface IF and / or the training interface TIF may in particular include a plurality of sub-interfaces that perform different steps of the corresponding method. In other words, the interface IF and / or the training interface TIF may also be considered as a plurality of interfaces IF or a plurality of training interfaces TIF. The computing unit CU and / or the training computing unit TCU may in particular include a plurality of computing sub-units that perform different steps of the corresponding method. In other words, the computing unit CU and / or the training computing unit TCU may also be considered as a plurality of computing units CU or a plurality of training computing units TCU.

[0145] Figure 8 A medical device 46 for controlling a medical object 44 based on a prediction data set PD is schematically shown. Here, the medical device 46 can advantageously include a processing unit 45 and / or the proposed provision unit PRVS. In addition, the medical device can include the proposed training unit TRS (not shown here). The medical object 44 can include, for example, an ablation needle and / or a laparoscope and / or an endoscope and / or an electrocautery device and / or a catheter.

[0146] Furthermore, the first image data set B1 can advantageously be recorded, in particular preoperatively, by means of a medical X-ray device 37. Figure 8, for this purpose, a medical C-arm X-ray device 37 is schematically shown by way of example. The medical C-arm X-ray device 37 comprises a detector unit 34, an X-ray source 33 and a processing unit 22. In order to record a first image data set B1, in particular a projection X-ray image and / or a Dyna-CT image, an arm 38 of the C-arm X-ray device 37 can be mounted movably about one or more axes. As a result, a first image data set B1 can be recorded, which in particular comprises a plurality of projection X-ray images, each of which has in particular different recording parameters, in particular acquisition geometries, from one another. Furthermore, the medical C-arm X-ray device 37 can comprise a moving device 39, which makes it possible to move the C-arm X-ray device 37 in space.

[0147] In order to record a first image data set B1 of an examination area UB of an examination object 31 arranged on a patient positioning device 32, the processing unit 22 can send a signal 24 to an X-ray source 33. Subsequently, the X-ray source 33 can emit an X-ray beam, in particular a cone beam and / or a fan beam. When the X-ray beam strikes the surface of the detector unit 34 after interacting with the examination area UB of the examination object 31, the detector unit 34 can send a signal 21 to the processing unit 22. The processing unit 22 can receive the first image data set B1, for example, based on the signal 21. Subsequently, the processing unit 22 can provide the first image data set B1 to the processing unit 45 of the medical device 46.

[0148] Furthermore, the medical C-arm X-ray device 37 may include an input unit 41 (e.g., a keyboard) and / or a display unit 42 (e.g., a monitor and / or a display). For example, in the case of a capacitive input display, the input unit 41 may preferably be integrated into the display unit 42. In this case, the proposed method and / or the medical C-arm X-ray device 37 may be controlled by inputs by an operator at the input unit 41. For example, a graphical representation of the first image data set B1 may be displayed on the display unit 42.

[0149] Furthermore, the medical device 46 can include an input unit 41' (e.g. a keyboard) and / or a display unit 42' (e.g. a monitor and / or display). In the case of a capacitive input display, the input unit 41' can preferably be integrated into the display unit 42'. In this case, the proposed method and / or the medical device 46 can be controlled by inputs by an operator at the input unit 41'. For example, a graphical representation of the prediction data set PD can be displayed on the display unit 42'.

[0150] The processing unit 45 can advantageously provide the received first image data set B1 to the provision unit PRVS. The provision unit PRVS can then advantageously carry out an embodiment of the proposed computer-implemented method for providing a prediction data set PD of the examination object 31. In this case, the provision unit PRVS can also provide the generated prediction data set PD PROV-PD to the processing unit 45 of the medical device 46. Based on the prediction data set PD, the processing unit 45 can advantageously realize regulation and / or control of the medical object 44.

[0151] Furthermore, at least one second image data set B2 can be recorded by means of an ultrasound device 82. For this purpose, the ultrasound device 82 can advantageously be arranged at the examination region UB of the examination object 31. Furthermore, the ultrasound device 82 can provide at least one second image data set B2, in particular recorded intraoperatively, to the processing unit 45 of the medical device 46. Furthermore, the processing unit 45 can advantageously provide at least one second image data set B2 to the provision unit PRVS.

[0152] Analogously to the acquisition and provision of the at least one second image data set B2 , at least one second training image data set can be acquired by means of the ultrasound device 82 and made available to the training unit TRS.

[0153] Furthermore, the ultrasound device 82 can also be designed for recording an elastography data set E1. Here, the recorded elastography data set E1 can be provided to the provision unit PRVS similarly to the at least one second image data set B2. Analogously to the recording and provision of the elastography data set E1, a training elastography data set can be recorded with the aid of the ultrasound device 82 and provided to the training unit TRS.

[0154] The schematic illustrations contained in the described figures are in no way intended to be to scale or parameter ratio.

[0155] Finally, it is pointed out again that the method described in detail above and the device shown are only embodiments, which can be modified in different ways by a person skilled in the art without departing from the scope of the present invention. In addition, the use of the indefinite article "a" or "an" does not exclude that the features involved can also exist multiple times. Similarly, the terms "unit" and "element" do not exclude that the components involved are composed of multiple cooperating subcomponents, which can also be spatially distributed if necessary.

Claims

1. A computer-implemented method for providing a prediction data set (PD) of an object under examination (31), comprising: - receiving a first image data set (B1) of an examination region (UB) of an examination object (31), - receiving at least one operating parameter (BP) of a medical object (44) arranged in an examination region (UB) of the examination object (31) and positioning information (P) of the medical object (44) arranged in the examination region (UB) of the examination object (31), - generating (DET-PD) a predicted data set (PD) by applying the trained function (TF) to the input data (ED), wherein the input data (ED) are based on the first image data set (B1) and at least one operating parameter (BP) and positioning information (P) of the medical object (44), wherein at least one parameter of the trained function (TF) is based on a comparison with a first comparative image data set (VPD), wherein the first comparative image data set (VPD) has changes in the examination region (UB) that are influenced by a medical object (44) relative to the first image data set (B1), - Provide (PROV-PD) prediction data set (PD), The prediction data set (PD) comprises information on the probability of the formation of fluid blisters and / or lesions in an examination region (UB) of the examination subject (31) with respect to the temporal course.

2. The method according to claim 1, further comprising: receiving an elastography data set (E1) of an examination region (UB) of the examination object (31) and registering (R1) the elastography data set (E1) to the first image data set (B1), In this case, the input data (ED) are also based on an elastography data set (E1).

3. The method according to claim 1, further comprising: receiving at least one second image data set (B2) of at least a part of an examination area (UB) of the examination object (31), wherein the at least one second image data set (B2) maps changes over time in an examination region (UB) of the examination subject (31) caused by the medical object (44), The input data (ED) are furthermore based on the at least one second image data set (B2). 4 . The method according to claim 3 , wherein the at least one second image data set ( B2 ) is recorded by means of an ultrasound device ( 82 ).

5. The method according to claim 3 or 4, characterized in that: receiving a plurality of second image data sets (B2) in chronological order, In this case, a prediction data set (PD) is respectively generated based on the second image data set (B2) received so far in chronological order.

6. The method according to claim 1, characterized in that The prediction data set (PD) has a validity range with respect to at least one operating parameter (BP).

7. The method according to any one of claims 1 to 4, characterized in that The first image data set (B1) is recorded by means of a magnetic resonance device and / or a medical X-ray device (37) and / or a computed tomography device.

8. The method according to claim 1, wherein: The prediction data set (PD) comprises probability information and / or behavior information regarding the temporal course of fluid blister formation and / or lesion formation in an examination region (UB) of the examination subject (31).

9. The method according to claim 1, wherein: In the prediction data set, the probability information about the temporal course of fluid blister formation and / or lesion formation is associated with at least one tissue region in the examination region.

10. The method according to claim 1, wherein: The probabilistic information comprises a temporal and / or spatial progression model of fluid blister formation and / or lesion formation in an examination region of the examination object.

11. The method according to claim 1, wherein: The prediction data set includes a volumetric model regarding the manifestation of fluid blister formation and / or lesion formation within the examination region.

12. A computer-implemented method for providing a trained function (TF), comprising: - receiving a first training image data set (TB1) of an examination region (UB) of an examination object (31), - receiving at least one training operating parameter (TBP) of a medical object (44) arranged in an examination region (UB) of the examination object (31) and training positioning information (TP) of the medical object (44) arranged in the examination region (UB) of the examination object (31), - receiving a further training image data set (WTB) of an examination region (UB) of an examination object (31), wherein the further training image data set (WTB) is recorded temporally after the first training image data set (TB1), wherein changes caused by the medical object (44) occur in an examination region (UB) of the examination object (31) after the recording of the first training image data set (TB1) and before the recording of the further training image data set (WTB), - generating (F) a comparative prediction data set (VPD) based on a further training image data set (WTB), wherein the comparative prediction data set (VPD) has, relative to the first training image data set (TB1), changes in the examination region (UB) that are influenced by the medical object (44), - generating a training prediction data set (TPD) by applying the trained function (TF) to the input data (ED), wherein the input data (ED) are based on the first training image data set (TB1) and at least one training operating parameter (TBP) and training positioning information (TP) of the medical object (44), - adjusting (ADJ-TF) at least one parameter of said trained function (TF) based on a comparison of said comparative prediction data set (VPD) with said training prediction data set (TPD), - Provide (PROV-TF) the trained function (TF).

13. The method according to claim 12, further comprising: receiving a training elastography data set of an examination region (UB) of the examination object (31) and registering the training elastography data set to the first training image data set (TB1), Therein, the input data (ED) are also based on the training elastography data set.

14. The method according to claim 12 or 13, further comprising: receiving at least one second training image data set of at least a part of an examination area (UB) of the examination object (31), wherein the at least one second training image data set maps changes over time in an examination region (UB) of the examination object (31) caused by the medical object (44), Therein, the input data (ED) are furthermore based on the at least one second training image data set.

15. The method according to claim 14, wherein: The at least one second training image data set is recorded by means of an ultrasound device (82).

16. The method according to claim 12, characterized in that The training prediction data set (TPD) comprises a validity range for at least one training run parameter (TBP).

17. The method according to claim 12, characterized in that The first training image data set (TB1) is recorded with the aid of a magnetic resonance device and / or a medical X-ray device (37) and / or a computed tomography device.

18. A provision unit (PRVS) for providing a prediction data set (PD), comprising a calculation unit (CU) and an interface (IF), - wherein the interface (IF) is designed to receive a first image data set (B1) of an examination region (UB) of an examination object (31), - wherein the interface (IF) is further designed to receive at least one operating parameter (BP) of a medical object (44) arranged in an examination region (UB) of the examination object (31) and positioning information (P) of the medical object (44) arranged in the examination region (UB) of the examination object (31), - wherein the computing unit (CU) is furthermore designed to generate (DET-PD) the prediction data set (PD) by applying a trained function (TF) to input data (ED), in, The input data (ED) are based on the first image data set (B1) and at least one operating parameter (BP) and positioning information (P) of the medical object (44), wherein at least one parameter of the trained function (TF) is based on a comparison with a first comparative image data set (VPD), wherein the first comparative image data set (VPD) has changes in the examination region (UB) that are influenced by the medical object (44) relative to the first image data set (B1), - wherein the interface (IF) is further designed to provide (PROV-PD) the prediction data set (PD), The prediction data set (PD) comprises information on the probability of the formation of fluid blisters and / or lesions in an examination region (UB) of the examination subject (31) with respect to the temporal course.

19. The providing unit according to claim 18, wherein: The prediction data set (PD) comprises probability information and / or behavior information regarding the temporal course of fluid blister formation and / or lesion formation in an examination region (UB) of the examination subject (31).

20. The providing unit according to claim 18, wherein: In the prediction data set, the probability information about the temporal course of fluid blister formation and / or lesion formation is associated with at least one tissue region in the examination region.

21. The providing unit according to claim 18, wherein: The probabilistic information comprises a temporal and / or spatial progression model of fluid blister formation and / or lesion formation in an examination region of the examination object.

22. The providing unit according to claim 18, wherein: The prediction data set includes a volumetric model regarding the manifestation of fluid blister formation and / or lesion formation within the examination region.

23. A medical device (45) comprising a providing unit (PRVS) according to claim 18, wherein: The medical device (45) is designed to control a medical object (44) based on a prediction data set (PD).

24. A training unit (TRS) for providing a trained function (TF), comprising a training interface (TIF) and a training computation unit (TCU), - wherein the training interface (TIF) is designed to receive a first training image data set (TB1) of an examination region (UB) of an examination object (31), - wherein the training interface (TIF) is further designed to receive at least one training operating parameter (TBP) of a medical object (44) arranged in an examination region (UB) of the examination object (31) and training positioning information (TP) of the medical object (44) arranged in the examination region (UB) of the examination object (31), - wherein the training interface (TIF) is also designed to receive a further training image data set (WTB) of the examination region (UB) of the examination object (31), in, the further training image data set (WTB) is recorded temporally after the first training image data set (TB1), wherein changes caused by the medical object (44) occur in an examination region (UB) of the examination object (31) after the recording of the first training image data set (TB1) and before the recording of the further training image data set (WTB), - the training computer unit (TCU) is furthermore designed to generate (F) a comparative prediction data set (VPD) based on the further training image data set (WTB), wherein the comparative prediction data set (VPD) has, relative to the first training image data set (TB1), changes in the examination region (UB) that are influenced by the medical object (44), - wherein the training computing unit (TCU) is furthermore designed to generate a training prediction data set (TPD) by applying a trained function (TF) to input data (ED), wherein the input data (ED) are based on the first training image data set (TB1) and at least one training operating parameter (TBP) and training positioning information (TP) of the medical object (44), - wherein the training computing unit (TCU) is further designed to adjust (ADJ-TF) at least one parameter of a trained function (TF) based on a comparison of the comparison prediction data set (VPD) with the training prediction data set (TPD), - wherein the training interface (TIF) is furthermore designed to provide (PROV-TF) the trained function (TF) The training prediction data set (TPD) comprises probability information on the temporal course of fluid blister formation and / or lesion formation in an examination region (UB) of the examination object (31).

25. The training unit according to claim 24, wherein: The training prediction data set (TPD) comprises probability information and / or behavior information regarding the temporal course of fluid blister formation and / or lesion formation in an examination region (UB) of the examination subject (31).

26. The training unit according to claim 24, wherein: In the training prediction data set (TPD), the probability information about the temporal course of fluid blister formation and / or lesion formation is associated with at least one tissue region in the examination region.

27. The training unit according to claim 24, wherein: The probabilistic information comprises a temporal and / or spatial progression model of fluid blister formation and / or lesion formation in an examination region of the examination object.

28. The training unit according to claim 24, wherein: The training prediction data set (TPD) includes a volumetric model of the behavior of fluid bleb formation and / or lesion formation within the examination region.

29. A computer program product comprising a computer program, which can be directly loaded into the memory of a providing unit (PRVS), and which has a program part so that when the providing unit (PRVS) executes the program part, it is used to perform all the steps of the method according to any one of claims 1 to 11; and / or the computer program can be directly loaded into the training memory (TMU) of a training unit (TRS), and which has a program part so that when the training unit (TRS) executes the program part, it is used to perform all the steps of the method according to any one of claims 12 to 17.

30. A computer-readable storage medium, on which a program part is stored that can be read and executed by a providing unit (PRVS), so that when the providing unit (PRVS) executes the program part, all the steps of the method according to any one of claims 1 to 11 are performed; and / or on which a program part is stored that can be read and executed by a training unit (TRS), so that when the training unit (TRS) executes the program part, all the steps of the method according to any one of claims 12 to 17 are performed.

Images