Medical image processing apparatus, method of operating medical image processing apparatus, and program

Abstract

To provide a medical image processing apparatus, a method of operating the medical image processing apparatus and a program that achieve stable performance in characteristics analysis of a medical image using contrast state.SOLUTION: A medical image (IIN) generated by performing contrast radiography is acquired (P1), a contrast state of the medical image is estimated (P2) based on analysis of the medical image, and characteristics analysis of a region of interest included in the medical image is executed (P3) based on characteristics in the region of interest using contrast state information (INF) representing the contrast state of the medical image.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The present invention relates to a medical image processing apparatus, an operating method for the medical image processing apparatus, and a program. [Background technology]

[0002] Dynamic contrast-enhanced CT is a well-known imaging technique that combines angiography using contrast agents with X-ray CT. For example, in dynamic contrast-enhanced CT of the liver, CT images are acquired by injecting a contrast agent and taking multiple images at different contrast phases, and then observing the changes in the density of lesions in the CT images. CT is an abbreviation for Computed Tomography.

[0003] Patent Document 1 describes an image discrimination device that automatically discriminates between contrast-enhanced and non-contrast-enhanced images. The device detects a first region from acquired image data that is not affected by a contrast agent, identifies a second region that is in a specified relative positional relationship with the first region and is affected by the contrast agent, and determines whether the image data is contrast-enhanced or not depending on whether the CT value of the second region is higher than a specified value.

[0004] Non-Patent Document 1 describes a deep learning model applied to lesion classification of liver tumors. The deep learning model described in this document takes as input images images in which tumor regions are cropped from images with known contrast phases (non-contrast, arterial phase, and equilibrium phase), and outputs classifications of five lesions: classical hepatocellular carcinoma, malignant tumors other than hepatocellular carcinoma, benign tumors, hemangiomas, and cysts. Note that benign tumors include images in which it is impossible to distinguish between benign and malignant. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2011-136030 [Non-patent literature]

[0006] [Non-Patent Document 1] Yasaka K, Akai H, Abe O, Kiryu S. Deep Learning with Convolutional Neural Network for Differentiation of Liver Masses at Dynamic Contrast-enhanced CT: A Preliminary Study. Radiology. 2018 Mar;286(3):887-896. doi: 10.1148 / radiol.2017170706. Epub 2017 Oct 23. PMID: 29059036. Summary of the Invention [Problem to be solved by the invention]

[0007] However, even in images captured at the same time elapsed from the start of contrast agent injection, the contrast state may differ depending on the subject's physique, physical condition, etc. In this case, when the contrast agent injection start time information included in the metadata is used to determine the contrast agent injection phase and perform characterization analysis, the performance of the characterization analysis may be degraded due to the variation in the contrast agent injection phase.

[0008] Furthermore, if the metadata does not include information about the injection start time of the contrast agent, the characterization will be performed without determining the contrast agent injection phase.

[0009] The system described in Patent Document 1 determines whether the acquired image data is contrast-enhanced imaging, and it is difficult to determine the contrast-enhanced phase of the captured image.

[0010] In the method described in Non-Patent Document 1, the contrast enhancement phase is assumed to be known. Furthermore, it is not possible to deal with variations in contrast enhancement phases that are affected by factors such as the physique of the subject, or cases where there is no information regarding the identification of the contrast enhancement phase.

[0011] The present invention has been made in consideration of the above circumstances, and aims to provide a medical image processing device, an operating method for the medical image processing device, and a program that achieve stable performance in analyzing the characteristics of medical images using contrast-enhanced conditions. [Means for solving the problem]

[0012] The medical image processing device according to the first aspect includes one or more processors and one or more memories storing programs to be executed by the one or more processors, and the one or more processors execute instructions of the programs to acquire medical images generated by performing contrast imaging, estimate the contrast state of the medical images based on an analysis of the medical images, and perform a characteristic analysis of a region of interest included in the medical images using contrast state information representing the contrast state of the medical images.

[0013] According to the medical image processing device of the first aspect, a medical image generated by performing contrast imaging is analyzed to estimate a contrast state, and the estimated contrast state is used to perform a characteristic analysis of a region of interest included in the medical image. This suppresses performance degradation of the characteristic analysis due to variations in the contrast state, and can achieve stable performance in the characteristic analysis of a medical image using the contrast state. Furthermore, even if the medical image does not contain information identifying the contrast state, the characteristic analysis of the medical image using the contrast state can be performed.

[0014] An example of estimating a contrast-enhanced state is estimating a contrast-enhanced phase. The contrast-enhanced state may include a non-contrast-enhanced state.

[0015] In a second aspect, in the medical image processing device of the first aspect, one or more processors select medical images using contrast enhancement status information.

[0016] According to this aspect, it is possible to select medical images corresponding to contrast conditions suitable for characterization.

[0017] In a third aspect, in the medical image processing device of the second aspect, the one or more processors select medical images having contrast state information suitable for characterization analysis.

[0018] According to this aspect, application of medical images that are not suitable for the characterization to the characterization is suppressed, thereby stabilizing the performance of the characterization.

[0019] In a fourth aspect, in the medical image processing device of the second or third aspect, the one or more processors select medical images according to restrictions on input images in the characteristic analysis.

[0020] According to this aspect, it is possible to select input images for the property analysis in accordance with the limitations on the input images for the property analysis.

[0021] In a fifth aspect, in a medical image processing device of any one of the second to fourth aspects, one or more processors select medical images for each contrast state information corresponding to each of two or more predefined types of contrast states.

[0022] According to this aspect, even when the contrast conditions suitable for the characterization are limited, it is possible to select medical images corresponding to the contrast conditions suitable for the characterization.

[0023] A sixth aspect is a medical image processing device according to any one of the second to fifth aspects, wherein the one or more processors select medical images having contrast state information corresponding to a contrast state excluding non-contrast.

[0024] According to this aspect, even when a characterization analysis that is not suitable for non-contrast imaging is applied, a medical image corresponding to a contrast imaging state that is suitable for characterization can be selected.

[0025] In a seventh aspect, in the medical image processing device according to any one of the first to sixth aspects, one or more processors extract a region of interest from the acquired medical image and perform a characteristic analysis of the region of interest.

[0026] According to this aspect, even when a medical image is acquired in which the region of interest has not yet been extracted, a property analysis can be performed on the region of interest.

[0027] In an eighth aspect, in the medical image processing device of any one of the first to seventh aspects, one or more processors estimate the contrast state of the acquired medical image using a trained learning model.

[0028] According to this aspect, it is expected that the accuracy of the contrast state estimation will be improved.

[0029] Examples of trained learning models include deep learning models such as neural networks.

[0030] In a ninth aspect, in a medical image processing device of any one of the first to eighth aspects, one or more processors perform a characteristic analysis of a region of interest contained in an acquired medical image using a trained learning model.

[0031] According to this aspect, the accuracy of the property analysis can be improved.

[0032] In a tenth aspect, in the medical image processing device of the ninth aspect, one or more processors use a trained learning model to extract features from each region of interest included in a medical image for each contrast state, and perform a characteristic analysis of the region of interest included in the medical image based on feature data in which the features of the regions of interest for each contrast state are concatenated.

[0033] According to this aspect, it is possible to perform a property analysis that takes into account the features of the region of interest included in medical images in a plurality of contrast conditions.

[0034] In an eleventh aspect, in the medical image processing device of the tenth aspect, one or more processors use, as a trained learning model, a feature extraction model that extracts features from each of the regions of interest included in the medical images for each contrast state, to extract features from each of the regions of interest included in the medical images for each contrast state, and link the features of the regions of interest included in the medical images for each contrast state, and perform a characteristic analysis of the regions of interest included in the medical images using, as a trained learning model, a classification model that classifies feature data in which the features of the regions of interest included in the medical images for each contrast state are linked.

[0035] According to this aspect, feature extraction and property analysis can be performed using a separate trained learning model for each process.

[0036] In a twelfth aspect, in the medical image processing device of the eleventh aspect, the one or more processors average some of the feature amounts of the region of interest included in the medical image for each contrast enhancement state.

[0037] According to this aspect, the processing load when linking feature amounts of regions of interest included in medical images for each contrast state is reduced.

[0038] In a thirteenth aspect, in the medical image processing device of the twelfth aspect, one or more processors calculate weights for each feature of the region of interest contained in the medical image for each contrast state using a weight calculation model as a trained learning model that calculates weights used when calculating a weighted average of the feature of the region of interest contained in the medical image for each contrast state.

[0039] According to this aspect, it is possible to estimate the contribution of the region of interest included in the medical image for each contrast state in the characterization analysis.

[0040] In a 14th aspect, in the medical image processing device of the 10th or 11th aspect, when a part of a region of interest included in a medical image for each contrast state is missing, one or more processors use a region of interest included in a medical image for a contrast state that has features similar to those of the region of interest included in the medical image for the missing contrast state instead of the region of interest included in the medical image for the missing contrast state.

[0041] According to this aspect, even if a region of interest in a part of the contrast enhancement state is missing, it is possible to perform a property analysis of the region of interest included in the medical image based on the contrast enhancement state.

[0042] The operating method of a medical image processing device according to the 15th aspect is a method of operating a medical image processing device to which a computer is applied, in which the medical image processing device executes the steps of: acquiring a medical image generated by performing contrast imaging; estimating the contrast state of the medical image based on analysis of the medical image; and performing a characteristic analysis of a region of interest included in the medical image using contrast state information representing the contrast state of the medical image.

[0043] According to the operating method of the medical image processing device of the fifteenth aspect, it is possible to obtain the same effects as the medical image processing device of the present disclosure. The constituent elements of the medical image processing device of the other aspects can be applied to the constituent elements of the operating method of the medical image processing device of the other aspects.

[0044] The program according to the sixteenth aspect is a program that enables a computer to realize the functions of: acquiring a medical image generated by performing contrast imaging; estimating the contrast state of the medical image based on analysis of the medical image; and performing a characteristic analysis of a region of interest included in the medical image using contrast state information that represents the contrast state of the medical image.

[0045] According to the program of the sixteenth aspect, it is possible to obtain the same effects as those of the medical image processing device of the present disclosure. The components of the medical image processing device of the other aspects may be applied to the components of the program of the other aspects. [Effects of the Invention]

[0046] According to the present invention, a medical image generated by performing contrast imaging is analyzed to estimate a contrast state, and the estimated contrast state is used to perform a characteristic analysis of a region of interest included in the medical image. This suppresses performance degradation of the characteristic analysis due to variations in the contrast state, and achieves stable performance in the characteristic analysis of medical images using the contrast state. Furthermore, even if the medical image does not contain information identifying the contrast state, the characteristic analysis of medical images using the contrast state can be performed. [Brief explanation of the drawings]

[0047] [Figure 1] Figure 1 is an explanatory diagram of dynamic contrast-enhanced CT. [Figure 2] FIG. 2 is a graph showing the relationship between the contrast enhancement phase and the CT value. [Figure 3] FIG. 3 is a conceptual diagram showing an outline of the processing applied to the property analyzer according to the first embodiment. [Figure 4] FIG. 4 is a functional block diagram showing an outline of the processing functions of the property analyzer according to the first embodiment. [Figure 5] FIG. 5 is a block diagram schematically illustrating an example of the hardware configuration of the property analyzer according to the first embodiment. [Figure 6] FIG. 6 is a flowchart showing the procedure of the property analysis method according to the first embodiment. [Figure 7] FIG. 7 is a conceptual diagram showing an outline of the processing applied to the property analyzer according to the second embodiment. [Figure 8] FIG. 8 is a functional block diagram showing an outline of the processing functions of the property analyzer according to the second embodiment. [Figure 9] FIG. 9 is a block diagram schematically illustrating an example of the hardware configuration of the property analyzer according to the second embodiment. [Figure 10] FIG. 10 is a flowchart showing the procedure of the property analysis method according to the second embodiment. [Figure 11] FIG. 11 is a conceptual diagram showing a specific example of the selection process. [Figure 12] FIG. 12 is a conceptual diagram showing an example of contrast phase estimation. [Figure 13] FIG. 13 is a conceptual diagram showing a specific example of property analysis. [Figure 14] FIG. 14 is a conceptual diagram showing a modified example of the property analysis process shown in FIG. [Figure 15] FIG. 15 is a conceptual diagram showing a specific example of the characteristic analysis process when CT images in some contrast phases are missing. [Figure 16] FIG. 16 is a conceptual diagram showing another specific example of the property analysis process. [Figure 17] FIG. 17 is a block diagram showing an example of the configuration of a medical information system in which a characteristic analyzer is used. DETAILED DESCRIPTION OF THE INVENTION

[0048] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings. In this specification, the same components are designated by the same reference numerals, and redundant explanations will be omitted where appropriate.

[0049] [About dynamic contrast-enhanced CT] Figure 1 is an explanatory diagram of dynamic contrast-enhanced CT. Figure 1 shows a schematic diagram of CT images taken at each contrast phase using dynamic contrast-enhanced CT. Dynamic contrast-enhanced CT is a technique in which a contrast agent such as an iodine contrast agent is injected into a vein in the arm, and then the same area is repeatedly photographed multiple times to observe changes over time in the CT images.

[0050] In other words, dynamic contrast-enhanced CT is an imaging method in which an organ is imaged multiple times at timings that take into account the hemodynamics of the organ. The imaging timing is determined depending on the organ. For example, when imaging the liver, imaging is performed at timings that take into account the hemodynamics of the liver. Note that in dynamic contrast-enhanced CT, imaging may be performed before the start of contrast agent injection, and non-contrast CT images similar to those obtained in plain CT may be obtained.

[0051] Figure 1 illustrates a non-contrast CT image I1, an arterial phase CT image I2, a portal venous phase CT image I3, and an equilibrium phase CT image I4. Figure 1 illustrates a CT image of any one slice image from a slice image group that includes multiple slice images in each contrast phase.

[0052] The horizontal axis in Figure 1 is the time axis, with the start time of contrast agent injection being 0 seconds, and the time axis is in seconds. Figure 1 shows an arterial phase CT image I2 taken approximately 35 seconds after the injection of the contrast agent, a portal venous phase CT image I3 taken approximately 80 seconds after the injection of the contrast agent, and an equilibrium phase CT image I4 taken approximately 150 seconds after the injection of the contrast agent.

[0053] Tumors appear differently in CT images due to differences in the contrast enhancement phase. Therefore, accurate information on the contrast enhancement phase is required for characterization. In the example shown in Figure 1, there is a high density in the arterial phase CT image I2, and the characterization result can be determined to be early density.

[0054] Here, the term "image" in this specification may include not only the image itself but also image data, which is a signal representing the image. Furthermore, the term "injection start time" may be read as "injection start timing."

[0055] Figure 2 is a graph showing the relationship between the contrast enhancement phase and the CT value. The horizontal axis of the graph shown in Figure 2 is the time axis, and the time axis is measured in seconds. The vertical axis shown in the same figure is the CT value axis. Curve 1 shows the change in CT value over time in the artery. Curve 2 shows the change in CT value over time in the portal vein. Curve 3 shows the change in CT value over time in the liver.

[0056] The contrast-enhanced phase is the state after a specific time has elapsed since the injection of the contrast agent. Dynamic contrast-enhanced CT of the liver defines the arterial phase, portal venous phase, and equilibrium phase. For example, the arterial phase indicates the state in which the contrast agent flows mainly into the arterial venous phase.

[0057] Contrast medium injected into a vein reaches the abdominal artery 30 to 40 seconds after the start of injection. Period t1 shown in Figure 2 corresponds to the arterial phase. Contrast medium injected into a vein reaches the portal vein 60 to 80 seconds after the start of injection. Period t2 shown in Figure 2 corresponds to the portal vein phase. Contrast medium injected into a vein reaches equilibrium between the intravascular and extracellular fluid concentrations 150 to 200 seconds after the start of injection. This state is the equilibrium phase. Period t3 shown in Figure 2 corresponds to the equilibrium phase.

[0058] While Figure 2 shows an example of an imaging phase for the liver, examples of imaging phases for the kidney include the corticomedullary phase, the parenchymal phase, and the excretory phase. As with the liver, the relationship between the imaging phase and the elapsed time from the start of injection of the contrast agent is also defined for the kidney.

[0059] [Property analysis issues] Analysis of the characteristics of CT images taken using dynamic contrast-enhanced CT requires accurate information on the contrast-enhanced phase, but identifying the contrast-enhanced phase presents challenges, such as missing metadata and differences in subject physique.

[0060] The metadata loss refers to the loss of information about the start time of contrast agent injection in the metadata. If the information about the start time of contrast agent injection in the metadata is missing or corrupted, it is difficult to determine the elapsed time from the start time of contrast agent injection, which makes it difficult to accurately determine the contrast agent injection phase.

[0061] In addition, differences in the subjects' physiques mean that the contrast agent spreads differently due to differences in the body size, heart rate, etc. of each subject, and even if the time when the contrast agent injection started is known, the contrast phase based on the elapsed time from the time when the contrast agent injection started cannot be accurately identified.

[0062] For example, the greater the amount of blood circulating per unit time, such as cardiac output, which represents the volume of blood pumped per unit time, the faster the contrast agent reaches the subject, the lower the maximum CT value, and the faster the time to reach the maximum CT value.

[0063] However, it is difficult to know the amount of blood circulating per unit time in advance, and there is a possibility that the amount of blood circulating per unit time may increase compared to when the subject is at rest due to tension, etc. If the amount of blood circulating per unit time varies in this way, the movement time of the contrast agent may also vary.

[0064] Below, we will explain a characteristic analysis device that does not rely on information about the start time of contrast agent injection in the metadata, but obtains accurate information about the contrast phase, which suppresses the influence of variations in the travel time of the contrast agent due to the subject's physique, etc., and performs characteristic analysis using the information about the contrast phase.

[0065] [Character analysis device according to the first embodiment] 3 is a conceptual diagram showing an outline of the processing applied to the characteristic analysis device according to the first embodiment. In the processing applied to the characteristic analysis device shown in the figure, 3D CT data generated by imaging using dynamic contrast-enhanced CT is acquired in CT image acquisition processing P1.

[0066] FIG. 3 shows the three-dimensional CT data acquired in the CT image acquisition process P1 as a two-dimensional CT image I IN Hereafter, CT image I IN The term can be interpreted as three-dimensional CT data.

[0067] In the CT image acquisition process P1, a CT image I in which the contrast phase is not specified is IN In the CT image acquisition process P1, a CT image I is acquired from a storage device that stores medical images. IN is obtained.

[0068] Figure 3 shows four types of CT images I acquired before the start of contrast injection, t1 seconds after the start of contrast injection, t2 seconds after the start of contrast injection, and t3 seconds after the start of contrast injection. IN Here is an example of how CT image I is acquired. IN is the time from the start of contrast injection to iIt is an arbitrary CT image in a CT image group containing multiple CT images acquired after 1 second. Note that i represents the number of times of imaging in chronological order, and an integer of 1 or more is used. Also, CT image I shown in Figure 3 IN includes non-contrast CT images.

[0069] In contrast phase estimation process P2, the CT image I acquired in CT image acquisition process P1 is IN Image analysis was performed on each of the CT images, and the contrast-enhanced phase was estimated. IN Contrast phase information for each NF The image analysis here may include processing using the pixel values ​​of the pixels that make up the image. NF To obtain the contrast phase information, NF This may include the meaning of the generation of

[0070] CT image I with estimated contrast phase IN is the contrast phase information I corresponding to the estimated contrast phase. NF Using the region of interest R OI The property analysis process P3 is performed, the information output process P4 is performed, and the analysis result A R Figure 3 shows the analysis result A of the property analysis. R Examples include early staining, washout, and filming.

[0071] In the CT image acquisition process P1 shown in FIG. 3, an arbitrary slice image among multiple slice images sampled at equal intervals from 3D CT data of a patient taken using a CT device is taken as a CT image I IN The slice image may be referred to as a tomographic image. In other words, the slice image may be understood as a cross-sectional image that is essentially a two-dimensional image.

[0072] 4 is a functional block diagram showing an outline of the processing functions of the property analysis device according to the first embodiment. The property analysis device 10 can be realized using computer hardware and software.

[0073] The characteristic analysis device 10 includes a CT image acquisition unit 12, a contrast phase estimation unit 14, a region of interest extraction unit 16, a characteristic analysis unit 18, and an information output unit 20.

[0074] The CT image acquisition unit 12 acquires the CT image I shown in FIG. IN The CT image acquisition unit 12 acquires the CT image I IN As the slice image group including a plurality of slice images generated for each of the specified number of imaging operations, one arbitrary slice image may be acquired, or a plurality of slice images may be acquired.

[0075] The contrast phase estimation unit 14 estimates the CT image I acquired by the CT image acquisition unit 12. IN Image analysis was performed on the CT image I IN The contrast enhancement phase estimation unit 14 estimates the contrast enhancement phase of the CT image I. IN Region of interest extracted from R OI Image analysis was performed on the CT image I IN The contrast enhancement phase may be estimated.

[0076] That is, the contrast phase estimation unit 14 estimates the CT image I IN For each of the above, the contrast phase information I shown in Figure 3 NF and obtain contrast phase information I NF CT image I IN and transmits it to the characteristic analysis unit 18. The contrast phase estimation unit 14 IN and contrast phase information I NF may be stored in association with each other.

[0077] In addition, CT image I IN The contrast phase of the CT image I IN The contrast-enhanced phase information I in the embodiment is an example of a contrast-enhanced state. NF is an example of contrast state information.

[0078] The region of interest extraction unit 16 extracts the CT image I acquired by the CT image acquisition unit 12. IN From the region of interest ROI The region of interest extraction unit 16 extracts the region of interest R OI The region of interest extraction unit 16 extracts a lesion region including a lesion such as a tumor from the CT image I. IN and the region of interest R OI Using a trained model that has learned the relationship between IN From the region of interest R OI The region of interest extraction unit 16 may extract the region of interest I from the CT image I designated by the user. IN Using the position information in CT image I IN From the region of interest R OI may be extracted.

[0079] An example of a learning model is a deep learning model such as a convolutional neural network, commonly referred to as CNN, which is an abbreviation for convolutional neural network.

[0080] The CT image acquisition unit 12 acquires the region of interest R OI CT image I from which IN When the region of interest R is acquired in advance, the processing of the region of interest extraction unit 16 is omitted. OI CT image I from which IN When acquiring the region of interest, the region of interest extraction unit 16 may not be provided.

[0081] The region of interest extraction unit 16 extracts the region of interest from the CT image I IN Region of interest in R OI The information representing the specified conditions is obtained, and the region of interest R OI Based on the specified conditions, CT image I IN From the region of interest R OI may be extracted.

[0082] The characteristic analysis unit 18 analyzes the CT image I IN Contrast Phase Information I NF Using CT image I IN Region of interest extracted from R OI The property analysis unit 18 performs a property analysis of the region of interest R OI and the region of interest R OIA trained learning model that has learned the relationship between the characteristics of the object and the target object may be applied. An example of the trained learning model is CNN. The characteristic analysis unit 18 to which the trained learning model is applied is expected to improve the accuracy of the characteristic analysis.

[0083] The information output unit 20 outputs the analysis results of the characteristic analysis performed by the characteristic analysis unit 18. The information output unit 20 outputs the acquired CT image I IN and the analysis results of the property analysis may be stored in the storage unit in association with each other.

[0084] The information output unit 20 outputs the CT image I IN Region of interest in R OI For example, the information output unit 20 functions as an output interface that outputs information representing the characteristics of the CT image I. IN Region of interest in R OI It may also function as an output interface that provides the properties of the sensor to other processing units.

[0085] The information output unit 20 may include at least one processing unit for generating data for display and converting data for external transmission, etc. The analysis results of the property analysis device 10 may be displayed using a display device, etc.

[0086] The characteristic analysis device 10 may be incorporated into a medical image processing device for processing medical images acquired in a medical institution such as a hospital. Furthermore, the processing function of the characteristic analysis device 10 may be provided as a cloud service.

[0087] [Explanation of medical images used as input] The DICOM standard, which defines the format and communication protocol for medical images, defines a series ID within a study ID, which is an identification code used to identify the type of examination. Note that ID is an abbreviation for identification. Also, medical images are synonymous with medical images.

[0088] For example, when performing contrast imaging of a patient's liver using dynamic contrast-enhanced CT, CT images of the area including the liver are taken multiple times at different imaging timings. Examples of multiple imaging times include the first imaging before contrast injection, the second imaging 35 seconds after contrast injection, the third imaging 80 seconds after contrast injection, and the fourth imaging 150 seconds after contrast injection.

[0089] These four scans are taken, and four types of CT data are obtained. The CT data referred to here is three-dimensional data made up of multiple consecutive slice images, and the collection of multiple slice images that make up the three-dimensional data is called an image series.

[0090] The four types of CT data obtained by performing a series of imaging including the above four imaging sessions are each assigned the same study ID and a separate series ID.

[0091] For example, for a liver contrast imaging test on a particular patient, Study 1 is assigned as the study ID, and Series 1 is assigned as the series ID for CT data obtained by imaging before contrast injection, Series 2 for CT data obtained by imaging 35 seconds after contrast injection, Series 3 for CT data obtained by imaging 80 seconds after contrast injection, and Series 4 for CT data obtained by imaging 150 seconds after contrast injection, and so on. A unique ID is assigned to each series.

[0092] Therefore, CT data can be identified by combining the study ID and series ID. However, in actual CT data, the correspondence between the series ID and the imaging timing may not be clearly understood. The imaging timing here may be interpreted as the elapsed time after the injection of the contrast agent.

[0093] Furthermore, because three-dimensional CT data has a large data size, when CT data is used as input data directly to perform processing such as estimating the contrast-enhanced phase, it may be difficult to process in terms of processing time, processing load, etc. Therefore, the characteristic analysis device 10 can estimate the contrast-enhanced phase based on image analysis using one or more slice images in the same image series as input.

[0094] [Example of hardware configuration for a property analyzer] 5 is a block diagram showing an example of the hardware configuration of the property analysis device according to the first embodiment. The property analysis device 10 can be realized by a computer system configured using one or more computers. Here, an example is shown in which one computer executes a program to realize various functions of the property analysis device 10.

[0095] The type of computer that functions as the property analyzer 10 is not particularly limited, and may be a server computer, a workstation, a personal computer, a tablet terminal, etc. The computer may also be a virtual machine.

[0096] The property analysis device 10 includes a processor 30, a computer-readable medium 32 which is a non-transitory tangible entity, a communication interface 34, an input / output interface 36, and a bus 38. Note that IF shown in Fig. 5 represents an interface.

[0097] The processor 30 includes a CPU (Central Processing Unit). The processor 30 may also include a GPU (Graphics Processing Unit). The processor 30 is connected to a computer-readable medium 32, a communication interface 34, and an input / output interface 36 via a bus 38. The processor 30 reads various programs, data, etc. stored in the computer-readable medium 32 and executes various processes.

[0098] The computer-readable medium 32 includes a memory 40, which is a main storage device, and a storage 42, which is an auxiliary storage device. The storage 42 may be configured using a hard disk drive, a solid-state drive, an optical disk, a magneto-optical disk, or a semiconductor memory. The storage 42 may be configured using an appropriate combination of hard disk drives and the like. The storage 42 stores various programs, data, and the like.

[0099] A hard disk drive may be referred to as an HDD, which is an abbreviation of the English term Hard Disk Drive, and a solid state drive may be referred to as an SSD, which is an abbreviation of the English term Solid State Drive.

[0100] The memory 40 is used as a working area for the processor 30 and as a storage unit that temporarily stores programs and various data read from the storage 42. A program stored in the storage 42 is loaded into the memory 40, and instructions of the program are executed by the processor 30, which functions as a processing unit that performs various processes defined by the program. The memory 40 stores a contrast phase estimation program 50, a region of interest extraction program 52, a characteristic analysis program 54, various data, and the like, which are executed by the processor 30.

[0101] The contrast phase estimation program 50 causes the processor 30 to execute the contrast phase estimation process performed using the contrast phase estimation unit 14 shown in Fig. 4. The contrast phase estimation program 50 may include a trained learning model.

[0102] The region of interest extraction program 52 causes the processor 30 to execute the region of interest extraction process performed using the region of interest extraction unit 16. The region of interest extraction program 52 may include a trained learning model.

[0103] The property analysis program 54 causes the processor 30 to execute the property analysis process performed using the property analysis unit 18. The property analysis program 54 may include a trained learning model. Each program shown in FIG. 5 includes one or more instructions. The processor 30 executes the instructions included in each program to realize the function corresponding to each program.

[0104] The communication interface 34 performs communication processing with an external device via a wired or wireless connection, and exchanges information with the external device. The property analysis device 10 is connected to a communication line via the communication interface 34. The communication line may be a local area network or a wide area network. The communication interface 34 can serve as a data acquisition unit that accepts input of data such as images. The communication line is not shown in the figure.

[0105] The property analysis device 10 may include an input device 60 and a display device 62. The input device 60 and the display device 62 are connected to the bus 38 via the input / output interface 36. Examples of the input device 60 include a keyboard, a mouse, a multi-touch panel, other pointing devices, and a voice input device. The input device 60 may be an appropriate combination of a keyboard and the like.

[0106] The display device 62 is an output interface that displays various types of information. Examples of the display device 62 include a liquid crystal display, an organic EL display, and a projector. The display device 62 may be an appropriate combination of liquid crystal displays and the like. Organic EL is referred to as OEL, which is an abbreviation of organic electroluminescence in English.

[0107] [Procedure for property analysis method] 6 is a flowchart showing the procedure of the characteristic analysis method according to the first embodiment. In the CT image acquisition step S10, the CT image acquisition unit 12 shown in FIG. 4 acquires the CT image I shown in FIG. INThe CT image acquisition step S10 shown in Fig. 6 corresponds to the CT image acquisition process P1 shown in Fig. 3. After the CT image acquisition step S10, the process proceeds to a contrast phase estimation step S12.

[0108] In the contrast phase estimation step S12, the contrast phase estimation unit 14 estimates the CT image I IN Based on the image analysis of the acquired CT image I IN The contrast enhancement phase estimation step S12 shown in Fig. 6 corresponds to the contrast enhancement phase estimation process P2 shown in Fig. 3. After the contrast enhancement phase estimation step S12, the process proceeds to a region of interest extraction step S14.

[0109] In the region of interest extraction step S14, the region of interest extraction unit 16 extracts the region of interest from the CT image I IN From the region of interest R OI After the region of interest extraction step S14, the process proceeds to the characteristic analysis step S16. The region of interest extraction step S14 may be performed in parallel with the contrast phase estimation step S12, or the order of the step S14 and the contrast phase estimation step S12 may be interchanged.

[0110] In the CT image acquisition step S10, the region of interest R OI CT image I from which IN is acquired, the region of interest extraction step S14 is omitted, and the process proceeds to the characteristic analysis step S16 after the contrast phase estimation unit 14.

[0111] In the property analysis step S16, the property analysis unit 18 analyzes the CT image I IN Region of interest R OI In the property analysis step S16, the CT image I IN The property analysis result may be stored in association with the property analysis step S16 shown in Fig. 6. The property analysis step S16 corresponds to the property analysis process P3 shown in Fig. 3. After the property analysis step S16, the process proceeds to an information output step S18.

[0112] In the information output step S18, the information output unit 20 outputs the analysis results of the property analysis performed in the property analysis step S16. The output of the analysis results in the information output step S18 may be performed in a manner that visualizes the analysis results, such as by displaying them on the display device 62. The information output step S18 shown in Fig. 6 corresponds to the information output processing P4 shown in Fig. 3. After the information output step S18, the procedure of the property analysis method ends.

[0113] After the information output step S18, the next CT image I IN Wait for input of the next CT image I IN When the input is made, each step from the CT image acquisition step S10 to the information output step S18 may be executed. After the information output step S18, the next CT image I IN The next CT image I is input during the specified period. IN If no input is made, the procedure of the property analysis method may be terminated.

[0114] The characteristic analysis method described in the embodiment is an example of an operation method of a medical image processing apparatus to which a computer is applied.

[0115] [Effects of the first embodiment] The property analysis device and property analysis method according to the first embodiment can achieve the following advantageous effects.

[0116] CT image I IN Image analysis was performed on the CT image I IN The contrast enhancement phase of CT image I is estimated. IN Contrast Phase Information I NF Using CT image I IN Region of interest R OI This allows for the analysis of the characteristics of the CT image I IN Even if the information on the start time of contrast injection is missing from the metadata, the contrast phase can be estimated and the CT image I IN Region of interest R OI The performance of characterization using the contrast phase for the stenosis is stabilized.

[0117] In addition, the variation in contrast phase due to differences in the subject's physique, etc. is suppressed, and the CT image I IN Region of interest R OI Contrast-enhanced time-phase information I NF The performance of the property analysis using the method is stable.

[0118] [Property analysis device according to the second embodiment] 7 is a conceptual diagram showing an outline of the processing applied to the property analyzer according to the second embodiment. Below, differences from the first embodiment will be mainly described, and descriptions of matters common to the first embodiment will be omitted as appropriate.

[0119] The processing function of the characteristic analysis device according to the second embodiment is to IN Contrast Phase Information I NF CT image I, which is the target of the characteristic analysis process P3, is IN In FIG. 7, the characteristic analysis process P3 selects non-contrast CT image I IN It is not compatible with non-contrast CT images. IN This shows the selection process P5 in which

[0120] In other words, as the target of the characteristic analysis process P3 shown in FIG. 7, a non-contrast CT image I IN , Arterial phase CT image I IN , Portal venous phase CT image I IN and CT image of equilibrium phase I IN Among them, arterial phase CT image I IN , Portal venous phase CT image I IN and CT image of equilibrium phase I IN are selected.

[0121] 8 is a functional block diagram showing an outline of the processing functions of the property analysis device according to the second embodiment. The property analysis device 10A shown in FIG. 8 is configured by adding a sorting unit 22 to the property analysis device 10 shown in FIG.

[0122] The selection unit 22 selects the CT image I estimated by the contrast phase estimation unit 14. IN Contrast phase information for each I NF The CT image I input to the characteristic analysis unit 18 isIN The selection unit 22 selects the CT image I based on a selection condition for the contrast phase set in advance. IN The selection unit 22 acquires information indicating the selection conditions for the contrast enhancement phase, and selects the CT image I based on the acquired selection conditions for the contrast enhancement phase. IN may be selected.

[0123] [Example of hardware configuration for a property analyzer] 9 is a block diagram showing an example of the hardware configuration of a property analysis device according to the second embodiment. In the property analysis device 10A shown in the figure, a memory 40A included in a computer-readable medium 32A stores a selection program 56.

[0124] The selection program 56 selects the contrast phase information I performed using the selection unit 22 shown in FIG. NF The characteristic analysis program 54 causes the processor 30 to execute a selection process based on the selected CT image I. IN Region of interest R OI The processor 30 executes a property analysis process for the

[0125] [Procedure for property analysis method] 10 is a flowchart showing the procedure of the property analysis method according to the second embodiment. The flowchart shown in this figure is obtained by adding a sorting step S13 to the flowchart shown in FIG.

[0126] That is, the process proceeds to the selection step S13 after the contrast phase estimation step S12. In the selection step S13, the selection unit 22 shown in FIG. IN Contrast phase information for each I NF CT image I applied to the property analysis step S16 using IN After the selection step S13, the process proceeds to a region of interest extraction step S14. The procedure of the region of interest extraction step S14 can be the same as that of the property analysis method according to the first embodiment, and therefore a description thereof will be omitted here.

[0127] The selection step S13 may be performed in a different order from the region of interest extraction step S14, or may be performed in parallel with the region of interest extraction step S14. That is, the selection step S13 may be performed after the contrast phase estimation step S12 and before the characterization analysis step S16.

[0128] In the selection process S13, CT image I IN is selected, in the information output step S18, the information output unit 20 may output the selection results in the selection step S13 when outputting the analysis results of the property analysis. The selection results output in the information output step S18 may be displayed on the display device 62.

[0129] [Examples of sorting processing] Fig. 11 is a conceptual diagram showing a specific example of the sorting process. 51 is the CT image I IN Contrast phase information for each I NF CT image I consisting of a set of contrast phases defined in advance using IN In other words, the selection process P 51 is contrast phase information I corresponding to each of two or more types of contrast states defined in advance. NF CT images of each IN In addition, the contrast phase information I NF Each is an example of each piece of contrast enhancement status information.

[0130] FIG. 11 shows the CT image I applied as the property analysis process P3 shown in FIG. IN is limited to two types, and CT image I IN Characterization analysis processing P with limited contrast phase 31 FIG. 11 shows an example of the property analysis process P 31 As a combination of input 1 and input 2, a combination of the arterial phase and the equilibrium phase and a combination of the arterial phase and the portal venous phase are exemplified.

[0131] FIG. 11 shows the contrast phase information I NF Two types of CT images with I IN The sorting process P 51The same contrast phase information I NF CT image type I IN may be input.

[0132] The selection process P5 shown in FIG. 7 is a process for selecting contrast-enhanced time phase information I NF CT image I was acquired IN One or more types of CT images I IN or contrast-enhanced phase information I excluding non-contrast-enhanced phase information NF CT image I was acquired IN One or more types of CT images I IN may be selected.

[0133] [Example of contrast phase estimation] FIG. 12 is a conceptual diagram showing an example of contrast phase estimation. 21 In the trained learning model L M is applied, and the trained learning model L M Let us take 3DCNN as an example. 3DCNN aggregates three-dimensional spatial information and performs three-dimensional convolution. The trained learning model L M The accuracy of contrast phase estimation using this method is expected to improve. Note that the 3D in 3DCNN stands for three dimensions.

[0134] Contrast phase estimation processing P 21 Now, CT image I IN is the learning model L M When the input is made to , the information of the class representing the contrast enhancement phase and the probability p for each class is derived, and the contrast enhancement phase with the highest probability p is the estimation result E R is output as

[0135] In Figure 12, the probability p for each class is calculated, where p is the probability of non-enhanced phase, p is 0.03, p is the probability of arterial phase, p is 0.87, p is the probability of portal venous phase, and p is 0.06, and p is the probability of equilibrium phase, and the probability p is 0.04. IN Estimated result of contrast phase E R The learning model L outputs the arterial phase as M Here is an example:

[0136] In addition, the contrast phase estimation process P shown in FIG. 21 The 3DCNN applied to is an example, and the contrast phase estimation process P2 shown in Figure 3 etc. is performed using the trained learning model L M Any classification model may be applied as

[0137] In addition, the contrast phase estimation is performed using CT image I IN Perform image analysis for each CT image I IN The system estimates a value indicating the ground truth, which is the time elapsed since the start of injection of the contrast agent for each CT image, and then calculates the time elapsed since the start of injection of the contrast agent by referring to a table that defines the correspondence between the time elapsed since the start of injection of the contrast agent and the contrast phase. IN The specific example of contrast enhancement phase estimation described with reference to Fig. 12 can also be applied to the first embodiment.

[0138] [Effects of the second embodiment] The property analyzing device and property analyzing method according to the second embodiment can achieve the following advantageous effects.

[0139] [1] The property analysis device 10A is a CT image I IN Contrast phase information for each I NF CT image I that matches the input of the characteristic analysis unit 18 is obtained using IN In this way, the CT image I that does not match the input of the characteristic analysis unit 18 is selected. IN This suppresses the input to the property analysis unit 18, and the performance of the property analysis in the property analysis unit 18 can be stabilized.

[0140] [2] The selection unit 22 selects non-contrast enhancement phase information I NF CT image I with IN CT image I to exclude IN As a result, when the characteristic analysis process P3 that does not support non-contrast is performed, the CT image I that does not match the input of the characteristic analysis unit 18 is selected. IN This suppresses the input to the property analysis unit 18, and the performance of the property analysis in the property analysis unit 18 can be stabilized.

[0141] [3] The selection unit 22 selects the contrast phase information I NF CT image I with IN As a result, when the characteristic analysis process P3 is performed in which the contrast phases that match the input are limited, the CT image I that does not match the input of the characteristic analysis unit 18 is selected. IN This suppresses the input to the property analysis unit 18, and the performance of the property analysis in the property analysis unit 18 can be stabilized.

[0142] [4] The selection unit 22 selects two types of contrast phase information I NF Two types of CT images I IN As a result, two types of CT images I corresponding to the two prescribed contrast phases are selected. IN is input to the characteristic analysis unit 18, and a CT image I that does not match the input of the characteristic analysis unit 18 is IN This suppresses the input to the property analysis unit 18, and the performance of the property analysis in the property analysis unit 18 can be stabilized.

[0143] [First modified example of input image of property analyzer] In the first and second embodiments, a three-dimensional image, which is three-dimensional CT data, is used as input, but slice images obtained by cutting out slices at equal intervals from the three-dimensional CT data may also be used as input.

[0144] Furthermore, instead of slice images, MIP images constructed at equal intervals and an average image generated from a plurality of slice images may be used, etc. Note that MIP is an abbreviation for Maximum Intensity Projection.

[0145] [Second modified example of input image of property analyzer] 6 and the like may be a combination of multiple types of data elements. For example, at least one type of image among a 3D image, a slice image, an MIP image, and an average image, which are partial images of CT data from the same image series, can be used as input. A combination of these multiple image types may be input to the CT image acquisition unit 12, and an output may be obtained from the information output unit 20.

[0146] For example, a combination of an average image and an MIP image may be input to the CT image acquisition unit 12, and an output may be obtained from the information output unit 20. The three-dimensional image here means a set of multiple slice images.

[0147] In this embodiment, dynamic contrast CT is used as an example of dynamic contrast imaging, but the property analysis shown in this embodiment can also be applied to modalities other than CT to which dynamic contrast imaging can be applied, such as dynamic contrast MRI.

[0148] [Specific example of property analysis processing] FIG. 13 is a conceptual diagram showing a specific example of the characterization analysis process. In the characterization analysis process shown in FIG. 13, the CT image I IN Region of interest extracted from R OI The feature values ​​of the CT images I for each contrast phase are extracted. IN Region of interest extracted from R OI The feature amounts are linked together, and a classification process for the feature is performed as the feature analysis. In this way, the feature analysis process is performed in consideration of the feature amounts of multiple contrast phases.

[0149] The property analysis device that performs the property analysis process shown in FIG. 13 is the property analysis device 10 shown in FIG. 4, and the property analysis unit 18 performs the property analysis process shown in FIG.

[0150] First, a CT image acquisition process P1 is performed, and a CT image I IN Next, a contrast phase estimation process P2 is performed, and a CT image I IN The contrast enhancement phase of CT image I is estimated. IN Contrast phase information for each I NFAlso, CT image I IN From the region of interest R OI The processing up to this point is the same as the CT image acquisition processing P1 and the contrast phase estimation processing P2 shown in FIG.

[0151] Region of interest R OI CT image I where no extraction is performed IN If CT image I is acquired, IN From the region of interest R OI is extracted. Region of interest R OI CT image I from which IN If CT image I is acquired, IN From the region of interest R OI The process of extracting is not performed.

[0152] Next, the feature extraction process P 101 was performed, and the contrast phase was estimated from the CT image I IN The region of interest R contained in each of OI , the CT image I is extracted using the feature extraction network 100. IN The features of each are extracted.

[0153] The feature extraction network 100 outputs a feature vector 102 representing non-enhanced features, a feature vector 104 representing arterial phase features, a feature vector 106 representing portal venous phase features, and a feature vector 108 representing equilibrium phase features.

[0154] The feature vectors 102, 104, 106, and 108 are one-dimensional feature vectors. 101 is the feature extraction network 100 extracting the input region of interest R OI This is a process of outputting the same number of feature vectors 102 as the number of

[0155] Next, the connection process P 102 is performed, and the feature vector 102, the feature vector 104, the feature vector 106, and the feature vector 108 are concatenated to generate feature data 110.

[0156] Furthermore, property analysis processing P 103 The classification network 120 performs classification processing of the properties based on the feature data 110. 103 corresponds to the property analysis process P3 shown in FIG.

[0157] Furthermore, information output processing P 104 The classification network 120 analyzes the classification results of the properties as the analysis result A. R The feature extraction network 100 and the classification network 120 may employ neural networks.

[0158] The property analysis process shown in FIG. 13 is performed by analyzing the region of interest R OI In addition, the processes of contrast phase estimation, feature extraction, and property classification are performed, and each region of interest R OI After feature vectors are extracted from the region of interest R, the feature vectors are concatenated, and the concatenated feature vectors are used to generate feature data 110. OI A configuration in which classification is performed may provide reduced effects of misalignment during the imaging phase, as well as reduced effects of patient breathing and patient movement.

[0159] Figure 13 shows the non-contrast region of interest R OI , the region of interest in the arterial phase R OI , the region of interest in the portal venous phase R OI and the region of interest R of the equilibrium phase OI Although an example has been given in which four feature extraction networks 100 are used for each of the four feature extraction networks, one feature extraction network 100 may be used in common. When comparing multiple feature vectors in a common feature space, it is desirable to use one feature extraction network.

[0160] 13 may be components that make up one trained model. For example, a network that performs feature extraction and a network that performs property analysis may be integrated to generate a region of interest ROI A single network can be configured to perform the property analysis.

[0161] Fig. 14 is a conceptual diagram showing a modified example of the characteristic analysis process shown in Fig. 13. In the characteristic analysis process shown in Fig. 14, the feature amounts of some contrast phases are averaged in comparison with the characteristic analysis process shown in Fig. 13.

[0162] Specifically, in the property analysis process shown in FIG. 14, the feature extraction process P 101 After that, an averaging process P is performed to generate a feature vector 109 by averaging the feature vector 106 of the portal venous phase and the feature vector 108 of the equilibrium phase. 105 will be added.

[0163] Portal venous phase region of interest R OI and the equilibrium phase region of interest R OI Since the feature amounts in are similar, the two feature vectors are averaged and combined into one. This makes it possible to reduce the number of feature vectors when generating the feature data 111. Note that although FIG. 14 illustrates an example of a CT image obtained by performing dynamic contrast-enhanced CT of the liver, similar processing can be performed on contrast-enhanced phases with similar feature amounts for other organs.

[0164] FIG. 15 is a conceptual diagram showing a specific example of the property analysis process when CT images in some contrast phases are missing. IN This example illustrates the case where is missing.

[0165] The connection process P 102 In the example, since the feature vector 108 in the equilibrium phase is missing, the feature vector 106 representing the features of the portal vein phase is treated as a feature vector obtained by averaging the feature vector 106 representing the features of the portal vein phase and the feature vector 108 in the equilibrium phase.

[0166] That is, the connection process P 102In the step S101, the feature vector 102 representing the non-enhanced features, the feature vector 104 representing the arterial phase features, and the feature vector 106 representing the portal venous phase features are concatenated to generate feature data 111A.

[0167] The characteristic analysis process shown in Figure 15 is performed on CT images I in some contrast phases. IN Even if the CT image is missing, it is possible to perform characterization analysis based on the estimated results of the contrast-enhanced phase. IN The example shows a case where the portal venous phase CT image I IN Equilibrium phase CT image I IN CT images other than I IN It is also possible that the part is missing.

[0168] 16 is a conceptual diagram showing another specific example of the characteristic analysis process. In the characteristic analysis process shown in the figure, the region of interest R OI The feature values ​​of the region of interest R for each contrast phase are extracted. OI The feature values ​​are weighted averaged and the properties are classified.

[0169] Specifically, the region of interest R for each contrast phase OI The weight W applied when weighting the feature values ​​of ei The weight calculation network 130 is a neural network.

[0170] The weight calculation network 130 calculates the region of interest R for each contrast phase. OI The weight W indicates how much the feature contributed to the classification of the property. ei The weight calculation process P 106 Weight calculation process P 106 Then, we set weights W for each feature separately. ei may be calculated.

[0171] The weight calculation network 130 calculates the region of interest R for each contrast phase. OI When the feature values ​​of are input, the region of interest R OI Dynamic weights W for the features of eiThe region of interest R for each contrast phase shown in Figure 16 is output. OI Weights W for the features of ei is an example and arbitrary values ​​are shown.

[0172] Weighted average processing P 112 Then, the region of interest R for each contrast phase OI The weight W ei The weighted average process is performed using the property analysis process P 103 Then, the classification network 120 calculates the region of interest R for each contrast phase. OI The classification process of the characteristics is carried out based on the weighted average of the feature values. 104 So, analysis result A R The output is carried out.

[0173] The feature extraction network 100, the weight calculation network 130, and the classification network 120 are connected to the region of interest R in which the contrast phase is estimated. OI The feature extraction network 100, the weight calculation network 130, and the classification network 120 may use backpropagation as a learning algorithm.

[0174] That is, the feature extraction network 100, the weight calculation network 130, and the classification network 120 perform learning in such a way that the loss of the output of the classification network 120 is minimized for each input.

[0175] For the training of the feature extraction network 100 and classification network 120 shown in Figures 13 to 15, backpropagation can be applied as a training algorithm, and training can be performed in such a way that the loss in the output of the classification network 120 for each input is minimized.

[0176] The feature extraction network 100 described in the embodiment is an example of a feature extraction model. The classification network 120 described in the embodiment is an example of a property analysis model. The weight calculation network 130 described in the embodiment is an example of a weight calculation model.

[0177] [Example of medical imaging system configuration] 17 is a block diagram showing an example of the configuration of a medical information system in which a characteristic analysis device is used. The characteristic analysis device 10 and the like described as the first and second embodiments can be incorporated into a medical image processing device 220 shown in FIG.

[0178] The medical information system 200 is a computer network established in a medical institution such as a hospital. The medical information system 200 includes a modality 230 for capturing medical images, a DICOM server 240, a medical image processing device 220, an electronic medical record system 244, and a viewer terminal 246. The elements of the medical information system 200 are connected via a communication line 248. The communication line 248 may be an in-house communication line within the medical institution. Alternatively, part of the communication line 248 may be a wide-area communication line.

[0179] Specific examples of the modality 230 include a CT device 231, an MRI device 232, an ultrasound diagnostic device 233, a PET device 234, an X-ray diagnostic device 235, an X-ray fluoroscopic diagnostic device 236, and an endoscope device 237. There may be various combinations of types of modality 230 connected to the communication line 248 depending on the medical institution. Note that MRI is an abbreviation for Magnetic Resonance Imaging. PET is an abbreviation for Positron Emission Tomography.

[0180] The DICOM server 240 is a server that operates in accordance with the DICOM specifications. The DICOM server 240 is a computer that stores and manages various data, including images captured using the modality 230. The DICOM server 240 is equipped with a large-capacity external storage device and a database management program.

[0181] The DICOM server 240 communicates with other devices via a communication line 248, sending and receiving various data including image data. The DICOM server 240 receives various data including image data generated using the modality 230 via the communication line 248, and stores and manages the data on a recording medium such as a large-capacity external storage device. The storage format of the image data and communication between devices via the communication line 248 are based on the DICOM protocol.

[0182] The medical image processing device 220 can acquire data from the DICOM server 240 or the like via a communication line 248. The medical image processing device 220 performs various processes such as image analysis on medical images captured using the modality 230. In addition to the processing functions of the characteristic analysis device 10, the medical image processing device 220 may be configured to perform various analytical processes such as computer-aided diagnosis, for example, a process for recognizing a lesion area from an image, a process for identifying a classification such as a disease name, and a segmentation process for recognizing an area of ​​an organ or the like. Note that computer-aided diagnosis may be referred to as CAD, which is an abbreviation of Computer Aided Diagnosis or Computer Aided Detection.

[0183] Furthermore, the medical image processing device 220 can send the processing results to the DICOM server 240 and the viewer terminal 246. The processing functions of the medical image processing device 220 may be installed in the DICOM server 240 or the viewer terminal 246.

[0184] Various pieces of information including various data stored in the database of the DICOM server 240 and processing results generated by the medical image processing device 220 can be displayed on the viewer terminal 246.

[0185] The viewer terminal 246 is a terminal for viewing images, called a PACS viewer or a DICOM viewer. Multiple viewer terminals 246 can be connected to the communication line 248. PACS is an abbreviation for Picture Archiving and Communication System. The form of the viewer terminal 246 is not particularly limited, and may be a personal computer, a workstation, a tablet terminal, or the like.

[0186] [About the programs that run the computer] A program that causes a computer to realize the processing functions of the property analysis device 10, etc., can be recorded on a computer-readable medium, which is a tangible, non-transitory information storage medium such as an optical disk, a magnetic disk, or a semiconductor memory, and the program can be provided through this information storage medium.

[0187] In addition, instead of providing the program by storing it on such a tangible, non-transitory computer-readable medium, it is also possible to provide the program signal as a download service using a telecommunications line such as the Internet.

[0188] Furthermore, some or all of the processing functions of the property analyzer 10 and the like may be realized by cloud computing, and may also be provided as a Software as a Service (SasS) service. SasS is an abbreviation for Software as a Service.

[0189] [Hardware configuration of each processing unit] The hardware structure of the processing unit that executes various processes, such as the CT image acquisition unit 12, the contrast phase estimation unit 14, the region of interest extraction unit 16, the characteristic analysis unit 18, and the information output unit 20 in the characteristic analysis device 10, is, for example, various processors as shown below.

[0190] There are various types of processors, including CPUs, which are general-purpose processors that execute programs and function as various processing units, GPUs, which are processors specialized for image processing, and FPGAs (Field Programmable Gate Arrays), which are processors whose circuit configuration can be changed after manufacturing. This includes dedicated electrical circuits such as programmable logic devices, ASICs, and the like, which are processors having circuitry designed specifically to execute specific processes.

[0191] A programmable logic device (PLD) is an abbreviation of Programmable Logic Device (ASIC), and Application Specific Integrated Circuit (ASIC).

[0192] A single processing unit may be configured with one of these various processors, or may be configured with two or more processors of the same or different types. For example, a single processing unit may be configured using multiple FPGAs, a combination of a CPU and an FPGA, or a combination of a CPU and a GPU.

[0193] Alternatively, multiple processing units may be configured as a single processor. Examples of multiple processing units configured as a single processor include, first, a form in which a single processor is configured using a combination of one or more CPUs and software, as typified by computers such as client and server computers, and this processor functions as multiple processing units. Second, a form in which a processor is used to realize the functions of an entire system including multiple processing units on a single IC chip, as typified by system-on-chip. A system-on-chip may be referred to as SoC, an abbreviation for System On a Chip. IC is an abbreviation for Integrated Circuit.

[0194] In this way, the various processing units are configured as hardware structures using one or more of the various processors described above. Furthermore, the hardware structure of these various processors is, more specifically, an electric circuit (circuitry) that combines circuit elements such as semiconductor elements. is.

[0195] The technical scope of the present invention is not limited to the scope described in the above embodiments. The configurations and the like in each embodiment can be appropriately combined with each other within the scope that does not deviate from the spirit of the present invention. [Explanation of symbols]

[0196] 1. Curve showing changes in CT values ​​over time in arteries 2 Curve showing the change in CT value over time in the portal vein 3 Curve showing changes in CT values ​​over time in the liver 10 Property analyzer 10A property analyzer 12 CT image acquisition unit 14 Contrast temporal phase estimation section 16 Region of interest extraction unit 18 Property analysis department 20 Information output section 22 Sorting Department 30 processors 32 Computer-readable medium 32A Computer-readable medium 34 Communication Interface 36 Input / Output Interface 38 Bus 40 memory 40A Memory 42 Storage 50 Contrast Phase Estimation Program 52 Region of interest extraction program 54 Property Analysis Program 56 Selection Program 60 Input Device 62 Display device 100 Feature Extraction Networks 102 Non-contrast feature vectors 104 Arterial Phase Feature Vectors 106 Portal venous phase feature vectors 108 Equilibrium phase feature vector 109 feature vectors 110 Feature Data 111 Feature Data 111A Feature Data 120 Classification Network 130 Weight Calculation Network 200 Medical Information Systems 220 Medical image processing equipment 230 Modalities 231 CT device 232 MRI machine 233 Ultrasound diagnostic equipment 234 PET equipment 235 X-ray diagnostic equipment 236 X-ray fluoroscopy equipment 237 Endoscopic Devices 240 DICOM Server 244 Electronic Medical Record System 246 viewer terminal 248 communication lines I1 Non-contrast CT image I2 arterial phase CT image I3 Portal venous phase CT image CT image of I4 equilibrium phase I IN CT images The period corresponding to the t1 arterial phase The period corresponding to the t2 portal venous phase t3 period corresponding to the equilibrium phase P1CT image acquisition processing P2 contrast phase estimation processing P3 property analysis processing P 31 Property analysis processing P4 information output processing P5 Sorting Process P 51 Sorting process P 101 Feature extraction processing P 102 Concatenation Processing P 103 Property analysis processing P 105 Averaging P 106 Weight calculation process P 112 Weighted Average Processing p probability W ei Weight S10~S18 Each step of the property analysis method

Claims

1. one or more processors; one or more memories in which programs to be executed by the one or more processors are stored; Equipped with The one or more processors execute instructions of the program, A medical image is acquired by performing contrast imaging. estimating a contrast state of the medical image based on an analysis of the medical image; A medical image processing apparatus that performs a characteristic analysis of a region of interest included in the medical image using contrast state information that indicates the contrast state of the medical image.

2. The medical image processing apparatus according to claim 1 , wherein the one or more processors select the medical images using the contrast state information.

3. The medical image processing apparatus according to claim 2 , wherein the one or more processors select the medical images having contrast condition information that matches the characterization analysis.

4. The medical image processing apparatus according to claim 2 , wherein the one or more processors select the medical images in accordance with a restriction on an input image in the characteristic analysis.

5. The medical image processing apparatus according to claim 2 , wherein the one or more processors select the medical images for each of contrast state information corresponding to each of two or more types of contrast states defined in advance.

6. The medical image processing apparatus according to claim 2 , wherein the one or more processors select the medical images having contrast state information corresponding to a contrast state other than non-contrast state.

7. The medical image processing apparatus according to claim 1 , wherein the one or more processors extract a region of interest from the acquired medical image and perform a characteristic analysis of the region of interest.

8. The medical image processing device according to claim 1 , wherein the one or more processors estimate the contrast state of the acquired medical image using a trained learning model.

9. The medical image processing apparatus according to claim 1 , wherein the one or more processors perform the characteristic analysis of the region of interest included in the acquired medical image using a trained learning model.

10. The one or more processors: extracting features from each region of interest included in the medical image for each contrast state using the trained learning model; The medical image processing apparatus according to claim 9 , further comprising: a characteristic analysis of the region of interest included in the medical image based on feature data obtained by linking feature amounts of the region of interest for each contrast enhancement state.

11. The one or more processors: extracting features from each region of interest included in the medical image for each contrast enhancement state using a feature extraction model that extracts features from each region of interest included in the medical image for each contrast enhancement state as the trained learning model; linking the feature amounts of the region of interest for each of the contrast enhancement states; The medical image processing device according to claim 10, wherein the trained learning model is a classification model that classifies feature data in which feature quantities of regions of interest included in the medical image for each contrast state are linked, and performs a characteristic analysis of the regions of interest included in the medical image.

12. The one or more processors: The medical image processing apparatus according to claim 10 or 11, wherein a part of feature amounts of the region of interest included in the medical image for each of the contrast enhancement states is averaged.

13. The one or more processors:

13. The medical image processing apparatus according to claim 12, wherein the trained learning model is a weight calculation model that calculates weights used when calculating a weighted average of feature amounts of a region of interest included in the medical image for each of the contrast conditions, and calculates the weights for each of the feature amounts of a region of interest included in the medical image for each of the contrast conditions.

14. The one or more processors:

12. The medical image processing apparatus according to claim 10, wherein, when a region of interest included in the medical image for each contrast state is missing, a region of interest included in a medical image for a contrast state having features similar to those of the missing region of interest included in the medical image for the contrast state is used instead of the missing region of interest included in the medical image for the contrast state.

15. A method for operating a medical image processing device to which a computer is applied, comprising: The medical image processing device, A step of acquiring medical images generated by performing contrast imaging; estimating a contrast state of the medical image based on an analysis of the medical image; A method for operating a medical image processing apparatus that executes a step of performing a characteristic analysis of a region of interest included in the medical image using contrast state information that indicates the contrast state of the medical image.

16. On the computer, A function to acquire medical images generated by performing contrast imaging; a function of estimating the contrast state of the medical image based on an analysis of the medical image; and A program that realizes a function of performing a characteristic analysis of a region of interest included in the medical image using contrast state information that indicates the contrast state of the medical image.

Images