Generation of artificial contrast-enhanced radiographic images

Abstract

This invention relates to the technical field for generating artificial contrast-enhanced radiographic images.

Description

Technical Field

[0001] The present disclosure relates to the technical field of generating artificial contrast radiographic images.

Background Art

[0002] WO 2019 / 074938 A1 discloses a method for reducing the amount of contrast agent in the generation of radiographic images with the aid of an artificial neural network.

[0003] In the disclosed method, in a first step, a training dataset is created. The training dataset includes, for each of a number of people, i) an original radiographic image (zero-contrast image), ii) a radiographic image after administration of a small amount of contrast agent (low-contrast image), and iii) a radiographic image after administration of a standard amount of contrast agent (full-contrast image). The standard amount is the amount recommended by the manufacturer and / or wholesaler of the contrast agent, and / or the amount approved by the regulatory authority, and / or the amount specified in the package insert for the contrast agent.

[0004] In a second step, the artificial neural network is trained for each person in the training dataset to predict an artificial radiographic image showing the acquisition region after administration of the standard amount of contrast agent based on the original image and the image after administration of an amount of contrast agent less than the standard amount. The radiographic image measured after administration of the standard amount of contrast agent serves as a reference (ground truth) in the training in any case.

[0005] In a third step, using the trained artificial neural network, based on the original image and the radiographic image after administration of an amount of contrast agent less than the standard amount, an artificial radiographic image showing the acquisition region that would be observed when the standard amount of contrast agent is administered for a new person can be predicted.

[0006] The method disclosed in WO 2019 / 074938 A1 has drawbacks.

[0007] The artificial neural network disclosed in International Publication No. 2019 / 074938A1 is trained to predict radiographic images after administration of a standard dose of contrast agent. The artificial neural network is not configured or trained to predict radiographic images after administration of less or more than the standard dose of contrast agent. The method described in International Publication No. 2019 / 074938A1 can, in principle, be trained to predict radiographic images after administration of doses different from the standard dose of contrast agent, but this requires further training data and further training.

[0008] Medical images generated by trained machine learning models may contain errors (see, for example, K. Schwarz et al.: On the Frequency Bias of Generative Models, https: / / doi.org / 10.48550 / arXiv.2111.02447).

[0009] Because physicians may make diagnoses and / or initiate treatment based on artificial medical images, such errors (artificial products) can be problematic. When scrutinizing artificial medical images, physicians need to know whether the features in the image are due to the real-world features of the subject being examined or are artificial products resulting from errors in predictions by trained machine learning models.

[0010] It is desirable to be able to generate radiographic images with variable contrast enhancement without requiring the generation of training data for each individual contrast enhancement, and without requiring the training of an artificial neural network. It is even more desirable to be able to generate radiographic images with variable contrast enhancement using a traceable deterministic process to generate variable contrast enhancement. This minimizes false negative and false positive results and facilitates the approval and use of corresponding medical procedures. Machine learning methods typically use statistical models with limited generalizability because they are based on a limited selection of training data. It is even more desirable to be able to generate radiographic images with variable contrast enhancement using a wide variety of contrast agents. It is even more desirable to be able to use a method for generating radiographic images with variable contrast enhancement using a wide variety of different contrast agents, regardless of their physical, chemical, physiological, or other properties. It is even more desirable to be able to generate radiographic images with variable contrast enhancement that have less error (artificial product). [Overview of the Initiative]

[0011] These and other objectives are achieved by the subject matter of the independent claims. Preferred embodiments of the present disclosure are found in the dependent claims, this specification and the drawings.

[0012] Therefore, in the first embodiment, this disclosure The first representation represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent. The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion target is used for a large number of criterion targets, i.e., (i) a reference representation generated by the first model that represents the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent, The training process for each standard subject is The reference representation generated by the first model is supplied to the second model. Receiving the modified reference representation from the aforementioned second model, The process includes reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters. To supply the third representation to the second model, Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, Outputting and / or storing the modified third representation, and / or transmitting the modified third representation to another computer system. The present invention provides a computer-aided method for generating a composite contrast-enhanced radiographic image, which includes the following steps.

[0013] This disclosure further states that Processor, and A computer system is provided, which includes a storage medium for storing an application program configured to perform an operation when executed by the aforementioned processor, and the operation is, The first representation represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent. The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion target is used for a large number of criterion targets, i.e., (i) a reference representation generated by the first model that represents the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent, The training process for each standard subject is The reference representation generated by the first model is supplied to the second model. Receiving the modified reference representation from the aforementioned second model, Supplying the third representation to the second model, including the step of reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters, Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, The process includes outputting and / or storing the modified third representation, and / or transmitting the modified third representation to another computer system.

[0014] This disclosure further provides a computer program that can be loaded into the working memory of a computer system, the computer program being used in the computer system The first representation represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent. The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion target is used for a large number of criterion targets, i.e., (i) a reference representation generated by the first model that represents the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent, The training process for each standard subject is The reference representation generated by the first model is supplied to the second model. Receiving the modified reference representation from the aforementioned second model, Supplying the third representation to the second model, including the step of reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters, Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, The process involves outputting and / or storing the modified third representation, and / or transmitting the modified third representation to another computer system.

[0015] This disclosure further states that To provide a first representation that represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent, The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion target is used for a large number of criterion targets, i.e., (i) a reference representation generated by the first model that represents the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent, The training process for each standard subject is The reference representation generated by the first model is supplied to the second model. Receiving the modified reference representation from the aforementioned second model, Supplying the third representation to the second model, including the step of reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters, Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, Outputting and / or storing the modified third representation, and / or transmitting the modified third representation to another computer system. The present invention provides the use of the contrast agent in a radiological examination method that includes the step of [the specified step].

[0016] This disclosure further states that To provide a first representation that represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent, The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion target is used for a large number of criterion targets, i.e., (i) a reference representation generated by the first model that represents the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent, The training process for each standard subject is The reference representation generated by the first model is supplied to the second model. Receiving the modified reference representation from the aforementioned second model, Supplying the third representation to the second model, including the step of reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters, Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, Outputting and / or storing the modified third representation, and / or transmitting the modified third representation to another computer system. The present invention provides a contrast agent for use in radiological examination methods that include the following steps.

[0017] This disclosure further provides a computer program product and a kit including a contrast agent, wherein the computer program product includes a computer program that can be loaded into the working memory of a computer system, and the computer program is installed on the computer system. To provide a first representation that represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent, The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion target is used for a large number of criterion targets, i.e., (i) a reference representation generated by the first model that represents the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent, The training process for each standard subject is The reference representation generated by the first model is supplied to the second model. Receiving the modified reference representation from the aforementioned second model, Supplying the third representation to the second model, including the step of reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters, Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, The process involves outputting and / or storing the modified third representation, and / or transmitting the modified third representation to another computer system. [Brief explanation of the drawing]

[0018] [Figure 1] As an example, in a schematic form, we demonstrate the generation of a third representation of the test domain under test, based on the first and second representations, with the help of the first model. [Figure 2] Further schematic examples are provided of generating a third representation of the test domain based on the first and second representations of the test domain, with the help of the first model. [Figure 3] As an example, the process for training a second model in a schematic form is shown. [Figure 4] As an example, we demonstrate the generation of a modified third representation based on a third representation, with the help of a trained second model, in a schematic form. [Figure 5] As an example, a model including the first model and the second model is shown in a schematic form. [Figure 6] As an example, a computer system relating to this disclosure is shown in a schematic form. [Figure 7] As an example, further embodiments of the computer system relating to this disclosure are shown in a schematic form. [Figure 8] As an example, in a schematic form, one embodiment of a computer implementation of the present disclosure in the form of a flowchart is shown. [Modes for carrying out the invention]

[0019] The subject matter of this disclosure will be further elaborated below without distinction of subject matter (methods, computer systems, computer programs (products), uses, contrast agents for use, kits). Rather, the subsequent clarification is intended to apply equally to all subject matter, regardless of the context in which the subject matter appears (methods, computer systems, computer programs (products), uses, contrast agents for use, kits).

[0020] Where steps are described in a particular order within this specification or in the claims, this does not necessarily mean that the disclosure is limited to that order. Instead, steps may be performed in a different order or in parallel with one another, with the exception that one step builds upon another, so that the step built upon the previous step must be performed next (although this will become clearer in the individual cases). Thus, the order described constitutes a preferred embodiment.

[0021] In some parts, the present invention will be illustrated in more detail with reference to the drawings. The drawings show specific embodiments having certain features and combinations of features that are intended to be used primarily for illustrative purposes, and the present invention should not be understood as being limited to the features and combinations of features shown in the drawings. Furthermore, the descriptions made in the drawings relating to features and combinations of features are intended to be generally applicable, i.e., transferable to other embodiments, and are not limited to the embodiments shown.

[0022] The disclosure includes means for generating one or more artificial radiographic images based on at least two representations of an area under examination after the addition / administration / use of various amounts of contrast agent, wherein the contrast can be varied between areas with and without the contrast agent.

[0023] The "subject of examination" is usually a living organism, preferably a mammal, and most preferably a human.

[0024] The "examination area" is a part of the object being examined, for example, the organ being examined, or a part of an organ, or multiple organs or another part of the organ being examined.

[0025] For example, the examination area may be the liver, kidneys, heart, lungs, brain, stomach, bladder, prostate, intestines, or a part of any of the aforementioned parts, or another part of the body of a mammal (e.g., a human).

[0026] In one embodiment, the examination area includes the liver or a portion of the liver, or the examination area is the liver or a portion of the liver of a mammal, preferably a human.

[0027] In further embodiments, the examination area includes the brain or a portion of the brain, or the examination area is the brain or a portion of the brain of a mammal, preferably a human.

[0028] In further embodiments, the examination area includes the heart or a portion of the heart, or the examination area is the heart or a portion of the heart of a mammal, preferably a human.

[0029] In further embodiments, the examination area includes or is a portion of the rib cage of a mammal, preferably a human.

[0030] In further embodiments, the examination area includes the stomach or a portion of the stomach, or the examination area is the stomach or a portion of the stomach of a mammal, preferably a human.

[0031] In further embodiments, the examination area includes the pancreas or a portion of the pancreas, or the examination area is the pancreas or a portion of the pancreas of a mammal, preferably a human.

[0032] In further embodiments, the examination area includes a kidney or a portion of a kidney, or the examination area is a mammalian, preferably human, kidney or a portion of a kidney.

[0033] In a further embodiment, the examination area includes one or both lungs or a portion of the lungs of a mammal, preferably a human.

[0034] In further embodiments, the examination area includes a breast or a portion of a breast, or the examination area is a breast or a portion of a breast of a female mammal, preferably a female human.

[0035] In further embodiments, the examination area includes the prostate or a portion of the prostate, or the examination area is the prostate or a portion of the prostate of a male mammal, preferably a male human.

[0036] The examination area, also known as the field of view (FOV), is the volume captured in a radiographic image. Typically, the examination area is defined by a radiologist, for example, on a localizer image. Of course, the examination area can also be defined alternatively or additionally in an automated manner, for example, based on a selected protocol.

[0037] The area to be examined is the area to undergo radiation examination.

[0038] "Radiology" is the field of medicine (including, for example, ultrasound) that uses electromagnetic and mechanical waves for diagnostic, therapeutic, and / or scientific purposes. In addition to X-rays, other ionizing radiation such as gamma radiation and electrons is also used. Because imaging is an important application, other imaging methods such as ultrasound and magnetic resonance imaging (nuclear magnetic resonance imaging) are also considered radiology, even if they do not use ionizing radiation. Therefore, in the context of this disclosure, the term "radiology" particularly includes the following imaging methods: computed tomography, magnetic resonance imaging, and ultrasound.

[0039] In one embodiment of this disclosure, the radiological examination is magnetic resonance imaging.

[0040] In a further embodiment, the radiological examination is computed tomography.

[0041] In a further embodiment, the radiological examination is an ultrasound examination.

[0042] In radiological examinations, contrast agents are commonly used to enhance contrast.

[0043] A "contrast agent" is a substance or mixture of substances that enhances the depiction of the structure and function of the body in radiological examinations.

[0044] In computed tomography, iodine-containing solutions are typically used as contrast agents. In magnetic resonance imaging (MRI), superparamagnetic materials (e.g., iron oxide nanoparticles, superparamagnetic iron-platinum particles (SIPP)) or paramagnetic materials (e.g., gadolinium chelate, manganese chelate, hafnium chelate) are typically used as contrast agents. In ultrasound examinations, liquids containing gas-filled microbubbles are typically administered intravenously.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 Nouh et al.: Radiographic and magnetic resonances contrast agents: Essentials and tips for safe practices, World J Radiol. 2017 Sep. 28; 9(9): 339-349; LC Abonyi et al.: Intravascular Contrast Media in Radiography: Historical Development & Review of Risk Factors for 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).

[0045] MRI contrast agents exert their effects in MRI examinations by altering the relaxation time of structures that take up the contrast agent. Two groups of materials can be distinguished: paramagnetic materials and superparamagnetic materials. Both groups of materials have unpaired electrons that induce a magnetic field around individual atoms or molecules. Superparamagnetic contrast agents mainly shorten T2, while paramagnetic contrast agents mainly shorten T1. The effect of contrast agents is indirect, as they do not emit a signal themselves, but merely affect the intensity of signals in their vicinity. An example of a superparamagnetic contrast agent is iron oxide nanoparticles (SPIO, superparamagnetic iron oxide). Examples of paramagnetic contrast agents include gadopentetate meglumine (trade name: Magnevist®, etc.), gadoteric acid (Dotarem®, Dotagita®, Cyclolux®), gadodiamide (Omniscan®), gadoteridol (ProHance®), gadobutrol (Gadovist®), gadopicrenol (Elucirem, Vueway), and gadoxetic acid (Primovist® / Eovist®), which are gadolinium chelates.

[0046] In one embodiment, the radiological examination is an MRI examination in which an MRI contrast agent is used.

[0047] In a further embodiment, the radiological examination is a CT examination in which a CT contrast agent is used.

[0048] In a further embodiment, the radiological examination is a CT scan in which an MRI contrast agent is used.

[0049] The generation of artificial radiographic images with variable contrast enhancement is based on at least two representations of the examination area, namely a first representation and a second representation.

[0050] The first expression represents the examination area without contrast agent, or after administration of a first amount of contrast agent. Preferably, the first expression represents the examination area without contrast agent (the original expression).

[0051] The second expression represents the examination area after administration of a second amount of contrast agent. The second amount is greater than the first amount (the first amount may be zero, as stated). The expression "after administration of a second amount of contrast agent" should not be understood as meaning that the first and second amounts are added together in the examination area. Therefore, the expression "the expression represents the examination area after administration of (the first or second) amount" should rather be understood as "the expression represents the examination area by (the first or second) amount" or "the expression represents the examination area containing (the first or second) amount."

[0052] In one embodiment, both the first and second amounts of contrast agent are less than the standard amount.

[0053] In a further embodiment, the second amount of contrast agent corresponds to a standard amount.

[0054] Furthermore, in this embodiment, the first amount of contrast agent is equal to zero, and the second amount of contrast agent is less than the standard amount.

[0055] In a further embodiment, the first amount of contrast agent is equal to zero, and the second amount of contrast agent corresponds to a standard amount.

[0056] The standard dose is typically the amount recommended by the manufacturer and / or wholesaler of the contrast agent, as well as the amount approved by the regulatory authority, and / or the amount specified in the package insert for the contrast agent.

[0057] For example, the standard dose of Primovist® is 0.025 mmol of Gd-EOB-DTPA disodium per kg of body weight.

[0058] In one embodiment of the present disclosure, the contrast agent is an agent comprising gadolinium(III) 2-[4,7,10-tris(carboxymethyl)-1,4,7,10-tetrazacyclododec-1-yl]acetic acid (also known as gadolinium-DOTA or gadoteric acid).

[0059] In further embodiments, the contrast agent is an agent containing gadolinium(III) ethoxybenzyldiethylenetriaminepentaacetic acid (Gd-EOB-DTPA), and preferably the contrast agent contains the disodium salt of gadolinium(III) ethoxybenzyldiethylenetriaminepentaacetic acid (also known as gadoxetic acid).

[0060] In one embodiment of the present disclosure, the contrast agent is an agent comprising gadolinium(III)2-[3,9-bis[1-carboxylat-4-(2,3-dihydroxypropylamino)-4-oxobutyl]-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-trien-6-yl]-5-(2,3-dihydroxypropylamino)-5-oxopentanoate (also known as gadopicrenol) (see, for example, International Publication No. 2007 / 042504 and International Publication No. 2020 / 030618 and / or International Publication No. 2022 / 013454).

[0061] In one embodiment of the present disclosure, the contrast agent is a compound containing dihydrogen[(±)-4-carboxy-5,8,11-tris(carboxymethyl)-1-phenyl-2-oxa-5,8,11-triazatridecane-13-oate(5-)]gadolinate(2-) (also known as gadobenic acid).

[0062] In one embodiment of the present disclosure, the contrast agent is an agent comprising tetragadolinium[4,10-bis(carboxylatomethyl)-7-{3,6,12,15-tetraoxo-16-[4,7,10-tris-(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]-9,9-bis({[({2-[4,7,10-tris-(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]propanoyl}aminoacetyl]amino}methyl)-4,7,11,14-tetraazaheptadecan-2-yl}-1,4,7,10-tetraazacyclododecane-1-yl]acetate (also known as gadoquatran) (e.g., J. Lohrke et al.: Preclinical Profile of Gadoquatrane: A Novel Tetrameric, Macrocyclic High Relaxivity Gadolinium-Based Contrast Agent. Invest Radiol., 2022, 1, 57(10): 629-638; see International Publication No. 2016193190).

[0063] In one embodiment of the present disclosure, the contrast agent is a compound of formula (I) Gd 3+ complex [ka] (I) In the formula, Ar is [ka] and [ka] It is a base selected from, # is a concatenation to X, X is a group selected from CH2, (CH2)2, (CH2)3, (CH2)4 and *-(CH2)2-O-CH2-#, Here, * represents linkage to Ar, and # represents linkage to an acetate residue. R 1 , R2 and R 3 is, independently of one another, a hydrogen atom or a group selected from C1-C3 alkyl, -CH2OH, -(CH2)2OH and -CH2OCH3, R 4 is a group selected from C2-C4 alkoxy, (H3C-CH2)-O-(CH2)2-O-, (H3C-CH2)-O-(CH2)2-O-(CH2)2-O- and (H3C-CH2)-O-(CH2)2-O-(CH2)2-O-(CH2)2-O-, R 5 is a hydrogen atom, and R 6 is a hydrogen atom, or a stereoisomer, tautomer, hydrate, solvate or salt thereof, or a mixture thereof is an agent comprising the same.

[0064] In one embodiment of the present disclosure, the contrast agent is a Gd complex of a compound of formula (II) 3+ complex

Chemical formula

Chemical formula

Chemical formula

[0065] The term "C1-C3 alkyl" refers to a linear or branched saturated monovalent hydrocarbon group having one, two, or three carbon atoms, such as methyl, ethyl, n-propyl, or isopropyl. The term "C2-C4 alkyl" refers to a linear or branched saturated monovalent hydrocarbon group having two, three, or four carbon atoms.

[0066] The term "C2-C4 alkoxy" refers to a linear or branched saturated monovalent group of the formula (C2-C4 alkyl)-O-, where "C2-C4 alkyl" is as defined above, for example, a methoxy group, an ethoxy group, an n-propoxy group, or an isopropoxy group.

[0067] In one embodiment of the present disclosure, the contrast agent is a compound comprising gadolinium 2,2',2''-(10-{1-carboxy-2-[2-(4-ethoxyphenyl)ethoxy]ethyl}-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate (see, for example, International Publication No. 2022 / 194777, Example 1).

[0068] In one embodiment of the present disclosure, the contrast agent is a compound comprising gadolinium 2,2',2''-{10-[1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate (see, for example, International Publication No. 2022 / 194777, Example 2).

[0069] In one embodiment of the present disclosure, the contrast agent is a compound comprising gadolinium 2,2',2''-{10-[(1R)-1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate (see, for example, International Publication No. 2022 / 194777, Example 4).

[0070] In one embodiment of the present disclosure, the contrast agent is an agent comprising gadolinium(2S,2'S,2''S)-2,2',2''-{10-[(1S)-1-carboxy-4-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}butyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}tris(3-hydroxypropanoate) (see, for example, International Publication No. 2022 / 194777, Example 15).

[0071] In one embodiment of the present disclosure, the contrast agent is a compound comprising gadolinium 2,2',2''-{10-[(1S)-4-(4-butoxyphenyl)-1-carboxybutyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate (see, for example, International Publication No. 2022 / 194777, Example 31).

[0072] In one embodiment of the present disclosure, the contrast agent is a compound comprising gadolinium 2,2',2''-{(2S)-10-(carboxymethyl)-2-[4-(2-ethoxyethoxy)benzyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate.

[0073] In one embodiment of the present disclosure, the contrast agent is a compound comprising gadolinium 2,2',2''-[10-(carboxymethyl)-2-(4-ethoxybenzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl]triacetate.

[0074] In one embodiment of the present disclosure, the contrast agent is an agent comprising gadolinium(III)5,8-bis(carboxylatomethyl)-2-[2-(methylamino)-2-oxoethyl]-10-oxo-2,5,8,11-tetraazadodecane-1-carboxylate hydrate (also known as gadodiamide).

[0075] In one embodiment of the present disclosure, the contrast agent is an agent comprising gadoteridol (also known as gadoteridol), which contains gadolinium(III) 2-[4-(2-hydroxypropyl)-7,10-bis(2-oxido-2-oxoethyl)-1,4,7,10-tetrazacyclododec-1-yl]acetate.

[0076] In one embodiment of the present disclosure, the contrast agent is an agent comprising gadolinium(III)2,2',2''-(10-((2R,3S)-1,3,4-trihydroxybutan-2-yl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl) triacetate (also referred to as gadobutrol or Gd-DO3A-butrol).

[0077] In the first step, the first expression and the second expression are provided, i.e., received or generated.

[0078] The term “receive” encompasses both the acquisition of an expression and the acceptance of an expression transmitted to, for example, the computer system of this disclosure. Expressions may be received from a computed tomography system, a magnetic resonance imaging system, or an ultrasound scanner. Radiographic images may be read from one or more data storage media and / or transmitted from another computer system.

[0079] The term “generate” preferably means that a representation is generated based on another (e.g., received) representation, or based on more than one other (e.g., received) representation. For example, the received representation may be a representation of the test region under test in real space. Based on this real-space representation, it is possible to generate a representation of the test region under test in frequency space by, for example, a transform operation (e.g., a Fourier transform). Further options for generating a representation based on one or more other representations are described herein.

[0080] The representation of the inspection area may be a representation in real space (image space), a representation in frequency space, a representation in projection space, or a representation in another space, for the purposes of this disclosure.

[0081] In a real-space representation, also referred to herein as a real-space depiction or real-space representation, the examination area is typically represented by a number of image elements (e.g., pixels or voxels or doxels), which may be, for example, a raster array, where each image element represents a portion of the examination area, and each image element may be assigned a color value or gray value. The color value or gray value represents the intensity of the signal, for example, the attenuation of X-rays. The DICOM format is a widely used format in radiology for storing and processing real-space representations. DICOM (Digital Images and Communications in Medicine) is an open standard for storing and exchanging information in medical image data management.

[0082] In the frequency-space representation, also referred to herein as frequency-space depiction or frequency-space representation, the test region is represented by a superposition of fundamental vibrations. For example, the test region may be represented by the sum of sine and cosine functions having different amplitudes, frequencies, and phases. The amplitude and phase may be plotted as functions of frequency, for example, in a two-dimensional or three-dimensional representation. Typically, the lowest frequency (origin) is located at the center. The further away from this center, the higher the frequency. Each frequency can be assigned an amplitude representing the frequency in the frequency-space depiction and a phase indicating the degree of deviation of each vibration from the sine or cosine vibration.

[0083] A representation in real space can be converted (transformed) into a representation in frequency space, for example, by a Fourier transform. Conversely, a representation in frequency space can be converted (transformed) into a representation in real space, for example, by an inverse Fourier transform.

[0084] Details on real-space and frequency-space descriptions, as well as their mutual conversions, are described in numerous publications; for example, see https: / / see.stanford.edu / materials / lsoftaee261 / book-fall-07.pdf.

[0085] The representation of the examination area in projection space is typically the result of computed tomography (CT) before image reconstruction. The projection space depiction can be understood as representing the raw data from the CT scan. In CT, the intensity or attenuation of X-rays is measured as they pass through the examination object. From this, projection values ​​can be calculated. In the second step, the object information encoded by the projection is converted into an image (real-space depiction) by computer-aided reconstruction. Reconstruction can be performed by Radon transform. The Radon transform describes the link between the unknown examination object and its associated projection.

[0086] For further details on the conversion of projection data to real-space representations, see numerous publications, for example, K. Fang: The Radon Transformation and Its Application in Tomography, Journal of Physics Conference Series 1903(1):012066.

[0087] The representation of the inspection region may also be a representation in Hough space. For the recognition of geometric objects in an image, after edge detection, a dual space is created in which all possible parameters of the geometric object are input for each point in the image where the edges are located; this is known as the Hough transform. Thus, each point in the dual space corresponds to the geometric object in image space. For a line, this may be, for example, the slope and y-intercept of the line, and for a circle, this may be the center and radius of the circle. Details about the Hough transform can be found in the literature (see, for example, AS Hassanein et al.: A Survey on Hough Transform, Theory, Techniques and Applications, arXiv:1502.02160v1).

[0088] There are further spaces in which representations of the inspection area may exist. For the sake of simplification and clarity, the present invention is described based on real-space representations for the majority of the description. However, this should not be understood as limiting. Those skilled in the art of image analysis will know how to apply appropriate parts of the description to representations other than real-space representations.

[0089] The first and second representations are supplied to the first model.

[0090] After jointly registering the first and second representations, it is possible to supply them to the first model. "Joint registration" (also known as "image registration" in the prior art) is used to make two or more real-space depictions of the same inspection region fit each other to the best possible state. One of the real-space depictions is defined as the reference image, and the other is called the target image. A compensatory transformation is calculated to best fit the target image to the reference image.

[0091] It is also possible to jointly register representations in frequency space. Here, it should be noted that translational movement in real space constitutes an additive linear phase ramp in frequency space. On the other hand, scaling and rotation are preserved in the Fourier transform and inverse Fourier transform. That is, scaling and rotation in frequency space are also scaling and rotation in real space (for example, S. Skare: Rigid Body Image Realignment in Image Space vs. k-Space, ISMRM Scientific Workshop on Motion Correction, 2014, https: / / cds.ismrm.org / protected / Motion_14 / Program / Syllabus / Skare.pdf).

[0092] The first model is configured to generate a third representation based on the first and second representations. The third representation represents the examination area after administration of a third amount of contrast agent.

[0093] The third amount of contrast agent is different from the first and second amounts. Preferably, the third amount of contrast agent is greater than the second amount, but it may also be less than the second amount.

[0094] For example, if the second quantity is less than the standard quantity, the third quantity may be equal to the standard quantity. However, the third quantity may also be greater than the standard quantity.

[0095] If the third quantity is less than the second quantity, the first model becomes the amount of contrast attenuation. This means that the contrast enhancement produced by the contrast agent in the second representation is attenuated (becomes less clear) in the third representation.

[0096] If the third quantity is greater than the second quantity, the first model results in contrast enhancement. This means that the contrast enhancement caused by the contrast agent in the second representation is amplified (becomes clearer) in the third representation.

[0097] The first model may be a machine learning model.

[0098] The term "machine learning model" can be understood as meaning a computer-implemented data processing architecture. Such a model is capable of receiving input data and supplying output data based on the input data and model parameters. Such a model is capable of learning the relationship between input and output data through training. During training, the model parameters may be adjusted to supply a desired output for a particular input.

[0099] During the training of such a model, the model is presented with training data that the model can learn from. The trained machine learning model is the result of the training process. In addition to the input data, the training data includes the exact output data (target data) that the model is intended to generate based on the input data. During training, patterns are identified that map the input data onto the target data.

[0100] During the training process, training data is input to the model, and the model generates output data. The output data is compared to target data. Model parameters are changed to reduce the difference between the output data and the target data to a (defined) minimum. Optimization methods such as gradient descent can be used to modify the model parameters to reduce the difference.

[0101] The difference can be quantified with the help of a loss function. This type of loss function can be used to calculate loss values ​​for a given pair of output and target data. The objective of the training process may also consist of changing (adjusting) the parameters of the machine learning model so that the loss values ​​for all pairs of the training dataset are reduced to a (defined) minimum.

[0102] For example, if the output data and target data are numbers, the loss function may be the absolute difference between these numbers. In this case, a high absolute loss value may mean that one or more model parameters need to be changed to a substantial extent.

[0103] For example, for output data in vector form, the loss function could be any of the following: a difference metric between vectors such as mean squared error, a norm of the difference vector such as cosine distance or Euclidean distance, Chebyshev distance, Lp norm of the difference vector, weighted norm, or any other type of difference metric between two vectors.

[0104] For higher-dimensional outputs, such as two-dimensional outputs or outputs of three dimensions or more, element-wise difference metrics may be used, for example. Alternatively or additionally, the output data may be converted to a one-dimensional vector, for example, before calculating the loss value.

[0105] The first model is, for example, the following publications, namely, International Publication No. 2019 / 074938A1, International Publication No. 2022 / 253687A1, International Publication No. 2022 / 207443A1, International Publication No. 2022 / 223383A1, International Publication No. 20227184297A1, International Publication No. 2022 / 179896A2, International Publication No. 2021 / 069338A1, European Patent Publication No. The machine learning model may be as described in one of the following publications: Patent Application Publication No. 22209510.1, European Patent Application Publication No. 23159288.2, PCT / European Patent Application Publication No. 2023 / 053324, PCT / European Patent Application Publication No. 2023 / 050207, Chinese Patent Application No. 110852993A, Chinese Patent Application No. 110853738A, US Patent Application Publication No. 2021150671A1, arXiv:2303.15938v1, doi:10.1093 / jrr / rrz030.

[0106] The first model may be a mechanistic (deterministic) model. Mechanistic models are based on fundamental principles and known relationships within a system. They often derive from scientific theories and domain-specific knowledge. Mechanistic models describe the underlying mechanisms of a system using mathematical formulas or physical laws. They aim to simulate the behavior of a system based on an understanding of its components and interactions.

[0107] On the other hand, machine learning models are data-driven, learned patterns and relationships derived from input data, without explicitly programming those relationships.

[0108] Therefore, while mechanism models are based on fundamental principles and aim to represent the underlying system mechanisms, machine learning models directly learn patterns and relationships from data without explicitly programming these relationships.

[0109] Therefore, the mechanism model is based on physical laws. In radiographic examinations, the signal produced by a contrast agent is usually dependent on the amount (e.g., concentration) of the contrast agent in the examination area. For example, the signal intensity may show a primary dependence or other form of dependence on the concentration of the contrast agent in the examination area across a defined concentration range. The functional dependence of signal intensity on concentration can be used to construct a mechanism model.

[0110] In one embodiment, the first model is a mechanism model in which, in the first step, the signal intensity distribution generated by the contrast agent in the examination area is determined based on the first and second representations, and in the second step, this α multiplier is added to the first or second representation, where α is the gain coefficient.

[0111] The generation of the signal intensity distribution generated by the contrast agent in the examination area may include, for example, subtraction of the first representation from the second representation. If the first representation represents the examination area of ​​the subject without contrast agent and the second representation represents the examination area of ​​the subject with contrast agent, subtracting the first representation from the second representation generates a representation of the examination area in which the signal intensity distribution is generated by the contrast agent alone. This is because signals not generated by the contrast agent are the same in both the first and second representations and are removed by the subtraction.

[0112] When this signal intensity distribution is added once (α=1) to the first representation, the second representation is obtained again.

[0113] When this signal intensity distribution is added to the first representation multiple times (α>1), a third representation of the examination region is obtained in which the contrast between the contrast-enhanced and non-contrast-enhanced regions is enhanced compared to the second representation. α can be an integer, or other real values ​​(e.g., 1.5 or 3.1416, or other values).

[0114] When the fractional part (0>α>1) of this signal intensity distribution is added to the first representation, a third representation of the examination region is obtained in which the contrast between the contrast-enhanced and non-contrast-enhanced regions is attenuated compared to the second representation.

[0115] Negative alpha values ​​are also possible; for example, a negative alpha value can be selected so that the portion of the investigation area where signal enhancement induced by the contrast agent occurs in the measurement-generated representation becomes completely dark (black) in the artificially generated representation.

[0116] Therefore, the gain coefficient α is a positive or negative real number. The gain coefficient α may be selected by the user, i.e., variable, or it may be predefined, i.e., predetermined. The gain coefficient α may be determined automatically, for example, based on the histograms of the first representation and / or the second representation, and / or based on the difference between the first representation and the second representation, and / or with the help of the initial model (see below).

[0117] Therefore, by changing the gain coefficient α, the contrast between the area with contrast agent and the area without contrast agent can be changed.

[0118] Therefore, with the help of the first model, it is possible to generate a third representation of the test area to be tested, representing the test area after administration of a third amount, based on the first and second representations, where the third amount may be different from the first and second amounts.

[0119] With the help of the first model, it is possible to generate a representation of the examination area with a larger-than-standard amount of contrast agent, based on a first representation of the examination area with no contrast agent or with a first amount of contrast agent, and a second representation of the examination area with a second amount of contrast agent that is less than or equal to the standard amount.

[0120] The above mechanism model is based on the assumption that the signal intensity, represented by gray or color values ​​in the representation of the examination area, exhibits a first-order dependency on the amount of contrast agent administered. This is particularly true in many MRI examinations. First-order dependency allows for changes in contrast by changing the gain coefficient α. Thus, a gain coefficient of α=2 means that a third amount of contrast agent is equivalent to twice the second amount.

[0121] The above mechanism model is disclosed in European Patent Publication No. 22207079.9, European Patent Publication No. 22207080.7, and European Patent Publication No. 23168725.2.

[0122] It should be noted that the mechanism model may be based on other dependencies instead of primary dependencies. These dependencies may be determined experimentally.

[0123] Figure 1 shows, as an example, the generation of a third representation of the test area under test based on the first and second representations, with the help of the first model, in a schematic form.

[0124] The subject of the test is pigs, and the test area includes the pig's liver.

[0125] The first representation, R1, is a magnetic resonance image representing the examination area in real space without contrast agent.

[0126] The second representation R2 represents the same examination area of ​​the same object being examined as the first representation R1 in real space. The second representation R2 is similarly a magnetic resonance image.

[0127] The second expression R2 represents the examination area after administration of a second amount of contrast agent. In this example, 25 μmol of hepatobiliary contrast agent per kg of body weight was administered intravenously to the subject. The second expression R2 represents the examination area in the so-called arterial phase (see, for example, DOI:10.1002 / jmri.22200).

[0128] Hepatobiliary contrast agents have the characteristic properties of being specifically taken up by liver cells (hepatocytes), accumulating in functional tissue (parenchyma), and enhancing contrast in healthy liver tissue. An example of a hepatobiliary contrast agent is disodium gadoxetate (Gd-EOB-DTPA disodium), which is described in U.S. Patent No. 6039931A and is commercially available under the trade names Primovist® and Eovist®. Further hepatobiliary contrast agents are described, in particular, in International Publication No. 2022 / 194777.

[0129] The first representation R1 and the second representation R2 are supplied to the first model M1. Based on the first representation R1 and the second representation R2, the first model M1 generates a third representation R3. In the example shown in Figure 1, the generation of the third representation R3 includes subtracting the first representation R1 from the second representation R2 (R2-R1). The generation of the third representation R3 further includes multiplying the difference between the second representation and the first representation by α and adding it to the first representation (R3=R1+α·(R2-R1)).

[0130] If negative gray / color values ​​result when subtracting the first representation R1 from the second representation R2, these negative values ​​can be avoided by setting them to zero (or another value).

[0131] The difference (R2-R1) represents the contrast enhancement (signal intensity distribution) generated in the examination area by the second amount of contrast agent.

[0132] The difference (R2-R1) is multiplied by the gain coefficient α, and the result of the multiplication is added to the first representation R1. This generates the third representation R3. In the example shown in Figure 1, the gain coefficient α = 3, that is, the difference (R2-R1) is added to the first representation R1 three times.

[0133] The third representation R3 may be normalized, that is, the gray / color values ​​can be multiplied by a coefficient such that the highest gray / color value is represented by, for example, gray tone / hue "white", and the lowest gray / color value is represented by, for example, gray tone / hue "black".

[0134] Therefore, in the example shown in Figure 1, the first model M1 consists of mathematical operations that perform subtraction, multiplication, and addition based on the gray value / color value of individual image elements (e.g., pixels, voxels).

[0135] Figure 2 schematically illustrates a further example of generating a third representation of the test domain based on the first and second representations of the test domain, with the help of the first model.

[0136] Figure 2 shows the inspection area of ​​the object being inspected in each form of representation.

[0137] First expression R1 I R1 represents the examination area in real space without contrast agent or after administration of a first amount of contrast agent. The examination area shown in Figure 2 includes a pig's liver. First representation R1 I This is a magnetic resonance image.

[0138] First real-space representation R1 I This is obtained by a transformation operation T, for example, the Fourier transform, which gives R1 the first representation of the inspection domain in frequency space. F It can be converted to the first frequency space representation R1 F Similarly, the first real-space representation R1 without contrast agent or after administration of a first amount of contrast agent. I This represents the same test subject and the same test area.

[0139] First frequency space representation R1 F This is the conversion operation T -1 For example, by the inverse Fourier transform, the first real-space representation R1 I It can be converted to. Conversion operation T -1 This is the inverse transform of the transformation operation T.

[0140] Second expression R2 I In real space, the first representation R1 I This represents the same test area as the same test subject. Second real-space representation R2 I This represents the examination area after administration of the second volume of contrast agent. The second volume is greater than the first volume (the first volume may be zero, as indicated). Second volume R2 I This is also a magnetic resonance image. As a contrast agent, disodium gadoxetate (Gd-EOB-DTPA disodium) was used in the example shown in Figure 2, where it was used as a contrast agent for hepatobiliary MRI.

[0141] Second real-space representation R2 I The conversion operation T results in a second representation R2 of the test domain in frequency space. F It can be converted to the second frequency space representation R2 F Similarly, the second real-space representation R2 after the administration of a second amount of contrast agent. I This represents the same test subject and the same test area.

[0142] Second frequency space representation R2 F This is the conversion operation T -1 The second real-space representation R2 I It can be converted to this.

[0143] In the example shown in Figure 2, the first frequency space representation R1 F and the second frequency space representation R2 F This is supplied to the first model M1. First frequency space representation R1 F and the second frequency space representation R2 F Based on this, the first model M1 is the third frequency space representation R3 F This generates the third frequency space representation R3. F This is the conversion operation T -1 (For example, the inverse Fourier transform) gives a third real-space representation R3 I It can be converted to this.

[0144] Model M1 shown in Figure 2 is the first real-space representation R1 Ito the first frequency space representation R1 F It then transitions to the second real-space representation R2 I to the second frequency space representation R2 F It does not include the conversion operation T that converts to. Similarly, the model M1 shown in Figure 2 is a third frequency space representation R3 F The third real-space representation R3 I Conversion operation T that converts to -1 It does not include the conversion operation T and / or the conversion operation T. -1 It is conceivable that these are (one or more) components of the first model M1, that is, the first model M1 is the transformation operation T and / or transformation operation T -1 It is possible to execute this.

[0145] The first model M1 is the first frequency space representation R1 F to the second frequency space representation R2 F Subtract from (R2 F -R1 F The results represent the signal intensity distribution in the frequency space generated by the contrast agent in the examination area.

[0146] difference R2 F -R1 F The fundamental frequency is multiplied by a weighting function WF that weights lower frequencies more highly than higher frequencies. In this case, the amplitude of the fundamental oscillation is multiplied by a weighting coefficient that increases as the frequency decreases. This step is an optional step that can be performed to increase the signal-to-noise ratio in the third representation, particularly for higher values ​​(e.g., greater than 3, 4, or 5) relative to the gain coefficient α. The result of this frequency-dependent weighting is the weighted representation (R²). F -R1 F ) W .

[0147] In frequency-space representation, contrast information is represented at low frequencies, while information about fine structure is represented at higher frequencies. Therefore, such weighting means that higher weighting is given to frequencies that contribute more to contrast than frequencies that contribute less. Image noise is usually evenly distributed across the frequency representation. The frequency-dependent weighting function has a filtering effect. Filtering increases the signal-to-noise ratio by reducing the spectral noise density at high frequencies.

[0148] The preferred weight functions are the Hann function (also known as the Hann window) and the Poisson function (Poisson window).

[0149] Other examples of weight functions can be found, for example, at https: / / de.wikipedia.org / wiki / Fensterfunktion#Beispiele_von_Fensterfunktionen; FJ Harris et al.: On the Use of Windows for Harmonic Analysis with the Discrete Fourier Transform, Proceedings of the IEEE, vol. 66, No. 1, 1978; https: / / docs.scipy.org / doc / scipy / reference / signal.windows.html; KMM Prabhu: Window Functions and Their Applications in Signal Processing, CRC Press, 2014, 978-1-4665-1583-3.

[0150] In the next step, the weighted difference (R2 F -R1 F ) W The gain coefficient α is multiplied to obtain the first frequency-space representation R1 F The result is added to the third representation R3 of the test domain under test in frequency space. F =R1 F +α·(R2 F-R1 F ) W In a further step, the third frequency space representation R3 F This is the conversion operation T -1 (For example, the inverse Fourier transform) gives a third representation R3 of the inspection domain in real space. I It will be converted to this.

[0151] Third expression R3 I This represents the examination area after administration of the third amount of contrast agent. The third amount depends on the gain coefficient α. For example, if the gain coefficient is 3 and the signal intensity distribution, represented by gray value / color value, shows a first-order dependency on the amount of contrast agent, then the third amount corresponds to three times the difference between the second and first amounts.

[0152] As shown in Figure 1, when the first model M1 is used to generate the third representation, not only is the contrast enhanced by a gain coefficient greater than 1 (α>1), but the noise is also enhanced to a similar degree. As shown in Figure 2, the first model M1 can achieve a constant reduction of noise by weighting it with a weighting function in the frequency domain.

[0153] To further reduce noise and / or other unwanted artificial products in the third representation, the second model is applied after the first model. Thus, the third representation produced by the first model is not a result of the generation of the synthetic contrast-enhanced representation of this disclosure. The second model helps to modify the third representation produced by the first model. In this case, the term “modification” may mean reducing or removing noise and / or artificial products.

[0154] Accordingly, according to this disclosure, two models, namely a first model and a second model, are used to generate a composite contrast-enhanced radiographic image. The first model is useful for contrast enhancement (α>1) or contrast reduction (α<1). The second model is useful for correction (e.g., noise suppression, artificial product suppression). The first model may also be referred to as the composite model, and the second model may also be referred to as the correction model.

[0155] In other words, according to this disclosure, two models, namely a first model and a second model, are used to generate a composite contrast-enhanced radiographic image. The first model generates an "suggestion" for the composite contrast-enhanced radiographic image, and the second model optimizes this "suggestion." The "suggestion" may be a first approximation of the composite contrast-enhanced radiographic image. The second model can modify (optimize) this approximation so that the result corresponds to an actual contrast-enhanced radiographic image.

[0156] The third representation generated by the first model is supplied to the second model as input data, and based on this input data and the model parameters, the second model generates a modified third representation. In addition to the third representation generated by the first model, further data may be supplied to the second model as input data (see below).

[0157] The second model is a machine learning model. The second model was trained on training data to generate a modified (e.g., altered and / or optimized) third representation based on the model parameters, which was based on the third representation generated by the first model.

[0158] The training data includes, for each of a number of reference targets, (i) a reference representation generated by a first model representing the reference region of the reference target after administration of a reference amount of contrast agent, and (ii) a measured reference representation of the reference region of the reference target after administration of a reference amount of contrast agent.

[0159] The term "many" means more than 10, preferably more than 100.

[0160] In this specification, the term “reference” is used to distinguish between the phase in which a trained second model is used for modifying the representation and the phase in which the second model is trained. Otherwise, the term “reference” has no limitation to its meaning. The term “(reference) representation” means that the corresponding description applies to both the representation under test and the reference representation of the reference. The “reference object” is the object from which the data (e.g., the reference representation) is used to train the second model. On the other hand, the data under test is used to use the second model, which has been trained (in combination with the first model) for prediction. Like the object under test, each reference object is usually a living organism, preferably a mammal, most preferably a human. The “reference region” is a part of the reference object. The reference region is usually (but not necessarily) the test region of the object under test. That is, if the test region is the organ or part of the organ under test (e.g., the liver or part of the liver), then the reference region of each such reference object is preferably the corresponding organ or the corresponding part of the organ of each reference object. The "reference dose" is the amount of contrast agent determined (defined) at least partially by the first model (for example, by the gain coefficient α). The "reference dose" may correspond to a third amount of contrast agent, or it may be different from the third amount.

[0161] Therefore, the training data on which the second model is trained includes (i) input data and (ii) target data. The second model is configured to generate output data based on the input data and model parameters. The output data is compared with the target data. The difference between the output data and the target data can be reduced by modifying the model parameters in an optimization method (e.g., gradient descent).

[0162] For each of the numerous reference objects, the input data includes a reference representation generated by a first model that represents the reference region of the reference object after administration of a reference dose of contrast agent. Thus, the input data is generated with the help of the first model. They are typically generated based on a first reference representation and a second reference representation. The first reference representation represents the reference region of each reference object without contrast agent or after administration of a first reference dose of contrast agent. The first reference dose may correspond to a first dose. The second reference representation represents the reference region of each reference object after administration of a second reference dose of contrast agent. The second reference dose is typically greater than the first reference dose. The second reference dose may correspond to a second dose. The first and second reference representations are fed into the first model, which generates a third reference representation. The third reference representation represents the reference region of each reference object after administration of a third reference dose of contrast agent. The third reference dose is typically greater than the second reference dose. The third reference quantity may correspond to the third quantity. The third reference representation is the reference representation supplied to the second model when training the second model.

[0163] The training data further includes, as target data, the measured reference representation for each reference target. The measured reference representation represents the reference region of each reference target after administration of a third reference dose of contrast agent. The measured reference representation is a measured representation. Therefore, it represents the reference region of each reference target, as it is actually after administration of the third reference dose of contrast agent (ground truth) and represents the reference region as the reference representation generated by the first model.

[0164] The training of the second model involves, for each of the numerous reference references, To supply the (third) reference representation generated by the first model to the second model, Receiving the revised reference representation from the second model, By modifying the model parameters, the difference between the modified reference representation and the measured reference representation can be reduced. Includes.

[0165] The training process can continue as long as the difference reaches a predefined minimum value, and / or until the difference can no longer be reduced by further modifying the model parameters.

[0166] Therefore, the second model is trained to modify the reference representations generated by the first model for different reference objects so as closely as possible to the respective measured reference representations.

[0167] Therefore, the second model is trained, for example, to reduce or eliminate noise and / or artificial products generated by the first model in the reference representation.

[0168] Figure 3 shows, as an example, the process for training the second model in a schematic form. The second model M2 is trained using the training data TD. The training data TD consists of a third reference representation RR3 of the reference region of each reference generated by the first model M1 for each of the numerous reference references, and the measured third reference representation RR3 of the reference region of the reference references. M This includes the following. The training data may optionally include additional data (see below).

[0169] In the example shown in Figure 3, only one set of training data TD for a single reference object is shown.

[0170] The reference population is humans, and the reference range includes the human lung.

[0171] Measured third reference representation RR3 M This represents the reference range of the baseline after administration of the third dose of contrast agent. Measured third baseline representation RR3 M This is an image generated by measurement, i.e., the measured third reference representation RR3 M This is the result of a radiation examination. For example, the measured third reference expression RR3 M This may be a CT image, MRI image, ultrasound image, or other radiographic image.

[0172] The third reference representation RR3 is generated by the first model M1. The first model M1 is configured to generate the third reference representation RR3 based on the first reference representation RR1 and the second reference representation RR2.

[0173] The first reference expression, RR1, represents the reference region of the reference subject without contrast agent or after administration of a first amount of contrast agent. The second reference expression, RR2, represents the reference region of the reference subject after administration of a second amount of contrast agent. The second amount is greater than the first amount. The third expression, RR3, represents the reference region of the reference subject after administration of a third amount of contrast agent.

[0174] The third reference representation RR3, generated by the first model M1, is supplied to the second model M2. The second model may optionally be supplied with further input data, such as the first reference representation RR1 and / or the second reference representation RR2 and / or additional / other data FD. This additional input data may include data specifying the contrast agent used, the first amount of contrast agent, the second amount of contrast agent, the third amount of contrast agent, gain coefficients, acquisition parameters for generating the first and / or second representations, reference objects, reference regions, and / or other properties / conditions. The optional use of additional input data is shown in Figure 3 with dashed arrows. If additional input data is used, these additional input data are also training data.

[0175] The second model M2 is based on the third reference representation RR3 and modified based on the model parameter MP (optionally, additional input data) to obtain the third representation RR3. C The second model M2 is configured to generate the measured third representation RR3. M A modified third representation, RR3, that approximates it as closely as possible. C It is trained to generate the modified third representation RR3. C This is the measured third representation RR3 M It is compared to the loss function LF, which is a modified third representation of RR3. CThe third expression RR3 was measured as follows: M It is used to quantify the difference between and . In an optimization method (e.g., gradient descent), the model parameter MP is modified to reduce (minimize) the difference, and therefore the calculated loss is determined using the loss function LF. The process is repeated to further train on further reference datasets until the difference is reduced to a predefined minimum and / or the difference can no longer be reduced by further modifying the model parameters.

[0176] The trained second model may be stored on a data storage medium, transmitted to another computer system (e.g., on a network), and / or used to generate a modified third representation of the test area under examination.

[0177] Figure 4 shows, as an example, the generation of a modified third representation based on a third representation with the help of a trained second model, in a schematic form.

[0178] For example, a trained second model M2 T The trained second model M2 may be trained as shown in Figure 3. T The third representation R3 of the test domain to be tested is supplied. The second trained model M2 T Further input data may be optionally supplied. This further input data may include data specifying the contrast agent used, the first amount of contrast agent, the second amount of contrast agent, the third amount of contrast agent, the gain coefficient, acquisition parameters for generating the first and / or second representations, the object being examined, the examination area, and / or other properties / conditions. The optional use of further input data is shown in Figure 4 by a dashed arrow. Further input data is used, in particular, if such data is also used to train a second model.

[0179] In the example shown in Figure 4, the subject of the examination is a human, and the examination area includes the human lung.

[0180] The third representation R3 is generated with the aid of the first model M1. The first model M1 is configured to generate a third representation based on the first representation R1 and the second representation R2. The first representation R1 represents an examination region of an examination subject without a contrast agent or after administration of a first amount of a contrast agent. The second representation R2 represents an examination region of the examination subject after administration of a second amount of a contrast agent. The second amount is larger than the first amount. The third representation R3 represents an examination region of the examination subject after administration of a third amount of a contrast agent. The third amount is different from the first amount and the second amount. In this embodiment, the third amount is larger than the second amount, that is, the first model generates a third representation R3 that exhibits contrast enhancement compared to the second representation R2.

[0181] The trained second model M2 T is configured and trained to generate a modified third representation R3 based on the third representation R3 (and optionally further input data) generated by the first model M1. The modified third representation R3 C is an artificial radiation image that is improved compared to the third representation R3 because it includes, for example, less noise and / or fewer artifacts. C

[0182] The second model may be an artificial neural network or may include such a network.

[0183] An “artificial neural network” includes at least three layers of processing elements, that is, a first layer having input neurons (nodes), an Nth layer having at least one output neuron (node), and N - 2 inner layers, where N is a natural number and is greater than 2.

[0184] Input neurons are responsible for receiving the input representation. Typically, there is one input neuron for each pixel or voxel of the input representation if the input representation is a real-space depiction in the form of raster graphics, or one input neuron for each frequency present in the input representation if the input representation is a frequency-space depiction. Additional input neurons may be present for additional input values ​​(e.g., information about the test region, information about the object being tested, information about the conditions that prevail during the generation of the input representation, information about the state represented by the input representation, and / or information about the time or time interval during which the input representation was generated).

[0185] The output neurons are responsible for producing the synthesized radiographic image.

[0186] The processing elements of the layer between the input neuron and the output neuron are connected to each other in a predetermined pattern having predetermined connection weights.

[0187] The artificial neural network may be a convolutional neural network (CNN), or may include such a network.

[0188] Convolutional neural networks can process input data in matrix form. This makes it possible to use digital radiographic images, represented in matrix form (e.g., width × height × color channels), as input data. For example, a normal neural network in the form of a multilayer perceptron (MLP) requires a vector as input; that is, to use a radiographic image as input, the pixels or voxels of the radiographic image must be extended sequentially in long chains. This means that a normal neural network cannot recognize an object in a radiographic image independently of its position in the image. The same object at different positions in the image will have completely different input vectors.

[0189] A CNN typically consists of an array of alternating filters (convolutional layers) and aggregated layers (reservoir layers) that terminate at one or more layers of "normally" fully connected neurons (crowded / fully connected layers).

[0190] An artificial neural network may have an autoencoder architecture, for example, an artificial neural network may have an architecture such as U-Net (see, for example, O. Ronneberger et al.: U-net: Convolutional networks for biomedical image segmentation, International Conference on Medical image computing and computer-assisted intervention, pages 234-241, Springer, 2015, https: / / doi.org / 10.1007 / 978-3-319-24574-4_28).

[0191] The artificial neural network may be a generative adversarial network (GAN) (see, for example, M.-Y. Liu et al.: Generative Adversarial Networks for Image and Video Synthesis: Algorithms and Applications, arXiv:2008.02793; J. Henry et al.: Pix2Pix GAN for Image-to-Image Translation, DOI: 10.13140 / RG.2.2.32286.66887).

[0192] The artificial neural network may be a generative adversarial network (GAN) specifically for image super-resolution (SR) (see, for example, C. Ledig et al.: Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network, arXiv:1609.04802v5).

[0193] The artificial neural network may also be a transformer network (see, for example, D. Karimi et al.: Convolution-Free Medical Image Segmentation using Transformers, arXiv:2102.13645 [eess.IV]).

[0194] If the first model is differentiable, the first and second models may be combined into a unified model. In this case, the first and second models are components of the unified model, where the third (reference) representation generated by the first model is directly supplied to the second model.

[0195] See, for example, Figures 1 and 2, the first model described herein is differentiable. It may be, for example, one or more layers prepending a second model designed as a neural network. The one or more layers may include computational operations having corresponding computational parameters (e.g., gain coefficient α) in the form of fixed (invariant) values, but the model parameters of the second model may be variable. This allows the first model to be included in the training of the second model, and the model parameters of the first model remain invariant during training. In such cases, the first model uses its invariant model parameters to specify the physical basis that the second model must observe / accept / allow. In contrast to a fully variable model for generating artificial contrast-enhanced radiographic images as described, for example, in International Publication No. 2019 / 074938A1, the method described herein has the advantage that the physical basis to which the model can move can be specified to the model as knowledge of the form of the first model. This makes the resulting synthetic radiographic image more realistic.

[0196] In a unified model designed as a neural network including a first model and a second model, one or more additional layers may be prepended to the first model. These one or more additional layers may include, for example, trainable (variable) model parameters that provide (improved) co-registration of the first (reference) representation and the second (reference) representation. Furthermore, the (improved) co-registration may be a component of training. The one or more additional layers prepending the first model may include trainable model parameters that learn one or more model parameters of the first model, e.g., optimized gain coefficients α and / or parameters of the weight function in the case of the weighting as shown in Figure 2. These one or more layers prepending the first model may have an architecture such as DenseNet (see, for example, G. Haung et al.: Densely Connected Convolutional Networks, arXiv:1608.06993v5).

[0197] The model, including the first and second models, can be trained through an end-to-end process.

[0198] It is further possible to supply the second model not only with the (reference) representation generated by the first model, but also with the first (reference) representation and / or the second (reference) representation. This allows the second model to be trained to reduce or remove artificial elements in the third (reference) representation due to inappropriate co-registration.

[0199] Figure 5 shows, as an example, a model that includes the first model and the second model in a schematic form.

[0200] Model M shown in Figure 5 has various processing layers L1, L2, L3, L4, L5, L6, L7, and L8. The number of processing layers was selected purely randomly, but the processing layers shown are for illustrative purposes only. Processing layers L4 and L5 form the first model M1, and processing layers L6, L7, and L8 form the second model M2. The first model M1 is prepended by processing layers L1, L2, and L3, which form the initial model M0.

[0201] The first representation R1 and the second representation R2 are supplied to the initial model M0. The first representation R1 represents the examination area of ​​the subject without contrast agent or after administration of a first amount of contrast agent, and the second representation R2 represents the examination area of ​​the subject after administration of a second amount of contrast agent, where the second amount is greater than the first amount, the subject is a human, and the examination area includes the human lung.

[0202] The initial model M0 may be configured to co-register a first representation R1 and a second representation R2, and / or determine the model parameters of the first model M1. The co-registered representations can be sent to layer L4 of the first model M1 via layer L3 of the initial M0 model. In this regard, it is equally possible to send the model parameters determined by the initial model M0 to the first model M1, but the first representation R1 and the second representation R2 are supplied to the first model M1 separately (see dashed arrow).

[0203] The first model M1 is configured to generate a third expression (not shown in Figure 5) based on (co-registered) expression R1 and (co-registered) expression R2. The third expression represents the examination area to be examined after administration of a third amount of contrast agent, the third amount being different from the first and second amounts, and preferably being greater than the second amount.

[0204] The third representation generated by the first model M1 is sent to layer L6 of the second model M2 via layer L5 of the first model M1.

[0205] The second model M2 can also be supplied with a first representation R1 and / or a second representation R2, both of which are optionally co-registered (see dashed arrow).

[0206] The second model M2 is based on the third expression (and optionally on the first expression R1 (jointly registered) and / or the second expression (jointly registered)), and is a modified third expression R3. C It is configured and trained to generate the modified third representation R3. C This has less noise and / or less artificiality compared to the third representation. Modified third representation R3 C This can be output via layer L8 of the second model M2.

[0207] Figure 6 shows, as an example, a schematic representation of the computer system relating to this disclosure.

[0208] A "computer system" is an electronic data processing system that processes data according to programmable computational rules. Such a system typically includes a "computer," which is a unit containing a processor for performing logical operations, and peripheral devices.

[0209] In computer technology, "peripheral devices" refer to all devices connected to a computer and used for controlling the computer, as well as / or as input and output devices. Examples include monitors (screens), printers, scanners, mice, keyboards, drives, cameras, microphones, and speakers. Internal ports and expansion cards are also considered peripheral devices in computer technology.

[0210] The computer system (1) shown in Figure 6 includes a receiving unit (11), a control calculation unit (12), and an output unit (13).

[0211] The control calculation unit (12) functions for the control of the computer system (1), the adjustment of the data flow between the units of the computer system (1), and the execution of calculations.

[0212] The control calculation unit (12) generates, or causes the receiving unit (11) to receive, a first representation such that the first representation represents an examination region of an examination subject without a contrast agent or after administration of a first amount of a contrast agent. generates, or causes the receiving unit (11) to receive, a second representation such that the second representation represents an examination region of the examination subject after administration of a second amount of the contrast agent, and the second amount is larger than the first amount. supply the first representation and the second representation to a first model configured to generate a third representation based on the first representation and the second representation, wherein the third representation represents an examination region of the examination subject after administration of a third amount of the contrast agent, and the third amount is different from the first amount and the second amount. The second model is trained in a training process based on training data. The training data for each reference object (i) a reference representation generated by a first model representing a reference region of a reference object after administration of a reference amount of a contrast agent, and (ii) a measured reference representation of the reference region of the reference object after administration of the reference amount of the contrast agent. The training process for each reference object includes supplying the reference representation generated by the first model to the second model, receiving a corrected reference representation from the second model, modifying model parameters to reduce the difference between the corrected reference representation and the measured reference representation, and supplying the third representation to the second model. receive, from the second model, a corrected third representation of the examination region of the examination subject. The output unit (13) is configured to output the corrected third expression, store it, and / or transmit it to another computer system. configured.

[0213] FIG. 7 shows, by way of example, a further embodiment of a computer system in schematic form. The computer system (1) includes a processing unit (21) connected to a storage medium (22). The processing unit (21) and the storage medium (22) form a control computing unit as shown in FIG. 6.

[0214] The processing unit (21) may include one or more processors, either alone or in combination with one or more storage media. The processing unit (21) may be conventional computer hardware capable of processing information such as digital images, computer programs, and / or other digital information. The processing unit (21) typically consists of an arrangement of electronic circuits, some of which may be designed as an integrated circuit or as a plurality of integrated circuits connected to each other (an integrated circuit may sometimes be referred to as a "chip"). The processing unit (21) may be configured to execute a computer program that may be stored in the working memory of the processing unit (21) or in the storage medium (22) of the same or a different computer system.

[0215] The storage medium (22) may be conventional computer hardware capable of temporarily and / or permanently storing information such as digital images (e.g., representations of inspection areas), data, computer programs, and / or other digital information. The storage medium (22) may include volatile and / or non-volatile storage media, and may be fixed in place or removable. Examples of suitable storage media include RAM (random access memory), ROM (read-only memory), hard disks, flash memory, replaceable computer floppy disks, optical disks, magnetic tapes, or combinations thereof. Examples of optical disks include compact disks with read-only memory (CD-ROM), compact disks with read / write functionality (CD-R / W), DVDs, Blu-ray discs, and similar types.

[0216] The processing unit (21) may be connected not only to the storage medium (22) but also to one or more interfaces (11, 12, 31, 32, 33) for displaying, transmitting, and / or receiving information. The interfaces may include one or more communication interfaces (11, 32, 33) and / or one or more user interfaces (12, 31). One or more communication interfaces may be configured to send and / or receive information to, for example, an MRI scanner, CT scanner, ultrasound camera, other computer systems, networks, data storage media, or the like. One or more communication interfaces may be configured to send and / or receive information via physical (wired) communication connections and / or wireless communication connections. One or more communication interfaces may include one or more interfaces for connecting to a network using, for example, cellular, Wi-Fi, satellite, cable, DSL, optical fiber, and / or the like. In some embodiments, one or more communication interfaces may include one or more near-field communication interfaces configured to connect devices having near-field communication technologies such as NFC, RFID, Bluetooth, Bluetooth LE, Zigbee, infrared (e.g., IrDA) or the like.

[0217] The user interface may include a display (31). The display (31) may be configured to display information to the user. Preferred examples of the display include liquid crystal displays (LCDs), light-emitting diode displays (LEDs), plasma display panels (PDPs), or the like. The user input interfaces (11, 12) may be wired or wireless and may be configured to receive information from the user of the computer system (1) for processing, storage and / or display, for example. Preferred examples of the user input interfaces include a microphone, an image recording device or a video recording device (e.g., a camera), a keyboard or keypad, a joystick, a touch-sensing surface (separate from the touchscreen or integrated into the touchscreen), or the like. In some examples, the user interface may include automated identification data acquisition (AIDC) technology for machine-readable information. This may include barcodes, radio frequency identification (RFID), magnetic strips, optical character recognition (OCR), integrated circuit cards (ICCs), and the like. The user interface may further include one or more interfaces for communication with peripheral devices such as printers and similar devices.

[0218] One or more computer programs (40) may be stored in a storage medium (22) and executed by a processing unit (21), thereby programming the processing unit (21) to perform the functions described herein. The retrieval, loading, and execution of instructions in the computer programs (40) may be performed sequentially, such that one instruction is retrieved, loaded, and executed at a time. However, the retrieval, loading, and / or execution may be performed in parallel.

[0219] The computer systems of this disclosure may be designed as laptops, notebooks, netbooks and / or tablet PCs, and may be components of MRI scanners, CT scanners or ultrasound diagnostic devices.

[0220] The present invention also provides computer program products. Such computer program products include a non-volatile data carrier, for example, a CD, DVD, USB stick, or another data storage medium. The computer program is stored on the data carrier. The computer program can be loaded into the working memory of a computer system (more specifically, into the working memory of the computer system of this disclosure). Thereafter, the computer system performs the following steps, namely: The first representation represents the examination area to be examined, either without contrast agent or after administration of a first amount of contrast agent. The second representation represents the examination area of ​​the subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount. The first model is configured to generate a third representation based on the first and second representations, wherein the third representation represents the examination area of ​​the examination subject after administration of a third amount of the contrast agent, and the first and second representations are supplied to the first model such that the third amount is different from the first and second amounts. The second model was trained during the training process based on the training data. The training data for each criterion is based on a large number of criterion objects, i.e. (i) A reference representation generated by a first model that represents the reference region of a reference target after administration of a reference dose of contrast agent, and (ii) including the measured reference expression of the reference area of ​​the reference subject after administration of a reference dose of contrast agent, The training process for each standard target is The standard representation generated by the first model is supplied to the second model. Receiving the revised reference representation from the second model, A third representation is supplied to a second model, including the step of reducing the difference between the modified reference representation and the measured reference representation by modifying the model parameters. Receiving a modified third representation of the inspection area of ​​the subject being inspected from the second model, The system outputs and / or stores the modified third representation, and / or transmits the modified third representation to another computer system.

[0221] Computer programs may be available for purchase as computer program products, for example, through web pages and / or application stores, as well as for download.

[0222] Computer program products may be sold in combination with (as a set with) a contrast agent. Such a set may also be referred to as a kit. Such a kit includes a contrast agent and a computer program product. Such a kit may also include a contrast agent and means that enable the purchaser to obtain the computer program, for example, to download the computer program from a webpage. These means may include a link, i.e., the address of a webpage from which the computer program can be obtained, for example, the address of a webpage from which the computer program can be downloaded to a computer system connected to the Internet. These means may include a code that gives the purchaser access to the computer program (for example, a string of letters and numbers or a QR code, or a data matrix code, or a barcode, or another optically and / or electronically readable code). Such links and / or codes may be printed, for example, on the packaging of the contrast agent and / or in the accompanying documentation for the contrast agent. Thus, a kit is a combination product that includes a contrast agent and a computer program, which can be purchased together (for example, in the form of access to the computer program, or in the form of executable program code in a data carrier).

[0223] Figure 8 shows, as an example, one embodiment of a computer implementation method in the form of a flowchart, in a schematic manner.

[0224] The method (100) includes: (110) providing the first representation that represents an examination region of an examination subject without a contrast agent or after administration of a first amount of the contrast agent; (120) providing the second representation that represents the examination region of the examination subject after administration of a second amount of the contrast agent, wherein the second amount is greater than the first amount; (130) supplying the first representation and the second representation to a first model configured to generate a third representation based on the first representation and the second representation, wherein the third representation represents an examination region of an examination subject after administration of a third amount of the contrast agent, and the third amount is different from the first amount and the second amount; (140) the second model being trained in a training process based on training data, wherein the training data for each reference subject includes: (i) a reference representation generated by a first model that represents a reference region of a reference subject after administration of a reference amount of the contrast agent, and (ii) a measured reference representation of the reference region of the reference subject after administration of the reference amount of the contrast agent, and the training process for each reference subject includes: supplying the reference representation generated by the first model to the second model; receiving a modified reference representation from the second model; modifying model parameters to reduce a difference between the modified reference representation and the measured reference representation, and (1): supplying the third representation to the second model; (150) receiving, from the second model, a modified third representation of the examination region of the examination subject; (160) outputting and / or storing the modified third representation and / or transmitting the modified third representation to another computer system.

[0225] The present invention can be used for a variety of purposes. Some examples of its use are shown below, but the invention is not limited to these examples.

[0226] The first use case relates to magnetic resonance imaging for identifying intraparenchymal tumors such as brain metastases and malignant gliomas. The invasive growth of these tumors makes it difficult to accurately distinguish between tumor and healthy tissue. However, determining the extent of the tumor is crucial for surgical resection. Distinguishing between tumor and healthy tissue is facilitated by the administration of extracellular material. After intravenous administration of gadobutrol, an extracellular MRI contrast agent, at a standard dose of 0.1 mmol per kg of body weight, intraparenchymal tumors can be identified much more easily. At higher doses, the contrast between lesions and healthy brain tissue increases further, and the detection rate of brain metastases increases linearly with the dose of contrast agent (see, for example, M. Hartmann et al.: Does the administration of a high dose of a paramagnetic contrast medium (Gadovist) improve the diagnostic value of magnetic resonance tomography in glioblastomas? doi: 10.1055 / s-2007-1015623).

[0227] A single triple dose or a second escalator may be administered up to a total dose of 0.3 mmol per kg of body weight. In this case, the patient and those around them will be additionally exposed to gadolinium, and additional costs will be incurred if a second scan is performed.

[0228] The present invention can be used to avoid exceeding the standard dose of contrast agent. It is possible to generate a first MRI image with no contrast agent or less than the standard dose, and a second MRI image with the standard dose. Based on these generated MRI images, it is possible to generate a composite MRI image in which the contrast between lesion and healthy tissue can be varied over a wide range by changing the gain coefficient α, as described in this disclosure. This makes it possible to achieve contrast that is only achievable by administering a larger-than-standard dose of contrast agent.

[0229] Another use case relates to the reduction of MRI contrast agents in magnetic resonance imaging. Gadolinium-containing contrast agents, such as gadobutrol, are used in a variety of examinations. They are used for contrast enhancement in cranial, spinal, thoracic, or other examinations. In the central nervous system, gadobutrol highlights areas of impaired blood-brain barrier and / or abnormal vascular regions. In breast tissue, gadobutrol allows for the visualization of the presence and extent of malignant breast disease. Gadobutrol is also used in contrast-enhanced magnetic resonance angiography for the diagnosis of stroke, detection of tumor hemoperfusion, and detection of focal cerebral ischemia.

[0230] Increased environmental pollution, the cost burden on healthcare systems, and concerns about the potential for acute side effects and long-term health risks, particularly in the case of repeated and prolonged exposure, have spurred efforts to reduce the dosage of gadolinium-containing contrast agents. This can be achieved by the present invention.

[0231] It is possible to generate a first MRI image without contrast agent and a second MRI image with a substandard amount of contrast agent. Based on these generated MRI images, it is possible to generate a composite MRI image in which the contrast can be varied over a wide range by changing the gain coefficient α, as described in this disclosure. This makes it possible to achieve the same contrast with a substandard amount of contrast agent as that obtained after administering a standard amount.

[0232] Another use case relates to the detection, identification, and / or characterization of liver lesions with the assistance of hepatobiliary contrast agents such as Primovist®.

[0233] Primovist® is administered intravenously (iv) at a standard dose of 0.025 mmol per kg of body weight. This standard dose is lower than the standard dose of 0.1 mmol per kg of body weight for extracellular MRI contrast agents. Unlike contrast-enhanced MRI using extracellular gadolinium-containing contrast agents, Primovist® enables dynamic multiphase T1w imaging. However, the lower dose of Primovist® and the possibility of transient motion artificial products occurring immediately after intravenous administration suggest that radiologists perceive Primovist®'s contrast enhancement in the arterial phase as inferior to that of extracellular MRI contrast agents. Nevertheless, contrast enhancement in the arterial phase and evaluation of vascular distribution in localized hepatic lesions are crucial for accurate characterization of lesions.

[0234] With the help of the present invention, it is possible to increase contrast, particularly in the arterial phase, without the need to administer higher doses.

[0235] It is possible to generate a first MRI image without contrast agent and a second MRI image in the arterial phase after administering a standard amount of contrast agent. Based on these generated MRI images, it is possible to generate a composite MRI image in which the contrast in the arterial phase can be varied over a wide range by changing the gain coefficient α, as described in this disclosure. This makes it possible to achieve contrast that is only achievable by administering a larger-than-standard amount of contrast agent.

[0236] Another use case relates to the use of MRI contrast agents in computed tomography (CT) scans.

[0237] In CT scans, MRI contrast agents typically offer less contrast enhancement than CT contrast agents. However, using MRI contrast agents in CT scans can be advantageous. An example is minimally invasive interventions in a patient's liver where a surgeon is monitoring the procedure using a CT scanner. Compared to magnetic resonance imaging, computed tomography (CT) has the advantage of allowing more significant surgical interventions within the examination area while simultaneously generating CT images of the area being examined. In contrast, only a limited number of surgical instruments and devices are compatible with MRI. Furthermore, patient access is restricted by the magnets used in MRI. Therefore, surgeons can perform procedures within the examination area while simultaneously visualizing the area with CT and tracking the procedure on a monitor.

[0238] For example, if a surgeon wants to perform a procedure on a patient's liver, such as a biopsy of a liver lesion or removal of a tumor, the contrast between liver lesions or tumors and healthy liver tissue is not as pronounced in CT images of the liver as it is in MRI images after administration of a hepatobiliary contrast agent. Currently, no hepatobiliary contrast agents specifically for CT are known and / or approved. Therefore, using MRI contrast agents, more specifically hepatobiliary MRI contrast agents, in computed tomography would allow for both the possibility of distinguishing between healthy and diseased liver tissue and the possibility of performing surgery while simultaneously visualizing the liver.

[0239] The relatively low contrast enhancement achieved by MRI contrast agents can be increased with the help of the present invention without the need to administer doses higher than the standard dose.

[0240] It is possible to generate a first CT image without MRI contrast agent and a second CT image after administering an amount of MRI contrast agent equivalent to a standard dose. Based on these generated CT images, it is possible to generate a composite CT image in which the contrast generated by the MRI contrast agent can be varied over a wide range by changing the gain coefficient α, as described in this disclosure. This makes it possible to achieve contrast that is only achievable by administering an amount of MRI contrast agent greater than the standard dose.

Claims

1. First expression (R1, R1 Ｉ R1 Ｆ ) represents the examination area to be examined without contrast agent, or after administration of a first amount of contrast agent, the first representation (R1, R1 Ｉ R1 Ｆ ) to provide Second expression (R2, R2 Ｉ , R2 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a second amount of the contrast agent, and the second expression (R2, R2) is greater than the first amount. Ｉ , R2 Ｆ ) to provide The first model (M1) is configured to generate a third representation (R3, R3 Ｉ , R3 Ｆ ) based on the first representation (R1, R1 Ｉ , R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ), and the third representation (R3, R3 Ｉ , R3 Ｆ ) represents the examination region of the examination object after administration of a third amount of the contrast agent, and the third amount is different from the first amount and the second amount, supplying the first representation (R1, R1 Ｉ , R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) to the first model (M1), The second model (M2) was trained during the training process based on the training data (TD). The training data (TD) for each reference target is used for a large number of reference targets, i.e., (i) a reference representation (RR3) generated by the first model representing the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) The measured reference expression of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent (RR3 Ｍ ), including, The training process for each standard subject is The reference representation generated by the first model (M1) is supplied to the second model (M2), Revised reference expression (RR3) Ｃ ) receive from the second model (M2), By modifying the model parameters, the modified reference representation (RR3 Ｃ ) and the measured reference expression (RR3 Ｍ The third representation (R3, R3) includes the step of reducing the difference between the two. Ｉ , R3 Ｆ ) to supply the above second model (M2), A modified third representation (R3) of the inspection area of ​​the subject being inspected, derived from the second model (M2). Ｃ ) receiving The aforementioned modified third expression (R3 Ｃ Outputting and / or storing the modified third representation (R3 Ｃ ) to send to another computer system, A computer-aided method including the following steps.

2. The method according to claim 1, wherein the first model (M1) is a mechanism model.

3. The third representation (R3, R3) by the first model (M1) Ｉ , R3 Ｆ The above generation of ) The first expression (R1, R1 Ｉ R1 Ｆ ) to the second expression (R2, R2 Ｉ , R2 Ｆ Subtracting from ) Multiply the subtraction result by the gain coefficient α. The multiplication result is expressed in the first representation (R1, R1 Ｉ R1 Ｆ The method according to any one of claims 1 or 2, comprising the step of adding to ).

4. The third representation (R1, R1) by the first model (M1) Ｉ R1 Ｆ The above generation of ) The first expression (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) Frequency space representation of the difference (R2 Ｆ -R1 Ｆ By multiplying ) by a frequency-dependent weighting function (WF), a weighted representation (R2) can be obtained. Ｆ -R1 Ｆ ) Ｗ To obtain The aforementioned weighted representation (R2 Ｆ -R1 Ｆ ) Ｗ Multiply by the gain coefficient α, The weighted representation (R2) obtained by multiplying by the gain coefficient α Ｆ -R1 Ｆ ) Ｗ to the first expression (R1, R1 Ｉ R1 Ｆ The method according to any one of claims 1 to 3, comprising the step of adding to ).

5. The method according to claim 4, wherein the frequency-dependent weighting function (WF) is a Hann window function or a Poisson window function.

6. The method according to any one of claims 3 to 5, wherein the gain coefficient α is greater than 1, preferably greater than 2.

7. The method according to any one of claims 3 to 5, wherein the gain coefficient α is greater than zero, and preferably less than 1.

8. The method according to any one of claims 3 to 5, wherein the gain coefficient α is less than zero.

9. The method according to any one of claims 1 to 8, wherein the third amount and the reference amount are greater than the standard amount of the contrast agent.

10. The method according to any one of claims 1 to 9, wherein the second model (M2) is an artificial neural network, and the first model (M1) includes one or more processing layers (L4, L5) that prepend the second model (M2).

11. The second model (M2) is prepended by one or more further processing layers (L1, L2, L3), the one or more further processing layers form the first model (M0), and the first model (M0) is the first representation (R1, R1 Ｉ R1 Ｆ ) and / or the second expression (R2, R2 Ｉ , R2 Ｆ The gain coefficient α and / or parameters of the frequency-dependent weighting function (WF) are determined based on the first representation (R1, R1 Ｉ R1 Ｆ ) and the second expression (R2, R2 Ｉ , R2 Ｆ The method according to claim 10, configured to jointly register )

12. The subject of the test and each reference subject are human or animal, preferably mammal. The method according to any one of claims 1 to 11, wherein the inspection area is a part of the object to be inspected, the reference area is a part of the reference object, and the reference area corresponds to the inspection area.

13. The first expression (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ The method according to any one of claims 1 to 12, wherein the result is a radiological examination, preferably an MRI and / or CT scan.

14. The aforementioned contrast agent, Gd of the compound of formula (I) ３＋ complex, 【Chemistry 1】 (I) In the formula, Ar is 【Chemistry 2】 and 【Transformation 3】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ - A base selected from #, Here, * represents linkage to Ar, and # represents linkage to an acetate residue. R １ , R ２ and R ３ Each of these is independently a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ４ is a group selected from C ２ to C ４ alkoxy, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and is a group selected from R ５ It is a hydrogen atom, Furthermore R ６ It is a hydrogen atom, or its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, Gd of the compound of formula (II) ３＋ complex 【Chemistry 4】 (II) In the formula, Ar is 【Transformation 5】 and 【Transformation 6】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ -A group selected from #, where * is linkage to Ar and # is linkage to an acetate residue. R ７ is a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ８ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ９ and R １０ These are each an independent hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, The contrast agent is the following substance, namely, Gadolinium(III) 2-[4,7,10-tris(carboxymethyl)-1,4,7,10-tetraazacyclododea-1-yl]acetic acid, Gadolinium(III) ethoxybenzyldiethylenetriaminepentaacetic acid, Gadolinium(III)2-[3,9-bis[1-carboxylat-4-(2,3-dihydroxypropylamino)-4-oxobutyl]-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-trien-6-yl]-5-(2,3-dihydroxypropylamino)-5-oxopentanoate, Dihydrogen [(±)-4-carboxy-5,8,11-tris(carboxymethyl)-1-phenyl-2-oxa-5,8,11-triazatridecane-13-oate(5-)]gadolinate(2-), Tetragadolinium [4,10-bis(carboxylatomethyl)-7-{3,6,12,15-tetraoxo-16-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]-9,9-bis({[({2-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]propanoyl}aminoacetyl]-amino}methyl)-4,7,11,14-tetraazahepta-decane-2-yl}-1,4,7,10-tetraazacyclododecane-1-yl]acetate, 2,2',2''-(10-{1-carboxy-2-[2-(4-ethoxyphenyl)ethoxy]ethyl}-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate, Gadolinium 2,2',2''-{10-[1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium 2,2',2''-{10-[(1R)-1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium (2S,2'S,2''S)-2,2',2''-{10-[(1S)-1-carboxy-4-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}butyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}tris(3-hydroxypropanoate), Gadolinium 2,2',2''-{10-[(1S)-4-(4-butoxyphenyl)-1-carboxybutyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium(III) 5,8-bis(carboxylatomethyl)-2-[2-(methylamino)-2-oxoethyl]-10-oxo-2,5,8,11-tetraazadodecane-1-carboxylate hydrate, Gadolinium(III) 2-[4-(2-hydroxypropyl)-7,10-bis(2-oxide-2-oxoethyl)-1,4,7,10-tetraazacyclododeca-1-yl]acetate, Gadolinium(III) 2,2',2''-(10-((2R,3S)-1,3,4-trihydroxybutan-2-yl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl) triacetate, Gadolinium-2,2',2''-{(2S)-10-(carboxymethyl)-2-[4-(2-ethoxyethoxy)benzyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium-2,2',2''-[10-(carboxymethyl)-2-(4-ethoxybenzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl]triacetate, The method according to any one of claims 1 to 13, comprising one of the above.

15. Receiving unit (11), Control calculation unit (12), and Includes an output unit (13), The control calculation unit (12) First expression (R1, R1 Ｉ R1 Ｆ ) represents the examination area to be examined without contrast agent, or after administration of a first amount of contrast agent, the first representation (R1, R1 Ｉ R1 Ｆ ) to generate or to send the first representation (R1, R1 Ｉ R1 Ｆ ) to receive Second expression (R2, R2 Ｉ , R2 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a second amount of the contrast agent, and the second expression (R2, R2) is greater than the first amount. Ｉ , R2 Ｆ ) to generate or the receiving unit (11) the second representation (R2, R2 Ｉ , R2 Ｆ ) to receive The first model (M1) is the first representation (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) based on the third expression (R3, R3 Ｉ , R3 Ｆ It is configured to generate the third representation (R3, R3 Ｉ , R3 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a third amount of the contrast agent, and the first expression (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) to be supplied to the first model (M1), The second model (M2) was trained during the training process based on the training data (TD). The training data (TD) for each reference target is used for a large number of reference targets, i.e., (i) a reference representation (RR3) generated by the first model (M1) representing the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) The measured reference expression of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent (RR3 Ｍ ), including, The training process for each standard subject is The reference representation (RR3) generated by the first model (M1) is supplied to the second model (M2). Revised reference expression (RR3) Ｃ ) receive from the second model (M2), By modifying the model parameters (MP), the modified reference expression (RR3) can be created. Ｃ ) and the measured reference expression (RR3 Ｍ The third representation (R3, R3) includes the step of reducing the difference between the two. Ｉ , R3 Ｆ ) to be supplied to the second model (M2), A modified third representation (R3) of the inspection area of ​​the subject being inspected, derived from the second model (M2). Ｃ ) to receive The output unit (13) receives the modified third representation (R3 Ｃ A computer system (1) configured to output, store, and / or transmit to another computer system.

16. A computer program product including a data carrier that stores a computer program (40) that can be loaded into the working memory (22) of a computer system (1), wherein the computer program (40) is loaded into the computer system (1) by the following steps, namely, First expression (R1, R1 Ｉ R1 Ｆ ) represents the examination area to be examined without contrast agent, or after administration of a first amount of contrast agent, the first representation (R1, R1 Ｉ R1 Ｆ ) to provide Second expression (R2, R2 Ｉ , R2 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a second amount of the contrast agent, and the second expression (R2, R2) is greater than the first amount. Ｉ , R2 Ｆ ) to provide The first model (M1) is the first representation (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) based on the third expression (R3, R3 Ｉ , R3 Ｆ It is configured to generate the third representation (R3, R3 Ｉ , R3 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a third amount of the contrast agent, and the first expression (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) to supply the above first model (M1), The second model (M2) was trained during the training process based on the training data (TD). The training data (TD) for each reference target is used for a large number of reference targets, i.e., (i) a reference representation (RR3) generated by the first model representing the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) The measured reference expression of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent (RR3 Ｍ ), including, The training process for each standard subject is The reference representation (RR3) generated by the first model (M1) is supplied to the second model (M2). Revised reference expression (RR3) Ｃ ) receive from the second model (M2), By modifying the model parameters (MP), the modified reference expression (RR3) can be created. Ｃ ) and the measured reference expression (RR3 Ｍ The third representation (R3, R3) includes the step of reducing the difference between the two. Ｉ , R3 Ｆ ) to supply the above second model (M2), A modified third representation (R3) of the inspection area of ​​the subject being inspected, derived from the second model (M2). Ｃ ) receiving The aforementioned modified third expression (R3 Ｃ Outputting and / or storing the modified third representation (R3 Ｃ ) to send to another computer system, A computer program product that executes a command.

17. First expression (R1, R1 Ｉ R1 Ｆ ) represents the examination area to be examined without contrast agent, or after administration of a first amount of contrast agent, the first representation (R1, R1 Ｉ R1 Ｆ ) to provide Second expression (R2, R2 Ｉ , R2 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a second amount of the contrast agent, and the second expression (R2, R2) is greater than the first amount. Ｉ , R2 Ｆ ) to provide The first model (M1) is the first representation (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) based on the third expression (R3, R3 Ｉ , R3 Ｆ It is configured to generate the third representation (R3, R3 Ｉ , R3 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a third amount of the contrast agent, and the first expression (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) to supply the above first model (M1), The second model (M2) was trained during the training process based on the training data (TD). The training data (TD) for each reference target is used for a large number of reference targets, i.e., (i) a reference representation (RR3) generated by the first model representing the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) The measured reference expression of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent (RR3 Ｍ ), including, The training process for each standard subject is The reference representation (RR3) generated by the first model (M1) is supplied to the second model (M2). Revised reference expression (RR3) Ｃ ) receive from the second model (M2), By modifying the model parameters (MP), the modified reference expression (RR3) can be created. Ｃ ) and the measured reference expression (RR3 Ｍ The third representation (R3, R3) includes the step of reducing the difference between the two. Ｉ , R3 Ｆ ) to supply the above second model (M2), A modified third representation (R3) of the inspection area of ​​the subject being inspected, derived from the second model (M2). Ｃ ) receiving The aforementioned modified third expression (R3 Ｃ Outputting and / or storing the modified third representation (R3 Ｃ ) to send to another computer system, The use of contrast agents in radiological examination methods that include the following steps.

18. The aforementioned radiological examination method is magnetic resonance imaging or computed tomography, and the contrast agent is Gd of the compound of formula (I) ３＋ complex, 【Transformation 7】 (I) In the formula, Ar is 【Transformation 8】 and 【Chemistry 9】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ - A base selected from #, Here, * represents linkage to Ar, and # represents linkage to an acetate residue. R １ , R ２ and R ３ Each of these is independently a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ４ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ５ It is a hydrogen atom, Furthermore R ６ It is a hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, Gd of the compound of formula (II) ３＋ complex 【Chemistry 10】 (II) In the formula, Ar is 【Chemistry 11】 and 【Chemistry 12】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ -A group selected from #, where * is linkage to Ar and # is linkage to an acetate residue. R ７ is a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ８ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ９ and R １０ These are each an independent hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, The contrast agent is the following substance, namely, Gadolinium(III) 2-[4,7,10-tris(carboxymethyl)-1,4,7,10-tetraazacyclododea-1-yl]acetic acid, Gadolinium(III) ethoxybenzyldiethylenetriaminepentaacetic acid, Gadolinium(III)2-[3,9-bis[1-carboxylat-4-(2,3-dihydroxypropylamino)-4-oxobutyl]-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-trien-6-yl]-5-(2,3-dihydroxypropylamino)-5-oxopentanoate, Dihydrogen [(±)-4-carboxy-5,8,11-tris(carboxymethyl)-1-phenyl-2-oxa-5,8,11-triazatridecane-13-oate(5-)]gadolinate(2-), Tetragadolinium [4,10-bis(carboxylatomethyl)-7-{3,6,12,15-tetraoxo-16-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]-9,9-bis({[({2-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]propanoyl}aminoacetyl]-amino}methyl)-4,7,11,14-tetraazahepta-decane-2-yl}-1,4,7,10-tetraazacyclododecane-1-yl]acetate, 2,2',2''-(10-{1-carboxy-2-[2-(4-ethoxyphenyl)ethoxy]ethyl}-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate, Gadolinium 2,2',2''-{10-[1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium 2,2',2''-{10-[(1R)-1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium (2S,2'S,2''S)-2,2',2''-{10-[(1S)-1-carboxy-4-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}butyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}tris(3-hydroxypropanoate), Gadolinium 2,2',2''-{10-[(1S)-4-(4-butoxyphenyl)-1-carboxybutyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium(III) 5,8-bis(carboxylatomethyl)-2-[2-(methylamino)-2-oxoethyl]-10-oxo-2,5,8,11-tetraazadodecane-1-carboxylate hydrate, Gadolinium(III) 2-[4-(2-hydroxypropyl)-7,10-bis(2-oxide-2-oxoethyl)-1,4,7,10-tetraazacyclododeca-1-yl]acetate, Gadolinium(III) 2,2',2''-(10-((2R,3S)-1,3,4-trihydroxybutan-2-yl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl) triacetate, Gadolinium-2,2',2''-{(2S)-10-(carboxymethyl)-2-[4-(2-ethoxyethoxy)benzyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, The use according to claim 17, comprising one of gadolinium-2,2',2''-[10-(carboxymethyl)-2-(4-ethoxybenzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl]triacetate.

19. First expression (R1, R1 Ｉ R1 Ｆ ) represents the examination area to be examined without contrast agent, or after administration of a first amount of contrast agent, the first representation (R1, R1 Ｉ R1 Ｆ ) to provide Second expression (R2, R2 Ｉ , R2 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a second amount of the contrast agent, and the second expression (R2, R2) is greater than the first amount. Ｉ , R2 Ｆ ) to provide The first model (M1) is the first representation (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) based on the third expression (R3, R3 Ｉ , R3 Ｆ It is configured to generate the third representation (R3, R3 Ｉ , R3 Ｆ ) represents the examination area of ​​the subject to be examined after administration of a third amount of the contrast agent, and the first expression (R1, R1 Ｉ R1 Ｆ ) and the second representation (R2, R2 Ｉ , R2 Ｆ ) to supply the above first model (M1), The second model (M2) was trained during the training process based on the training data (TD). The training data (TD) for each reference target is used for a large number of reference targets, i.e., (i) a reference representation (RR3) generated by the first model representing the reference region of the reference target after administration of a reference amount of the contrast agent, and (ii) The measured reference expression of the reference area of ​​the reference subject after administration of the reference amount of the contrast agent (RR3 Ｍ ), including, The training process for each standard subject is The reference representation (RR3) generated by the first model (M1) is supplied to the second model (M2). Revised reference expression (RR3) Ｃ ) receive from the second model (M2), By modifying the model parameters (MP), the modified reference expression (RR3) can be created. Ｃ ) and the measured reference expression (RR3 Ｍ The third representation (R3, R3) includes the step of reducing the difference between the two. Ｉ , R3 Ｆ ) to supply the above second model (M2), A modified third representation (R3) of the inspection area of ​​the subject being inspected, derived from the second model (M2). Ｃ ) receiving The aforementioned modified third expression (R3 Ｃ Outputting and / or storing the modified third representation (R3 Ｃ ) to send to another computer system, A contrast agent for use in radiological examination methods that include the following steps.

20. The aforementioned radiological examination method is magnetic resonance imaging or computed tomography, and the contrast agent is Gd of the compound of formula (I) ３＋ complex, 【Chemistry 13】 (I) In the formula, Ar is 【Chemistry 14】 and 【Chemistry 15】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ - A base selected from #, Here, * represents linkage to Ar, and # represents linkage to an acetate residue. R １ , R ２ and R ３ Each of these is independently a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ４ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ５ It is a hydrogen atom, Furthermore R ６ It is a hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, Gd of the compound of formula (II) ３＋ complex 【Chemistry 16】 (II) In the formula, Ar is 【Chemistry 17】 and [Chemistry 18] It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ -A group selected from #, where * is linkage to Ar and # is linkage to an acetate residue. R ７ is a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ８ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ９ and R １０ These are each an independent hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, The contrast agent is the following substance, namely, Gadolinium(III) 2-[4,7,10-tris(carboxymethyl)-1,4,7,10-tetraazacyclododea-1-yl]acetic acid, Gadolinium(III) ethoxybenzyldiethylenetriaminepentaacetic acid, Gadolinium(III)2-[3,9-bis[1-carboxylat-4-(2,3-dihydroxypropylamino)-4-oxobutyl]-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-trien-6-yl]-5-(2,3-dihydroxypropylamino)-5-oxopentanoate, Dihydrogen [(±)-4-carboxy-5,8,11-tris(carboxymethyl)-1-phenyl-2-oxa-5,8,11-triazatridecane-13-oate(5-)]gadolinate(2-), Tetragadolinium [4,10-bis(carboxylatomethyl)-7-{3,6,12,15-tetraoxo-16-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]-9,9-bis({[({2-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]propanoyl}aminoacetyl]-amino}methyl)-4,7,11,14-tetraazahepta-decane-2-yl}-1,4,7,10-tetraazacyclododecane-1-yl]acetate, 2,2',2''-(10-{1-carboxy-2-[2-(4-ethoxyphenyl)ethoxy]ethyl}-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate, Gadolinium 2,2',2''-{10-[1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium 2,2',2''-{10-[(1R)-1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium (2S,2'S,2''S)-2,2',2''-{10-[(1S)-1-carboxy-4-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}butyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}tris(3-hydroxypropanoate), Gadolinium 2,2',2''-{10-[(1S)-4-(4-butoxyphenyl)-1-carboxybutyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium(III) 5,8-bis(carboxylatomethyl)-2-[2-(methylamino)-2-oxoethyl]-10-oxo-2,5,8,11-tetraazadodecane-1-carboxylate hydrate, Gadolinium(III) 2-[4-(2-hydroxypropyl)-7,10-bis(2-oxide-2-oxoethyl)-1,4,7,10-tetraazacyclododeca-1-yl]acetate, Gadolinium(III) 2,2',2''-(10-((2R,3S)-1,3,4-trihydroxybutan-2-yl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl) triacetate, Gadolinium-2,2',2''-{(2S)-10-(carboxymethyl)-2-[4-(2-ethoxyethoxy)benzyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, A contrast agent for use according to claim 18, comprising one of gadolinium-2,2',2''-[10-(carboxymethyl)-2-(4-ethoxybenzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl]triacetate.

21. A kit comprising the computer program product and contrast agent according to claim 16, wherein the contrast agent is preferably Gd of the compound of formula (I) ３＋ complex, 【Chemistry 19】 (I) In the formula, Ar is 【Chemistry 20】 and 【Chemistry 21】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ - A base selected from #, Here, * represents linkage to Ar, and # represents linkage to an acetate residue. R １ , R ２ and R ３ Each of these is independently a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ４ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-, (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ )-O-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ５ It is a hydrogen atom, Furthermore R ６ It is a hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, Gd of the compound of formula (II) ３＋ complex 【Chemistry 22】 (II) In the formula, Ar is 【Chemistry 23】 and 【Chemistry 24】 It is a base selected from, # is a concatenation to X, X is CH ２ , (CH ２ ) ２ , (CH ２ ) ３ , (CH ２ ) ４ and *-(CH ２ ) ２ -O-CH ２ -A group selected from #, where * is linkage to Ar and # is linkage to an acetate residue. R ７ is a hydrogen atom, or C １ ~C ３ Alkyl, -CH ２ OH, - (CH ２ ) ２ OH and -CH ２ OCH ３ It is a base selected from, R ８ C ２ ~C ４ Alkoxy, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-, (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O- and (H ３ C-CH ２ O)-(CH ２ ) ２ -O-(CH ２ ) ２ -O-(CH ２ ) ２ It is a base selected from -O-, R ９ and R １０ These are each an independent hydrogen atom, or comprising its stereoisomers, tautomers, hydrates, solvates or salts, or mixtures thereof, The contrast agent is the following substance, namely, Gadolinium(III) 2-[4,7,10-tris(carboxymethyl)-1,4,7,10-tetraazacyclododea-1-yl]acetic acid, Gadolinium(III) ethoxybenzyldiethylenetriaminepentaacetic acid, Gadolinium(III)2-[3,9-bis[1-carboxylat-4-(2,3-dihydroxypropylamino)-4-oxobutyl]-3,6,9,15-tetraazabicyclo[9.3.1]pentadeca-1(15),11,13-trien-6-yl]-5-(2,3-dihydroxypropylamino)-5-oxopentanoate, Dihydrogen [(±)-4-carboxy-5,8,11-tris(carboxymethyl)-1-phenyl-2-oxa-5,8,11-triazatridecane-13-oate(5-)]gadolinate(2-), Tetragadolinium [4,10-bis(carboxylatomethyl)-7-{3,6,12,15-tetraoxo-16-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]-9,9-bis({[({2-[4,7,10-tris(carboxylatomethyl)-1,4,7,10-tetraazacyclododecane-1-yl]propanoyl}aminoacetyl]-amino}methyl)-4,7,11,14-tetraazahepta-decane-2-yl}-1,4,7,10-tetraazacyclododecane-1-yl]acetate, 2,2',2''-(10-{1-carboxy-2-[2-(4-ethoxyphenyl)ethoxy]ethyl}-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetate, Gadolinium 2,2',2''-{10-[1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium 2,2',2''-{10-[(1R)-1-carboxy-2-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}ethyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium (2S,2'S,2''S)-2,2',2''-{10-[(1S)-1-carboxy-4-{4-[2-(2-ethoxyethoxy)ethoxy]phenyl}butyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}tris(3-hydroxypropanoate), Gadolinium 2,2',2''-{10-[(1S)-4-(4-butoxyphenyl)-1-carboxybutyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, Gadolinium(III) 5,8-bis(carboxylatomethyl)-2-[2-(methylamino)-2-oxoethyl]-10-oxo-2,5,8,11-tetraazadodecane-1-carboxylate hydrate, Gadolinium(III) 2-[4-(2-hydroxypropyl)-7,10-bis(2-oxide-2-oxoethyl)-1,4,7,10-tetraazacyclododeca-1-yl]acetate, Gadolinium(III) 2,2',2''-(10-((2R,3S)-1,3,4-trihydroxybutan-2-yl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl) triacetate, Gadolinium-2,2',2''-{(2S)-10-(carboxymethyl)-2-[4-(2-ethoxyethoxy)benzyl]-1,4,7,10-tetraazacyclododecane-1,4,7-triyl}triacetate, A kit containing one of the following: gadolinium-2,2',2''-[10-(carboxymethyl)-2-(4-ethoxybenzyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl]triacetate.

Images