JP5047945B2 - X-ray CT scan simulator apparatus, X-ray CT apparatus, and X-ray CT scan simulator program - Google Patents

Description

  The present invention relates to an X-ray CT scan simulator apparatus, an X-ray CT apparatus, and an X-ray CT scan simulator program, and particularly to a technique useful for image noise simulation and prior examination of appropriate imaging conditions.

  This application is a patent application claiming priority based on Japanese Patent Application No. 2006-103547 and Japanese Patent Application No. 2007-059408 based on the Japanese Patent Law. No. 2006-103547 and No. 2007-059408 This application is incorporated by reference to enjoy the benefits of the issue.

  Conventional X-ray CT devices detect X-rays radiated onto a fan shape from an X-ray source with a single-row detector, but in recent years, X-rays radiated in a cone shape are detected with a multi-row detector. Detection multi-slice CT has been put into practical use. Not only the number of detectors but also the thinning is remarkable.

  Technological innovations in X-ray CT systems are not limited to detectors, and the speed of gantry rotation has been increasing, enabling a wide range of data to be collected in a short time. Since a much wider range can be scanned compared to the conventional case, the increase in the exposure dose becomes remarkable when photographing with the same tube current as in the conventional case. Even when the imaging range is relatively narrow, examinations with high local dose are increasing due to the long imaging time, such as CT Perfusion and ECG.

  For these reasons, it is important to reduce exposure by low-dose imaging such as low-tube voltage imaging and low-tube current imaging. However, low-dose imaging causes an increase in image noise. Since image noise depends on the slice thickness, photographing with a thin slice thickness causes an increase in image noise.

  A technique called Auto Exposure Control (AEC) as shown in Patent Document 1 is put into practical use in order to obtain uniform image quality regardless of the subject's body shape while suppressing unnecessary X-ray exposure. It has become. In AEC, usually, an operator designates a target image SD, and a tube voltage value and a tube current value so as to become the target SD are automatically set.

  The number of SD images that can be used for diagnosis depends on the type and size of the disease. Even for the same disease, the required image quality differs depending on the purpose of the examination (whether it is a medical examination or a close examination). For this reason, it is not easy to set the optimal imaging conditions according to the examination purpose, the disease, and the size of the subject while suppressing the exposure dose. Therefore, it is clinically useful to know what kind of image can be obtained under certain imaging conditions before the examination, and simulation techniques for examining the optimal imaging conditions have been proposed (Patent Document 2, Patent Document 3). reference).

JP 2004-073865 JP JP 2004-329661 A JP 2004-57831 A

  With the technique disclosed in Patent Document 1, an image with a desired noise amount and SD value can be obtained. However, it is not easy to set an optimal target SD value for obtaining an image suitable for the inspection purpose while suppressing unnecessary exposure, and there is a problem that abundant experience and advanced knowledge are required.

  With the technique disclosed in Patent Document 2, it is possible to simulate what kind of image is obtained under any photographing condition before photographing. Since the image quality can be visually confirmed while viewing the simulation image displayed on the screen, the setting of the optimum shooting condition is simplified compared to the technique disclosed in Patent Document 1. However, in the technique disclosed in Patent Document 2, since it is necessary to prepare sample images having various SDs for each imaging region in advance, the sample images to be prepared are enormous. If the sample image does not contain an SD image that is expected from the shooting conditions, a simulation image is created by interpolation processing from two images that have SD close to it. For this reason, there is a possibility that the error may increase depending on the photographing conditions, and there is a possibility that the reliability of the simulation accuracy may be lacking.

  In the technique disclosed in Patent Document 3, a simulation image is created by performing reconstruction after adding quantum noise and electrical noise to measured projection data. Therefore, unlike the technique disclosed in Patent Document 2, interpolation processing is not included, and it is considered that the accuracy of simulation is improved. However, the creation of a simulation image requires actually measured projection data, so that there is a problem that it can be applied only when there is a history of photographing the subject in the past.

  Further, in the technique disclosed in another embodiment of Patent Document 3, a method of adding a noise pattern image created in advance to a reconstructed image is shown. In the case of this method, it is necessary to prepare various noise pattern images in advance according to shooting conditions. Each shooting parameter such as tube voltage, tube current, and FOV has many patterns, so there is a problem that a huge amount of noise pattern images must be prepared to handle all shooting conditions. is there.

  An object of the present invention is to provide an X-ray CT scan simulator device, an X-ray CT scan simulator device, and an X-ray CT simulation device capable of highly accurate simulation regardless of whether or not there is a history of taking a subject in the past without preparing enormous sample images and noise pattern images. To provide a CT apparatus and an X-ray CT scan simulator program.

  In order to solve the above problem, an X-ray CT scan simulator apparatus according to the present invention includes an image storage unit that stores a reference image, a target noise value setting unit that sets a noise target value of a desired image, and the set Noise image generation means for generating a noise image based on a target noise value, simulation image generation means for generating a simulation image by synthesizing the generated noise image and the reference image, and display means for displaying the simulation image It is characterized by providing.

  Further, the X-ray CT apparatus according to the present invention, an X-ray source that irradiates a subject with X-rays, an X-ray detector that is disposed opposite to the X-ray source and detects X-rays transmitted through the subject, A tomographic image of the subject is reproduced based on a rotating device that is mounted with the X-ray source and the X-ray detector and rotates around the subject, and transmitted X-ray doses in a plurality of directions detected by the X-ray detector. An X-ray CT apparatus comprising: an image reconstruction device to be configured; an imaging condition input device for inputting the X-ray irradiation conditions and image reconstruction conditions; and an image display device for displaying the tomographic image. Further, the X-ray CT scan simulator device is further mounted.

  The X-ray CT scan simulator program according to the present invention includes a step of acquiring a reference image, a step of setting a noise target value of a desired image, and a step of generating a noise image based on the set target noise value. And causing the computer to execute a step of creating and outputting a simulation image by combining the generated noise image and the reference image.

  According to the present invention, in the X-ray CT scan simulator, before measuring the subject by simulating what kind of image is obtained under the condition of the target SD value regardless of the presence or absence of the past imaging history There is an effect that it is possible to examine the optimum shooting conditions.

  Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

<First embodiment>
[Hardware configuration]
FIG. 1 is a diagram showing a preferred embodiment of an X-ray CT apparatus equipped with an X-ray CT scan simulator according to the present invention. An X-ray CT apparatus 1 equipped with an X-ray CT scan simulator according to the present invention includes a gantry 2, and the gantry 2 includes an X-ray source 3, a collimator 4, and a detector array 5 located on the opposite surface of the gantry 2. Have.

  The detector array 5 is formed by detector elements 6 that detect X-rays transmitted through a subject on a bed (not shown). The detector elements 6 are arranged in rows or in a plurality of parallel rows. Each detector element 6 generates an electrical signal representing the intensity of the incident X-ray beam, in other words, attenuation when the X-ray beam passes through the subject.

  X-ray projection data is collected by rotating the gantry 2 around the rotation center 8 in a state where X-rays 7 are irradiated from the X-ray source 3. The country 2 and the X-ray source 3 are controlled by the control unit 9 of the X-ray CT apparatus 1.

  The control unit 9 includes an X-ray control means 10, a gantry control means 11, and a DAS (data acquisition system) 12, and an analog signal from the detector element 6 is converted into a digital signal by the DAS 12. The digitized X-ray data is reconstructed by the reconstruction means 19 in the arithmetic processing means 13 and stored in the storage means 22 in the arithmetic processing means 13.

  Arithmetic processing means 13 is an arithmetic processing unit such as a computer, and body thickness estimation means 14 that obtains the body thickness of the subject, and an approximation that obtains a water phantom having an X-ray absorption equivalent to the subject by converting the body thickness to a water equivalent thickness Model calculation means 15, image SD calculation means 16 for obtaining an image SD in an image taken in the past, and addition noise for obtaining a noise amount to be added to an original image (also referred to as a reference image) in order to create an image of a desired SD A quantity calculation means 17, a water approximation model projection data creation means 18 that generates projection data of an approximate model having an X-ray absorption equivalent to the subject, a reconstruction means 19 that performs image reconstruction from the projection data, and noise Noise image creation means 20 for obtaining a noise image from the projection data of the approximate model added and the projection data of the approximate model not adding noise, and a simulation image by adding the noise image to the original image Simulation image generating means 21 for obtaining a storage means 22 such as a hard disk, temporary storage means 23 such as a memory, and an input means 24 such as a mouse or keyboard.

  Further, at least one of a digital signal processor (DSP), a micro processor unit (MPU), and a central processing unit (CPU) (not shown) is provided. The image display means 25 is a display device such as a display integrated with the arithmetic processing means 13 or independent. In FIG. 1, the control unit 9 and the arithmetic processing unit 13 are separated, but both may be integrated. The reconstruction means 19 may be an arithmetic unit independent of the arithmetic processing means 13.

[Configuration of X-ray CT scan simulator]
FIG. 2 is a configuration diagram of the X-ray CT scan simulator 200 of the present embodiment. The X-ray CT scan simulator 200 includes a storage unit 22 that is an image storage unit, an input unit 24 that is a target noise value setting unit, a noise image generation unit 30, and a simulation image generation unit 21.

  The storage unit 22 stores a reference image serving as a reference for a simulation image created by the X-ray CT scan simulator 200. The reference image is a past image obtained by CT imaging of the subject in the past or a human phantom image obtained by CT imaging of a human phantom that faithfully reproduces the internal tissue of the human body. Sufficient X-ray dose should be used to capture human phantom images so that clear images can be obtained.

  The input means 24 is for inputting a target noise value, and specifically, is a keyboard, a mouse or the like.

  The noise image generation means 30 generates and outputs a noise image based on the target noise value input by the input means 24. The noise image generating means 30 includes a body thickness estimating means 14, an approximate model calculating means 15, an approximate model projection data creating means 18, an image SD calculating means 16, an added noise amount calculating means 17, and a noise image creating means 20.

  The body thickness estimation means 14 acquires the body thickness of the subject. The body thickness may be manually input by the operator or may be obtained from a scanogram. If the reference image is a past image, it may be automatically estimated by a method described later.

  The approximate model calculating means 15 calculates an approximate model based on the body thickness estimated by the body thickness estimating means 14. The approximate model may be a model approximated with water so as to have the same amount of X-ray absorption as the subject (hereinafter referred to as water approximation model), and the amount of X-ray absorption equivalent to the subject obtained from the average CT value of the subject A model using a virtual substance having the following (hereinafter referred to as an average CT value substance approximate model) may be used. A method for calculating the approximate model will be described later.

  The approximate model projection data creating means 18 creates projection data by irradiating the approximate model calculated by the approximate model calculating means 15 in a simulated manner with X-rays. A method for creating projection data will be described later.

  The image SD calculation means 16 calculates the noise amount of the reference image. A method for calculating the noise amount of the reference image will be described later.

  Based on the target noise value set by the input unit 24 and the noise amount calculated by the image SD calculation unit 16, the addition noise amount calculation unit 17 calculates a noise amount to be added to the reference image.

  The noise image creating means 20 creates a noise image by reconstructing the added projection data obtained by adding the amount of noise to be added to the approximate model projection data.

  The simulation image creating unit 21 creates a simulation image by synthesizing the noise image generated by the noise image generating unit 30 and the reference image.

[Process flow]
FIG. 3 is a processing flow showing a method for creating a simulation image in the present embodiment. First, in the subject to be simulated, it is searched whether there is a history of imaging the same part in the past (step S201), and if there is an imaging history, an image captured in the past (past image) is the original image (Step S202).

  If there is no shooting history, an image obtained by shooting a phantom (human body phantom) that faithfully reproduces the internal composition of the human body is set as the original image (step S203). Prepare an image of the human phantom in advance. In the publicly known example, it is necessary to prepare various SD images for human body phantoms for each imaging region, but in the present invention, an image captured with a sufficient dose is obtained for each region so that a clear image can be obtained. If so, in other words, it is sufficient to prepare one image with a small SD for each part.

  As for the FOV, it is only necessary to take a photograph with a representative FOV for each part, and to enlarge or reduce this appropriately to match the FOV in the simulation. Therefore, it is not necessary to prepare an enormous amount of human phantom images for each SD and FOV as in the known example.

  Next, the body thickness of the subject is automatically estimated (step S204). The body thickness may be obtained from a scanogram, or a method as described later may be used. Although the operator may directly input the body thickness in step S204, it is desirable to automatically estimate the body thickness in order to reduce the burden on the operator.

  Next, a water equivalent thickness is obtained based on the body thickness obtained in step S204, and an approximate model having this water equivalent thickness is calculated (step S205). The calculation method of the water equivalent thickness and the approximate model will be described later. Next, the image SD in the original image is calculated (step S206). A method for calculating the image SD in the original image will be described later.

  Next, the target image SD value is input (step S207). The SD value input here is an arbitrary SD value desired by the operator. In step S207, the SD value may be input directly, or X-ray conditions (tube voltage, tube current, slice thickness, etc.) may be input. In this case, the input X-ray condition and the size of the approximate model may be converted into SD.

  Next, the SD value in the original image obtained in step S206 is compared with the target SD value input in step S207, and the amount of noise to be added to the original image is calculated (step S208).

  Next, projection data of the approximate model is created (step S209). Two types of projection data are created: an ideal system that does not include noise, and a projection data in which noise is added by the amount of noise obtained in step S208.

  Next, a noise image is created from the noise-free projection data obtained in step S209 and the projection data obtained by adding the noise (step S210). The noise image is obtained by subtracting the difference between the reconstructed image of the projection data added with noise and the reconstructed image of the projection data not adding noise, or the difference between the projection data added with noise and the projection data not adding noise. Obtained by reconfiguration. Next, the noise image obtained in step S210 is added to the original image set in step S202 or step S203 to create a simulation image (step S211).

[Body thickness estimation method]
The body thickness estimation method in S204 will be described. Fig. 4 shows the body thickness estimation method. First, a mask image 31 in which the pixel value of the pixel in the subject is 1 and the pixel values of the other pixels are 0 is created from the original image 30, and the inertia main axis 32 in the mask image 31 is obtained. As a method for creating the mask image 31, a method as disclosed in Japanese Patent Application Laid-Open No. 2004-097665 may be used, or any other known method may be used.

  Here, the distance between the intersections P1, P2 of the inertia main axis 32 and the contour of the mask image 31 is the body thickness in the inertia main axis direction. The inertial main axis 32 is obtained according to the following formula 1, where I (x, y) is the pixel at the inertial main axis q and the coordinates (x, y).

{Number 1}

    In the formula 1, a, b, and c are expressed by the following formula 2.

{Number 2}

  There is also a method using a scanogram for estimating the body thickness. The scanogram is analyzed, and the body thickness in the AP direction (or LR direction) is obtained from the maximum attenuation amount of X-rays attenuated in the subject.

[Method of calculating water equivalent thickness and approximate model from body thickness]
A method of calculating the water equivalent thickness and the approximate model from the body thickness of S205 will be described. The body thickness can be obtained from the method described above with reference to FIG. Here, when the pixel value of the pixel located in the original image 31 on the inertial main axis 32 shown in FIG. 4 is scanned and the average CT value is CA, the water permeation length L1 in the inertial main axis direction is obtained by the following equation (3). It is done.

{Number 3}

  In Equation 3, l represents the distance between P1 and P2. Assuming that the subject is water equivalent, L1 obtained by Equation 3 is the water equivalent thickness in the inertial main axis direction. The water equivalent thickness in the direction perpendicular to the inertial main axis may be obtained assuming that the cross section of the subject is an ellipse. If the average pixel value of all the pixels in the mask is CS, the water equivalent thickness L2 in the direction perpendicular to the principal axis of inertia can be obtained by the following equation (4).

{Number 4}

  In Equation 4, S represents the number of pixels in the mask, that is, the area. The approximate model is an ellipse having the same center of gravity as the original image, the major axis length L1 (or L2), and the minor axis length L2 (or L1).

  When the body thickness is obtained from the scanogram, the water depth for an arbitrary amount of attenuation is known, so if the subject is assumed to be water equivalent, the AP direction from the body thickness in the AP direction (or LR direction) described above is used. The water equivalent thickness (or LR direction) can be calculated. The water equivalent thickness in the LR direction (or AP direction) may be obtained assuming that the cross section of the subject is an ellipse.

  The sum of the area of the ellipse and the attenuation is considered to be proportional, and the area of the ellipse is proportional to the product of the major and minor axes of the ellipse, so the sum of the attenuation (= the area of the ellipse) and the AP direction (or LR direction) It is possible to calculate the water equivalent thickness (= ellipse major axis or minor axis) in the LR direction (or the ellipse major axis) from the water equivalent thickness (= ellipse minor axis or major axis). The ellipse thus calculated becomes an approximate model. The method for obtaining the water equivalent thickness from the scanogram is not limited to this, and any known method may be used.

  As described above, the approximate model may approximate the entire object to be imaged with a water phantom based on the water equivalent thickness, but the object to be imaged includes those with greatly different X-ray absorption characteristics such as bone and soft tissue. In this case, it is desirable to segment the object to be imaged for each organ and approximate it with different X-ray absorbing materials for each segmented region. What is necessary is just to assign the absorption coefficient (micro | micron | mu) (alpha) obtained by following Formula at the time of the projection data preparation mentioned later to each segmented part.

{Equation 5}

  Here, μw is the water X-ray absorption coefficient, and ΔCT is the difference between the CT value of each segmented part and the CT value of air. Any known technique may be used for segmentation. This segmentation assigns a calcium absorption coefficient in order to approximate the inner area of a predetermined width from the outer edge of the ellipse to the bone (skull), for example, when photographing the head, and the area further inside than that area is soft tissue. May be used to assign a water absorption coefficient to approximate

[Method of calculating image SD in original image]
A method for calculating the image SD in the original image in S206 will be described. The number of views used for reconstructing the CT image Img (z) at the slice position z is M, and the convenient view number m is m = 0 to M−1. When the number of views per rotation is N, the number of used views M is not necessarily equal to the number of views N per rotation. Here, the X-ray attenuation index T can be expressed as a function T (m) of the view number to be used. Assuming that the maximum value of the X-ray attenuation index T in view number m = 0 to M-1 is Tmax (0: M-1), and the reference tube current value i_ref is assumed to correspond at this time, the tube current for view number m The value iv (m) is expressed by the following equation (6).

{Equation 6} iv (m) = i_ref * exp (T (m)-Tmax (0: M-1))
On the other hand, assuming that the time trot for which the scanner makes one revolution is equal to the reference time trot_ref, and the X-ray attenuation index T is a constant value during that time, xv is used as the tube voltage, and the reference tube current value i_ref is used as the tube current value i. The image noise variance value V when the image thickness thk is reconstructed as the reference image thickness thk_ref using the reconstruction filter function g with equal weighting on the rotating view number N_ref is a function of the X-ray attenuation index T As shown in the following equation (7).

{Number 7}

  The image noise variance prediction value V * when the tube current value iv (m) expressed by the above-described equation 6 is used is expressed by the following equation 8.

{Number 8}

  Here, w (m) in Expression 8 is a view direction weight applied to each view. Since the parameters such as tube voltage, tube current, reconstruction filter function, and image thickness are known in the original image, the image noise variance prediction value V * in the original image can be calculated according to Equation 8, and the square root of V * To obtain the SD value in the original image.

[Method for calculating the amount of additive noise]
A method of calculating the added noise amount in S208 will be described. Let X be the energy of X-ray particles (photons), and P be the number of X-ray particles (photons). E is the tube voltage kV, P is an amount determined depending on the tube voltage kV and tube current mA, and the dispersion of photon noise (SD of photons) Nq is expressed by the following equation (9).

{Number 9}

  Here, β is a constant determined depending on the apparatus, and β is obtained empirically from the signal intensity when photographing without a subject. g is the energy conversion efficiency of the detector, and η is the X-ray detection efficiency of the detector. The photon energy, the number of photons, the photon noise, and the system noise in the X-ray conditions when the original image (SD is σ0) is taken as E0, P0, Nq0, and Ns0, respectively, and the photon energy, the number of photons in the simulation image (SD is σt), Assuming that photon noise is Et, Pt, Nqt, and Nst, respectively, the variance of the photon noise amount Nadd that should be added to the original image to obtain a simulation image is expressed by the following equation (10).

{Number 10}

  For simplification of calculation, the term of system noise in Formula 10 may be ignored, but it is desirable to consider not only photon noise but also the influence of system noise in order to perform a detailed simulation. System noise under each imaging condition can be obtained by scanning the X-ray tube with a sufficiently thick lead plate that does not transmit X-rays.

  As a method for calculating the amount of added noise, the following method is also conceivable. SD in the original image is σ0, and the target image SD is σt. When a noise image is added to the original image to create a simulation image, SDσa in the noise image is obtained by the following equation (11).

{Number 11}

  FIG. 5 is an example showing the relationship between transmission length and SD in a water phantom. The curves shown in FIG. 5 can be obtained by photographing water phantoms of various diameters and measuring SD at that time. It is desirable to prepare the curve shown in FIG. 5 for every tube voltage that can be set by the CT scanner to be used. Since the transmission length (transmission thickness) of the approximate model is obtained by using the method as described above in step S205, referring to the curve shown in FIG. 5, the mAs value for obtaining σa calculated by Equation 11 Is obtained. Since the noise amount at an arbitrary mAs value can be obtained empirically by actually measuring the phantom, this may be used as the added noise amount.

  The curve shown in FIG. 5 may be prepared in advance by preparing only a curve at a typical mAs value, and curves at other mAs values may be calculated using the fact that SD is proportional to the square root of the mAs value. However, in this case, an error may increase under imaging conditions in which system noise is conspicuous. Therefore, it is desirable to prepare curves in advance for all mAs values that can be set in the CT apparatus to be used.

[How to create projection data of approximate model]
The approximate model projection data creation method of S209 will be described. FIG. 6 is an explanatory diagram of a method for creating projection data. When the energy intensity of X-rays with energy E and I _E, I _E is expressed by the following equation.

{Number 12}

  Where N is the Avogadro number, ρ and A are the target density and atomic weight, m0 is the mass of the electron, c is the speed of light, T0 is the energy of the incident electron, and Q is the energy intensity of the X-ray emitted from one electron. is there. Q is an approximate expression determined by the ratio E / T of photon energy E and electron energy T. In Non-Patent Document 1 below, an approximate expression is required.

R. Bitch and M. Marshall, Computation of Bremsstrahlung X-ray Spectra and Comparison with Spectra Measured with a Ge (Li) Detector, Phys. Med. Biol., Vol24, No.3, 505-517, 1979 dT / dl This is the amount of energy loss per unit length and is called stopping power. The stopping power is described in Non-Patent Document 2 for each element.

Katsuhiko Yamada, Hiroki Nohara, Radiation Metrology, Chapter 2, Japan Society of Radiological Technology, 1983. μE is the X-ray absorption coefficient of the target in the X-ray tube, and l is the electron incident distance. Equation 12 is an energy intensity distribution, and IE / E may be calculated if a photon number distribution is obtained. Equation 12 represents the X-ray spectrum by bremsstrahlung, and the characteristic X-ray Ich emitted from the excited atom is approximated by the following equation (13).

{Number 13}

  TK and L are energy required to remove electrons from the K and L electron orbits, and can be obtained by comparing the absolute intensity of Ich with the value described in Non-Patent Document 3.

HPA. Catalog of Spectctral Data for Diagnostic X-rays, SRS-30, Hospital Physicist's Association, 1979 The approximate model projection data can be calculated using linear attenuation characteristics and transmission distance. As shown in the following formula 14, for each photon energy E, the X-ray absorption coefficient μn of each X-ray absorbing material n is calculated, and the projection value at each photon energy is calculated based on the multiplication value with the transmission path length ln. Projection data is created by the product sum of the contribution rate W (E) of each energy obtained from the effective spectrum distribution and the projection value.

{Number 14}

  Here, Nadd is the above-described additive noise. FIG. 7 shows an outline of calculation of the transmission path length. Considering a three-dimensional coordinate system with an arbitrary point (e.g., IISO center) as the origin, the geometric coordinates of the X-ray source, detector, and approximate model can be obtained. Thus, an X-ray path incident on an arbitrary detector can be expressed by a linear equation.

  Since the method for obtaining the intersection of the ellipse and the straight line is known, the coordinates of the two intersections can be easily obtained, and the distance between the intersections, that is, the transmission path length (L1 and L2) can be obtained. Such an operation may be performed for all X-ray paths. In the example of FIG. 7, for the sake of simplicity, an elliptical simulated organ with an X-ray absorption coefficient μ2 is arranged in an elliptical simulated human body with an X-ray absorption coefficient μ1, but a larger number of organs and different organs are shown. Those skilled in the art will readily understand that projection data can be obtained in the same manner even when a shape-shaped organ is arranged.

  In order to perform a more accurate simulation, it is desirable to consider the attenuation of X-rays in the compensation filter and scintillator when creating projection data. In the present embodiment, projection data is obtained by creating pseudo-projection data of an approximate model. However, if there is an imaging history in the past, projection data may be obtained by reprojecting a CT image. .

[Screen display]
For the display of the simulation image, only the simulation image of the target image SD value may be displayed on the screen, but the display as shown in FIG. 8 may be performed. KVpt and mAst in FIG. 8 are tube voltage and mAs value for obtaining a simulation image of the target image SD value. In the example of FIG. 8, the tube image and mAs value are centered on the simulation image 80 of the target image SD value. A simulation image 81 in the case where the above is appropriately changed is displayed side by side.

  The target image SD value is a parameter arbitrarily designated by the operator, but the set target image SD value is not necessarily optimal for reaching the inspection purpose. Therefore, in a user interface that displays only the simulation image of the target image SD value (hereinafter referred to as [UI]), the simulation image can be displayed while changing the target image SD value many times in the process of examining the optimum shooting conditions. A rework process is required.

  However, as shown in Fig. 8, not only the simulation image of the target image SD value but also the simulation image when the X-ray conditions are appropriately changed are displayed side by side, so that not only the process of recreating the simulation image can be greatly reduced, but also each SD These images can be easily compared, and it is easy to set optimal shooting conditions.

  In the example of FIG. 8, the case where the mAs value and the tube voltage value are changed is shown, but the parameters to be changed are not limited to these two, and any arbitrary parameters such as the imaging slice thickness, the reconstruction slice thickness, the reconstruction function, the helical pitch, etc. It can be a parameter. In the example of FIG. 8, a total of nine images are displayed side by side in the vertical and horizontal directions, but the number and layout of the images are not limited to this and may be arbitrary.

  If an arbitrary image is selected on the screen, the X-ray condition at that time is registered as an imaging protocol, and the imaging protocol is automatically set as the imaging condition when imaging a patient that is a simulation target. Such a system would increase clinical utility.

<Second embodiment>
8 shows an example in which simulation images with several SDs are displayed side by side, but when one of the simulation images arranged in FIG. 8 is selected, the parameters of the two axes displayed in parallel are fixed, and the others A plurality of simulation images with different parameters may be displayed side by side.

  FIG. 9 is a screen display example showing this embodiment. In FIG. 9 (a), the vertical axis is the tube voltage kVp, the horizontal axis is the tube voltage mAs, the simulation image 80 generated in step S211 is the center, and the simulation image 81 when the tube voltage and mAs value are appropriately changed is surrounded. Are displayed side by side. Here, when the user clicks and selects the simulation image 80 with the mouse, the tube voltage, tube current kVpt, and mAst of the simulation image 80 are fixed, and other parameters such as the helical pitch (HP) and the reconstruction function (F1, A simulation image 82 in which F2 and F3) are changed is generated.

  Then, the simulation image 80 selected in FIG. 9 (a) is displayed at the center, and the simulation image 82 having the same tube current and tube voltage, different HP, and using the reconstruction function is displayed side by side ( Fig. 9 (b)).

  Thereby, it is possible to examine optimum X-ray imaging conditions while first setting two parameters and then viewing a simulation image when further parameters are changed.

  In the above description, the combination of the tube current and the tube voltage is first determined, and then the helical pitch and the reconstruction function are determined. However, the combination of parameters and the order of determination are not limited to this.

<Third embodiment>
In the third embodiment, a simulation image is confirmed, and the X-ray imaging conditions at that time are used as imaging conditions for scan imaging.

  For example, a “shooting condition determination” icon 90 is provided in FIGS. 8 and 9B. In FIG. 8 and FIG. 9 (b), one simulation image is selected by clicking with the mouse, and then the shooting condition determination icon 90 is clicked with the mouse, and then the shooting condition (FOV, shooting slice) of the selected simulation image is selected. Information indicating (thickness, reconstruction function, HP, image slice thickness, tube voltage, tube current) is transmitted to the control unit 9 of the X-ray CT apparatus 1. The X-ray control means 10 of the control unit 9 performs scan imaging according to the received imaging conditions.

  When the tube voltage and tube current are determined as shown in FIG. 8 and all other imaging conditions such as FOV and imaging slice thickness are not transmitted from the simulator, X-ray CT The control unit 9 of the apparatus 1 automatically inputs the received shooting conditions and causes the user to input the missing shooting conditions.

  According to the present embodiment, the imaging conditions determined while viewing the simulation image can be transmitted to the X-ray CT apparatus, and the imaging conditions can be automatically input to perform scanning imaging.

<Fourth embodiment>
SD may be displayed side by side with the same simulation images with the same noise pattern. Due to the randomness of noise, the appearance of noise, that is, the noise pattern is different even when the same subject is imaged under exactly the same X-ray conditions. Therefore, the disease may be difficult to see even if the image is taken under the same X-ray conditions because of the balance between the size and position of the disease and the noise pattern. Therefore, particularly in close examinations, it is necessary to set imaging conditions that can reliably detect a disease regardless of the noise pattern.

  In such cases, a UI that displays simulation images side by side with the same SD but different noise patterns would be clinically useful. FIG. 10 is a screen display example in which nine simulation images 85 with the same SD value and different noise patterns are displayed in parallel. The number of dots indicating noise (one large dot and eight small dots) is the same, but the positions where dots are scattered are different in each simulation image. By displaying such a simulation image, it is possible to determine the shooting condition in consideration of the noise pattern.

  In recent years, a technique has been introduced that reduces the exposure dose by improving the SN of an image using an image filter that can remove noise while suppressing a decrease in resolution as shown in Non-Patent Document 4. .

Video Information Medical Separate Volume Multi-Slice CT Book2005 vol37 (7) 145-149 Therefore, when examining exposure reduction imaging protocols using image filters, an image filter as shown in Non-Patent Document 4 is applied to simulation images. A UI that can perform noise removal processing with any arbitrary image filter will be useful clinically.

  For example, in the same way as in the case of FIG. 8 described above, when arbitrary parameters are selected on the vertical axis and the horizontal axis and the simulation images after filtering processing when the X-ray conditions are appropriately changed are displayed side by side, When the processing parameters (smoothing level, sharpening level, etc.) are changed, it is desirable to switch the display of the simulation image after the filtering process accordingly.

  Further, as shown in FIG. 11, a UI that displays the simulation image 111 after the filtering process when the smoothing level and the sharpening level are changed with respect to the same simulation 110 image may be clinically useful.

<Fifth embodiment>
[Configuration of X-ray CT scan simulator]
FIG. 12 is a configuration diagram of the X-ray CT scan simulator 1200 of the present embodiment. The X-ray CT scan simulator 1200 includes image storage means 1210, target noise value setting means 1230, noise image generation means 1240, simulation image creation means 1244, and sample noise image editing means 1250.

  The image storage unit 1210 stores a reference image serving as a reference for a simulation image created by the X-ray CT scan simulator 1200 and a plurality of sample noise images. The reference image is a past image obtained by CT imaging of the subject in the past or a human phantom image obtained by CT imaging of a human phantom that faithfully reproduces the internal tissue of the human body. Sufficient X-ray dose should be used to capture human phantom images so that clear images can be obtained. In addition, the human phantom image may be prepared by taking a representative FOV for each part. The sample noise image is an image including only noise components. It is desirable to prepare the sample noise image according to the part, the FOV, the ratio of the FOV to the subject size, and the reconstruction function used at the time of image reconstruction. The method for creating the sample noise image has already been described.

  The target noise value setting means 1230 sets the noise target value of the simulation image created by the X-ray CT scan simulator 1200. The target noise value setting unit 1230 includes a body thickness acquisition unit 1231, an approximate model generation unit 1232, and a target noise value calculation unit 1233. The target noise value setting unit 1230 may be an image noise amount input unit that inputs an image noise amount. Note that the image noise amount is usually defined as a standard deviation that is a variation in CT values of a homogeneous phantom image, and may be abbreviated as an image SD (Standard Deviation).

  The body thickness acquisition unit 1231 acquires the body thickness of the subject. The body thickness may be manually input by the operator, may be obtained from a scanogram, or may be automatically estimated by the method already described when the reference image is a past image.

  The approximate model generation unit 1232 generates an approximate model based on the body thickness acquired by the body thickness acquisition unit 1231. The approximate model may be a model approximated with water so as to have the same amount of X-ray absorption as the subject (hereinafter referred to as water approximation model), and the amount of X-ray absorption equivalent to the subject obtained from the average CT value of the subject It may be a model using a virtual substance having the following (hereinafter referred to as an average CT value substance approximate model)

  The target noise value calculation unit 1233 calculates the amount of image noise based on the input shooting conditions and the approximate model generated by the approximate model generation unit 1232.

  The noise image generation means 1240 generates and outputs a noise image based on the target noise value set by the target noise value setting means 1230. The noise image generating unit 1240 includes a sample noise image selecting unit 1220, a reference image noise amount calculating unit 1241, an added noise amount calculating unit 1242, and a noise amplitude changing unit 1243.

  The sample noise image selection means 1220 selects one sample noise image from a plurality of sample noise images according to the shooting condition of the reference image. To select a sample noise image, reference is made to FOV, the FOV to subject size ratio, the reconstruction function, etc. from among the shooting conditions.

  The reference image noise amount calculation unit 1241 calculates the noise amount of the reference image. The method for calculating the noise amount of the reference image has already been described.

  Based on the target noise value set by the target noise value setting unit 1230 and the noise amount calculated by the reference image noise amount calculation unit 1241, the addition noise amount calculation unit 1242 calculates the noise amount to be added to the reference image. .

  The noise amplitude changing unit 1243 changes the noise amplitude of the sample noise image selected by the sample noise image selecting unit 1220 based on the added noise amount calculated by the added noise amount calculating unit 1242.

  The simulation image creating unit 1244 creates a simulation image by synthesizing the noise image generated by the noise image generating unit 1240 and the reference image.

  The sample noise image editing unit 1250 is a unit for performing editing such as addition / deletion of a sample noise image stored in the image storage unit 1210. Using the sample noise image editing means 1250, the operator designates at least one of the shape of the virtual phantom, the CT value, the barycentric coordinates, and the field of view for imaging, and adds the sample noise image.

[Process flow]
FIG. 13 is a diagram showing a processing flow until a simulation image is generated according to the present embodiment. A rough flow is setting of a reference image in steps S1301 to S1303, setting of a target noise value in steps S1304 to S1306, generation of a noise image in steps S1307 to S1311, and synthesis of a simulation image in step S1312. Each step will be described below.

(Step S1301)
The X-ray CT scan simulator 1200 searches the image storage means 1210 for past images of the subject to be simulated, and determines whether there is a history of imaging the same part in the past. If it is determined that there is a past image, the process proceeds to step S1302, and if not, the process proceeds to step S1303.

(Step S1302)
The X-ray CT scan simulator 1200 sets a past image as a reference image.
(Step S1303)
The X-ray CT scan simulator 1200 sets a human phantom image as a reference image.

(Step S1304)
The body thickness acquisition unit 1231 automatically estimates the body thickness of the subject from a scanogram or a past image. The estimation method has already been described. Although the operator may directly input the body thickness, it is better to automatically estimate the body thickness in order to reduce the burden on the operator.

(Step S1305)
The approximate model generation means 1232 generates an approximate model based on the body thickness obtained in S1306. The approximate model may be a water approximate model or an average CT value substance approximate model. The method for calculating the approximate model has already been described.

(Step S1306)
The target noise value setting unit 1230 sets a target noise value. The target noise value may be set manually by an operator, or may be set by the target noise value calculation means 1233 based on the approximate model generated in S1307 and the shooting conditions. good.

(Step S1307)
The X-ray CT scan simulator 1200 acquires the FOV, the FOV and subject size ratio, and the reconstruction function from the imaging conditions of the reference image set in S1302 or S1303.

(Step S1308)
The sample noise image selection unit 1220 selects one sample noise image from the plurality of sample noise images in the image storage unit 1210 based on the imaging condition acquired in S1305.

(Step S1309)
The reference image noise amount calculation unit 1241 calculates the image noise amount of the reference image. The method for calculating the amount of image noise has already been described.

(Step S1310)
The added noise amount calculation means 1242 calculates the amount of noise to be added to the reference image based on the target noise value set in S1308 and the image noise amount of the reference image calculated in S1309. The method for calculating the amount of added noise has already been described.

(Step S1311)
The noise amplitude changing unit 1243 changes the noise amplitude of the sample noise image selected in S1305 according to the added noise amount calculated in S1310.

(Step S1312)
The simulation image creating means 1244 creates a simulation image by synthesizing the reference image set in S1302 or S1303 and the sample noise image whose noise amplitude has been changed in S1311.

  It is desirable to prepare the sample noise image according to the part, the FOV, the ratio of the FOV to the subject size, and the reconstruction function. The human body has a vertically long organ such as the head and a horizontally long organ such as the abdomen. Here, the X-axis diameter of the subject is Rx, and the Y-axis diameter is Ry. Sample noise images are created for each part. Rx / Ry is about 0.787 for vertical organs such as the head, and Ry / Rx is about 0.787 for horizontal organs such as the abdomen. It is desirable that a simulated organ is arranged in the above-described virtual phantom so that the sample noise image is generated, and it is desirable to create a sample noise image using such a virtual phantom.

  For FOV, it is desirable to create sample noise images for all possible FOVs in all imaging protocols. FIG. 14 is an explanatory diagram of the ratio between the FOV and the subject size. For example, assume that a subject 902 is included in an image 901 as shown in FIG. Here, assuming that the X-axis diameter of the subject is Rx and the Y-axis diameter is Ry, the ratio of FOV and subject size described in the present invention is the ratio of Rx / FOV or Ry / FOV, and is 0 to 1. It is a real number. Depending on the part to be photographed, the subject size may be almost the same as FOV, or it may be only about 30% of FOV. Therefore, it is desirable to create sample noise images in steps of about 0.05 when the ratio of FOV to subject size is in the range of 0.3 to 1.00. Even for data captured under the same X-ray conditions, the noise pattern varies depending on the reconstruction function. Therefore, it is desirable to create sample noise images for all reconstruction functions used in the clinic.

  The above-mentioned sample noise image is not only prepared in advance, but a function that can be arbitrarily created and added by an operator is clinically useful. A GUI that allows the operator to arbitrarily specify the shape, CT value, barycentric coordinates, noise amount (SD), and FOV of the above-mentioned virtual phantom so that the operator can add sample noise images as shown in Fig. 15. It is desirable to be provided. In addition, it is desirable to provide a GUI that displays a list of stored sample noise images on the screen and allows the operator to arbitrarily delete noise images that the operator has determined to be unnecessary.

  Through the above process, when displaying the created simulation image, the reference image set in S1302 or S1303, the sample noise image selected in S1305, and the sample noise image after amplitude change obtained in S1311 are displayed simultaneously with the simulation image. It is desirable to display all of them in a row or for any combination desired by the operator. Moreover, you may perform a display as described in 1st embodiment-4th embodiment.

  FIG. 16 shows an application example of the present invention. FIG. 16 (a) is a head CT image 120 of SD 4.0 taken at 100kV, 200mAs. FIG. 16B is a noise image 121 that is added to create a simulation image with Sd of 5.6 from the original image 120 in FIG. FIG. 16C is a simulation image 122 obtained by adding the noise image 121 of FIG. 16B to the original image 120 of FIG. 16A, and SD is 5.6. Therefore, this corresponds to a case where the mAs value is reduced to about ½ compared to the photographing condition when the original image is obtained.

  Those skilled in the art will readily understand that the simulation image can be created in the same manner when the SD in the simulation image is changed or in other parts.

  In the above embodiment, an X-ray CT apparatus equipped with an X-ray CT scan simulator has been described. However, an X-ray CT scan simulator apparatus equipped with only a simulator function without an X-ray imaging function may be configured.

  Further, the functions described in the above embodiments may be configured as a scan simulator program that causes a computer to execute the functions, and the above functions may be realized by installing this program in a personal computer or a workstation. In that case, the simulator image does not necessarily have to be displayed on the monitor of the personal computer or workstation on which the scan simulator program is installed, and the simulator image is transmitted to a terminal device connected via a network such as a LAN. A simulator image may be displayed.

The block diagram of the X-ray CT apparatus carrying the X-ray CT scan simulator of this invention. The block diagram of 1st embodiment. The processing flow of a first embodiment. Explanatory drawing of a body thickness estimation method. The figure which shows the permeation | transmission length in a water phantom, and SD. Explanatory drawing of the production method of projection data. Explanatory drawing of the calculation method of transmission length. The screen display example of 1st embodiment. The screen display example of 2nd embodiment. The screen display example of 4th embodiment. The screen display example of 4th embodiment. The block diagram of 5th embodiment. Processing flow of the fifth embodiment. Explanatory drawing of ratio of FOV and subject size. GUI example for the operator to create and add sample noise images. The example which applied this invention clinically.

Explanation of symbols

  1 X-ray CT system, 2 gantry, 3 X-ray source, 4 collimator, 5 detector array, 6 detector element, 7 X-ray, 8 rotation center, 9 control unit, 10 X-ray control means, 11 gantry control means, 12 DAS, 13 calculation processing means, 14 body thickness estimation means, 15 approximate model calculation means, 16 image SD calculation means, 17 additive noise amount calculation means, 18 approximate model projection data creation means, 19 reconstruction means, 20 noise image creation Means, 21 simulation image creation means, 22 storage means, 23 temporary storage means, 24 input means, 25 image display means

Claims (18)

Image storage means for storing a reference image;
Target noise value setting means for setting a noise target value of a desired image;
Noise image generation means for generating a noise image based on the set target noise value;
A simulation image creating means for creating a simulation image by combining the generated noise image and the reference image;
Display means for displaying the simulation image;
Equipped with a,
The noise image generation means includes
Body thickness acquisition means for acquiring the body thickness of the subject;
An approximate model generating means for generating an approximate model of the subject based on the body thickness;
Reference image noise amount calculating means for calculating the noise amount of the reference image;
Added noise amount calculating means for calculating an added noise amount to be added to the reference image based on the target noise value and the noise amount of the reference image;
Approximate model projection data creating means for creating reference projection data that is projection data of the approximate model, and addition projection data obtained by adding the amount of added noise to the reference projection data;
Noise image reconstruction means for reconstructing a noise image using the addition projection data;
An X-ray CT scan simulator device characterized by comprising: In the X-ray CT scan simulator device of claim 1 ,
The noise image creation means creates a noise image by subtracting the reconstructed image based on the addition projection data and the reconstructed image based on the reference projection data, or creates a difference between the addition projection data and the reference projection data An X-ray CT scan simulator characterized by generating differential projection data by processing, and reconstructing the differential projection data to create a noise image. In the X-ray CT scan simulator device of claim 1 ,
The approximate model generation means calculates an approximate model composed of a virtual water phantom having a water equivalent thickness estimated from the body thickness of the subject, or for each portion obtained by dividing the imaging region of the subject, An X-ray CT scan simulator device that calculates an approximate model to which an X-ray absorption coefficient calculated from the difference between the CT value of the part and the CT value of air and the X-ray absorption coefficient of water is assigned. In the X-ray CT scan simulator device of claim 1 ,
The approximate model projection data creating means virtually generates an X-ray energy spectrum, and for each photon energy, a projection value at each photon energy based on a multiplication value of an X-ray absorption coefficient and a transmission path length of each X-ray absorbing material. And calculating the reference projection data and the addition projection data based on the product sum of the contribution ratio of each energy and the projection value, or reprojecting the past image of the subject and calculating the reference projection data and the addition. An X-ray CT scan simulator characterized by creating projection data. In the X-ray CT scan simulator device of claim 1 ,
The X-ray CT scan simulator apparatus characterized in that the reference image noise amount calculating means calculates an image noise amount included in the reference image based on an imaging condition when the reference image is captured. In the X-ray CT scan simulator device of claim 1,
The simulation image creating means further includes a simulation image group consisting of a plurality of simulation images in which at least one of the imaging conditions of the simulation image, an image filter, and the scattered pattern of the image noise are different. The X-ray CT scan simulator apparatus characterized in that the display means creates and displays the simulation image and the simulation image group side by side. In the X-ray CT scan simulator device of claim 6 ,
The image processing apparatus further includes selection means for selecting one of the simulation images displayed side by side on the display means, and the display means displays the imaging conditions and a plurality of simulation images in correspondence with each other, and the simulation is performed by the selection means. When one of the images is selected, the simulation image creating unit is configured to select a plurality of simulations in which the shooting conditions displayed on the display unit are the same among the shooting conditions of the selected simulation image, and the other shooting conditions are different. create a simulation image group including images again, the display means, you and displaying side by side and the simulation image is the selected simulation images created again the X-ray CT scan simulator device. In the X-ray CT scan simulator device of claim 6 ,
Further comprising designation means for designating one simulation image of the simulation image or the simulation image group displayed on the display means, and the designated image is designated when the simulation image displayed on the display means is designated. An X-ray CT scan simulator apparatus, wherein imaging conditions for obtaining a simulation image are output as imaging conditions at the time of scanning imaging of the subject. In the X-ray CT scan simulator device of claim 1,
The image storage means further stores a plurality of sample noise images, and the noise image generation means selects a sample noise image from the plurality of sample noise images according to a shooting condition; and
Reference image noise amount calculating means for calculating the noise amount of the reference image;
Added noise amount calculating means for calculating a noise amount to be added to the reference image based on the target noise value and the noise amount of the reference image;
Noise amplitude changing means for changing the noise amplitude of the selected sample noise image based on the amount of added noise;
An X-ray CT scan simulator device characterized by comprising: In the X-ray CT scan simulator device of claim 9 ,
The reference image stored in the image storage means is a past image obtained by imaging a subject in the past or a human phantom image obtained by imaging a human phantom. Scan simulator device. In the X-ray CT scan simulator device of claim 9 ,
The target noise value setting means includes a body thickness acquisition means for acquiring the body thickness of the subject,
An approximate model generating means for generating an approximate model of the subject based on the body thickness;
Target noise value calculating means for calculating the target noise value based on the photographing condition and the approximate model;
An X-ray CT scan simulator device comprising: In the X-ray CT scan simulator device of claim 11 ,
The body thickness acquisition means is a means for estimating the body thickness of the subject based on a scanogram image obtained by scanogram imaging, or a body thickness based on past images obtained by imaging the subject in the past. An X-ray CT scan simulator device characterized by being an estimation means. In the X-ray CT scan simulator device of claim 11 ,
The X-ray CT scan simulator device characterized in that the approximate model generation means is means for calculating a water equivalent thickness equivalent to the body thickness and calculating an approximate model having the water equivalent thickness. In the X-ray CT scan simulator device of claim 9 ,
The plurality of sample noise images stored in the image storage means are sample noise images having different parts, imaging field of view, ratio of imaging field of view to subject size, and reconstruction function used at the time of image reconstruction. Line CT scan simulator device. In the X-ray CT scan simulator device of claim 9 ,
An X-ray CT scan simulator device further comprising editing means for specifying at least one of a virtual phantom shape, CT value, barycentric coordinates, and field of view, and adding and deleting sample noise images. Equipped with an X-ray source that irradiates the subject with X-rays, an X-ray detector that is disposed opposite to the X-ray source and detects X-rays transmitted through the subject, and the X-ray source and the X-ray detector A rotating device that rotates around the subject, an image reconstruction device that reconstructs a tomographic image of the subject based on transmitted X-ray doses in a plurality of directions detected by the X-ray detector, In an X-ray CT apparatus comprising an imaging condition input device that inputs irradiation conditions and image reconstruction conditions, and an image display device that displays the tomographic image,
An X-ray CT apparatus comprising the X-ray CT scan simulator apparatus according to any one of claims 1 to 15 . Obtaining a reference image;
Setting a desired noise value for the desired image;
Generating a noise image based on the set target noise value;
Synthesizing the generated noise image and the reference image to generate and output a simulation image;
An X-ray CT scan simulator program Ru cause the computer to execute,
The noise image generation step includes
Obtaining the body thickness of the subject;
Generating an approximate model of the subject based on the body thickness;
Calculating a noise amount of the reference image;
Calculating an addition noise amount to be added to the reference image based on the target noise value and the noise amount of the reference image;
Creating reference projection data that is projection data of the approximate model, and addition projection data obtained by adding the amount of added noise to the reference projection data;
Reconstructing a noise image using the added projection data;
X-ray CT scan simulator program characterized by causing a computer to execute . In the X-ray CT scan simulator program of claim 17 ,
The noise image generation step includes
Selecting one sample noise image from a plurality of sample noise images according to shooting conditions;
Calculating a noise amount of the reference image;
Calculating a noise amount to be added to the reference image based on the target noise value and the noise amount of the reference image;
Changing the noise amplitude of the selected sample noise image based on the amount of added noise;
X-ray CT scan simulator program characterized by causing a computer to execute.

Images