CT system and storage media

Abstract

This technology provides the ability to generate virtual monochromatic X-ray images at various energies (keV). [Solution] The CT system reconstructs a first CT image 31 based on data obtained by scanning a subject under scanning conditions in which a first tube voltage (120kV) is applied to the X-ray tube; infers a virtual monochromatic X-ray image 32 of a second energy (50keV) that is different from a first energy (70keV) corresponding to the first tube voltage based on the first CT image; and determines a first CT value (v) corresponding to the first energy. i Based on ), the second CT value (v) corresponding to the second energy is determined. j ) and the third CT value (v) corresponding to the third energy (40 keV) k Based on a first coefficient (scaling coefficient a) that defines the relationship with ), the CT value (v) of the virtual monochromatic X-ray image 33 at the third energy (40 keV) is determined. k ) execute the decision.

Description

[Technical Field]

[0001] The present invention relates to a CT system that generates a virtual monochromatic X-ray image of a subject (such as a patient), and a storage medium that stores commands for generating a virtual monochromatic X-ray image of the subject. [Background technology]

[0002] CT scanners are well-known medical devices that non-invasively image subjects. Because CT scanners can acquire tomographic images of subjects in a short scan time, they are widely used in hospitals and other medical facilities.

[0003] A CT scanner generates X-rays by applying a predetermined voltage to the cathode-anode tube of the X-ray tube. The generated X-rays pass through the subject and are detected by a detector. The CT scanner reconstructs a CT image of the subject based on the data detected by the detector. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] Patent No. 6031618 [Overview of the project] [Problems that the invention aims to solve]

[0005] SECT (Single Energy CT) is a well-known imaging technique for CT scanners. SECT is a method of obtaining a CT image of a subject by applying a predetermined voltage (e.g., 120kVp) to the cathode-anode tube of an X-ray tube to generate X-rays. However, with SECT, the CT values ​​can be similar even for different materials, making it difficult to distinguish between different substances.

[0006] Therefore, DECT (Dual Energy CT) technology is being researched and developed. DECT is a technology that uses X-rays in different energy ranges to discriminate between materials, and CT scanners compatible with DECT are commercially available. DECT technology has a wide range of applications; for example, it can be used in techniques to reconstruct virtual monochromatic X-ray images at each energy level.

[0007] On the other hand, many medical institutions do not have CT scanners that support DECT. Therefore, in order to enable medical institutions that do not have DECT-compatible CT scanners to provide diagnoses using virtual monochromatic X-ray images, research and development are underway to use deep learning to infer virtual monochromatic X-ray images from images acquired by SECT.

[0008] For example, a technique has been developed to infer a 50 keV image based on an image obtained when a 120 kV tube voltage is applied to an X-ray tube. This technique uses a pre-trained neural network created using AI technology to infer a 50 keV virtual monochromatic X-ray image. This pre-trained neural network is created, for example, as follows.

[0009] To create such a pre-trained neural network, a large amount of training data is prepared using 120kV CT images and 50keV virtual monochromatic X-ray images, and the neural network is trained using this data. This training allows for the creation of a pre-trained neural network that can infer 50keV virtual monochromatic X-ray images from 120kV CT images. The pre-trained neural network created in this way can infer 50keV virtual monochromatic X-ray images from 120kV CT images obtained using single-energy techniques. CT images obtained with a 120kV tube voltage are generally known to exhibit characteristics similar to 70kV virtual monochromatic X-ray images, although this depends on the type of CT scanner. Therefore, by inferring 50keV virtual monochromatic X-ray images using the pre-trained neural network, users can compare 120kV CT images (70keV virtual monochromatic X-ray images) with 50keV virtual monochromatic X-rays. Thus, an improvement in diagnostic accuracy can be expected.

[0010] Furthermore, while the above technology can only perform inference on 50keV images, it is expected that being able to perform inference on images of other keVs would further contribute to improving diagnostic accuracy. For example, by creating a trained neural network that can perform inference on images of other keVs in addition to the trained neural network that performs inference on 50keV images, it becomes possible to perform inference on images of other keVs as well as 50keV, which is expected to further contribute to improving diagnostic accuracy.

[0011] However, the above method requires creating a trained neural network for each virtual monochromatic X-ray image of each energy. Therefore, training data must be prepared for each virtual monochromatic X-ray image of the energy to be inferred, and the neural network must be trained accordingly. This leads to the problem of increased development time.

[0012] Therefore, there is a need for a technology that can generate virtual monochromatic X-ray images of various energies with minimal development effort. [Means for solving the problem]

[0013] A first aspect of the present invention is a CT system, An X-ray tube to which tube voltage is applied, One or more processors, Reconstructing a first CT image based on data obtained by scanning a subject under scanning conditions in which a first tube voltage is applied to the X-ray tube. Based on the first CT image, a virtual monochromatic X-ray image of a second energy different from the first energy corresponding to the first tube voltage is inferred. The CT value of the virtual monochromatic X-ray image of the third energy is determined based on a first coefficient that defines the relationship between the second CT value corresponding to the second energy and the third CT value corresponding to the third energy, using the first CT value corresponding to the first energy as a reference. One or more processors that execute A CT system, including a CT scanner.

[0014] Furthermore, a second aspect of the present invention is a non-temporary computer-readable storage medium in which instructions are stored, wherein when an instruction is executed by the one or more processors, the instructions are sent to the one or more processors. Reconstructing a first CT image based on data obtained by scanning a subject under scanning conditions in which a first tube voltage is applied to the X-ray tube. Based on the first CT image, a virtual monochromatic X-ray image of a second energy different from the first energy corresponding to the first tube voltage is inferred. The CT value of the virtual monochromatic X-ray image of the third energy is determined based on a first coefficient that defines the relationship between the second CT value corresponding to the second energy and the third CT value corresponding to the third energy, using the first CT value corresponding to the first energy as a reference. It is a storage medium that performs operations including [specific actions]. [Effects of the Invention]

[0015] In this invention, by using a first coefficient that defines the relationship between the second CT value corresponding to the second energy and the third CT value corresponding to the third energy, based on a first CT value corresponding to the first energy, a virtual monochromatic X-ray image of any energy desired by the user (e.g., 40 keV) can be calculated. Therefore, since it is not necessary to create a trained neural network for each virtual monochromatic X-ray image of each energy, development man-hours can be reduced, and development costs can be significantly reduced. [Brief explanation of the drawing]

[0016] [Figure 1] This is a block diagram of the CT system 100 in the first embodiment. [Figure 2] This is a schematic diagram of a curve representing the change in CT value with respect to energy (keV). [Figure 3] This is an explanatory diagram of how to create curve 12. [Figure 4] This is a diagram showing curve 12. [Figure 5] This diagram shows the location of 50 keV indicated by the dashed line 22. [Figure 6] This diagram shows the location of 40 keV indicated by the dashed line 23. [Figure 7] This is an explanatory diagram of the relationship between CT values. [Figure 8] This is an explanatory diagram illustrating the relationship between the CT values ​​of curve 11. [Figure 9] This is an explanatory diagram for calculating the CT value of a 45keV virtual monochromatic X-ray image. [Figure 10] This is an explanatory diagram of the lookup table LUT1 stored in the memory device. [Figure 11] This is a flowchart for acquiring a virtual monochromatic X-ray image of a subject at any energy. [Figure 12] This figure schematically shows a CT image 31 obtained by scanning the subject. [Figure 13]This figure schematically shows the inferred virtual monochromatic X-ray image 32 at 50 keV. [Figure 14] This is an explanatory diagram of the formula used to create a virtual monochromatic X-ray image at 40 keV. [Figure 15] This is an explanatory diagram for calculating the CT value v40 using equation (9). [Figure 16] This is a flowchart of step ST3 in the second embodiment. [Figure 17] This is an explanatory diagram for Step ST3. [Figure 18] This diagram illustrates the problems that arise when scanning a subject with a tube voltage of 100kV. [Figure 19] In the third embodiment, this is an explanatory diagram illustrating the principle of generating a 50 keV virtual monochromatic X-ray image from a 100 kV CT image. [Figure 20] This is an explanatory diagram of the lookup table LUT2 stored in the memory device. [Figure 21] This is a flowchart for acquiring a virtual monochromatic X-ray image at 50 keV. [Figure 22] This is an explanatory diagram of the flow shown in Figure 21. [Modes for carrying out the invention]

[0017] The following describes embodiments for carrying out the invention, but the present invention is not limited to the following embodiments.

[0018] Figure 1 is a block diagram of the CT system 100 in the first embodiment. The CT system 100 has a gantry 102. The gantry 102 has an opening, and the subject (object to be scanned) 112 moves into the opening, and the scan of the subject 112 is performed.

[0019] The gantry 102 is equipped with an X-ray tube 104, a filter unit 103, a pre-collimator 105, and an X-ray detector 108, among other components.

[0020] The X-ray tube 104 generates X-rays when a predetermined voltage is applied to the cathode-anode tube. The filter section 103 includes, for example, a flat plate filter and / or a bowtie filter. The pre-collimator 105 is a component used to narrow the X-ray irradiation range so that X-rays are not irradiated into unwanted areas.

[0021] The X-ray detector 108 includes multiple detector elements 202. These detector elements 202 detect the X-ray beam 106 that is irradiated from the X-ray tube 104 and passes through a subject 112, such as a patient. Therefore, the X-ray detector 108 can acquire projection data for each view.

[0022] Projection data detected by the X-ray detector 108 is collected by the DAS 214. The DAS 214 performs predetermined processing on the collected projection data, including sampling and digital conversion. The processed projection data is transmitted to the computer 216. The computer 216 stores the data from the DAS 214 in the storage device 218. The storage device 218 includes one or more storage media for storing programs, instructions executed by the processor, etc. The storage media can be, for example, one or more non-temporary computer-readable storage media. The storage device 218 can include, for example, a hard disk drive, a floppy disk drive, a compact disc read / write (CD-R / W) drive, a digital multipurpose disc (DVD) drive, a flash drive, and / or a solid-state storage drive.

[0023] Computer 216 includes one or more processors. Computer 216 uses one or more processors to output commands and parameters to DAS 214, X-ray controller 210, and / or gantry motor controller 212, and to control system operations such as data acquisition and / or processing.

[0024] The computer 216 is connected to an operator console 220. By operating the operator console 220, the operator can input predetermined operator inputs related to the operation of the CT system 100 into the computer 216. The computer 216 receives operator inputs, including commands and / or scan parameters, via the operator console 220 and controls system operation based on these inputs. The operator console 220 may include a keyboard (not shown) or a touchscreen for the operator to specify commands and / or scan parameters.

[0025] The X-ray controller 210 controls the X-ray tube 104 based on control signals from the computer 216. The gantry motor controller 212 also controls the gantry motor based on control signals from the computer 216.

[0026] Figure 1 shows only one operator console 220, but two or more operator consoles may be connected to the computer 216.

[0027] Furthermore, the CT system 100 may be configured to connect multiple remotely located displays, printers, workstations, and / or similar devices, for example, via a wired network and / or a wireless network.

[0028] In one embodiment, for example, the CT system 100 may include or be coupled to an image archiving and communication system (PACS) 224. In an exemplary embodiment, the PACS 224 may be coupled to a radiology information system, a hospital information system, and / or a remote system such as an internal or external network (not shown).

[0029] The computer 216 supplies commands to the table motor controller 118 for controlling the table 116. The table motor controller 118 can control the table 116 based on the commands it receives. In particular, the table motor controller 118 can move the table 116 so that the subject 112 is properly positioned within the opening of the gantry 102.

[0030] As described above, the DAS214 samples and digitally converts the projection data acquired by the detector element 202. The image reconstructor 230 then reconstructs the image using the sampled and digitally converted data. The image reconstructor 230 includes one or more processors that can perform the image reconstruction process. In Figure 1, the image reconstructor 230 is shown as a separate component from the computer 216, but the image reconstructor 230 may form part of the computer 216. Alternatively, the computer 216 may perform one or more functions of the image reconstructor 230. Furthermore, the image reconstructor 230 may be located away from the CT system 100 and operably connected to the CT system 100 using a wired or wireless network.

[0031] The image reconstructor 230 can store the reconstructed image in the storage device 218. The image reconstructor 230 may also transmit the reconstructed image to the computer 216. The computer 216 can transmit the reconstructed image and / or patient information to a display 232 which is communicatively coupled to the computer 216 and / or the image reconstructor 230.

[0032] The various methods and processes described herein can be stored as executable instructions in a non-temporary storage medium within the CT system 100 or in an external storage medium communicably connected to the CT system 100. These executable instructions may be stored in a single storage medium or distributed across multiple storage media. One or more processors provided in the CT system 100 execute the various methods, steps, and processes described herein in accordance with the instructions stored in the storage medium. For example, in this embodiment, the processor executes methods, steps, and processes related to the trained neural network 30 (the trained neural network 30 is a neural network that infers a 50 keV virtual monochromatic X-ray image from a 120 kV CT image; see, for example, Figures 13 and 15). The CT system 100 is configured as described above.

[0033] As described above, the CT system 100 can use a pre-trained neural network 30. The pre-trained neural network 30 is a neural network that infers a 50 keV virtual monochromatic X-ray image from a 120 kV CT image. The neural network itself that infers a 50 keV virtual monochromatic X-ray image from a 120 kV CT image is known, and for example, it is introduced in some CT devices available from GE Healthcare under the name "True Enhance DL". As described above, the pre-trained neural network 30 can infer a 50 keV virtual monochromatic X-ray image from a 120 kV CT image. Therefore, even in CT devices that do not have dual-energy technology, a 50 keV virtual monochromatic X-ray image can be obtained from a 120 kV CT image obtained by scanning a subject, so an improvement in diagnostic performance is expected.

[0034] Furthermore, while the trained neural network 30 can only infer from 50 keV images, it is expected that further improvements in diagnostic performance could be achieved if another trained neural network capable of inferring from images at other keV levels could be prepared. For example, if a trained neural network capable of inferring from images at other keV levels were created in addition to the trained neural network capable of inferring from 50 keV images, it would be possible to infer from images at other keV levels as well as 50 keV images, which is expected to greatly contribute to further improvements in diagnostic performance.

[0035] Therefore, it is conceivable to create a pre-trained neural network for virtual monochromatic X-ray images at various energies and infer virtual monochromatic X-ray images at different energies. For example, if we consider making a diagnosis based on images at 40keV, 50keV, 60keV, 70keV, 80keV, 90keV, 100keV, 110keV, 120keV, 130keV, and 140keV, we can create a pre-trained neural network for each of these energies and infer virtual monochromatic X-ray images at these energies. Thus, since we can infer virtual monochromatic X-ray images at various energies, it becomes possible to further improve diagnostic performance.

[0036] However, the above method requires creating a pre-trained neural network for each virtual monochromatic X-ray image at each energy level. Therefore, training data must be prepared for each virtual monochromatic X-ray image at the energy level to be inferred, and the neural network must be trained accordingly. This leads to the problem of increased development time.

[0037] Therefore, the inventors of this application have diligently conducted research and devised a method for generating virtual monochromatic X-ray images at various energies without having to create a pre-trained neural network for each energy level. The basic concept of this method will be explained below with reference to Figure 2.

[0038] Figure 2 is a schematic diagram of the curve representing the change in CT value with respect to energy (keV). Figure 2 shows curves 11-15 obtained for each part of the human body.

[0039] Curve 11 represents the change in kidney CT value in response to energy (keV). Note that curve 11 also represents the change in CT value when contrast agent is injected into the kidney. Curve 12 represents the change in bone CT value in response to energy (keV). Curve 13 represents the change in liver CT value in response to energy (keV). Curve 14 represents the change in CT value of renal cysts in response to energy (keV). Curve 15 is a curve that shows the change in the CT value of fat in response to energy (keV).

[0040] These curves 11-15 can be created by actually scanning a subject with a CT scanner equipped with dual-energy technology and based on the data obtained from the scan. Alternatively, a phantom containing common substances found in the human body (e.g., water, iodine, calcium) may be scanned, and CT value curves for each part may be created based on the data obtained from the phantom. Furthermore, CT value curves may be created based on data obtained from both the subject scan and the phantom scan.

[0041] Below, we will take curve 12 as a representative example from curves 11 to 15 and explain how to create curve 12, referring to Figure 3.

[0042] To create curve 12, first, virtual monochromatic X-ray images in the energy range of 40 keV to 140 keV are acquired based on data obtained by scanning multiple subjects and / or phantoms using dual-energy technology. In this embodiment, virtual monochromatic X-ray images in the range of 40 keV to 140 keV are acquired every 5 keV, but virtual monochromatic images of any keV can be acquired. For example, virtual monochromatic X-ray images may be acquired every 1 keV, or every 10 keV. Note that, due to space limitations, only the virtual monochromatic X-ray images S1 to Sn for 40 keV energy obtained by scanning are shown here, but multiple virtual monochromatic X-ray images are also acquired for other energies. Then, the CT value of the bone is identified for each energy (keV) from the virtual monochromatic X-ray images obtained in this way. For example, focusing on 40 keV, the CT value of the bone is identified from the virtual monochromatic X-ray images S1 to Sn. In Figure 3, bar B1 is shown at 40 keV, representing the range of variation in CT values. This bar B1 represents the distribution range of CT values ​​for multiple virtual monochromatic X-ray images S1 to Sn. Furthermore, CT values ​​are identified for other energies in the same way as at 40 keV. In Figure 3, bars B2 to B21 are shown for 45 keV to 140 keV, representing the range of variation in CT values.

[0043] Then, a curve 12 is drawn so as to pass through the range of bars B1 to B21 indicated for each energy (keV). The curve 12 may be created to pass through the midpoint of the range of CT values ​​defined for each bar, or it may be created to pass through the point where the CT values ​​are most densely concentrated within the range of CT values ​​defined for each bar.

[0044] Although Figure 3 illustrates the method for creating curve 12, the other curves 11, 13, 14, and 15 can be created using the same method.

[0045] Therefore, curves 11-15 allow us to understand how the CT value changes with respect to energy (keV). Furthermore, as explained earlier, techniques have been developed to infer a virtual monochromatic X-ray image at 50 keV from a 120 kV CT image.

[0046] The inventors of this application have conducted diligent research and discovered that by utilizing the CT values ​​obtained from the curve and a technique for inferring a 50 keV virtual monochromatic X-ray image from a 120 kV CT image, it is possible to generate virtual monochromatic X-ray images of energies other than 50 keV from a single-energy CT image without having to create a pre-trained neural network for each energy (keV) virtual monochromatic X-ray image. The principle of generating virtual monochromatic X-ray images of energies other than 50 keV by utilizing the CT values ​​obtained from the curve and a technique for inferring a 50 keV virtual monochromatic X-ray image from a 120 kV CT image will be explained in detail below with reference to Figures 4 to 9.

[0047] Figure 4 shows curve 12. First, we determine the energy (keV) of the virtual monochromatic X-ray image corresponding to the tube voltage used for single-energy imaging. While various voltage values ​​can be used for the tube voltage in single-energy imaging, here we will consider the case where 120kV is used as the tube voltage for single-energy imaging. Generally, the characteristics of a 120kV CT image (e.g., contrast) can be considered sufficiently similar to those of a 70keV virtual monochromatic X-ray image. Therefore, we assume that the energy of the virtual monochromatic X-ray image corresponding to 120kV is 70keV.

[0048] Next, we identify the position representing 70 keV on the horizontal axis (keV axis) of curve 12. In Figure 4, the position of 70 keV is shown by the dashed line 21.

[0049] Next, from the horizontal axis (keV axis) of the curve 12, identify the position of the energy (keV) of the virtual monochromatic X-ray image inferred by the trained neural network. In this embodiment, since the energy of the virtual monochromatic X-ray image inferred by the trained neural network is 50 keV, identify the position representing 50 keV on the horizontal axis (keV axis) of the curve 12. In FIG. 5, the position of 50 keV is indicated by a dashed line 22.

[0050] Next, consider the case of generating a virtual monochromatic X-ray image with an energy different from that of the 50 keV virtual monochromatic X-ray image. Here, consider the case of generating a virtual monochromatic X-ray image with an energy of 40 keV. Therefore, identify the position representing 40 keV on the horizontal axis (keV axis) of the curve 12. In FIG. 6, the position of 40 keV is indicated by a dashed line 23.

[0051] Next, examine the relationship between the CT values at 70 keV, 50 keV, and 40 keV (see FIG. 7).

[0052] FIG. 7 is an explanatory diagram of the relationship between CT values. In FIG. 7, the CT value at 70 keV is represented by "v 70 ", and here, v 70 ≈100 HU. Also, the CT value at 50 keV is represented by "v 50 ", and here, v 50 ≈170 HU. Further, the CT value at 40 keV is represented by "v 40 ", and here, v 40 ≈240 HU. Note that these CT values v 70 , v 50 , and v 40 may vary slightly depending on the method of creating the curve 12. For example, the CT value at 70 keV may deviate from 100 HU. However, such a deviation in CT value is a sufficiently small value that can be ignored in explaining the effects of this embodiment. Therefore, in the following explanation, it will be explained that the CT value is 100 HU at 70 keV, 170 HU at 50 keV, and 240 HU at 40 keV.

[0053] First, the CT value v at 70 keV 70 and the CT value v at 50 keV 50 Calculate the difference ΔCT1. ΔCT1 is expressed by the following formula. ΔCT1=v 50 -v 70 (1)

[0054] CT value v 70 = 100HU, and CT value v 50 Since = 170HU, ΔCT1 can be calculated using the following formula. ΔCT1=v 50 -v 70 =170HU-100HU =70HU

[0055] Next, the CT value v at 70 keV 70 and the CT value v at 40 keV 40 Calculate the difference ΔCT2. ΔCT2 can be calculated using the following formula. ΔCT²=v 40 -v 70 (2)

[0056] CT value v 70 = 100HU, and CT value v 40 Since = 240HU, ΔCT2 can be calculated using the following formula. ΔCT²=v 40 -v 70 =240HU-100HU =140HU

[0057] Therefore, it can be considered that ΔCT1 and ΔCT2 have the following relationship. ΔCT2 = 2 * ΔCT1 (3)

[0058] From equation (3), we can see that ΔCT2 can be expressed as twice ΔCT1. In other words, using the CT value of 70 keV as the base, the CT value of 40 keV is twice the CT value of 50 keV.

[0059] Next, we examine the relationship between CT values ​​at 70 keV, 50 keV, and 40 keV for a curve other than curve 12 (see Figure 8).

[0060] Figure 8 is an explanatory diagram illustrating the relationship between the CT values ​​of curve 11. At 70 keV, the CT value is approximately 120 HU; at 50 keV, it is approximately 225 HU; and at 40 keV, it is approximately 330 HU.

[0061] Next, we calculate the difference ΔCT1 between the CT value of 120HU at 70keV and the CT value of 225HU at 50keV. ΔCT1 can be calculated using the following formula. ΔCT1 = 225HU - 120HU = 105HU

[0062] Furthermore, we calculate the difference ΔCT2 between the CT value of 120HU at 70keV and the CT value of 330HU at 40keV. ΔCT2 can be calculated using the following formula. ΔCT2 = 330HU - 120HU =210HU

[0063] Therefore, it can be considered that ΔCT1 and ΔCT2 have the following relationship. ΔCT2 = 2 * ΔCT1

[0064] Therefore, it can be seen that for curve 11 as well, ΔCT2 can be expressed as twice ΔCT1. Furthermore, although a detailed explanation is omitted, for the other CT curves 13-15, ΔCT1 and ΔCT2 can generally be expressed by the relationship in equation (3).

[0065] From the above considerations, it was found that ΔCT2 can be calculated by multiplying ΔCT1 by a scaling factor of "2", regardless of the imaging site.

[0066] Furthermore, as shown in equations (1) and (2), ΔCT1 and ΔCT2 can be expressed by the following equations. ΔCT1=v 50 -v70 ΔCT²=v 40 -v 70

[0067] Therefore, substituting equations (1) and (2) into equation (3), we obtain the following equation. ΔCT2 = 2 * ΔCT1 v 40 -v 70 =2(v 50 -v 70 ) v 40 =2(v 50 -v 70 )+v 70 (4)

[0068] Therefore, the CT value of a 70 keV virtual monochromatic X-ray image v 70 And the CT value of a 50 keV virtual monochromatic X-ray image v 50 By substituting this into equation (4), we can see that the CT value of a 40 keV virtual monochromatic X-ray image can be calculated.

[0069] Note that in equation (4), the CT value v of the 40 keV virtual monochromatic X-ray image is used. 40 While an example of calculating the CT value for virtual monochromatic X-ray images at other energies is explained, the CT values ​​for virtual monochromatic X-ray images at other energies can also be calculated according to the above explanation. For example, we will focus on a 45 keV virtual monochromatic X-ray image as an example of a virtual monochromatic X-ray image at an energy other than 40 keV, and explain how to calculate the CT value of the 45 keV virtual monochromatic X-ray image. Figure 9 is an explanatory diagram for calculating the CT value of a 45 keV virtual monochromatic X-ray image.

[0070] As explained earlier, the CT value v at 70 keV 70 is v 70 = 100HU, and the CT value v at 50keV 50 is v 50 = 170 HU. Also, the CT value v at 45 keV 45 is v 45 = 200 HU.

[0071] CT value v at 70 keV70 (=100HU) and the CT value v at 50keV 50 The difference ΔCT1 from (=170HU) can be calculated using the following formula. ΔCT1 = 170HU - 100HU =70HU

[0072] Also, the CT value v at 70 keV 70 (=100HU) and the CT value v at 45keV 45 The difference ΔCT2 from (=200HU) can be calculated using the following formula. ΔCT2 = 200HU - 100HU =100HU

[0073] Therefore, it can be considered that ΔCT1 and ΔCT2 have the following relationship. ΔCT2 ≈ 1.4 * ΔCT1 (5)

[0074] From equation (5), it can be seen that in the case of a 45 keV virtual monochromatic X-ray image, ΔCT2 can be calculated by multiplying ΔCT1 by a scaling factor of 1.4.

[0075] Therefore, the CT value of a 45 keV virtual monochromatic X-ray image v 45 It can be calculated using the following formula. v 45 =1.4(v 50 -v 70 )+v 70 (6)

[0076] Therefore, comparing equation (4) and equation (6), simply changing the scaling factor in equation (4) from "2" to "1.4" will change the CT value v of a 45keV virtual monochromatic X-ray image. 45 It can be seen that this can be calculated.

[0077] Therefore, in equations (4) and (6), the CT value v 40 and v 45 "v" represents the CT value of a virtual monochromatic X-ray image of any energy E. EBy replacing this with ", and further replacing the scaling factors "2" and "1.4" with the scaling factor "a", equations (4) and (6) can be generalized to the following equations. v E =a(v 50 -v 70 )+v 70 (7)

[0078] In equation (7), v 70 This is the CT value of a 70 keV virtual monochromatic X-ray image, and v 50 V is the CT value of a virtual monochromatic X-ray image at 50 keV. Furthermore, a is a scaling factor, which defines the relationship between the CT value corresponding to 40 keV and the CT value corresponding to 50 keV, using the CT value corresponding to 70 keV as a reference. Since the scaling factor a is determined by the energy E, simply changing the value of the scaling factor a will change the CT value v of a virtual monochromatic X-ray image at any energy E. E It can be seen that this can be calculated.

[0079] As explained earlier, a 70 keV virtual monochromatic X-ray image corresponds to a 120 kV CT image. Therefore, the CT value of the 120 kV CT image is given by the symbol "v tube_120 When expressed as ", the v in equation (7) 70 is, v tube_120 It can be replaced by . In other words, equation (7) can be expressed as follows: v E =a(v 50 -v tube_120 )+v tube_120 (8) Here, v tube_120 CT value of 120kV CT image v 50 CT values ​​of a virtual monochromatic X-ray image at 50 keV v E CT value of a virtual monochromatic X-ray image of energy E

[0080] Therefore, the CT value v of the 120kV CT image tube_120and the CT value v of the virtual monochromatic X-ray image at 50 keV 50 Once the CT value v of the virtual monochromatic X-ray image at energy E is known E it can be understood that it can be calculated. In this embodiment, the storage device stores a look-up table representing the correspondence between the energy E of the virtual monochromatic X-ray image and the scaling coefficient a (see FIG. 10).

[0081] FIG. 10 is an explanatory diagram of the look-up table LUT1 stored in the storage device. The look-up table LUT1 has columns for the energy E (keV) of the virtual monochromatic X-ray image and the scaling coefficient a. In FIG. 10, as examples of the energy E (keV), for the sake of explanation, 40 keV, 60 keV, 80 keV, 100 keV, 120 keV, and 140 keV are shown, but they are not limited to these energies, and any energy within the range of 40 keV to 140 keV can be used as an example of the energy E (keV) in the look-up table LUT1. Further, energies less than 40 keV and / or greater than 140 keV can also be used as examples of the energy E (keV) in the look-up table LUT1.

[0082] Also, in the column of the scaling coefficient a, the scaling coefficients a corresponding to each energy E are respectively shown as "a 40 ", "a 60 ", "a 80 ", "a 100 ", "a 120 ", and "a 140 ". For example, the scaling coefficient a 40 for energy E = 40 keV is a 40 = 2 as described above. Here, the other scaling coefficients "a 60 ", "a 80 ", "a 100 ", "a 120 ", and "a 140While specific values ​​for "" are not shown, the values ​​of these other scaling factors can also be determined in the same way as described with reference to Figures 7 to 9.

[0083] Therefore, the lookup table LUT1 has a virtual monochromatic X-ray image with an energy of 40 keV to 140 keV and a scaling factor a 40 ~a 140 This represents the correspondence between the two. In this embodiment, the generation of a virtual monochromatic X-ray image of any energy of the subject is achieved using equation (8) and the lookup table LUT1. The following describes the flow for generating a virtual monochromatic X-ray image of any energy of the subject in this embodiment.

[0084] Figure 11 is a flowchart for acquiring a virtual monochromatic X-ray image of any energy, and Figures 12 to 15 are explanatory diagrams for each step performed in the flowchart of Figure 11.

[0085] In step ST1, the subject is scanned under scanning conditions in which a tube voltage of 120kV is applied to the X-ray tube. The processor reconstructs a CT image of the subject based on the data obtained by scanning the subject. Figure 12 schematically shows the CT image 31 obtained by scanning the subject. After scanning the subject, the process proceeds to step ST2.

[0086] In step ST2, the processor uses the trained neural network 30 to infer a 50keV virtual monochromatic X-ray image from the CT image 31. Figure 13 schematically shows the inferred 50keV virtual monochromatic X-ray image 32. The processor can input the CT image 31 into the trained neural network 30 to infer the 50keV virtual monochromatic X-ray image 32. Alternatively, the processor may perform preprocessing on the CT image 31 as needed, and then input the preprocessed CT image 31 into the trained neural network 30 to infer the 50keV virtual monochromatic X-ray image 32. After inferring the 50keV virtual monochromatic X-ray image 32, the process proceeds to step ST3.

[0087] In step ST3, the processor generates a virtual monochromatic X-ray image with an energy different from that of the 50 keV virtual monochromatic X-ray image 32 based on the CT image 31 and the 50 keV virtual monochromatic X-ray image 32. Step ST3 will be described with reference to FIGS. 14 and 15.

[0088] FIG. 14 is an explanatory diagram of an equation used to generate a virtual monochromatic X-ray image with an energy different from that of the 50 keV virtual monochromatic X-ray image 32. Hereinafter, although 40 keV is considered as an example of an energy different from others, the same description can be applied to energies other than 40 keV.

[0089] First, the processor selects a scaling factor a corresponding to 40 keV from a plurality of scaling factors a 40 ~a 140 stored in the storage device. The scaling factor a corresponding to 40 keV is a = a 40 . Therefore, the processor selects the scaling factor a = a 40 from the look-up table LUT1. After selecting the scaling factor a 40 , the selected scaling factor a 40 is substituted into a in Equation (8). Since a 40 = 2, a = a 40 = 2 is substituted into a in Equation (8). Therefore, the following equation is obtained. v E = a(v 50 - v tube_120 ) + v tube_120 = 2(v 50 - v tube_120 ) + v tube_120 (9)

[0090] Here, since it is considered to generate a 40 keV virtual monochromatic X-ray image, v E in Equation (9) can be replaced with v 40 . Therefore, Equation (9) becomes the following equation. v 40 =2(v 50 -v tube_120 )+v tube_120 (10)

[0091] Therefore, the processor uses equation (10) to determine the CT value v of the 40 keV virtual monochromatic X-ray image. 40 The CT value v can be calculated as follows. 40 Calculate.

[0092] Figure 15 shows the CT value v using equation (10). 40 This is an explanatory diagram for calculating the calculations. The processor identifies the pixel Pi of interest from the CT image 31. Then, it calculates the CT value v of the pixel Pi of interest. i The following is read out. Since CT image 31 is an image obtained with a tube voltage of 120kV, the processor reads out v in equation (10). tube_120 ni, v i Substitute this value.

[0093] Furthermore, the processor selects the pixel P of the CT image 31 from the 50 keV virtual monochromatic X-ray image 32. i Pixel P at the same position j CT value v j Read out the CT value v of the 50keV virtual monochromatic X-ray image 32. j This is the CT value v in equation (10). 50 Since this is a value representing v in equation (10), the processor is v 50 ni, v j Substitute this value.

[0094] Therefore, the processor processes the pixels P of the CT image 31 in the 40 keV virtual monochromatic X-ray image 33. i Pixel P at the same position k CT value v k (=v 40 ) can be calculated. For example, v i = 95HU, and v j If =160HU, v k It can be calculated as =225HU.

[0095] In the above explanation, pixel P of the 40 keV virtual monochromatic X-ray image 33 k CT value v k The procedure for calculating the first image has been described, but the same method can be used to calculate the other pixels of the 40keV virtual monochromatic X-ray image 33. Therefore, the 40keV virtual monochromatic X-ray image 33 can be calculated based on the CT image 31 and the 50keV virtual monochromatic X-ray image 33.

[0096] In the above explanation, we described an example of calculating a 40keV virtual monochromatic X-ray image 33 as a virtual monochromatic X-ray image at a different energy than the 50keV virtual monochromatic X-ray image. However, by using equation (8) and the lookup table LUT1, it is possible to calculate virtual monochromatic X-ray images at other energies in addition to the 40keV virtual monochromatic X-ray image 33. For example, if you want to calculate a 60keV virtual monochromatic X-ray image, you can calculate the 60keV virtual monochromatic X-ray image based on the following equation.

[0097] v 60 =a 60 (v 50 -v tube_120 )+v tube_120 (11) Equation (11) above is the same as v in equation (8). E v 60 Substitute this and change the scaling coefficient a in equation (8) to the scaling coefficient a corresponding to the energy of 60 keV. 60 This is the formula obtained by substituting the given terms.

[0098] The processor is v in equation (11) tube_120 Substitute the CT value of CT image 31 into equation (11) and get v 50 By substituting the CT value of the 50 keV virtual monochromatic X-ray image 32 into the formula, the CT value of the 60 keV virtual monochromatic X-ray image can be calculated.

[0099] Similarly, virtual monochromatic X-ray images of other energies can be calculated. The display device can display not only the 50 keV virtual monochromatic X-ray image 32, but also virtual monochromatic X-ray images of any energy other than 50 keV. Therefore, users can visually confirm virtual monochromatic X-ray images of various energies. In this way, the flow shown in Figure 11 is completed.

[0100] In the first embodiment, a virtual monochromatic X-ray image of any energy desired by the user (e.g., 40 keV) can be calculated by utilizing a lookup table LUT1 (see Figure 14) that represents the correspondence between energy E (keV) and scaling coefficient a. Therefore, in the first embodiment, a virtual monochromatic X-ray image of any energy desired by the user can be calculated by utilizing the lookup table LUT1, without having to create a pre-trained neural network 30 for each energy in the range of 40 keV to 140 keV. Thus, a virtual monochromatic X-ray image of any energy can be generated without having to prepare training data for the neural network for each energy in the range of 40 keV to 140 keV, or without training the neural network for each energy. Therefore, the method of the first embodiment that utilizes the lookup table LUT1 can reduce development effort and significantly reduce development costs compared to the method of creating a pre-trained neural network 30 for each energy in the range of 40 keV to 140 keV.

[0101] In this embodiment, the scaling coefficient a is selected from the lookup table LUT1 stored in the memory device. However, instead of storing the lookup table LUT1, a group of CT value data (for example, one of the curves 11 to 15) representing the change in CT value with respect to the energy (keV) of the virtual monochromatic X-ray image may be stored, and the scaling coefficient a may be calculated based on the stored CT value data.

[0102] In this embodiment, the trained neural network 30 infers a 50 keV virtual monochromatic X-ray image 32 from the CT image 31. However, the inferred virtual monochromatic X-ray image is not necessarily limited to a 50 keV image; instead, a virtual monochromatic X-ray image of a different energy may be inferred. If the energy of the inferred virtual monochromatic X-ray image is extended to any energy other than 50 keV, equation (8) can be generalized to the following equation. v E =a(v E_inf -v tube_120 )+v tube_120 (12) Here, v in equation (12) E_inf This is an arbitrary energy of the inferred virtual monochromatic X-ray image.

[0103] Therefore, the inferred virtual monochromatic X-ray image is not limited to a 50 keV virtual monochromatic X-ray image; even if a virtual monochromatic X-ray image at an energy other than 50 keV is inferred, the CT value v of the virtual monochromatic X-ray image at any energy E will be determined. E It is possible to calculate the CT value v of a virtual monochromatic X-ray image of any energy E desired by the user. For example, instead of a trained neural network 30 that infers a 50 keV virtual monochromatic X-ray image from a 120 kV CT image, if a trained neural network that infers a 60 keV virtual monochromatic X-ray image from a 120 kV CT image is prepared, the CT value v of a virtual monochromatic X-ray image of any energy E desired by the user can be calculated using the 120 kV CT image and the inferred 60 keV virtual monochromatic X-ray image. E This can be calculated. In this case, equation (12) can be expressed as follows: v E =a(v E_inf -v tube_120 )+v tube_120 =a(v 60 -v tube_120 )+v tube_120 Here, v 60 This represents the CT value of the inferred 60 keV virtual monochromatic X-ray image.

[0104] Thus, the inferred virtual monochromatic X-ray image is not limited to a 50 keV virtual monochromatic X-ray image; a virtual monochromatic X-ray image of any energy other than 50 keV may also be inferred.

[0105] In this embodiment, the case where the tube voltage of the X-ray tube is 120kV is described. However, the tube voltage of the X-ray tube is not limited to 120kV, and a different tube voltage may be used instead of 120kV. If the tube voltage of the X-ray tube is extended to any tube voltage other than 120kV, equation (12) can be generalized to the following equation. v E =a(v E_inf -v tube_x )+v tube_x (13) Here, v tube_x This represents the CT value of the CT image obtained when the tube voltage is x (kV).

[0106] Therefore, the tube voltage of the X-ray tube is not limited to 120kV; even with a different tube voltage than 120kV, the CT value of a virtual monochromatic X-ray image at any energy desired by the user can be calculated.

[0107] Thus, the tube voltage is not limited to 120kV, and tube voltages other than 120kV can also be used.

[0108] (2) Second embodiment The second embodiment will be described in the same manner as the first embodiment, following the flow shown in Figure 11. Note that steps ST1 and ST2 in the second embodiment are the same as steps ST1 and ST2 in the first embodiment. Therefore, steps ST1 and ST2 will be briefly described, and step ST3 will be described in detail.

[0109] First, steps ST1 and ST2 are performed to acquire a CT image 31 and a 50 keV virtual monochromatic X-ray image 32. The CT image 31 and the 50 keV virtual monochromatic X-ray image 32 acquired in steps ST1 and ST2 are shown in Figure 13. After acquiring the CT image 31 and the 50 keV virtual monochromatic X-ray image 32, the process proceeds to step ST3.

[0110] In step ST3, a 40keV virtual monochromatic X-ray image is generated based on the CT image 31 and the 50keV virtual monochromatic X-ray image 32. Note that step ST3 in the second embodiment is different from step ST3 in the first embodiment, so step ST3 in the second embodiment will be described below (see Figure 16).

[0111] Figure 16 is a flowchart of step ST3 in the second embodiment, and Figure 17 is an explanatory diagram of step ST3.

[0112] In step ST31, the processor generates a difference image 34 between the CT image 31 and the virtual monochromatic X-ray image 32. After generating the difference image 34, the process proceeds to step ST32.

[0113] In step ST32, the processor calculates a scaling factor a corresponding to the 40 keV energy from the lookup table LUT1. 40 Read out (=2). Then, set the scaling factor a to the CT value of each pixel in the difference image 34. 40 Multiply by 2 to create multiplied image 35. After creating multiplied image 35, proceed to step ST33.

[0114] In step ST33, the processor adds the CT image 31 to the multiplied image 35. In this way, a 40keV virtual monochromatic X-ray image 36 can be created.

[0115] In the second embodiment, as in the first embodiment, a virtual monochromatic X-ray image of any energy desired by the user can be calculated by utilizing the lookup table LUT1. Therefore, in the second embodiment as well, there is no need to create a trained neural network 30 for each virtual monochromatic X-ray image for each energy (keV), thus reducing development effort and significantly lowering development costs.

[0116] (3) Third Embodiment In the first and second embodiments, examples of scanning a subject with the X-ray tube voltage set to 120kV were described. In the third embodiment, an example of scanning a subject with the X-ray tube voltage set to a different voltage than 120kV will be described. In the following, 100kV will be used as an example of a tube voltage different from 120kV, and an example of scanning a subject with a tube voltage of 100kV will be described, however, the tube voltage may be a different voltage other than 100kV.

[0117] Before specifically describing the third embodiment, we will first point out the problems that arise when scanning a subject with a tube voltage different from 120kV (100kV). After clarifying these problems, we will then specifically describe the third embodiment.

[0118] Figure 18 illustrates the problems that arise when scanning a subject with a tube voltage of 100kV. The 120kV CT images described in the first and second embodiments correspond to virtual monochromatic X-ray images at 70keV (see dashed line 21). On the other hand, CT images obtained with a tube voltage of 100kV can generally be considered to correspond to virtual monochromatic X-ray images at 64keV. In Figure 18, the position of 64keV is indicated by the dashed line 41. In other words, when comparing 100kV and 120kV, the energy (keV) is not the same, and a shift of about 6keV occurs. Therefore, when a CT image (64keV) obtained with a tube voltage of 100kV is input to the trained neural network 30, the inferred virtual monochromatic X-ray image will be a virtual monochromatic X-ray image with a different energy, offset by a certain energy ΔE from 50keV. In Figure 18, the energy of the virtual monochromatic X-ray image output by inputting a CT image (64keV) obtained with a tube voltage of 100kV to the trained neural network 30 is indicated by the dashed line 42. In other words, the energy of the inferred virtual monochromatic X-ray image is shifted by ΔE from 50 keV, which means that it is not possible to obtain a virtual monochromatic X-ray image at 50 keV.

[0119] One way to address this problem is to create a new pre-trained neural network 30 that infers a 50keV virtual monochromatic X-ray image from a 100kV CT image, separate from the pre-trained neural network 30 that infers a 50keV virtual monochromatic X-ray image from a 120kV CT image. However, this method requires the preparation of both a pre-trained neural network 30 for 100kV and a pre-trained neural network 30 for 120kV, which increases development time. Therefore, in the third embodiment, a 50keV virtual monochromatic X-ray image is generated from a 100kV CT image without creating a pre-trained neural network 30 for 100kV. This method is described below.

[0120] Figure 19 is an explanatory diagram illustrating the principle of generating a 50 keV virtual monochromatic X-ray image from a 100 kV CT image in a third embodiment.

[0121] First, we calculate the difference ΔCT3 between the CT value at 50 keV and the CT value at 70 keV. Since the CT value at 50 keV is 170 HU and the CT value at 70 keV is 100 HU, ΔCT3 can be calculated using the following formula. ΔCT3 = 170HU - 100HU =70HU

[0122] Next, we calculate the difference ΔCT4 between the CT value at 64keV and the CT value at 70keV. Since the CT value at 64keV is 115HU and the CT value at 70keV is 100HU, ΔCT4 can be calculated using the following formula. ΔCT4 = 115HU - 100HU = 15HU

[0123] Therefore, it can be considered that ΔCT3 and ΔCT4 have the following relationship. ΔCT4 / ΔCT3 = 15 / 70

[0124] Thus, the CT value at 64 keV is shifted by ΔCT4 / ΔCT3 = 15 / 70 compared to the CT value at 70 keV. Therefore, by shifting the inferred CT value vx of the virtual monochromatic X-ray image at energy Ex by vx(15 / 70), we can obtain (or approximate) the CT value of the virtual monochromatic X-ray image at 50 keV.

[0125] Therefore, ΔCT4 / ΔCT3 can be used as a conversion coefficient b to convert the CT value vx of the inferred energy Ex virtual monochromatic X-ray image to the CT value of the 50 keV virtual monochromatic X-ray image. In the example above, the conversion coefficient b is b = 15 / 70, but if necessary, the value (15 / 70)w, obtained by multiplying 15 / 70 by the weighting coefficient w, can also be used as the conversion coefficient b. In the above explanation, the conversion coefficient b was explained using curve 12, but approximately the same value of the conversion coefficient b can be obtained using curves 11 and 13-15 for other parts of the body. Therefore, regardless of the imaging site, using the above conversion coefficient b makes it possible to convert the CT value vx of the inferred virtual monochromatic X-ray image to the CT value of the 50 keV virtual monochromatic X-ray image.

[0126] Furthermore, although the above description described a CT image with a tube voltage of 100kV, the corresponding conversion coefficient b can be determined for other tube voltages according to the above description. In the third embodiment, the storage device stores a lookup table that represents the correspondence between the tube voltage x of the X-ray tube and the conversion coefficient b (see Figure 20).

[0127] Figure 20 is an explanatory diagram of the lookup table LUT2 stored in the memory device. The lookup table LUT2 has columns for the tube voltage x of the X-ray tube and the conversion coefficient b. In Figure 20, 80kV, 90kV, 100kV, 110kV, 120kV, and 130kV are shown as examples of tube voltage x, but it is not limited to these energies, and any tube voltage can be used as an example of tube voltage x in the lookup table LUT2. Furthermore, in the column for the conversion coefficient b, the conversion coefficient b corresponding to each tube voltage is listed as "b 80 "b 90 "b 100 "b 110 "b 120 ", and "b 130 This is indicated by "[...]."

[0128] Therefore, the lookup table LUT2 has a tube voltage x = 80kV to 130kV and a conversion coefficient b = b 80 ~b 130 This shows the correspondence between the two.

[0129] The following describes the flow for acquiring a 50 keV virtual monochromatic X-ray image using the lookup table LUT2 in a third embodiment.

[0130] Figure 21 is a flowchart for acquiring a 50 keV virtual monochromatic X-ray image, and Figure 22 is an explanatory diagram of the flowchart in Figure 21.

[0131] In step ST51, the subject is scanned under scanning conditions in which a tube voltage of 100kV is applied to the X-ray tube. The processor reconstructs a CT image of the subject based on the data obtained by scanning the subject. Figure 22 schematically shows the CT image 51 obtained by scanning the subject. After scanning the subject, the process proceeds to step ST52.

[0132] In step ST52, the processor uses the trained neural network 30 to infer a virtual monochromatic X-ray image 52 from the CT image 51, as shown in Figure 22. In the third embodiment, the tube voltage used to acquire the CT image 51 is 100kV. The trained neural network 30 is trained to infer a 50keV virtual monochromatic X-ray image from a 120kV CT image. Therefore, when the trained neural network 30 infers a virtual monochromatic X-ray image 52 from a 100kV CT image 51, as explained earlier, the inferred virtual monochromatic X-ray image 52 is not a 50keV virtual monochromatic X-ray image, but a different virtual monochromatic X-ray image 52 with an energy Ex offset by a certain energy ΔE from 50keV. Thus, simply inferring a virtual monochromatic X-ray image 52 from the CT image 51 using the trained neural network 30 does not yield a 50keV virtual monochromatic X-ray image. Therefore, in the third embodiment, the CT values ​​of the virtual monochromatic X-ray image 52 inferred by the trained neural network 30 are converted to the CT values ​​of a 50 keV virtual monochromatic X-ray image. To perform this conversion, the process proceeds to step ST53. Step ST53 uses the lookup table LUT2 to convert the CT value vx of the virtual monochromatic X-ray image at energy Ex to the CT value v50 at 50 keV. Step ST53 is described below. Since step ST53 includes steps ST531 to ST532, each step will be explained in order.

[0133] In step ST531, the processor converts the lookup table LUT2 to the conversion coefficient b, as shown in Figure 22. 80 ~b 130 From the options, select the conversion coefficient b corresponding to 100kV. For 100kV, the conversion coefficient b = b 100 Therefore, the processor uses the conversion coefficient b corresponding to 100kV as b=b 100 Select b=b 100 After making your selection, proceed to step ST532.

[0134] In step ST532, the processor selects the conversion coefficient b 100 Based on this, the CT value vx of the virtual monochromatic X-ray image at energy Ex is converted to the CT value v50 at 50 keV. Specifically, the CT value is converted as follows.

[0135] As shown in Figure 22, the processor identifies the pixel Pj of interest from the virtual monochromatic X-ray image 52 of energy Ex (keV). Then, it determines the CT value v of the pixel Pj of interest. j Read out the conversion coefficient b 100 ni v j Multiply by this. This gives the pixel P of the 50keV virtual monochromatic X-ray image 53. k CT value v k It is possible to calculate this.

[0136] In the above explanation, the CT value v of pixel Pj in the virtual monochromatic X-ray image 52 with energy Ex (keV) is shown. j We have explained an example of converting one pixel, but other pixels can be converted in the same way. Therefore, a virtual monochromatic X-ray image 53 of 50 keV can be obtained. This is how the flow ends.

[0137] In the third embodiment, after obtaining a CT image 51 with a tube voltage of 100kV, a virtual monochromatic X-ray image 52 is inferred from the CT image 51 using a trained neural network 30. However, the trained neural network 30 is trained to infer a 50keV virtual monochromatic X-ray image from a 120kV CT image. Therefore, when the CT image 51 obtained with a tube voltage of 100kV is input to the trained neural network 30, a virtual monochromatic X-ray image 52 offset by a certain amount of energy from 50keV is output. Thus, in the third embodiment, in order to obtain a 50keV virtual monochromatic X-ray image 53, a conversion coefficient b is used to convert the CT value of the virtual monochromatic X-ray image 52 output from the trained neural network 30 to the CT value of a 50keV virtual monochromatic X-ray image. Therefore, in the third embodiment, a 50 keV virtual monochromatic X-ray image 53 can be generated without having to prepare a new trained neural network for 100 kV in addition to the trained neural network 30 for 120 kV. As a result, there is no need to prepare a trained neural network 30 for each X-ray tube voltage, which reduces development effort and significantly lowers development costs.

[0138] Furthermore, after generating a 50 keV virtual monochromatic X-ray image 53 in step ST530, virtual monochromatic X-ray images of other energies other than 50 keV can be calculated using the lookup table LUT1 (see Figure 14), similar to the first embodiment.

[0139] In the third embodiment, the 100kV CT image 51 is described as an image corresponding to a 64keV virtual monochromatic X-ray image. However, when an X-ray is irradiated onto a subject, the energy characteristics of the X-rays absorbed by the subject differ depending on the size of the subject. Therefore, depending on the size of the subject, the 100kV CT image 51 does not necessarily correspond to a 64keV virtual monochromatic X-ray image. For example, generally, the larger the size of the subject, the more the energy of the virtual monochromatic X-ray image corresponding to the 100kV CT image shifts to the higher energy side, and the smaller the size of the subject, the more the energy of the virtual monochromatic X-ray image corresponding to the 100kV CT image shifts to the lower energy side. Therefore, it is desirable to determine the energy (keV) of the virtual monochromatic X-ray image corresponding to the 100kV CT image based on the size of the subject. An example of the size of the subject is the diameter of the imaging area. For example, the processor can determine the water equivalent diameter Dw based on the CT image of the subject (e.g., a scout image), and then determine the diameter of the imaging area based on the determined water equivalent diameter Dw. Therefore, the energy of the virtual monochromatic X-ray image corresponding to the diameter of the imaging area can be determined. Specifically, by storing in the memory a plurality of diameters of the imaging area (e.g., 20cm, 25cm, 30cm, and 35cm, etc.) and a third lookup table corresponding to each diameter, which represents the correspondence between a plurality of tube voltages and a plurality of conversion coefficients b, the third lookup table corresponding to the determined diameter of the imaging area can be selected. Therefore, since the conversion coefficient corresponding to the tube voltage can be identified based on the selected third lookup table, the accuracy of the estimated virtual monochromatic X-ray image can be further improved. [Explanation of Symbols]

[0140] 11, 12, 13, 14, 15 curves 21, 22, 23, 41, 42 Dashed lines 30 trained neural networks 31, 51 CT images 32, 33, 36, 52, 53 Virtual monochromatic X-ray images 34 difference pixels 35 Multiply Image 100 CT System 102 Gantry 103 Filter section 104 X-ray tube 105 Pre-collimator 106 X-ray beam 108 X-ray detectors 112 subjects 116 Tables 118 Table Motor Controller 202 detector element 210 X-ray controller 212 Gantry Motor Controller 214 DAS 216 Computers 218 Storage device 220 Operator Console 224 PACS 230 Image Reconstructors 232 displays

Claims

1. It is a CT system, An X-ray tube to which tube voltage is applied, One or more processors, Reconstructing a first CT image based on data obtained by scanning a subject under scanning conditions in which a first tube voltage is applied to the X-ray tube. Based on the first CT image, a virtual monochromatic X-ray image of a second energy different from the first energy corresponding to the first tube voltage is inferred. The CT value of the virtual monochromatic X-ray image of the third energy is determined based on a first coefficient that defines the relationship between the second CT value corresponding to the second energy and the third CT value corresponding to the third energy, using the first CT value corresponding to the first energy as a reference. One or more processors that execute A CT system, including [a specific feature / feature].

2. The one or more processors described above are The CT system according to claim 1, wherein a trained neural network is used to infer a virtual monochromatic X-ray image of the second energy from the first CT image.

3. The one or more processors described above are The CT system according to claim 1, wherein the CT value of the virtual monochromatic X-ray image of the third energy is determined using a first lookup table that represents the correspondence between multiple energies of a virtual monochromatic X-ray image and multiple first coefficients.

4. The one or more processors described above are Selecting a first coefficient corresponding to the first energy from among the plurality of first coefficients in the first lookup table, To generate a virtual monochromatic X-ray image of a third energy based on a selected first coefficient, the first CT image, and the virtual monochromatic X-ray image of a second energy. The CT system according to claim 3, which performs the following.

5. The generation of the third energy virtual monochromatic X-ray image is To generate a difference image between the first CT image and the second energy virtual monochromatic X-ray image, The CT value of the difference image is multiplied by a first coefficient selected from the plurality of first coefficients to generate a multiplied image, and By adding the multiplied image to the first CT image, a virtual monochromatic X-ray image of the third energy is generated. The CT system according to claim 3, including the following:

6. The one or more processors described above are Selecting the value of the first coefficient corresponding to the first energy from among the plurality of first coefficients in the first lookup table, The CT value of the difference image is multiplied by the selected first coefficient to generate a multiplied image. The CT system according to claim 5, which performs the following.

7. The aforementioned one or more processors are Reconstructing a second CT image based on data obtained by scanning a subject under scanning conditions in which a second tube voltage is applied to the X-ray tube. Using the aforementioned trained neural network, a virtual monochromatic X-ray image with a different energy than the second energy is inferred from the second CT image. Using a second coefficient, the CT value of the virtual monochromatic X-ray image of the other energy is converted to the CT value of the virtual monochromatic X-ray image of the second energy. The CT system according to claim 1, which performs the following.

8. The aforementioned one or more processors are The CT system according to claim 7, wherein a second lookup table representing the correspondence between multiple tube voltages and multiple second coefficients is used to convert the CT value of the virtual monochromatic X-ray image of the other energy to the CT value of the virtual monochromatic X-ray image of the second energy.

9. The one or more processors described above are Selecting a second coefficient corresponding to the second tube voltage from among the plurality of second coefficients in the second lookup table, The CT system according to claim 8, which converts the CT value of the virtual monochromatic X-ray image of the other energy to the CT value of the virtual monochromatic X-ray image of the second energy based on a selected second coefficient.

10. The one or more processors described above are The CT system according to claim 7, wherein the CT value of the virtual monochromatic X-ray image of the other energy is converted to the CT value of the virtual monochromatic X-ray image of the second energy using the size of the subject and a second coefficient.

11. The CT system according to claim 10, wherein the size of the subject is the diameter of the imaging area of ​​the subject.

12. The one or more processors described above are The CT system according to claim 11, wherein the diameter of the imaging area is determined based on a scout image of the subject.

13. The storage device contains, Multiple diameters of the area being scanned, A third lookup table corresponding to each diameter, the third lookup table representing the correspondence between multiple tube voltages and multiple second coefficients It is remembered, The one or more processors described above are The CT system according to claim 12, wherein a second coefficient corresponding to the second tube voltage is identified based on a third lookup table corresponding to the diameter of the imaging area determined based on the scout image.

14. The one or more processors described above are The CT system according to claim 12, wherein the water equivalent diameter is determined based on a scout image of the subject, and the diameter of the imaging area is determined based on the determined water equivalent diameter.

15. One or more non-temporary computer-readable storage media in which instructions are stored, wherein when an instruction is executed by one or more processors, the instructions are sent to one or more processors. Reconstructing a first CT image based on data obtained by scanning a subject under scanning conditions in which a first tube voltage is applied to the X-ray tube. Based on the first CT image, a virtual monochromatic X-ray image of a second energy different from the first energy corresponding to the first tube voltage is inferred. The CT value of the virtual monochromatic X-ray image of the third energy is determined based on a first coefficient that defines the relationship between the second CT value corresponding to the second energy and the third CT value corresponding to the third energy, using the first CT value corresponding to the first energy as a reference. A storage medium that performs operations including those mentioned above.

Images