Machine learning in contrast radiology

JP2024512369A5Active Publication Date: 2024-12-04BAYER AG
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
JP2023554810
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Priority Date
2021-04-12
Filing Date
2021-11-29
Publication Date
2024-12-04
Estimated Expiration
2041-11-29

AI Technical Summary

Technical Problem

Radiological examinations that track the diffusion of contrast agents over time for focal liver lesions are lengthy, requiring patient immobility and are uncomfortable due to prolonged examination times.

Method used

A machine learning model predicts synthetic radiological images by processing a time series of measured images in frequency space, allowing for the generation of images at different time points without the need to capture all images, thus reducing examination time.

Benefits of technology

The method shortens examination time, enhances patient comfort, and reduces computational complexity by separating contrast information from detail information, making the process more efficient.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

The present invention relates to the technical field of generating artificial contrast radiological images by machine learning methods.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to the technical field of generating artificial contrast radiological images using machine learning methods. [Background technology]

[0002] Radiology is the branch of medicine that deals with the processing of images for diagnostic and therapeutic purposes.

[0003] Although X-rays and X-ray sensitive films were formerly used primarily in medical imaging, today radiology includes a variety of different imaging methods, such as, for example, computed tomography (CT), magnetic resonance imaging (MRI) or ultrasound.

[0004] All of these methods may employ substances that facilitate the delineation or demarcation of particular structures within the subject, said substances being called contrast agents.

[0005] In computed tomography, iodine-containing solutions are usually used as contrast agents.In magnetic resonance imaging (MRI), superparamagnetic substances (e.g., iron oxide nanoparticles, superparamagnetic iron platinum particles (SIPP)) or paramagnetic substances (e.g., gadolinium chelates, manganese chelates) are usually used as contrast agents.

[0006] Based on their pattern of spread within tissue, contrast agents can be broadly divided into the following categories: extracellular contrast agents, intracellular contrast agents, and intravascular contrast agents.

[0007] Extracellular MRI contrast agents include, for example, the gadolinium chelates gadobutrol (Gadovist®), gadoteridol (Prohance®), gadoteric acid (Dotarem®), gadopentetic acid (Magnevist®) and gadodiamide (Omnican®). The highly hydrophilic properties of the gadolinium chelates and their low molecular weight result in rapid diffusion into the interstitial space after intravenous administration. After a certain relatively short period of circulation in the blood circulation system, they are excreted via the kidneys.

[0008] Intracellular contrast agents are taken up to some extent into the cells of the tissue and subsequently excreted.

[0009] Intracellular MRI contrast agents based on gadoxetic acid are distinguished, for example, by the fact that they are proportionally and specifically taken up by liver cells, hepatocytes, accumulate in functional tissues (parenchyma) and enhance contrast in healthy liver tissue before being excreted in the feces via the bile duct. Examples of such contrast agents based on gadoxetic acid are described in US 6,039,931 A; they are commercially available, for example, under the trade names Primovist® or Eovist®. Furthermore, an MRI contrast agent with low uptake in hepatocytes is gadobenate dimeglumine (Multihance®).

[0010] The contrast enhancing effect of Primovist® / Eovist® is mediated by the stable gadolinium complex Gd-EOB-DTPA (gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid). DTPA forms a complex with very high thermodynamic stability with the paramagnetic gadolinium ion. The ethoxybenzyl (EOB) radical is the mediator of hepatobiliary uptake of the contrast agent.

[0011] Intravascular contrast agents are distinguished by a significantly longer residence time in the blood circulation compared to extracellular contrast agents. Gadofosveset is, for example, an intravascular MRI contrast agent based on gadolinium. It is used in the trisodium salt monohydrate form (Ablavar®). It binds to serum albumin, thereby achieving a long residence time of the contrast agent in the blood circulation (half-life in blood of about 17 hours).

[0012] There are a number of radiological examinations in which a contrast agent is administered to the patient and the dynamic spread of the contrast agent in the body is followed by imaging methods, for example the detection and differential diagnosis of focal liver lesions by dynamic contrast-enhanced magnetic resonance imaging.

[0013] Primovist® can be used to detect tumors in the liver. The blood supply to healthy liver tissue is achieved mainly via the portal vein (portal vein), whereas the liver artery (arteria hepatica) supplies most primary tumors. Thus, after intravenous injection of a bolus of contrast agent, it is possible to observe a time delay between the signal increase of the healthy liver parenchyma and that of the tumor.

[0014] In addition to malignant tumors, frequently found in the liver are benign lesions such as cysts, hemangiomas and focal nodular hyperplasia (FNH). Proper planning of treatment requires distinguishing these from malignant tumors. Primovist® can be used to identify benign and malignant focal liver lesions. T1-weighted MRI provides information on the characteristics of said lesions. Differentiation is achieved by taking advantage of the different blood supplies to the liver and tumors, as well as the temporal profile of contrast enhancement.

[0015] The contrast enhancement achieved by Primovist® can be divided into at least two phases: the dynamic phase (including the so-called arterial, portal and late phases) and the hepatobiliary phase, in which significant uptake of Primovist® into hepatocytes already occurs.

[0016] In the case of contrast enhancement achieved by Primovist® during the distribution phase, what is observed is a typical perfusion pattern that provides information for lesion characterization. Delineating the vascularization helps to characterize the lesion type and determine the spatial relationship between the tumor and the blood vessels.

[0017] In T1-weighted MRI images, Primovist® leads to a clear signal enhancement in healthy liver parenchyma 10-20 min after injection (hepatobiliary phase), whereas lesions containing no or only few hepatocytes, e.g. metastases or moderately to poorly differentiated hepatocellular carcinoma (HCC), appear as darker areas. Summary of the Invention [Problem to be solved by the invention]

[0018] Thus, tracking the diffusion of contrast agent over time throughout the dynamic and hepatobiliary phases offers good possibilities for detection and differential diagnosis of focal liver lesions, but the examination spans a relatively long time span. To minimize motion artifacts in the MRI images, movement by the patient should be largely avoided over the above-mentioned time span. Long restrictions on movement can be uncomfortable for the patient.

[0019] This problem and further problems are solved by the present invention. The invention provides means that allow one or more radiological images to be synthetically generated. The synthetically generated radiological images are predicted by a machine learning model based on a time series of measured radiological images showing contrast agent enhancement that varies over time. This has the advantage that the radiological examination can be expedited due to the fact that not all radiological images important for the diagnosis have to be measured; one or more radiological images can be predicted (calculated) based on the measured radiological images; and thus the examination time can be reduced. Furthermore, the prediction of the radiological images is performed in frequency space (and not in real space, as is usual). As a result, it is possible to separate the contrast information from the detailed information in the radiological representation of the examination area, to limit the training and prediction to the contrast information, and then reintroduce the detailed information after the prediction. This procedure for example reduces the computational complexity. Furthermore, working in frequency space means a higher tolerance to imperfections in the image registration. [Means for solving the problem]

[0020] In a first aspect, the present invention provides a method for producing a composition comprising the steps of: - receiving a plurality of first representations of an examination region of an examined object in frequency space, where at least some of the first representations represent the examination region during a first time span after administration of a contrast agent; - feeding the plurality of first representations to a predictive model, where the predictive model is trained based on the first and second reference representations of the examination area of ​​the multiple examination objects to generate one or more second reference representations representative of the examination area during a second time span in frequency space from the first reference representations, at least some of which are representative of the examination area during a first time span after administration of a contrast agent in frequency space; - receiving one or more predicted representations of the test region in frequency space from the prediction model, where the one or more predicted representations represent the test region for a second time span; - transforming the one or more predicted representations into one or more representations of the examination area in real space; - outputting and / or storing one or more representations of the examination area in real space; The present invention provides a computer-implemented method, comprising:

[0021] The present invention further comprises: A receiving unit; a control and calculation unit; and Output unit and Equipped with Wherein the control and calculation unit is - prompting the receiving unit to receive a plurality of first representations of an examination region of the examined object in frequency space, where at least some of the first representations represent the examination region during a first time span after administration of a contrast agent; - supplying a plurality of first representations to a predictive model, where the predictive model has been trained to generate, based on the first and second reference representations of the examination area of ​​the multiple examination objects, one or more second reference representations representative of the examination area during a second time span in frequency space from the first reference representations, at least some of which are representative of the examination area during a first time span after administration of a contrast agent in frequency space; - receiving one or more predicted representations of the test region in frequency space from the prediction model, where the one or more predicted representations represent the test region for a second time span; - transforming the one or more predicted representations into one or more representations of the examination area in real space; - prompting the output unit to output and / or store one or more representations of the examination area in real space. To provide a system, the system being configured.

[0022] The present invention further provides a computer program product comprising a computer program loadable into the memory of a computer system, which causes the computer system to: - receiving a plurality of first representations of an examination region of an examined object in frequency space, where at least some of the first representations represent the examination region during a first time span after administration of a contrast agent; - feeding the plurality of first representations to a predictive model, where the predictive model is trained based on the first and second reference representations of the examination area of ​​the multiple examination objects to generate one or more second reference representations representative of the examination area during a second time span in frequency space from the first reference representations, at least some of which are representative of the examination area during a first time span after administration of a contrast agent in frequency space; - receiving one or more predicted representations of the test region in frequency space from the prediction model, where the one or more predicted representations represent the test region for a second time span; - transforming the one or more predicted representations into one or more representations of the examination area in real space; - outputting and / or storing one or more representations of the examination area in real space; Prompt them to do so.

[0023] The present invention further provides the use of an imaging agent in a method for predicting at least one radiological image, the method comprising the steps of: - generating a plurality of first representations of an examination region of the examined object in frequency space, where at least some of the first representations represent the examination region during a first time span after administration of a contrast agent; - feeding the plurality of first representations to a predictive model, where the predictive model is trained based on the first and second reference representations of the examination area of ​​the multiple examination objects to generate one or more second reference representations representative of the examination area during a second time span in frequency space from the first reference representations, at least some of which are representative of the examination area during a first time span after administration of a contrast agent in frequency space; - receiving one or more predicted representations of the test region in frequency space from the prediction model, where the one or more predicted representations represent the test region for a second time span; - transforming the one or more predicted representations into one or more representations of the examination area in real space; - outputting and / or storing one or more representations of the examination area in real space; Includes.

[0024] There is further provided an imaging agent for use in a method for predicting at least one radiological image, the method comprising: - administering a contrast agent, where the contrast agent spreads within an examination region of the examined subject; - generating a plurality of first representations of an examination region of the examined object in frequency space, where at least some of the first representations represent the examination region during a first time span after administration of a contrast agent; - feeding the plurality of first representations to a predictive model, where the predictive model is trained based on the first and second reference representations of the examination area of ​​the multiple examination objects to generate one or more second reference representations representative of the examination area during a second time span in frequency space from the first reference representations, at least some of which are representative of the examination area during a first time span after administration of a contrast agent in frequency space; - receiving one or more predicted representations of the test region in frequency space from the prediction model, where the one or more predicted representations represent the test region for a second time span; - transforming the one or more predicted representations into one or more representations of the examination area in real space; - outputting and / or storing one or more representations of the examination area in real space; Includes.

[0025] Additionally, a kit is provided comprising an imaging agent and a computer program product according to the present invention.

[0026] The invention will be described in more detail below without distinguishing between the subject matter of the invention (method, system, computer program product, use, imaging agent for use, kit). On the contrary, the following description is intended to apply equally to all subject matter of the invention, regardless of the context (method, system, computer program product, use, imaging agent for use, kit) in which it occurs.

[0027] When steps are described in a sequence in this specification or claims, this does not necessarily mean that the invention is limited to the sequence described. Instead, unless one step is built on another, this absolutely requires that the step that builds on the previous step is performed successively (however clear in the individual case), but it is also conceivable that the steps are performed in a different sequence or in parallel with each other. Therefore, the sequence described is a preferred embodiment of the invention.

[0028] The invention makes it possible to shorten the time span of a radiological examination of an object to be examined.

[0029] A "test subject" is usually a living organism, preferably a mammal, very particularly preferably a human being.

[0030] The "examination area" is usually a part of the examination object, for example an organ or a part of an organ. The "examination area", also called image volume (field of view, FOV), is in particular the volume that is imaged in the radiological image. The examination area is typically defined by a radiologist, for example on an overview image (localizer). Of course, the examination area can alternatively or additionally also be defined automatically, for example based on a selected protocol.

[0031] The term "radiological examination" is understood to mean all imaging methods that allow insight into the examined object using electromagnetic rays, particle radiation or mechanical waves for diagnostic, therapeutic and / or scientific purposes. The term "radiology" in the context of the present invention particularly encompasses the following examination methods: computed tomography, magnetic resonance imaging, ultrasound examination, positron emission tomography, echocardiography, scintigraphy.

[0032] In a preferred embodiment of the invention, the radiological examination is a magnetic resonance imaging examination.

[0033] Magnetic resonance imaging (abbreviated MRI) is an imaging method used especially in medical diagnosis to depict the structure and function of tissues and organs within the human or animal body.

[0034] In MRI, the magnetic moments of protons in the object are aligned in a basic magnetic field, resulting in a macroscopic magnetization along the longitudinal direction. This is then deflected from its rest position by irradiation with a radio frequency (HF) pulse (excitation). The return from the excited state to the rest position (relaxation), i.e. the dynamics of the magnetization, is then detected as a relaxation signal by one or more HF receiving coils.

[0035] For spatial encoding, rapidly switched gradient fields are superimposed on the basic magnetic field. The captured relaxation signals, or the detected and spatially resolved MRI data, initially exist as raw data in frequency space and can be transformed into real space (image space), for example, by a subsequent inverse Fourier transformation.

[0036] For native MRI, tissue contrast is created by different relaxation times (T1 and T2) and proton densities. T1 relaxation represents the transition of longitudinal magnetization to its equilibrium state, where T1 is the time it takes to reach 63.21% of the equilibrium magnetization before resonant excitation. This is also called the longitudinal relaxation time or spin-lattice relaxation time. T2 relaxation similarly represents the transition of transverse magnetization to its equilibrium state.

[0037] In radiological examinations, contrast agents are commonly used for contrast enhancement.

[0038] An "imaging agent" is a substance or mixture of substances that improves the depiction of body structures and functions in imaging procedures such as diagnostic x-rays, magnetic resonance imaging, and ultrasound.

[0039] Examples of contrast agents can be found in the literature (e.g., ASL Jascinth et al.: Contrast Agents in computed tomography: A Review, Journal of Applied Dental and Medical Sciences, 2016, Vol. 2, Issue 2, 143-149; H. Lusic et al.: X-ray-Computed Tomography Contrast Agents, Chem. Rev. 2013, 113, 3, 1641-1666; https: / / www.radiology.wisc.edu / wp-content / uploads / 2017 / 10 / contrast-agents-tutorial.pdf, MR Nough et al.: Radiographic and magnetic resonances contrast agents: Essentials and tips for safe practices, World J Radiol. 2017 Sep 28; 9(9): 339-349; LCA Bonyi et al.: Intravascular Contrast Media in Radiography: Historical Development & Review of Risk Factors for (See Adverse Reactions, South American Journal of Clinical Research, 2016, Vol. 3, Issue 1, 1-10; ACR Manual on Contrast Media, 2020, ISBN: 978-1-55903-012-0; A. Ignee et al.: Ultrasound contrast agents, Endosc Ultrasound. 2016 Nov-Dec; 5(6): 355-362).

[0040] MRI contrast agents exert their effect by modifying the relaxation time of the structures that incorporate them. Two groups can be distinguished: paramagnetic and superparamagnetic substances. Both groups of substances have unpaired electrons that induce a magnetic field around individual atoms or molecules. Superparamagnetic contrast agents cause a predominant shortening of T2, whereas paramagnetic contrast agents cause a predominant shortening of T1. The effect of said contrast agents is indirect, since the agent itself does not emit a signal, but merely affects the intensity of the signal in its vicinity. An example of a superparamagnetic contrast agent is iron oxide nanoparticles (SPIO, superparamagnetic iron oxide). Examples of paramagnetic contrast agents include gadolinium chelates such as gadopentetate dimeglumine (trade name: Magnevist (registered trademark), etc.), gadoteric acid (Dotarem (registered trademark), Dotagita (registered trademark), Cyclolux (registered trademark)), gadodiamide (Omniscan (registered trademark)), gadoteridol (ProHance (registered trademark)), and gadobutrol (Gadovist (registered trademark)).

[0041] Using the invention it is possible to predict one or more composite radiographic images of an examination region, the prediction being performed using a prediction model.

[0042] The predictive model is a computer-aided model configured to predict one or more second representations of an examination region of an examination object in frequency space based on a plurality of first representations of the examination region of an examination object in frequency space, at least some of the plurality of first representations representing the examination region during a first time span after administration of a contrast agent, and at least one second representation representing the examination region during a second time span.

[0043] The term "plurality" means a number of at least 2. The first representation of the plurality used in prediction is typically 10 or less.

[0044] The second time span can come before or after the first time span. It is also possible for the time spans to at least partially overlap, or for one time span to be within the other.

[0045] Figure 1(a) (Fig.1(a)), Figure 1(b) (Fig.1(b)), Figure 1(c) (Fig.1(c)) and Figure 1(d) (Fig.1(d)) are for illustration purposes. Figure 1(a), Figure 1(b), Figure 1(c) and Figure 1(d) respectively show a timeline. Defined time points are marked on the timeline. Time t0 characterizes the time point when the contrast agent is administered to the examination subject. The dashed frames indicate the time span during which the examination subject undergoes a radiological examination, i.e. the time the examination subject spends in, for example, a magnetic resonance imaging system or a computed tomography system.

[0046] FIG. 1(a) shows a schematic diagram of a typical profile of a radiological examination. The examination object is introduced into the tomography apparatus. At a time t -1 At time t -1 At time instant (t -1 At time t0, contrast agent has not yet been administered to the subject, i.e. the representation is a contrast-free (native) representation. At time t0, contrast agent is administered to the subject, which is located in the tomography apparatus. At time t1, t2 and t3, further representations of the subject are generated. The subject then leaves the tomography apparatus. The subject is then imaged over a relatively long time span T t , was located in the tomography device. During this time, four representations of the examination area were generated based on the measurements.

[0047] 1(b) shows a schematic diagram of a profile of a radiological examination according to the invention. An examination object is introduced into the tomography apparatus. At a time t -1 At time t -1 At time (t -1), contrast agent has not yet been administered to the subject under examination, i.e. the representation is a contrast agent-free (native) representation. At time t0, contrast agent is administered to the subject under examination, which is located in the tomography apparatus. At times t1 and t2, further representations of the subject under examination are generated. The subject then leaves the tomography apparatus. A representation generated based on a measurement at time t3 in the case of FIG. 1(a) is predicted (calculated) in the case of the profile shown in FIG. 1(b). The subject under examination is subjected to a time span T a during the time span T a is the time span T in Fig. 1(a). t Therefore, in the case of Fig. 1(b), the radiological examination does not last as long as in the case of Fig. 1(a) and is more comfortable for the subject; however, what is produced in both cases is t -1 , t1, t2 and t3. -1 The representations generated based on measurements at t1, t2 and t3 are in the context of the present invention initial representations of the examination region of the examination object, where at least some of them represent the examination region during a first time span after administration of the contrast agent, i.e. at the times t1, t2 and t3; -1 The expression at represents the examination region in the time span before the administration of the contrast agent. The expression at time t represents the examination region in the time span before the administration of the contrast agent. -1 , t1 and t2, which represents the inspection area during a second time span, said second time span coming after the first time span in the case of FIG. 1(b).

[0048] Fig. 1(c) shows diagrammatically a further profile of a radiological examination according to the invention. A contrast agent is administered to the examination object at time t0. At said time (t0), the examination object is not yet located in the tomography apparatus. Only after the contrast agent has been administered is the examination object introduced into the tomography apparatus. At times t1, t2 and t3 (in a first time span after the administration of the contrast agent) representations of the examination area of ​​the examination object are generated. These represent the examination area in the first time span after the administration of the contrast agent. After the generation of the representation at time t3, the examination object can leave the tomography apparatus and the radiological examination is finished. From the representations at times t1, t2 and t3, it is possible to determine the area of ​​examination of the examination object at time t -1 It is possible to predict the expression at time t -1 The expression in represents the inspection area during a second time span, which comes before the first time span. b during the time span T b is the time span T in Fig. 1(a). t Therefore, in the case of Fig. 1(c), the radiological examination does not last as long as in the case of Fig. 1(a) and is more comfortable for the subject, but what is produced in both cases is t -1 , t1, t2 and t3 are representations of the examination area representing the examination area at time instants t1, t2 and t3.

[0049] Fig. 1(d) shows a further profile of a radiological examination according to the invention in a schematic way. This example is intended to make it clear that the invention is not limited to individual administrations of contrast agent. For example, it is conceivable to carry out two or more administrations. In the case of individual administrations, it is not necessary to administer the same contrast agent, but instead different contrast agents can be administered. A (first) contrast agent is administered to the examination object at a time t0. At said time (t0), the examination object is not yet located in the tomography apparatus. It is only after the administration of the contrast agent that the examination object is introduced into the tomography apparatus. At a time t2, a first representation of the examination area of ​​the examination object is generated. The representation at the time t2 represents the examination area in a first time span after the administration of the (first) contrast agent. At a time t3, a (second) contrast agent is administered. At said time (t3), the examination object is located in the tomography apparatus. After administration of the (second) contrast agent, two further representations of the examination region are generated, one at time t4 and another at time t5. The representations generated at time t2, t4 and t5 are representations that represent the examination region during a first time span after administration of the contrast agent. These representations are -1 and / or the expression of the examination area at time t -1 and the representation of the examination area at t1 represents the examination area during a second time span, said second time span coming before the first time span.

[0050] Combinations of the profiles shown in Figures 1(b), 1(c) and 1(d) with further profiles / variations are possible as well.

[0051] In a preferred embodiment of the invention, the first time span starts before or with the administration of a contrast agent. It is advantageous if one or more representations of the examination area are generated showing the examination area without contrast agent (native images), since the radiologist can already obtain important information about the health of the examined subject from such images. For example, the radiologist can identify bleeding in native MRI images.

[0052] In order for a predictive model according to the present invention to be able to make the predictions described herein, it must be properly configured beforehand.

[0053] Here, the term "prediction" means that at least one representation of the examination area representative of the examination area during a second time span in frequency space is calculated using a plurality of first representations of the examination area in frequency space, where at least some of the plurality of first representations represent the examination area during a first time span after administration of a contrast agent.

[0054] The predictive model is preferably created (constructed, trained) using a self-learning algorithm in a supervised machine learning process. Training data is used for learning. The training data includes a plurality of representations of the examination area for each examination object of a large number of examination objects. The examination area is usually the same for all examination objects (e.g., a part of a human body, an organ or a part of an organ). The representations of the training data set are also referred to herein as reference representations. The term "large number" preferably means more than 10, even more preferably more than 100.

[0055] For each examination subject, the training data includes i) a plurality of first reference representations of the examination region in frequency space, at least some of which represent the examination region during a first time span after administration of a contrast agent, and ii) one or more second reference representations of the examination region in frequency space representing the examination region during a second time span.

[0056] A predictive model is trained to predict (calculate) for each test subject one or more second referring expressions from a plurality of first referring expressions.

[0057] During machine learning, a self-learning algorithm generates a statistical model based on training data. This means that the examples are not simply memorized, but the algorithm "recognizes" patterns and regularities in the training data. Thus, the predictive model can also evaluate unknown data. Validation data can be used to test the quality of the evaluation of unknown data.

[0058] The predictive model can be trained by supervised learning, i.e., a pair of data sets (first and second representations) are presented to the algorithm one after the other, and the algorithm then learns the relationship between the first and second representations.

[0059] Self-learning systems trained by supervised learning have been widely described in the prior art (see, for example, C. Perez: Machine Learning Techniques: Supervised Learning and Classification, Amazon Digital Services LLC-Kdp Print Us, 2019, ISBN1096996545, 9781096996545).

[0060] Preferably, the predictive model is or comprises an artificial neural network.

[0061] An artificial neural network comprises at least three layers of processing elements: a first layer having input neurons (nodes), an Nth layer having at least one output neuron (node), and N-2 hidden layers, where N is a natural number greater than 2.

[0062] The input neurons are operative to receive a first representation, and the output neurons are operative to output one or more second representations for a plurality of the first representations.

[0063] The processing elements of the layer between the input neurons and the output neurons are connected to each other in a predetermined pattern with predetermined connection weights.

[0064] Preferably, the artificial neural network is a so-called Convolutional Neural Network (abbreviated as CNN).

[0065] Convolutional neural networks are able to process input data in the form of matrices. A CNN essentially consists of alternating filters (convolutional layers) and aggregation layers (pooling layers), and finally one or more layers of "normal" fully connected neurons (dense / fully connected layers).

[0066] The training of the neural network can be carried out, for example, by the backpropagation method. The objective here for the network is the maximum reliability of mapping given input data to given output data. The mapping quality is described by a loss function. The objective is to minimize the loss function. In the backpropagation method, the artificial neural network is trained by modifying the connection weights.

[0067] In the trained state, the connection weights between the processing elements include information about the relationship between the first representation and one or more second representations, and this information can be used to predict one or more second representations indicative of the examination region during a second time span for a new plurality of first representations (e.g., of a new subject), at least some of which indicative of the examination region during the first time span after administration of a contrast agent.

[0068] Cross-validation methods can be used to split the data into training and validation data sets. The training data set is used for backpropagation training of the network weights. The validation data set is used to check the accuracy of predictions that the trained network can make when applied to unknown data.

[0069] Further details regarding the construction and training of artificial neural networks can be gleaned from the prior art (see, for example, S. Khan et al.: A Guide to Convolutional Neural Networks for Computer Vision, Morgan & Claypool Publishers 2018, ISBN1681730227, 9781681730226, WO2018 / 183044A1, WO2018 / 200493, WO2019 / 074938A1, WO2019 / 204406A1, WO2019 / 241659A1).

[0070] Preferably, the predictive model is a generative adversarial network (GAN) (see, for example, http: / / 3dgan.csail.mit.edu / ).

[0071] In addition to the representation, further information about the inspection object, inspection area, inspection conditions and / or radiological inspection can also be used for training, validation and prediction.

[0072] Examples of information about the examined subject are sex, age, weight, height, medical history, nature and duration and amount of medicines already taken, blood pressure, central venous pressure, respiratory rate, serum albumin, total bilirubin, blood glucose, iron content, respiratory capacity, etc. These can be read, for example, from a database or an electronic patient file.

[0073] Examples of information on the area of ​​examination are: pre-existing conditions, surgery, partial resection, liver transplant, iron liver, fatty liver, etc.

[0074] As already explained, the representation of the test domain used for training, validation and prediction is a representation of the test domain in frequency space (also called spatial-frequency space or Fourier space or frequency domain or Fourier representation).

[0075] In magnetic resonance imaging, raw data usually arises as so-called k-space data by the above-mentioned measurement methods. Said k-space data is a representation of the examination area in frequency space, i.e. such k-space data can be used for training, validation and prediction. If a representation in real space exists, such a representation in real space can be converted (transformed) into a representation in frequency space by Fourier transformation; conversely: a representation in frequency space can be converted (transformed) into a representation in real space, for example, by inverse Fourier transformation.

[0076] Thus, if a radiological image of the examination area exists in the form of a two-dimensional image in real space, this representation of the examination area can be transformed by a 2D Fourier transform into a two-dimensional representation of the examination area in frequency space.

[0077] A 3D image (volume rendering) of the examination area can be treated as a stack of 2D images. It is also possible to convert the 3D image into a 3D representation of the examination area in frequency space by a 3D Fourier transform.

[0078] It is possible to use transforms other than the Fourier transform to convert the real space representation into a frequency space representation. Three main properties that must be satisfied by such a transform are: a) The existence of a clear inverse transform (a clear link between the real space representation and the frequency space representation) b) Locality of contrast information c) Robustness with respect to defective image registration.

[0079] The details of transforming from one representation to another are described in numerous textbooks and publications (see, for example, W. Burger, MJ Burger: Digital Image Processing: An Algorithmic Introduction Using Java, Texts in Computer Science, 2nd edition, Springer-Verlag, 2016, ISBN: 9781447166849; W. Birkfellner: Applied Medical Image Processing, Second Edition: A Basic Course, Verlag Taylor & Francis, 2014, ISBN: 9781466555570; R. Bracewell: Fourier Analysis and Imaging, Verlag Springer Science & Business Media, 2004, ISBN: 9780306481871).

[0080] FIG. 2 exemplarily and diagrammatically illustrates the generation of a representation of an examination region in real space and in frequency space.

[0081] FIG. 2 shows a timeline. At three different time points t1, t2 and t3, a representation of the examination area is generated based on the measurements. The examination area is a human lung. At time point t1, a first representation is generated. This can be a representation of the examination area (lungs) in real space (O1) or a representation of the examination area (lungs) in frequency space (F1). The representation of the examination area in real space (O1) can be transformed by a Fourier transform FT into a representation of the examination area in frequency space (F1). The representation of the examination area in frequency space (F1) can be transformed by an inverse Fourier transform iFT into a representation of the examination area in real space (O1). The two representations (O1) and (F1) contain the same information about the examination area, only in different representations. At time point t2, a further representation is generated. This can be a representation of the examination area (lungs) in real space (O2) or a representation of the examination area (lungs) in frequency space (F2). The representation of the examination area in real space (O2) can be transformed by a Fourier transform FT into a representation of the examination area in frequency space (F2). The representation of the examination area in frequency space (F2) can be transformed by an inverse Fourier transform iFT into a representation of the examination area in real space (O2). The two representations (O2) and (F2) contain the same information about the examination area, only in different representations. At time t3, a further representation is generated. This can be a representation of the examination area (lungs) in real space (O3) or a representation of the examination area (lungs) in frequency space (F3). The representation of the examination area in real space (O3) can be transformed by a Fourier transform FT into a representation of the examination area in frequency space (F3). The representation of the examination area in frequency space (F3) can be transformed by an inverse Fourier transform iFT into a representation of the examination area in real space (O3). The two representations (O3) and (F3) contain the same information about the examination area, only in different representations.

[0082] Representations (O1), (O2) and (O3) of the examination region in real space are familiar representations for humans; they can be immediately grasped by humans. Representations (O1), (O2) and (O3) show how the contrast agent spreads dynamically in the veins. The same information is contained in representations (F1), (F2) and (F3), although it is more difficult for humans to grasp.

[0083] Figure 3 (Fig. 3) shows, in a schematic and exemplary manner, how the representations of the test regions (F1), (F2) and (F3) in frequency space as generated in Figure 2 can be used to train a predictive model (PM). The representations (F1), (F2) and (F3) form a set of training data of test objects. Training is performed using multiple training data sets of multiple test objects.

[0084] Representations (F1) and (F2) are of a plurality (here two) of first reference representations of the examination area in frequency space, at least some of which represent the examination area during a first time span after administration of a contrast agent. Representation (F3) is a second reference representation of the examination area in frequency space representing the examination area during a second time span. In FIG. 3, a prediction model is trained to predict a representation (F3) of the examination area in frequency space from the representations (F1) and (F2) of the examination area in frequency space. The representations (F1) and (F2) are input to a prediction model (PM), and the prediction model calculates a representation (F3*) from the representations (F1) and (F2). The asterisk (*) indicates that the representation (F3*) is a predicted representation. The calculated representation (F3*) is compared with the representation (F3). The deviations can be used in a backpropagation method to train a prediction model to reduce the deviations to a defined minimum value. If a predictive model is trained based on a large number of training datasets of a large number of test subjects and the prediction reaches a defined accuracy, the trained predictive model can be used for prediction, which is exemplarily and diagrammatically shown in FIG.

[0085] Figure 4 (Fig.4) shows the predictive model (PM) trained in Figure 3. The predictive model is used to generate one or more second representations of the examination area in frequency space based on a plurality of first representations of the examination area in frequency space, at least some of which represent the examination area in a first time span after administration of a contrast agent, said one or more second representations representing the examination area in a second time span.

[0086] In this example, two first representations of the examination area in frequency space (F ~ 1) and (F ~ 2) is input to a prediction model (PM), which in turn generates a second representation (F ~ 3*). The tilde (~) indicates that the expression is a new expression to be tested, and typically, no expression exists that was used in the training method to train the predictive model. The asterisk (*) indicates that the expression (F ~ 3*) is the predicted representation. The representation of the test region in frequency space (F ~ 3*) is expressed in real space by the inverse Fourier transform iFT ( ~ 3*).

[0087] The use of a representation of the examination area in frequency space has an advantage over the use of a representation of the examination area in real space (also called image space): with a representation of the examination area in frequency space, contrast information, which is important for training and prediction, can be separated from detailed information (fine structure). This allows, in the case of training, to focus on the information that is to be learned by the predictive model, and, in the case of prediction, also on the information that is to be predicted by the predictive model: the contrast information.

[0088] While the contrast information in a representation of the examination area in real space is typically distributed throughout the representation (each pixel / voxel essentially has information about the contrast), the contrast information in a representation of the examination area in frequency space is encoded in and around the center of the frequency space. In other words, the low frequencies in the representation in frequency space are responsible for the contrast, while the high frequencies contain information about the fine structure.

[0089] It is therefore possible to isolate the contrast information, restrict training and prediction to the contrast information, and reintroduce information about the fine structure after training / prediction.

[0090] In a preferred embodiment, the method according to the invention comprises the steps of: - receiving a plurality of first representations of an examination region of an examination object in frequency space, where at least a portion of the first representations represent the first examination region during a first time span after administration of a contrast agent; - identifying a region in the first representation, where the identified region includes a center of frequency space; - reducing the first representation to the identified region, where a plurality of reduced first representations are obtained; - feeding the plurality of reduced first representations to a predictive model; - receiving one or more second representations of the test region in frequency space from the predictive model, where the one or more second representations represent the test region during a second time span; - supplementing the one or more second representations with one or more regions of the received first representation that are outside the identified region, whereby one or more supplemented second representations are obtained; - transforming the one or more captured second representations into one or more representations of the examination area in real space; - outputting and / or storing one or more representations of the examination area in real space; Includes.

[0091] Identification of the region in the first representation can for example be achieved by a user of the computer system according to the invention by inputting one or more parameters into the computer system according to the invention and / or by selecting from a list defining the shape and / or size of the region. However, it is also conceivable that the identification is performed automatically, for example by a computer system according to the invention which is suitably configured to select a given region in the representation of the examination region.

[0092] The identified region is typically smaller than the frequency space filled by the first representation, but in any case includes the center of the frequency space.

[0093] The region of the frequency space that includes the center of the frequency space (also called the origin or zero point) contains the contrast information relevant to the method according to the invention. If the identified region is smaller than the frequency space filled by the first representation, the result is a lower computational complexity for the subsequent prediction (this also applies in particular to the training of the predictive model). Thus, the choice of the size of the region can have a direct impact on the computational complexity.

[0094] In principle, it is also possible to identify a region corresponding to the entire frequency space filled by the first representation; in such a case there is no reduction of the frequency space into sub-regions and the computational complexity is at its maximum.

[0095] Thus, by specifying a region around the center of frequency space, a user of a computer system according to the invention can decide for himself whether he wants a complete representation of the examination region in frequency space to form the basis for training and prediction, or whether he wants to reduce the computational complexity, where the user can directly influence the required computational complexity through the size of the specified region.

[0096] The identified region typically has the same dimensions as the frequency space: in the case of a 2D representation in a 2D frequency space, the identified region typically is an area, and in the case of a 3D representation in a 3D frequency space, the identified region typically is a volume.

[0097] The identified region can in principle have any shape, and thus can be, for example, circular and / or angular, concave and / or convex. Preferably, the region is rectangular or cubic in the case of a 3D frequency space in a Cartesian coordinate system, and rectangular or square in the case of a 2D frequency space in a Cartesian coordinate system. However, it can also be spherical or circular or have other shapes.

[0098] Preferably, the geometric centroid of the identified region coincides with the centre of the frequency space.

[0099] The representations used for training, validation and prediction are reduced to the identified regions. The term "reduced" here means that all parts of the representation that are not in the identified range are cut off (discarded) or covered by masking. In the case of masking, the regions outside the identified regions are covered with a mask, so that only the identified regions remain uncovered, and in the case of masking, the color values ​​of the corresponding pixels / voxels can be set to zero (black), for example.

[0100] The representation thus obtained is also referred to herein as the reduced representation.

[0101] The first representation obtained after reduction (reduced first representation) is fed to a prediction model, which has been previously trained in a training method for learning the dynamic influence of the amount of contrast agent on the representation of the examination region in frequency space, which training likewise preferably utilizes the reduced representations (reduced first representation and reduced second representation).

[0102] In this way, the predictive model can learn what dynamic effect the contrast agent had on the representation of the examination region and apply this learned "knowledge" to predict one or more (reduced) second representations based on the (reduced) first representation.

[0103] The one or more predicted second representations represent the examination region in frequency space during a second time span.

[0104] One or more predicted second representations are calculated and output by a predictive model based on the (reduced) first representation.

[0105] If at least one predicted second representation is generated based on a reduced first representation, it is preferable to add back previously discarded parts (cut out or covered with a mask) in order to lose as little information as possible about the fine structure in the final artificially generated image.

[0106] In order to reuse previously discarded (cropped or masked) portions, the at least one predicted second representation may be superimposed with the at least one received first representation such that the at least one predicted second representation replaces a corresponding superimposed frequency region of the at least one initially received first representation. Preferably, the predicted second representation replaces a corresponding frequency region of the initially received first representation that represents the examination region without contrast agent.

[0107] The permutation of the superimposed frequency domains corresponds to the supplementation of the received first representation with one or more regions of frequency space that were omitted when reducing the first representation to the identified region.

[0108] In other words: the frequency space of the at least one predicted second representation of the examination area is filled by an area of ​​the at least one initially received first representation, whereby the at least one initially received first representation is larger than the predicted second representation.

[0109] Thus, by using a representation of the examination region in frequency space, it is possible to separate the contrast information from the detailed information, restrict training and prediction to the contrast information, and then reintroduce the detailed information after training and / or prediction. As already explained, this procedure reduces the computational complexity during training, validation and prediction.

[0110] However, working in frequency space has yet another advantage over working in real space: in frequency space, coregistration of individual representations is not as important as in real space. "Co-registration" (also called "image registration" in the prior art) is an important process in digital image processing, which serves to match two or more images of the same, or at least similar, scene with each other in the best possible way. One of the images is defined as the reference image, the others are called object images. A compensation transformation is calculated to optimally match said object image with the reference image. The images to be aligned are different from each other, since they were acquired from different positions, at different times and / or with different sensors.

[0111] In the present case, the individual representations of the plurality of first representations of the examination region are firstly generated at different times; secondly, they differ with regard to the content and diffusion of the contrast agent in the examination region.

[0112] Therefore, the use of a representation of the examination area in frequency space has an advantage over the use of a representation of the examination area in real space, and the training, validation and prediction methods are more tolerant with respect to errors in image registration. In other words: if the representation in frequency space is not exactly superimposed, this has less impact than if the representation in real space is not superimposed to pixel / voxel accuracy. This is due to the properties of the Fourier transform: as already explained, the contrast information of a Fourier transformed image is always mapped near the origin (center) of the Fourier space. Turns or rotations in image space (real space) result in image information (e.g. visible structures) being localized in different areas of the transformed image. However, in Fourier space, these transformations do not change the areas where the contrast information relevant to the present invention is encoded.

[0113] FIG. 5 exemplarily illustrates diagrammatically the steps in training a predictive model according to a preferred embodiment of the present invention.

[0114] What is received are two first representations (F1) and (F2) of the object to be examined in frequency space and a second representation (F3) of the object to be examined in frequency space. In the representations (F1), (F2) and (F3), respectively, the same area A is identified. The area A includes the center of the frequency space, in this case having a square shape, the geometric center of gravity of the square coinciding with the center of the frequency space. The representations (F1), (F2) and (F3) are each reduced to the identified area A: three reduced representations (F1 red ), (F2 red ) and (F3 red ) The reduced representation is used for training. The predictive model is then trained on the reduced representation (F1 red ) and (F2) reduced to (F3 red ) is trained to predict the reduced representation (F1 red ) and (F2 red ) is fed to the predictive model (PM), which then uses the reduced representation (F3 red ) as close as possible to the reduced representation (F3*red ) to calculate

[0115] FIG. 6 exemplarily illustrates generally how the predictive model trained in FIG. 5 can be used for prediction.

[0116] In this example, two first representations of the examination area in frequency space (F ~ 1) and (F ~ 2) are received and each is reduced to a specific region A. The result is two reduced first representations (F ~ 1 red ) and (F ~ 2 red ). The reduced first representation (F ~ 1 red ) and (F ~ 2 red ) is fed to a trained predictive model (PM), which then uses the reduced first representation (F ~ 1 red ) and (F ~ 2 red ) to the second representation (F ~ 3 red Calculate the reduced second representation (F ~ 3 red *) in a further step is to convert the received first representation (F ~ The first representation received (F ~ 1) area (F ~ 1 DI ) (with a region outside the specified region A). As explained, the first representation received (F ~ Alternatively or in addition to part 1), a second representation (F ~ 2) into a third representation (F ~ 3 red* ) can also be added.

[0117] Supplementary expression (F ~ 3 red *)+(F ~ 1 DI) to obtain the expression of the inspection area in real space (O ~ 3*) can be generated by inverse Fourier transform.

[0118] It should be noted that other methods can also be used to convert the frequency space representation into a real space representation, such as, for example, iterative reconstruction methods.

[0119] The method according to the invention can be implemented by means of a computer system. The invention further provides a computer system configured to carry out the method according to the invention (e.g. by means of a computer program according to the invention).

[0120] 7 (Fig. 7) shows a schematic and exemplary embodiment of a computer system according to the invention. The computer system (10) comprises a receiving unit (11), a control and calculation unit (12) and an output unit (13).

[0121] A "computer system" is a system for electronic data processing that processes data by means of programmable computational rules. Such a system usually comprises a control and computation unit, often called a "computer", which comprises a processor for carrying out logical operations, a memory for loading computer programs, and peripheral devices.

[0122] In computer technology, a "peripheral" refers to any device that is connected to a computer and used to control the computer and / or as an input and output device. Examples are monitors (screens), printers, scanners, mice, keyboards, joysticks, drives, cameras, microphones, speakers, etc. Built-in ports and expansion cards are also considered peripherals in computer technology.

[0123] Today's computer systems are generally subdivided into desktop PCs, portable PCs, laptops, notebooks, netbooks and tablet PCs, as well as so-called handhelds (e.g. smartphones); all such systems may be used for the practice of the present invention.

[0124] Input into the computer system (e.g. for control by a user) is achieved via input means such as, for example, a keyboard, a mouse, a microphone, a touch-sensitive display, etc. Output is achieved via an output unit (13), which may in particular be a monitor (screen), a printer and / or a data storage medium.

[0125] A computer system (10) according to the present invention is configured to predict one or more second representations of an examination area of ​​an object in frequency space from a plurality of first representations of the examination area of ​​the object in frequency space representing the examination area during a first time span after administration of a contrast agent, where the one or more second representations represent the examination area during the second time span.

[0126] The control and calculation unit (12) serves to control the receiving unit (11) and the output unit (13), to coordinate the data and signal flow between the various units, to process the representation of the examination area and to generate the artificial radiological image. It is conceivable that there are several control and calculation units.

[0127] The receiving unit (11) serves to receive a representation of the examination area. The representation may for example be transmitted from a magnetic resonance imaging system, from a computed tomography system or may be read from a data storage medium. The magnetic resonance imaging system or the computed tomography system may be a component of the computer system according to the invention. However, it is also conceivable that the computer system according to the invention is a component of the magnetic resonance imaging system or the computed tomography system. The representation may be transmitted via a network connection or a direct connection. The representation may be transmitted via wireless communication (WLAN, Bluetooth, mobile communication, etc.) and / or via a cable. It is conceivable that there are several receiving units. The data storage medium may also be a component of the computer system according to the invention or may be connected thereto, for example via a network. It is conceivable that there are several data storage media.

[0128] The representation and possibly further data (eg information about the examination object, image acquisition parameters, etc.) are received by a receiving unit and transmitted to a control and calculation unit.

[0129] The control and computing unit is configured to generate an artificial radiation image based on the received data.

[0130] Via an output unit (13), the artificial radiation image can be displayed (e.g. on a monitor), output (e.g. via a printer) or stored on a data storage medium. It is conceivable that there are several output units.

[0131] FIG. 8 exemplarily illustrates in flow chart form a preferred embodiment of the method according to the invention for training a predictive model.

[0132] The method (100) comprises the steps of: (110) receiving training data, where for each examination object of the number of examination objects, the training data includes: i) a plurality of first reference representations of the examination area in frequency space, at least some of which represent the examination area during a first time span after administration of a contrast agent, and ii) one or more second reference representations of the examination area in frequency space representing the examination area during a second time span; (120) for each test subject: feeding a plurality of first referring expressions to a predictive model, where the predictive model is trained to generate one or more second referring expressions based on the plurality of first referring expressions, where the training includes minimizing a loss function, where the loss function quantifies deviation of the generated second referring expressions from the one or more received second referring expressions; (130) outputting and / or storing the trained prediction model and / or feeding the trained prediction model to a method for predicting one or more representations of the test region of a new test object; Includes.

[0133] FIG. 9 exemplarily illustrates in the form of a flow chart a further preferred embodiment of the method according to the invention for training a predictive model.

[0134] The method (200) includes the steps of: (210) receiving training data, where the training data includes, for each examination object of the number of examination objects, i) a plurality of first reference representations of the examination area in frequency space, at least some of which represent the examination area during a first time span after administration of a contrast agent, and ii) one or more second reference representations of the examination area in frequency space representing the examination area during a second time span; (220) identifying a region in a first reference expression, where the identified expression includes a center in frequency space; (230) reducing a referring expression to a specified region, where a plurality of reduced first referring expressions and one or more reduced second referring expressions are obtained for each test object; (240) for each test subject: feeding the plurality of reduced first referring representations to a predictive model, where the predictive model is trained to generate one or more reduced second referring representations based on the plurality of reduced first referring representations, where the training includes minimizing a loss function, where the loss function quantifies deviation of the generated reduced second referring representations from the one or more reduced second referring representations in the training data; (250) outputting and / or storing the trained predictive model and / or feeding the trained predictive model to a method for predicting one or more representations of the test region of a new test object. Includes.

[0135] FIG. 10 exemplarily illustrates in flow chart form a preferred embodiment of a method according to the present invention for predicting one or more representations.

[0136] The method (300) includes the steps of: (310) providing a predictive model, where the predictive model has been trained according to the method (100) described above; (320) receiving a plurality of first representations of an examination region of an examination object in frequency space, where at least a portion of the first representations represent the examination region during a first time span after administration of a contrast agent; (330) providing the plurality of first representations to a predictive model; (340) receiving one or more predicted representations of the test region in frequency space from the prediction model, where the one or more predicted representations represent the test region for a second time span; (350) converting the one or more predicted representations into one or more representations of the examination area in real space; (360) Outputting and / or storing a representation of one or more inspection regions in real space. Includes.

[0137] FIG. 11 exemplarily illustrates in flow chart form a further preferred embodiment of a method according to the present invention for predicting one or more representations.

[0138] The method (400) includes the steps of: (410) providing a predictive model, where the predictive model has been trained according to the method (200) described above; (420) receiving a plurality of first representations of an examination region of an examination object in frequency space, where at least a portion of the first representations represent the examination region during a first time span after administration of a contrast agent; (430) identifying a region in the first representation, where the identified region includes a center of frequency space; (440) reducing the first representation to the identified region, where a plurality of reduced first representations are obtained; (450) feeding the plurality of reduced first representations to a predictive model; (460) receiving one or more second representations of the test area in frequency space from the prediction model, where the one or more second representations represent the test area during a second time span; (470) supplementing one or more second representations with one or more regions of the received first representation that are outside the identified region, where one or more supplemented second representations are obtained; (480) Transforming the one or more supplemented second representations into one or more representations in real space; (490) Outputting and / or storing one or more representations of the examination area in real space. Includes. EXAMPLES

[0139] Below are some examples of how the present invention can be used to generate artificial radiation images.

[0140] Example 1 In one embodiment, the present invention is used to simulate intravascular contrast agents (also called blood pool contrast agents).

[0141] When generating radiological images with relatively long acquisition / scan times, for example in the case of free-breathing image acquisition of the chest and abdomen to depict the vasculature (e.g. diagnosis of pulmonary embolism in free-breathing MRI), the extracellular contrast agent is cleared from the vasculature relatively quickly, which means that the contrast decays quickly. However, it would be advantageous if the contrast could be maintained for a longer time.

[0142] To solve this problem, multiple first representations of the examination region in frequency space are generated / received in a first step, where at least some of the first representations represent the examination region after administration of a contrast agent.

[0143] The administered imaging agent may be an extracellular imaging agent and / or an intracellular imaging agent.

[0144] The contrast agent is introduced, preferably using weight-based dosing, into the subject's blood vessel, for example an arm vein, from where it travels with the blood along the blood circulation system.

[0145] The "circulatory system" refers to the pathways covered by blood in the body of humans and most animals. It is the system of blood flow formed by the heart and by the network of blood vessels (cardiovascular system, vascular system).

[0146] The extracellular contrast agent circulates within the blood circulation for a period of time that depends on the subject, the contrast agent and the dose administered, while being continuously cleared from the blood circulation via the kidneys.

[0147] At least one first representation of the vasculature or a portion thereof is captured while the contrast agent spreads and / or circulates within the examined vasculature. Multiple first representations may be captured representing different phases of the diffusion of the contrast agent in the vasculature or a portion thereof (e.g., a distribution phase, an arterial phase, a venous phase, etc.). The capture of multiple images allows for later differentiation of blood vessel types.

[0148] The measured representations represent the vasculature or a part thereof with contrast enhancement relative to the surrounding tissue. Preferably, at least one representation represents an artery with contrast enhancement (arterial phase), while at least one further representation represents a vein with contrast enhancement (venous phase).

[0149] The artificial representation is generated based on the measured representation. The artificial representation preferably shows the same examination area as the measured representation. If multiple measured representations of the examination area are captured at different times after administration of the contrast agent, the later representations will in particular show the blood vessels with a gradually decreasing contrast to the surrounding tissue, since the contrast agent is gradually cleared from the blood vessels. In contrast, the artificial representation shows the blood vessels with a consistently high contrast to the surrounding tissue.

[0150] This is achieved by the measured representations being fed into a predictive model according to the present invention, which has been pre-trained to predict multiple representations indicative of vascular contrast enhancement that are constant over time, based on measured representations indicative of vascular contrast enhancement that changes over time.

[0151] The reference data used for training and validating such predictive models typically include measured representations of the examined area after administration of an extracellular or intracellular contrast agent. The reference data may further include a representation of the examined area after administration of a blood pool contrast agent. Such reference data may be, for example, confirmed in a clinical study. An intravascular contrast agent that may be used in such clinical trials is, for example, Ferumoxytol. Ferumoxytol is a colloidal iron-carbohydrate complex approved for the parenteral treatment of iron deficiency in chronic kidney disease where oral therapy cannot be performed. Ferumoxytol is administered as an intravenous injection. Ferumoxytol is commercially available as a solution for intravenous injection under the trade name Rienso® or Ferahme®. Iron-carbohydrate complexes exhibit superparamagnetism and can therefore be used for contrast enhancement in MRI studies (off-label) (see, for example, LP Smits et al.: Evaluation of ultrasmall superparamagnetic iron-oxide (USPIO) enhanced MRI with ferumoxytol to quantify arterial wall inflammation, Atherosclerosis 2017, 263:211-218).

[0152] It is equally conceivable to use the representation after administration of the intravascular contrast agent Ablavar® as training data.

[0153] Similarly, it is conceivable to synthetically generate a reference representation showing blood vessels in the examination region with constant contrast enhancement over time, for example by a segmentation method based on the first representation. Segmentation methods are widely described in the literature. The following publications can be mentioned as examples: F. Conversano et al.: Hepatic Vessel Segmentation for 3D Planning of Liver Surgery, Acad Radiol 2011,18:461-470; S. Moccia et al.: Blood vessel segmentation algorithms-Review of methods, datasets and evaluation metrics, Computer Methods and Programs in Biomedicine 158(2018)71-91; M. Marcan et al.: Segmentation of hepatic vessels from MRI images for planning of electroporation-based treatments in the liver, Radiol Oncol 2014;48(3):267-281; ​​TA Hope et al.: Improvement of Gadoxetate Arterial Phase Capture With a High Spatio-Temporal Resolution Multiphase Three-Dimensional SPGR-Dixon Sequence, Journal of Magnetic Resonance Imaging 38:938-945(2013); WO2009 / 135923A1, US6754376B1, WO2014 / 162273A1, WO2017 / 139110A1, WO2007 / 053676A2, EP2750102A1).

[0154] Based on the first representation provided, the trained predictive model then generates a second representation that shows vascular contrast enhancement that is constant over time.

[0155] Example 2 In a preferred embodiment, the present invention is used to generate (predict) one or more artificial MRI images in dynamic contrast-enhanced magnetic resonance imaging.

[0156] In the following text, the term "image" is used. An "image" is a representation in the context of the present invention. An image can be a representation in real space or in frequency space. For training and prediction of a prediction model, a representation in frequency space is always used, i.e., for example, k-space data. However, if a representation in real space is generated based on measurements, they can be transformed, for example by Fourier transformation, into a representation in frequency space before being introduced into the training and / or prediction.

[0157] The examination region is introduced into a basic magnetic field. The examination region is subjected to an MRI method in the course of which a number of MRI images are generated that are indicative of the examination region during a first time span. These MRI images generated based on measurements during the first time span are also referred to herein as first MRI images.

[0158] The term "multiple" means that at least two (first) MRI images are generated, preferably at least three (first) MRI images, very particularly preferably at least four (first) MRI images.

[0159] The subject is administered a contrast agent which spreads over the examination region. The contrast agent is preferably administered intravenously as a bolus (e.g., into an arm vein) using weight-based dosing.

[0160] Preferably, the contrast agent is a hepatobiliary contrast agent, such as, for example, Gd-EOB-DTPA or Gd-BOPTA. In a particularly preferred embodiment, the contrast agent is a substance or mixture of substances which has gadoxetic acid or a salt of gadoxetic acid as the contrast-enhancing active substance. The disodium salt of gadoxetic acid (disodium Gd-EOB-DTPA) is quite particularly preferred.

[0161] The first time span preferably comprises a contrast agent distributed in the examination region.Preferably, the first time span comprises an arterial phase and / or a portal venous phase and / or a late phase in dynamic contrast-enhanced magnetic resonance imaging of the liver or part of the liver of the examination subject.The phases are defined and described, for example, in the following publications: J.Magn.Reson.Imaging,2012,35(3):492-511,doi:10.1002 / jmri.22833;Clujul Medical,2015,Vol.88 no.4:438-448,DOI:10.15386 / cjmed-414;Journal of Hepatology,2019,Vol.71:534?542,http: / / dx.doi.org / 10.1016 / j.jhep.2019.05.005).

[0162] Figure 12 (Fig. 12) shows diagrammatically the time profile of the concentration of the contrast agent in the hepatic artery (A), hepatic vein (V) and healthy liver cells (P) after administration of a hepatic biliary contrast agent in the arm vein of a human. The concentration is shown in the form of signal intensity I in the described areas (hepatic artery, hepatic vein, hepatic cells) in the magnetic resonance measurement as a function of time t. Upon intravenous bolus injection, the concentration of the contrast agent first rises in the hepatic artery (A) (dashed curve). The concentration passes through a maximum value and then decreases. The concentration in the hepatic vein (V) rises more slowly than in the hepatic artery and reaches its maximum value (dotted curve). The concentration of the contrast agent in healthy liver cells (P) rises slowly (continuous curve) and reaches its maximum value only at a very later time point (maximum value is not shown in Fig. 12). Several characteristic time points can be defined: At time point TP0, the contrast agent is administered intravenously as a bolus. At time TP1, the concentration (signal intensity) of the contrast agent in the hepatic artery reaches its maximum value. At time TP2, the curves of the signal intensity of the hepatic artery and the hepatic vein intersect. At time TP3, the concentration (signal intensity) of the contrast agent in the hepatic vein passes its maximum value. At time TP4, the curves of the signal intensity of the hepatic artery and the hepatocytes intersect. At time T5, the concentration in the hepatic artery and the hepatic vein has decreased to a level that no longer causes measurable contrast enhancement.

[0163] In a preferred embodiment, the first time span comprises: (i) showing an examination area that does not contain a contrast agent; (ii) showing the examination region during the arterial phase, in which the contrast agent spreads through the arteries to the examination region; (iii) showing the examination region during the portal venous phase, in which contrast agent enters the examination region via the portal vein, and (iv) showing the examination area during the late phase, in which the concentration of contrast agent in the arteries and veins decreases and the concentration of contrast agent in the hepatic cells increases; A selection is made to generate an MRI image of the liver or part of the liver being examined.

[0164] Preferably, the first time span begins with a time span of 1 minute to 1 second before or with administration of the contrast agent and lasts for a time span of 2 to 15 minutes, preferably 2 to 13 minutes, and even more preferably 3 to 10 minutes from administration of the contrast agent. Because contrast agents undergo very slow renal and biliary excretion, the second time span can last up to 2 hours or more after administration of the contrast agent.

[0165] In a preferred embodiment, the first time span includes at least time points TP0, TP1, TP2, TP3 and TP4.

[0166] In a preferred embodiment, at least MRI images of all the following phases are generated (based on the measurements): in the time span before TP0, in the time span from TP0 to TP1, in the time span from TP1 to TP2, in the time span from TP2 to TP3, and in the time span from TP3 to TP4.

[0167] It is contemplated that one or more MRI images are generated (based on measurements) prior to TP0, during the timespans TP0-TP1, TP1-TP2, TP2-TP3, and TP3-TP4, respectively.

[0168] Based on the (first) MRI image generated (based on the measurements) during the first time span, a second MRI image or a number of second MRI images are predicted showing the examination area during the second time span. The predicted MRI image for the second time span is also referred to herein as the second MRI image.

[0169] In a preferred embodiment of the present invention, the second time span follows the first time span.

[0170] The second time span is preferably a time span within the hepatobiliary phase; preferably a time span beginning at least 10 minutes after administration of the contrast agent, preferably at least 20 minutes after administration of the contrast agent.

[0171] At least one second representation representative of the examination region during a second time span is predicted using a prediction model according to the invention, the prediction model being pre-trained to predict one or more MRI images showing the examination region during a second time span based on a plurality of first MRI images showing the examination region during the first time span.

[0172] The embodiment described here is also shown diagrammatically in FIG. 1(b).

[0173] Example 3 In a further preferred embodiment of the invention, the invention is used to distinguish liver lesions from blood vessels. In T1-weighted MRI images, Primovist® leads to clear signal enhancement in healthy liver parenchyma 10-20 minutes after injection (in the hepatobiliary phase), whereas lesions that do not contain hepatocytes or contain only a few hepatocytes, e.g. metastases or moderately to poorly differentiated hepatocellular carcinoma (HCC), appear as darker areas. However, in the hepatobiliary phase, blood vessels also appear as dark areas, meaning that in MRI images generated during the hepatobiliary phase, it is not possible to distinguish liver lesions from blood vessels based on contrast alone.

[0174] The present invention can be used to generate an artificial MRI image of a subject's liver or part of the liver, in which the contrast between blood vessels and liver cells in the liver has been artificially minimized to make liver pathologies more easily identifiable.

[0175] An "image" is a representation in the context of the present invention. An image can be a representation in real space or in frequency space. For training and prediction of a predictive model, a representation in frequency space is always used. However, a representation in real space can be generated based on measurements and then transformed into a representation in frequency space, for example by Fourier transformation, before being introduced into the training and / or prediction.

[0176] The plurality of first representations includes at least one representation of the examination region in which blood vessels are identifiable, preferably depicted with contrast enhancement by a contrast agent (vascular representation).

[0177] When using paramagnetic contrast agents, blood vessels in such representations are characterized by high signal intensity due to contrast enhancement (high signal representation). Thus, structures (that are continuous) in such representations that have signal intensities within an empirically ascertainable range can be attributed to blood vessels. This means that with such representations, there is information about where blood vessels are depicted in the real space representation, or which structures can be attributed to blood vessels (arteries and / or veins) in the real space representation.

[0178] The plurality of first representations further includes at least one representation of the examination region in which healthy liver cells are depicted with contrast enhancement (hepatocyte representation), for example a representation of the examination region acquired during a hepatobiliary phase.

[0179] Information from at least one vascular representation via blood vessels is combined with information from at least one hepatocyte representation, which includes (artificially) generating (calculating) at least one representation in which contrast differences between structures attributable to blood vessels and structures attributable to healthy hepatocytes are equalized.

[0180] Here, the term "smoothing" means "matching" or "minimizing." The goal of smoothing is to make the boundaries between blood vessels and healthy liver cells disappear in the artificially generated representation, so that blood vessels and healthy liver cells in the artificially generated representation appear as homogenous tissues, whereas liver lesions stand out structurally due to the different contrasts.

[0181] Typically, an artificial representation of 1(number=1) is predicted based on a vascular representation of 1(number=1) and a hepatocyte representation of 1(number=1).

[0182] It is contemplated that in addition to the at least one vascular representation and the at least one hepatocyte representation, the at least one native representation is additionally used to predict the at least one artificial representation.

[0183] In one embodiment, the generation of the artificial representation comprises: - feeding at least one blood vessel representation and at least one hepatocyte representation to a predictive model, where the predictive model has been trained by supervised learning to generate at least one artificial representation from the at least one reference blood vessel representation and the at least one reference hepatocyte representation based on the reference representation, where differences in contrast between structures attributable to blood vessels and structures attributable to healthy hepatocytes are leveled in the at least one artificial representation; - receiving at least one artificial representation as output from the predictive model; Includes.

[0184] Example 4 In a further embodiment, the present invention is used to generate a native MRI image of the liver, where one or more artificial MRI images of the liver or part of the liver of the subject are generated showing the liver or part of the liver without contrast enhancement caused by a contrast agent, where the artificial MRI image(s) are all created based on an MRI image acquired with contrast enhancement caused by a contrast agent.

[0185] An "image" is a representation in the context of the present invention. An image can be a representation in real space or in frequency space. For training and prediction of a predictive model, a representation in frequency space is always used. However, a representation in real space can be generated based on measurements and then transformed into a representation in frequency space, for example by Fourier transformation, before being introduced into training and / or prediction.

[0186] The examination region is introduced into the basic magnetic field. A contrast agent that permeates the examination region is administered to the examination subject. The contrast agent is preferably administered intravenously (e.g., in an arm vein) as a bolus using weight-based dosing. Preferably, the contrast agent is a hepatobiliary contrast agent, such as, for example, Gd-EOB-DTPA or Gd-BOPTA. In a particularly preferred embodiment, the contrast agent is a substance or substance mixture with gadoxetic acid or a salt of gadoxetic acid as the contrast-enhancing active substance. The disodium salt of gadoxetic acid (disodium Gd-EOB-DTPA) is quite particularly preferred.

[0187] A plurality of first representations of the examination region are generated representing the examination region in a first time span after administration of the contrast agent. Preferably, the plurality of first representations are T1 weighted representations.

[0188] Preferably, the plurality of first representations includes at least one representation of the examination region that represents the examination region during a dynamic phase, for example, at least one representation that represents the examination region during an arterial phase, during a venous phase, and / or during a late phase (see, for example, FIG. 12 and related description in Example 2). When using a paramagnetic contrast agent, blood vessels in such representations are characterized by high signal intensity due to contrast enhancement (high signal depiction).

[0189] Preferably, the plurality of first representations further comprises at least one representation of the examination region representative of the examination region during a hepatobiliary phase, during which healthy liver tissue (parenchyma) is visualized with contrast enhancement.

[0190] MRI examinations in the dynamic and hepatobiliary phases are performed over a relatively long time span. Over said time span, movements by the patient should be avoided in order to minimize movement artifacts in the radiological images. Long restrictions on movement can be uncomfortable for the patient. Therefore, shortened MRI procedures are currently being established, in which contrast agent is already administered to the examination subject over a certain time span (i.e. 10-20 minutes) prior to MRI image acquisition, in order to be able to directly acquire MRI images in the hepatobiliary phase. Then, after administration of a second dose of contrast agent, MRI images of the dynamic phase are acquired in the same MRI process. Compared to conventional MRI processes, the residence time of the patient or the examination subject in the MRI is consequently clearly shorter. Therefore, according to the invention, it is preferred to record at least one representation of the liver or parts of the liver in the hepatobiliary phase after a (first) administration of a first contrast agent to the examination subject, and to record at least one further representation of the same liver or parts of the same liver in the dynamic phase after administration of a second contrast agent or a second administration of the first contrast agent to the same examination subject. The first imaging agent is a hepatobiliary, paramagnetic imaging agent. The second imaging agent may be an extracellular, paramagnetic imaging agent.

[0191] The first representation of the examination region is then fed to a prediction model according to the invention. The prediction model is pre-trained to predict one or more second representations showing the liver or part of the liver of the examination subject without contrast enhancement caused by a contrast agent based on the received first representation. The prediction model is preferably created with the aid of a self-learning algorithm in a supervised machine learning process. What is used for learning is training data comprising a plurality of representations of the examination region during the dynamic and hepatobiliary phases of the liver or part of the liver of a number of examination subjects. Furthermore, the training data also comprises representations of the examination region in which there was no contrast enhancement, i.e. generated without administration of a contrast agent.

[0192] The embodiment described here is also shown diagrammatically in FIG. 1(c).

[0193] Example 5 In a further preferred embodiment, the present invention is used to reduce patient examination times in dynamic contrast enhanced magnetic resonance imaging of the liver.

[0194] Here, the contrast agent is administered in the form of two boluses. The first administration is performed at a time when the examination subject is not yet located in the MRI scanner. In the case of the first administration, a first contrast agent is administered. The first contrast agent is preferably administered intravenously (e.g., into the arm vein) as a bolus with a weight-based dosing. The first contrast agent is preferably a hepatobiliary contrast agent, such as, for example, Gd-EOB-DTPA or Gd-BOPTA. In a particularly preferred embodiment, the first contrast agent is a substance or substance mixture with gadoxetic acid or a salt of gadoxetic acid as a contrast-enhancing active substance. The disodium salt of gadoxetic acid (disodium Gd-EOB-DTPA) is very particularly preferred.

[0195] After administration of the first contrast agent, the subject may be introduced into the MRI scanner and a time span may be waited before a first MRI image is generated at a first time point.

[0196] An "image" is a representation in the context of the present invention. An image can be a representation in real space or in frequency space. For training and prediction of a predictive model, a representation in frequency space is always used. However, a representation in real space can be generated based on measurements and then transformed into a representation in frequency space, for example by Fourier transformation, before being introduced into training and / or prediction.

[0197] The time span between the first administration and the generation of the first MRI image is preferably in the range of 5 minutes to 1 hour, even more preferably in the range of 10 minutes to 30 minutes, and most preferably in the range of 8 minutes to 25 minutes.

[0198] The first MRI image represents the liver or a portion of the liver of the subject during a hepatobiliary phase after administration of the first contrast agent. Healthy liver cells are depicted with contrast enhancement in the first MRI image as a result of administration of the first contrast agent.

[0199] The hepatobiliary phase during which the first MRI image is generated is also referred to herein as the first hepatobiliary phase. The first contrast agent has reached healthy liver cells, which, in the case of a paramagnetic contrast agent, results in contrast enhancement, signal enhancement, of healthy liver cells. During the arterial phase, portal venous phase, and the later phase that occurs after administration of the first contrast agent, no MRI images are generated. The arterial phase, portal venous phase, and the later phase that occurs after administration of the first contrast agent are also referred to herein as the first arterial phase, the first portal venous phase, and the first later phase.

[0200] It is conceivable that multiple MRI images are generated during the first hepatobiliary phase.

[0201] After generating one or more first MRI images during the first hepatobiliary phase, a contrast agent is administered a second time. The second contrast agent is a second contrast agent. The second contrast agent can be the same contrast agent as the first contrast agent; however, the second contrast agent can also be a different contrast agent, preferably an extracellular contrast agent. Similarly, the second contrast agent is preferably administered intravenously (e.g., into an arm vein) as a bolus using weight-based dosing.

[0202] The administration of the first contrast agent is also referred to herein as the first administration; the administration of the second contrast agent is also referred to herein as the second administration. If the first contrast agent and the second contrast agent are the same, then what is administered is a first administration of the hepatobiliary contrast agent, and at a later time, a second administration of the hepatobiliary contrast agent. If the first contrast agent and the second contrast agent are different, then what is administered is a first administration of the first contrast agent, said first contrast agent being a hepatobiliary contrast agent, and at a later time, a second administration of the second (different) contrast agent.

[0203] At the time of the second administration (or at the time of the administration of the second contrast agent), the subject is preferably already located in the MRI scanner. After the administration of the second contrast agent, the arterial phase, the portal venous phase and the late phase are again passed through. The arterial phase, the portal venous phase and the late phase are also referred to herein as the second arterial phase, the second portal venous phase and the second late phase. In the second arterial phase and / or the second portal venous phase and / or the second late phase, an MRI image or a number of MRI images are generated. The MRI images are referred to as the second, third, fourth, etc. in the order of acquisition.

[0204] In a preferred embodiment, a second MRI image is generated during the second arterial phase, a third MRI image is generated during the second portal venous phase, and a fourth MRI image is generated during the second late phase. Such second MRI image is particularly representative of arteries with contrast enhancement; such third MRI image is particularly representative of veins with contrast enhancement.

[0205] It is further contemplated that more than one MRI image may be generated during the above-mentioned phases.

[0206] An artificial MRI image can be calculated from MRI images generated during one or more phases after administration of the first and second contrast agents.

[0207] The purpose of generating an artificial MRI image from the measured MRI images is to enhance the contrast between healthy liver tissue and other regions. When using a hepatobiliary paramagnetic contrast agent as the first contrast agent, the signal intensity of the healthy liver tissue during the second arterial phase, the second portal venous phase and the second late phase still rises as a result of the administration of the first contrast agent. The second contrast agent spreading in the second phase described above similarly leads to a high signal in the tissues in which it spreads. This means that there is only low contrast in the MRI image between the healthy liver tissue and the remaining tissues, which are contrast enhanced by the (second) contrast agent. To enhance this contrast, what is generated with the aid of the predictive model is at least one artificial MRI image that shows how the examination area looks in the dynamic phase after the administration of the first contrast agent or how the examination area looks if only the second contrast agent is administered: the blood vessels are depicted with contrast enhancement as a result of the administration of the second contrast agent, whereas the healthy liver cells are not depicted with contrast enhancement as a result of the administration of the first contrast agent. In other words, what is produced is an artificial MRI image that looks like a second MRI image, with the difference that the contrast enhancement of healthy liver cells caused by the administration of the first contrast agent is subtracted (removed) from the second MRI image.

[0208] The embodiment described here is also shown diagrammatically in FIG. 1(d). [Brief description of the drawings]

[0209] [Figure 1(a)] FIG. 1(a) shows a schematic of a typical profile of a radiological inspection. [Figure 1(b)] FIG. 1(b) shows a schematic diagram of a radiological inspection profile according to the present invention. [Figure 1(c)] FIG. 1(c) shows diagrammatically a further profile of a radiological examination according to the invention. [Figure 1(d)] FIG. 1(d) shows diagrammatically a further profile of a radiological examination according to the invention. [Diagram 2]FIG. 2 exemplarily and diagrammatically illustrates the generation of a representation of an examination region in real space and in frequency space. [Diagram 3] FIG. 3 (Fig. 3) shows, diagrammatically and exemplarily, how the representation of the test regions (F1), (F2), and (F3) in frequency space as generated in FIG. 2 can be used to train a predictive model (PM). [Figure 4] Figure 4 (Fig. 4) shows the prediction model (PM) trained in Figure 3. [Diagram 5] FIG. 5 exemplarily illustrates diagrammatically the steps in training a predictive model according to a preferred embodiment of the present invention. [Figure 6] FIG. 6 exemplarily illustrates generally how the predictive model trained in FIG. 5 can be used for prediction. [Figure 7] FIG. 7 illustrates, in a schematic and exemplary manner, one embodiment of a computer system according to the present invention. [Figure 8] FIG. 8 exemplarily illustrates in flow chart form a preferred embodiment of the method according to the invention for training a predictive model. [Figure 9] FIG. 9 exemplarily illustrates in the form of a flow chart a further preferred embodiment of the method according to the invention for training a predictive model. [Figure 10] FIG. 10 exemplarily illustrates in flow chart form a preferred embodiment of a method according to the present invention for predicting one or more representations. [Figure 11] FIG. 11 exemplarily illustrates in flow chart form a further preferred embodiment of a method according to the present invention for predicting one or more representations. [Figure 12] Figure 12 (Fig. 12) shows a schematic of the time profile of the concentration of contrast agent in the hepatic artery (A), hepatic vein (V) and healthy liver cells (P) after administration of hepatic biliary contrast agent into the arm vein of a human.

Claims

1. 1. A computer-implemented method comprising: - receiving a plurality of first representations of an examination region of an object of examination in frequency space, at least some of said first representations representing said examination region during a first time span after administration of a contrast agent; - feeding said plurality of first representations to a predictive model, said predictive model being trained based on first reference representations of said examination area of ​​a number of examination objects, at least some of said first reference representations being trained to generate from said first reference representations representing said examination area during a first time span after administration of a contrast agent in frequency space one or more second reference representations representing said examination area during a second time span in frequency space; - receiving one or more predicted representations of the examination area in frequency space from the prediction model, the one or more predicted representations representing the examination area for a second time span; - transforming said one or more predicted representations into one or more representations of said examination area in real space; - outputting said one or more representations of said examination area in real space; The computer-implemented method comprising:

2. 2. The method of claim 1, wherein the plurality of first representations comprises: - at least one representation of said examination area in frequency space representative of said examination area before said administration of said contrast agent; and at least one representation of said examination area in frequency space representative of said examination area in said first time span after said administration of said contrast agent; and the one or more predicted representations represent the examination area in the second time span, the second time span following the first time span.

3. 2. The method of claim 1, wherein the plurality of first representations includes at least two representations of the examination area in frequency space representing the examination area in the first time span after the administration of the contrast agent, and the one or more predicted representations represent the examination area in the second time span, the second time span preceding the first time span.

4. 2. The method of claim 1, wherein the plurality of first representations comprises: at least one representation of said examination area in frequency space representative of said examination area in said first time span after said administration of a first contrast agent; at least one representation of the examination region in frequency space representing the examination region in the first time span after the administration of a second contrast agent, the first contrast agent and the second contrast agent being the same or different, the second contrast agent being administered after the first contrast agent, and the one or more predicted representations represent the examination area in the second time span, the second time span preceding the first time span.

5. The method of claim 1 , wherein the multiple predicted representations represent the examination area in the second time span with a contrast enhancement that is constant over time.

6. The method according to any of the preceding claims, wherein the predictive model is or comprises an artificial neural network.

7. The method according to any of the preceding claims, wherein the first representation of the examination region in frequency space is k-space data of a magnetic resonance imaging examination.

8. 7. The method according to claim 1, wherein in a first step a plurality of radiographic images of the examination area in real space are received, and these received radiographic images are transformed into the first representation of the examination area in frequency space by a Fourier transform.

9. A method according to any one of claims 1 to 8, comprising: - receiving a plurality of first representations of an examination region of an object of examination in frequency space, at least some of said first representations representing said examination region during a first time span after administration of a contrast agent; - identifying a region in said first representation, said identified region including a centre of said frequency space; - reducing said first representation to said identified region; - feeding a plurality of said reduced first representations to a predictive model; receiving one or more second representations of the examination area in frequency space from the prediction model, the one or more second representations representing the examination area during a second time span; - supplementing said one or more second representations with one or more regions of the received first representation that are outside said identified region; - transforming said one or more captured second representations into one or more representations of said examination area in real space; - outputting said one or more representations of said examination area in real space; The method comprising:

10. The method according to any of the preceding claims, wherein the one or more supplemented second representations are transformed into one or more representations of the examination area in real space by an inverse Fourier transformation.

11. 1. A system comprising: a receiving unit; a control and calculation unit; and ・ Output unit and Equipped with The control and computing unit comprises: - configured to prompt said receiving unit to receive a plurality of first representations of an examination region of an examination object in frequency space, at least some of said first representations representing said examination region during a first time span after administration of a contrast agent; - configured to feed said plurality of first representations to a predictive model, said predictive model being trained based on first reference representations of said examination area of ​​a number of examination objects to generate from said first reference representations, at least some of said first reference representations representing said examination area during a first time span after administration of a contrast agent in frequency space, one or more second reference representations representing said examination area during a second time span in frequency space, - configured to receive one or more predicted representations of the examination area in frequency space from the prediction model, the one or more predicted representations representing the examination area during a second time span; configured to transform said one or more predicted representations into one or more representations of said examination area in real space; - configured to prompt said output unit to output said one or more representations of said examination area in real space; , the system.

12. 1. A computer program product comprising a computer program capable of being loaded into a memory of a computer system, the computer program product comprising: - receiving a plurality of first representations of an examination region of an object of examination in frequency space, at least some of said first representations representing said examination region during a first time span after administration of a contrast agent; - feeding said plurality of first representations to a predictive model, said predictive model being trained based on first reference representations of said examination area of ​​a number of examination objects, at least some of said first reference representations being trained to generate from said first reference representations representing said examination area during a first time span after administration of a contrast agent in frequency space one or more second reference representations representing said examination area during a second time span in frequency space; - receiving one or more predicted representations of the examination area in frequency space from the prediction model, the one or more predicted representations representing the examination area for a second time span; - transforming said one or more predicted representations into one or more representations of said examination area in real space; - outputting one or more representations of said examination area in real space; the product causing the computer system to execute the

13. 1. Use of a contrast agent in a method for predicting at least one radiological image, said method comprising: - administering said contrast agent, said contrast agent spreading over an examination area of ​​an examination subject; - generating a plurality of first representations of the examination region of the object in frequency space, at least some of said first representations representing said examination region during a first time span after administration of the contrast agent; - feeding said plurality of first representations to a predictive model, said predictive model being trained based on first reference representations of said examination area of ​​a number of examination objects, at least some of said first reference representations being trained to generate from said first reference representations representing said examination area during a first time span after administration of a contrast agent in frequency space one or more second reference representations representing said examination area during a second time span in frequency space; - receiving one or more predicted representations of the examination area in frequency space from the prediction model, the one or more predicted representations representing the examination area for a second time span; - transforming said one or more predicted representations into one or more representations of said examination area in real space; - outputting said one or more representations of said examination area in real space; The use comprising:

14. 1. A contrast agent for use in a method for predicting at least one radiological image, said method comprising: - administering said contrast agent, said contrast agent spreading over an examination area of ​​an examination subject, - generating a plurality of first representations of the examination region of the object in frequency space, at least some of said first representations representing said examination region during a first time span after administration of the contrast agent; - feeding said plurality of first representations to a predictive model, said predictive model being trained based on first reference representations of said examination area of ​​a number of examination objects, at least some of said first reference representations being trained to generate from said first reference representations representing said examination area during a first time span after administration of a contrast agent in frequency space one or more second reference representations representing said examination area during a second time span in frequency space; - receiving one or more predicted representations of the examination area in frequency space from the prediction model, the one or more predicted representations representing the examination area for a second time span; - transforming said one or more predicted representations into one or more representations of said examination area in real space; - outputting one or more representations of said examination area in real space; The contrast agent comprising:

15. A kit comprising an imaging agent and the computer program product of claim 12.