JP6958790B2 - Mammography interpretation device and mammography interpretation method - Google Patents

Description

The present invention relates to a workstation equipped with an indicator and an input device for interpreting a mammographic image of a breast collected by X-ray mammography.

In recent years, the importance of breast cancer screening by X-ray mammography has increased. This is because the prevalence of breast cancer and the number of deaths are increasing not only in Japan but also in the world, and breast cancer is the leading cause of death among women aged 30 to 64 years.

Among these, in Europe and the United States, the morbidity rate of breast cancer is increasing, but the mortality rate is decreasing, which is contributed by the introduction of breast cancer screening by mammography and the early detection of breast cancer by the high screening rate of 60 to 80%. It is believed that there is. On the other hand, in Japan, although the national government recommends regular screening, the breast cancer screening rate is low and the mortality rate is increasing year by year.

Compared to Westerners, Japanese have a higher proportion of high-concentration mammary glands, and with the spread of breast cancer screening by mammography in recent years, the problem of high-concentration mammary glands, that is, lesions are hidden by high-concentration mammary glands, and mass lesions are discovered. The problem of difficulty is highlighted. In other words, in mammography, the risk of lesions being hidden in normal mammary glands differs depending on the composition of the mammary glands (fatty, scattered mammary glands, heterogeneous high concentration, extremely high concentration), especially tumor lesions compared to calcified lesions. Will be higher.

In recent years, an imaging device compatible with tomosynthesis has appeared, and it has become possible to obtain tomographic images with less overlap of mammary glands, and it is possible to determine whether or not the mammary glands are normal with higher accuracy. However, a device compatible with tomosynthesis is expensive, and the burden on the patient increases due to an increase in imaging time and exposure dose, and the number of tomographic images generated is large, which imposes a heavy burden on the image interpreting doctor.

In such a situation, the medical image display device and the mammography device described in Patent Document 1 have been proposed for the purpose of improving the efficiency of the image interpretation work of the mammography image. The medical image display device of Patent Document 1 is integrally equipped with a mammography device. As a result, the mammography apparatus displays graph information indicating the correspondence between the tomographic position and the mammary gland amount for each tomographic image, and superimposes and displays the graph marker GM representing the tomographic position of the displayed tomographic image on the graph information G1. Therefore, an image reference person such as an image interpreter can visually recognize the position where the graph marker GM is displayed and perform image interpretation while easily grasping the mammary gland amount of the displayed tomographic image. This is intended to improve the interpretation efficiency.

JP-A-2017-047080

However, in the medical image display device described in Patent Document 1 described above, although the graph marker GM representing the tomographic position of the displayed tomographic image is superimposed and displayed on the graph information G1, the image interpreter displays a mass or tumor in the shadow of the mammary gland. From the viewpoint of whether or not the tumors such as these are mixed in or hidden, it does not contribute to the efficiency and facilitation of interpretation by visually observing the mammography images one by one.

Of course, "improving and facilitating image interpretation" here means not only shortening the image interpretation work, but also reliably interpreting even smaller lesions and improving diagnostic accuracy. .. For this reason, it is desirable to also have a marking mechanism so as not to overlook "lesions of at least larger size". Therefore, it is desired to have a mechanism capable of ensuring higher image interpretation accuracy, increase the throughput of the image interpretation work, and reduce the labor burden of the image interpretation work of the image interpretation doctor.

Therefore, in view of the situation of the conventional mammography image interpretation work described above, the present invention can secure higher image interpretation accuracy, increase the throughput of the image interpretation work for the image interpretation doctor, and interpret the image. An object of the present invention is to provide an image interpretation device and an image interpretation method for mammography that can reduce the labor burden of work.

In order to achieve the above object, the mammography image interpretation device according to the present invention is
With a monitor
An image display means for displaying a two-dimensional X-ray image of the human breast taken by an X-ray mammography apparatus on the screen of the monitor, and an image display means.
On the screen of the monitor on which the X-ray image is displayed by the image display means, a simulated area having a specific shape and size and simulating an assumed lesion portion is movable on the screen. a simulated area setting display device for setting display,
Pixel specifying means for specifying a plurality of pixels forming the simulated area on the screen, and
A combination of a plurality of original luminances of at least a part of the plurality of pixels specified by the pixel specifying means and a luminance obtained by subjecting the at least a part of the plurality of pixels to a process for characterizing the spread of the simulated area. Luminance synthesis means and
The simulated area, and table Shimesuru superimposed display means to the screen of the monitor so as to be displayed superimposed on the X-ray image with a synthesized luminance by the luminance synthesizing unit,
It is characterized by being equipped with.

In this way, the simulated area that imitates the lesion can be superimposed and displayed on the X-ray image of the breast with the brightness that characterizes the extent of the simulated area. For this reason, the user (such as an image interpreter) moves the simulated area superposed in this way to a site of interest, so that the simulated area expands at that site, that is, a size that does not want to overlook a lesion such as a tumor. It is possible to compare and observe (spread) with the spread of the site of interest in the mammary gland tissue. Specifically, the lesion is not hidden in a part of the mammary gland tissue of interest, or a state equivalent to that of a ruler can be easily and quickly obtained on the image. Therefore, the interpretation of the mammography image can be performed more accurately and quickly, which can contribute to the improvement of the interpretation throughput and also contribute to the reduction of the labor of the interpretation work of the interpreter.

As a preferred example, the simulated area is set to a desired size preset so that the operator prevents the lesion from being overlooked on the X-ray image.

Further, as another preferable example, the simulated area setting means includes an area changing means capable of changing at least one of the size and shape of the simulated area displayed on the screen of the monitor.

Further, in the mammography interpretation method according to the present invention, a two-dimensional X-ray image of a human breast taken by an X-ray mammography apparatus is displayed on a monitor screen, and the screen of the monitor on which the X-ray image is displayed is displayed. in the simulated area showing simulated the size of and assumed lesion has a specific shape (and spread), the set Mr. so as to be movable on the screen in order to explore the X-ray image display Te, and the identifying a plurality of pixels forming the simulated region, a plurality of original brightness of at least some of said plurality of pixels said identified, to the at least a portion of the plurality of pixels, the simulated area synthesizing the luminance obtained by applying the process of characterizing the spread, the simulated area, displayed on the screen of the monitor so as to be displayed superimposed on the X-ray image with the synthesized intensity,
It is characterized by that.

This image interpretation method can also exert an effect equivalent to the effect that the above-mentioned image interpretation device can provide.
It should be noted that the above-described mammography interpretation apparatus and mammography interpretation method-related features can be functionally provided as a recording medium on which a mammography interpretation program can be recorded in advance and the program thereof.

In the attached drawing
FIG. 1 is a configuration diagram including a partial functional block of the mammography image interpretation device and the workstation on which the mammography image interpretation method is carried out according to the first embodiment. FIG. 2 is a block diagram illustrating an outline of an arithmetic unit mounted on the workstation according to the embodiment. FIG. 3 is a diagram for explaining the position of the simulated region placed on the X-ray mammography image. FIG. 4 is a screen view for explaining a size comparison between the simulated area and a part of the mammary gland tissue (whether or not the simulated area is hidden by the mammary gland tissue). FIG. 5 is a screen view showing an example of a simulated region generated based on the luminance synthesis by the first synthesis pattern. FIG. 6 is a screen view showing an example of a simulated region generated based on the luminance synthesis by the second composite pattern. FIG. 7 is a screen view showing an example of a simulated region generated based on the luminance synthesis by the third composite pattern. FIG. 8 is a graph for explaining the proper use of synthetic formulas with a threshold value as a boundary according to the third synthetic pattern. FIG. 9 is a screen view illustrating the relationship between the pseudo-lesion of luminance synthesis, the mammary gland tissue, and the display, which is faced when using the fourth synthetic pattern. FIG. 10 is a schematic flowchart showing an example of an image interpretation process accompanied by a simulated area, which is executed by an arithmetic unit (CPU). FIG. 11 is a configuration diagram including a partial functional block of the workstation on which the mammography interpretation device according to the second embodiment is implemented. FIG. 12 is an explanatory diagram illustrating interlocking display of simulated regions of two mammography images, which are paired images, according to the second embodiment. FIG. 13 is a diagram illustrating various examples of the shape of a specific region (simulated region) simulating a lesion portion.

Hereinafter, the mammography image interpretation device and the mammography image interpretation method according to the embodiment of the present invention will be described with reference to the accompanying drawings.

[First Embodiment]
The mammography image interpretation device and the mammography image interpretation method according to the first embodiment will be described with reference to FIGS. 1 to 10 and 13.

As shown in FIG. 1, the mammography image interpretation device and the mammography image interpretation method are performed by an image interpretation workstation (hereinafter, simply referred to as a workstation) 3 that receives a mammography image taken by the mammography device 1 via an image server 2. It has been implemented.

The workstation 3 includes an arithmetic unit 11 that calculates various default programs by a computer, and an input unit 12 and a display 13 that are communicably connected to the arithmetic unit 11 and function as a man-machine interface.

The mammography apparatus 1 has a conventionally well-known configuration, and collects an X-ray transmission image obtained by performing X-ray irradiation and X-ray detection in the compression direction in a state where the breast of the human body is compressed by a compression plate. It is configured to do. This X-ray transmission image is a two-dimensional image in which the degree of X-ray attenuation exhibited by various tissues in the breast is expressed by, for example, gray level gradation. In this X-ray transmission image, not only a simple two-dimensional image but also a plurality of frame data collected by tomosynthesis imaging (captured multiple times at different X-ray incident angles while the breast is compressed) are shifted and added. It also includes reconstructed images that have been reconstructed using techniques such as shift and add and filtered back projection, and focused on a region of constant fault thickness at that location.

In general, this mammography is often performed in two directions, MLO (Mediao-Lateral Skull) and CC (Cranio-Caudal), in order to minimize the blind area of the breast. In the case of two-way imaging, since the direction in which the mammary gland tissue is compressed and expanded is different, the possibility that the lesion site is hidden by the mammary gland tissue during image interpretation and is overlooked can be reduced to some extent. In addition, at the time of image interpretation, an abnormality can be detected by comparing the left and right mammary gland structures.

The mammography image taken by the mammography apparatus 1 is sent to the image server 2 together with the examination information (patient name, patient ID, shooting date and time, shooting conditions, etc.) via the communication network CM1 (Internet, etc.) as an example, and the image is taken. It is stored in the server 2. The image server 2 is configured as, for example, a PACS (picture archiving and communication system). The image server 2 also receives examination information and mammography images from one or a plurality of medical facilities. Transmission. The DICOM (Digital Imaging and Communications in Medicine) standard is adopted as an example for this transmission and storage.

Further, the workstation 3 is communicably connected to the image server 2 via the communication network CM2 (for example, the Internet). Therefore, the workstation 3 can download the mammography image matching the search condition from the image server 2 together with the inspection information by transmitting the search condition to the image server 2.

Further, in the workstation 3, the arithmetic unit 11 specifically has a CPU (central processing unit) 21 and a non-volatile CPU (central processing unit) 21, which form the central configuration of the computer, as shown in FIG. 2 (ROM (read-only). A first storage device 22 composed of a memory) or the like, a second storage device 23 composed of a volatile RAM (random access memory) or the like, and examination information and image data such as patient information and examination information. A dedicated image storage device 24 for storing and is connected to each other via a bus 25 so as to be able to communicate with each other. An input / output interface (I / O) 26 is connected to the bus 25, and the input / output interface (I / O) 26 is connected to the bus 25. As described above, the 26 is also communicably connected to the input device 12 operated by an operator (such as an image interpreting doctor), the display 13 for image interpretation observation, and the image server 2.

The CPU 21 can call an application program stored in advance in the first storage device 22 into its own work area, execute the instructions sequentially according to the steps of the program, and if necessary, execute the instructions. It is configured to be able to communicate with the outside via the input / output interface 26. The second storage device 23 is used for temporarily writing and reading data accompanying the operation of the CPU 21.

Here, the first storage device 22 is provided as a non-transitory computer-readable recoding medium. Therefore, by executing the image interpretation program stored in advance in the first storage device 22 by the arithmetic unit 11 (specifically, the CPU 21), the area setting unit 11A, the pixel identification unit 11B, and the brightness are functionally executed. The synthesis unit 11C, the display control unit 11D, and the movement control unit 11E are obtained. These functional parts will be clarified by the explanation of the flowchart of FIG. 10 described later.

The input device 12 is composed of devices such as a mouse and a keyboard operated by an operator (such as an image interpreter, hereinafter simply referred to as an operator). As an example, the display 13 is usually composed of a color monitor and a high-definition monitor.

Therefore, the inspection information stored in the image server 2 can be accessed at the request of the operator, and a list of inspection information can be displayed on the color monitor. Further, the operator selects the desired inspection information on the displayed inspection information list, obtains the mammography image data of the inspection from the image server 2, and can display it on the high-definition monitor of the display 13. Has been done.

In this workstation 3, as schematically shown in FIG. 3, a two-dimensional mammography image IM _m collected by the mammography apparatus 1 is displayed on a display 13 (particularly, a high-definition monitor: hereinafter referred to as a display). The first feature is that the mammography image IM _{m is displayed by superimposing the region R sp} of a specific shape on the position according to the operator's request.

Further, for _{the brightness (that is, pixel value) of each of the plurality of pixels Pn forming the region of the specific shape R sp} , the original brightness (original brightness: the brightness of the collected mammography image as it is) and the spread of the original brightness are displayed. The brightness obtained from the composite pattern that defines the characterizing process (or is determined in advance by default) is combined, and the region R _{sp of the} specific shape is superimposed and displayed on the _{mammography image IM m according to the combined brightness.} This is a second feature.

By this display, the specially processed region R _sp of a specific shape is superimposed and displayed on the mammography image IM _m. The region R _sp having a specific shape can be freely moved on the screen of the display 13 in response to an instruction via the operator's input device 12. Since the mammography image IM _m is fixedly displayed on the screen, _{only the region R sp} of a specific shape is moved according to the movement instruction, the composite luminance is calculated based on the pixel of the movement destination, and the display 13 is displayed. It can be updated and displayed according to the display synchronization.

<Characteristics of this embodiment>
Here, features useful for the image interpretation work performed on the workstation 3 according to the present embodiment will be described.

<About the first feature>
The shape of the region R _sp of this specific shape is set as, for example, a circle, but it does not have to be a circle, and various other shapes can be adopted. For example, this region R _sp may be a polygon such as a pentagon or a hexagon, or may be an ellipse. A modified example of this shape will be described later.

The operator can _{apply this region R sp} to a suspected lesion on the mammography image to determine if the lesion can be hidden at that site. On the mammography image, for example, when there is a part that looks like a tumor visually because the display brightness of a part of the mammary gland structure is high, a region R _{sp of a} specific shape is applied to that part. This is to confirm whether the local region R _sp is large enough to hide the lesion. That is, since _{applying this region R sp} is a judgment by a ruler for confirming the size of the suspicious portion, the region R _sp having a specific shape is hereinafter referred to as a "simulated region R _{sp" from the functional aspect.} I will decide. It should be noted that this "simulated area R _sp " can be said to be an area of interest (ROI) from the perspective of an image interpreter.

As an example of the use of this "simulated area R _sp ", when the "simulated area R _{sp " is applied on the screen of the mammography image IM m} , the operator's input device 12 is used from the initial position or the part just determined. According to the instruction, the "simulated area R _sp " is moved to the next (suspicious) part to be seen.

Since this simulated region R _sp primarily has a ruler function as described above, the size of the _{simulated region R sp is extremely important.} It is desirable that this size be set to the minimum value of "don't miss a larger one" of the lesion (eg, mass) to be found in the interpretation. Furthermore, the size of the simulated region R _sp is also the minimum size of the lesion required to be detected.

Therefore, as an example, the simulated region R _sp is set as a circular region having a diameter of 1 cm. In the case of this example, it is possible to visually determine whether or not a tumor lesion having a diameter of 1 cm can be hidden in the mammary gland structure, and it is possible to support interpretation.

Representing this schematically, in the case of FIG. 4A, since a part of the mammary gland structure NS _{is smaller than the simulated area R sp,} it can be determined that a tumor lesion having a diameter of 1 cm or more cannot be hidden in this part. .. On the contrary, as shown in FIG. 3B, another part of the mammary gland structure NS to _{which the simulated area R sp is applied is larger than the area R sp} , so that the tumor lesion is brightly hidden in that part. It can be determined that it exists or may exist. _{For this reason, the operator moves the simulated area R sp} from the figure (A) to the figure (B), and whether or not there is a tumor lesion at each destination position, and where it is hidden by the mammary gland structure NS. You can read the image while paying attention to whether or not it is present.

<About the second feature>
_{When the "simulated area R sp} " is applied (set) to the local part of interest on the screen of the mammography image IM _{m and} the compositing process described later is instructed, the brightness of the local part is synthesized by the brightness compositing pattern. Processing is done. Thereby, in addition to the above-mentioned ruler function, the original luminance (original pixel value of the mammography image) of the pixels forming the _{simulated region R sp can be replaced with the composite luminance according to the characteristics. Therefore, the simulated region R sp} corresponding to the characteristics of the original luminance is displayed on the display 13 in a state of being superimposed on the mammography image IM _m. This display is also updated in synchronization with the screen display of the display device 13.

The compositing process is performed according to one or more compositing patterns (luminance compositing mode) that can be selected by the operator or determined by default settings. The synthesis process is based on the properties exhibited by the original brightness of the pixels constituting the simulated region R _sp, lesion corresponding to the simulated region R _sp is to visually demonstrate whether or not hidden mammary tissue It is a process. Therefore, since the lesion is hidden determination of the target corresponding to the simulated region R _sp, likened the simulated area R _sp full moon, mammary tissue can be assumed that a cloud reflected in the full moon. Therefore, if the cloud covers the full moon, the full moon behind the cloud cannot be seen. In order to visually clarify this hidden state, four types of composite patterns (brightness composite mode) are used in the present embodiment. ) Is available for selection.

<4 types of synthetic patterns>
-First composite pattern In the first composite pattern, the _{brightness is synthesized only at the marginal pixels of the simulated area R sp} , and the inside is output with the pixel value of the mammography image itself, that is, the original pixel value. This first composite pattern is effective when it is desired to accurately check the original pixel value inside the _{simulated region R sp.}

When synthesizing the output brightness of the peripheral pixels, the output may be output at a constant brightness, or the value obtained by XORing the original brightness and the intermediate brightness may be used as the output value so as not to be buried in the surrounding display brightness.

FIG. 5 shows an example _{of the simulated region R sp} based on this first synthetic pattern. As shown in the figure, _{only the line L sp} indicating the edge of the _{simulated region R sp} is drawn.

Even in the synthetic process using this first synthetic pattern, in the mammography image IM _{m , it is determined whether or not the simulated region R sp} simulated to correspond to the tumor site is hidden in the mammary gland region. Effective for interpretation.

-Second composite pattern The second composite pattern uses the _{original brightness as an alpha value for the pixels in the region of the simulated area R sp} , and outputs the original brightness and the specific value by alpha blending. As a result, the inside of the region can be expressed with a sense of sheerness corresponding to the brightness.

When this second synthetic pattern is used, the synthetic formula can be provided in the following two ways, 1 and 2.

Synthesis formula (1): out = r * org + (1-r) * a
Synthesis formula (2): our = r * org + (1-r) * a * δ _{i, j}
here,
Org: Pixel original brightness,
r = org / gradation width,
a: Constant (eg a = 128)
δ _{i, j} : A coefficient that simulates the gradation of a sphere, depending on the position (i, j) of the simulated region R _sp. As a result, based on the composite formula (1), for example, as shown in FIG. 6 (A), the simulated region R _{sp is superimposed and displayed on the mammography image IM m} , and based on the composite formula (2), for example, FIG. 6 ( As shown in B), the simulated region R _sp is superimposed and displayed on _{the mammography image IM m.}

-Third synthesis pattern _{For each pixel constituting the simulated region R sp} , the composite brightness is calculated by switching between two synthesis formulas depending on whether the original brightness is less than or equal to the threshold value. The threshold is set to, for example, 108 when the original luminance of the pixel is represented by 256 gradations. This threshold can be changed.

を使用する。ここで、org: 原輝度、a, b, c: 変更可能な合成パラメータである。これらのパラメータのうち、aは疑似病変の規定の輝度値を成し且つ基本的な明るさを制御するパラメータ（このパラメータには、上述したδ_i,j：模擬領域Ｒ_spの位置に依存した係数を掛けてもよい）、bはシグモイド関数のゲインで収束する位置を制御するパラメータ、ｃはシグモイド関数のオフセットで収束度合いを制御するパラメータである。例えば、a=32, b=5, c=0.７である。 As an example, when the original brightness of each pixel is less than the threshold value,
Synthesis formula (3): out = (1-r _u ) * org + r _u * 192 + 128
r _u = org / 255
If the original brightness of each pixel is greater than or equal to the threshold

To use. Where org: original brightness, a, b, c: modifiable composite parameters. Of these parameters, a is the parameter that forms the specified brightness value of the pseudo-lesion and controls the basic brightness (this parameter depends on the position of _{δ i, j} : simulated region R _{sp described above.} A coefficient may be multiplied), b is a parameter that controls the position of convergence by the gain of the sigmoid function, and c is a parameter that controls the degree of convergence by the offset of the sigmoid function. For example, a = 32, b = 5, c = 0.7.

The setting of the threshold value used in this third composite pattern has a great influence on the display result of the composite brightness. Since the main purpose of this function is to support the determination of whether a mass lesion can be hidden in the mammary gland structure, the threshold value for separating the mammary gland tissue and the adipose tissue is specified as the initial value. ing. Of course, this threshold can be changed by the operator in advance or during interpretation.

Therefore, in the present embodiment, the initial threshold value used for this third synthetic pattern is automatically determined by using the Otsu method, which is a binarization discrimination method based on the histogram in the breast region. It is configured to. The threshold determined by this method is generally set to a value that separates mammary gland tissue and adipose tissue. Of course, such an automatically set threshold value is configured to be changeable by an operator's operation (mouse wheel operation, etc.).

_{FIG. 7 shows a similar mammography image IM m} when the threshold value is reduced from the mammography image IM _{m (Fig. (A)) in which the simulated region R sp} based on the automatically set threshold value is superimposed and displayed (Fig. 7 (Fig. 7)). B)) and a similar mammography image IM _m (FIG. 3C) when the threshold value is increased are illustrated. By increasing the threshold value, the number of pixels calculated by the composite formula (3) increases, so that the display brightness of the original pixel having a brightness lower than the threshold value becomes higher and the visual effect is strengthened (Fig. (C)). ).

In addition, in FIGS. 8A and 8B, the output luminance (synthetic luminance) and the pseudo lesion (pseudo-lesion) with respect to the input luminance (original luminance: original luminance = 108 = threshold value) based on the synthetic formulas (3) and (4) are shown. An example of the characteristics of each of the brightness of the pixels in the area) is shown.

As a result, for example, as shown in FIG. 7, the simulated region R _{sp is superimposed and displayed on the mammography image IM m,} and the part of the full moon (the part simulating the mass part) that appears to appear from between the clouds has a predetermined threshold value. Since it has a brightness of less than that, it is displayed at a high brightness value so that it is clearly defined that it is below the threshold value. On the other hand, the rest of the full moon covered by the cloud (mammary gland tissue) is a pixel having a brightness equal to or higher than the threshold value. Therefore, based on the above synthetic formula (4), the sigmoid function is used, and the term of the pseudo-lesion in which the output value converges as the brightness increases is added to the original brightness. As a result, the synthetic brightness is calculated and displayed so that the _{lesion (simulated area R sp} ) exists on the back side of the mammary gland structure (cloud).

As a result, it is possible to create a sense of sheerness in which the full moon is hidden behind the clouds, and it is possible to exert a visual effect of converting the shading of the mammary gland structure into a deep appearance.

-Fourth synthetic pattern In the fourth synthetic pattern, _{the average line absorption coefficient of each pixel in the simulated area R sp} is estimated from the original brightness of each pixel, and a model (simulated area) simulating a simulated lesion is embedded. The combined line absorption coefficient is calculated, and their brightness is combined and output.

Specifically, a value obtained by combining the average line absorption coefficient estimated from the original brightness of each pixel forming the simulated region R _sp and the line absorption coefficient of the tumor is calculated, and the combined brightness is calculated using that value as a parameter.

9 (A), (B), and (C) show an example of a mammography image of a phantom in which a model of a pseudo-lesion is implanted and the fat / mammary gland ratio is changed. As shown in FIGS. (A), (B), and (C), from adipose tissue (100% fat) to average breast tissue (50% fat, 50% mammary gland) and dense breast tissue (100% fat). As an example, as the proportion of the mammary gland increases to 30% fat and 70% mammary gland), the synthetic component of the pseudo-lesion becomes smaller and is displayed as being more integrated into the mammary gland region.

Therefore, the line absorption coefficient synthesized by embedding the pseudo-lesion is inversely transformed and used as the output brightness to obtain the output of the pseudo-lesion. The inverse transformation is an exponential function approximated by a combination of a particular display luminance and the line absorption coefficient. Therefore, it is necessary to appropriately cope with the variation in image quality due to the difference in the photographing device and the difference in the image processing.

In order to deal with variations in image quality due to differences in imaging equipment and image processing, the composite component of the pseudo-lesion is calculated using the combined line absorption coefficient as a parameter and added to the original display brightness to synthesize the output brightness.

The synthesized line absorption coefficient is converted into a normalized parameter of 0 to 1.0 and applied to the calculation formula of the synthetic component of the pseudo-lesion to calculate the output brightness.

The effect of breast thickness on the synthetic components of pseudolesions is mitigated by adjusting the amount of exponential offset. Further, in order to simulate the spherical shadow of the pseudo lesion, the synthesized line absorption coefficient is calculated by a cylindrical shape having a uniform thickness, and the parameter of the spherical shape is applied to the display correction amount of the synthetic component.

In the workstation 3 according to the present embodiment, the fourth synthetic pattern is omitted, and only the first to third synthetic patterns described above are adopted, and the first to third synthetic patterns are used. It may be configured to be selectable or switchable as appropriate.

Next, an example of the image interpretation process executed on the workstation 3 will be described with reference to FIG. This interpretation processing is executed by the CPU 21 of the arithmetic unit 11.

In step S10, the CPU 21 interactively sends the patient information and the examination information to the image server 2 via the interface 26 together with the image transmission request with the operator. In response to this, the corresponding mammography image is transmitted from the image server 2, and this image data is stored in, for example, a dedicated image storage device 24.

Next, in step S11, the CPU 21 interactively determines the display mode of the received mammography image with the operator. As this display mode, one mammography image (independent display of either the left or right breast, or juxtaposed display of the left and right breasts (in the case of this juxtaposed display, a display mode in which the left and right pair images are juxtaposed in the same direction). Select a display mode that juxtaposes symmetrically).

Next, in step S12, the CPU 21 causes the display 13 to display the mammography image in the determined display mode. Now, it is assumed that the mammography image of either the left or right breast is independently displayed on the screen of the display 13.

After this display is completed, the CPU 21 determines the initial position of _{the simulated area R sp} described above by a command of an operator (doctor or the like) looking at the display screen of the display 13 or by an automatic setting based on an image recognition operation. (Step S13).

Further, the CPU 21 _{interactively determines the shape and / or size of the simulated region R sp} in response to a command from the operator (step S14). Therefore, this step S14 functionally serves as an area changing means.

For example, in _{order to avoid the situation where the simulated area R sp} is hidden by the mammary gland tissue and cannot be visually detected, the width and size of the mammary gland tissue on the screen and lesions such as a mass that should be read without being overlooked. It is configured to be selectable and set in consideration of the shape and / or size of the part. As an example, these shapes and / or sizes are stored in a table (for example, the first storage device 22) in advance, so that the operator may refer to and select them. Of course, the operator may specify an arbitrary size on the spot.
For example, as a shape, _{the ruler function to be provided in the simulated region R sp} is considered first, and as shown in FIG. 13, the simulated region R _sp-1 of the shape (round) is suitable (FIG. 13 (A)). reference). Considering the interpretation that additionally considers the shape of the lesion, the elliptical (oval) simulated area R _sp-2 (see Fig. 13 (B)) and the polygonal (polygonal) simulated area R _{sp- 3} (see Fig. 13 (C)), polygonal simulated area R _sp-4 (see Fig. 13 (D)), irregular simulated area R _sp-5 (see Fig. 13 (E)). One or more of the shapes (see) may be set to be selectable. These simulated areas are described in "Mammography Guidelines, 3rd Size Augmentation, Editing (Japanese Society of Radiological Technology) / Japanese Society of Radiological Technology, Igaku-Shoin", "Chapter 6 Mammogram Finding Terms, page 40, Figure". It is a basic shape area based on "6-1 Shamer of mass shape". In reality, this basic shape collapses and deforms, and in some cases, spicula protrudes from the periphery.
In the present invention, even these variously shaped simulated regions R _sp (R _sp-1 , R _sp-2 , R _sp-3 , R _sp-4 , R _sp-5 ) are, for example, the longest. It suffices to keep the diameter as a known size (diameter R1 in the case of a circle) (see diameters R2, R3, R4, R5 in FIG. 13), whereby a ruler function can also be obtained.
Therefore, these simulated regions R _sp are stored in advance in the first storage device 22, for example, as data indicating the shape and / size for each shape and / or size (diameter). In this case, the size (diameter) may be interactively set by the reader at the time of reading.
Of course, only the circular simulated region R _sp-1 may be held in a settable manner, and only the size (diameter) thereof may be set to be continuously or stepwise changeable within a certain range (for example, 5 mm to 15 mm).

For example, as such a shape and size, a circular simulated region R _sp with a diameter of 1 cm is set as a physical (in real space) dimension within the actual breast. Of course, the shape and size may be other than this. For example, a circular or polygon with a diameter of 8 mm or 5 mm may be set in an attempt to interpret a lesion having a small size, that is, a smaller lesion. The smaller the simulated area R _sp is set in the real space, the smaller the size of the lesion simulated by it. It is decided by the operator in consideration of the balance.

Then, CPU 21 selects a synthetic pattern to be used for the synthesis of required when to superimpose the simulated region R _sp mammography image IM _m, the luminance of the pixels of the mammographic image IM _m (pixel value) (step S15 ). For example, since the above-mentioned first to third types of synthetic patterns are prepared as the synthetic patterns, the operator can specify any of the synthetic patterns by interactively selecting them.

The various preparation steps described above are not necessarily limited to the order illustrated, and may be performed in a different order, or a desired order and determination contents may be set in advance by default.

When the preparation is completed in this way, the CPU 21 _{moves while designating the simulated area R sp} and waits for a command as to whether or not to perform image interpretation (step S16). When the operator instructs the image reading start (step S16, YES), CPU 21 is first simulated region R _sp is placed in the initial position of the mammographic image IM _m, mammographic image IM constituting the simulated region R _sp at its initial position _{A plurality of pixels Pn of m} are determined (step S17). As shown in FIG. 3, the plurality of pixels Pn are specified by positions (i, j) on the vertical and horizontal x and y coordinates.

When this is completed, the CPU 21 designates the first target pixel among the plurality of pixel Pn by the coordinate value (step S18). Next, the CPU 21 performs a luminance synthesis process on the target pixel according to a processing method based on any of the selected first to third synthesis patterns, and temporarily stores the processing result in, for example, the second storage device 23. Store (step S19). The processing method is as described above according to each synthetic pattern.

After that, the CPU 21 determines whether or not the target pixel to be brightness-combined still remains (step S20), and when it determines that it still remains (step S20, YES), proceeds to the process in step S18. The process of steps S19 and S20 described above is repeated. This is repeated until all the target pixels constituting the simulated region R _{sp are processed.}

Therefore, when the determination of NO is made in step S20, it means that the processing of all the target pixels constituting the _{simulated area R sp has been completed.} Therefore, the CPU 21 updates the display frame data to create the display data (step S21). This display data is sent by the CPU 21 to the display 13 via the interface 26 and displayed on the screen (step S22).

As a result, the simulated area R _{sp is superimposed and displayed on the display 13 at the initial position of the mammography image IM m in place} of or as a part of the screen of the preparatory work up to that point (step S22), for example, FIG. See 5, 6 and 7).

Each brightness (each pixel) of the superimposed simulated region R _sp is the original brightness by the above-mentioned first to third composite patterns, and at least a part of the pixels are composited by the processing method of each composite pattern. It is modulated) and does not simply replace or overwrite the entire area with pixels of different brightness. That is, the simulated region R _sp has region information simulating a lesion such as a tumor while having at least a part of the luminance information of the original pixel.

The CPU 21 further determines whether or not the operator has moved the simulated region R _{sp on the mammography image IM m} (step S23). Further, when the determination result is NO, the process _{waits while determining the movement of the simulated area R sp} until there is an end instruction, and if there is an end instruction, the interpretation is ended (step S24). On the other hand, when a YES determination is made in step S23, the process is returned to step S17, and pixel determination, synthesis processing, and display at the updated simulated area position are performed (steps S17 to S22). As a result, for example, _{the state of the simulated region R sp of FIG. 4 (A) is moved to the state of the simulated region R sp} of the same (B), and the above-described processing is performed. This series of processing is _{repeated until the operator stops the movement of the simulated area R sp} and issues a command to end the interpretation.

It should be noted that a processing procedure to be executed by the CPU 21 should be established so that the display mode, the shape and size of the simulated area, the composite pattern to be used (that is, the luminance composite mode), etc. can be arbitrarily switched at an arbitrary timing during the interpretation. Of course you can.

Here, functionally, step S14 corresponds to the area setting unit 11A, step S17 corresponds to the pixel identification unit 11B, steps S18 to S20 correspond to the luminance synthesis unit 11C, and step S21 corresponds to the display control unit 11D. The step S23 corresponds to the movement control unit 11E, and the step S11 plays a part of the interlocking control unit 11F.

Functionally, step S22 corresponds to the image display means, step S15 constitutes the selection means, and step S23 constitutes the area moving means.

As described above, according to the workstation 3 according to the present embodiment, the operator (doctor or the like) can determine a desired synthetic pattern from the plurality of prepared synthetic patterns. Therefore, the brightness (pixel value) of at least a part of the pixels of _{the simulated region R sp} is modulated based on the processing method defined by the synthesis pattern, and the modulated brightness is combined with the original original brightness and displayed. A simulated region R _sp is created for this. The created simulated area R _sp is displayed superimposed on the _{mammography image IM m at} a desired position specified by the operator. This superimposed display is made every time the operator moves the simulated area R _sp , at a changed position accompanying this movement.

As can be seen from the simulation images illustrated in FIGS. 5, 6 and 7, the operator visually interprets the screen of _{the mammography image IM m displayed on the display 13.} According to this screen _{, the simulated region R sp} is displayed as a round full moon-shaped region at a desired position in _{the mammography image IM m} , and at least the edges thereof are drawn. According to the second and third composite patterns, _{not only the edge of the simulated region R sp} but also the region inside the simulated region R sp is processed and displayed with the luminance value. Furthermore, the luminance information of all or part of the mammary gland tissue formed by the pixels on the mammography image IM _m corresponding to the simulated region R _{sp is also left at least partially.} As a result, the simulated region R _sp that imitates a lesion such as a tumor and other, for example, mammary gland tissue are appropriately superimposed and displayed.

That is, since this simulated region R _sp imitates the shape and size of the lesion, if this simulated region R _sp is hidden and invisible in a part of the mammary gland tissue, a tumor or the like is formed in a part of the mammary gland tissue. The lesion may be hidden. That is, the operator _{can quickly determine whether or not the simulated region R sp} is a portion that hides the lesion by hitting the region of interest (suspected) while moving the simulated region R sp. That is, the operator _{can obtain a ruler function for searching the simulated region R sp} , that is, a portion that brightly hides or makes it difficult to see the lesion, taking into consideration its intuition.

Therefore, the operator can quickly determine the size and spread state _{of the simulated region R sp while moving the simulated region R sp to the region of interest.} Of course, when a site where the simulated area R _sp is hidden is found, it is necessary to confirm and diagnose whether or not there is a lesion in that site by another method. Compared with the conventional method that does not use the _{simulated area R sp} having such a ruler function, in the case of the present embodiment, it can be expected that the image interpretation work will be remarkably speeded up and the throughput will be increased by the auxiliary means by _{the simulated area R sp.} .. In addition, it is possible to reduce the operator's interpretation work.

Further, by moving the simulated region R _sp to the region of interest, it is easy to focus the operator's attention during visual observation on the region, so to speak, a secondary effect of pointing confirmation can be obtained.

[Second Embodiment]
Subsequently, the workstation according to the second embodiment will be described with reference to FIGS. 11 and 12, and FIG. 10 described above. In this second embodiment, the same reference numerals are used for the same or similar components as those in the first embodiment, and the description thereof will be omitted or simplified.

As shown in FIG. 11, the second workstation has an interlocking control unit 11F by the program processing of the image interpretation processing given by the CPU 21 of the arithmetic unit 11.

Explaining this interlocking control, as schematically shown in FIGS. 12A and 12B, the positions of the simulated regions that are pseudo-lesions are at the same position in the pair images in the same direction, and in the pair images of the left and right pairs. Initially display at symmetrical positions. After that, according to the operation of the operator, the update display is interlocked on the pair image while maintaining the initial positional relationship. This function is obtained in each of steps S13, S16 to S22 of FIG. 10 when those processes are performed in parallel with each pair of images. Therefore, steps S13 and S16 to S23 functionally correspond to the interlocking control unit shown in FIG.

_{Specifically, the two mammography images IM m} collected by photographing the left and right breasts with the X-ray mammography apparatus 1 are symmetrically displayed on the screen of the display 13 (see FIG. 12 (A)) or in the same direction (see FIG. 12 (A)). (See FIG. 12B) Displayed side by side. Further, in each of the two displayed mammography images IM _m , the simulated regions R _sp (R _{sp (L)} , R _{sp (R)} ) are initially symmetrical or have the same positional relationship in the same direction. Is set to.

Further, in the two mammography images IM _m , the two simulated regions whose brightness is combined move and are displayed in conjunction with each other while maintaining the same positional relationship in the symmetrical or the same direction. NS.

As a result, the simulated areas are symmetrically arranged and displayed on the left and right pair of mammography images. Therefore, in addition to the effects enjoyed in the first embodiment described above, when evaluating the local asymmetric shadow (FAD), the position of the simulated region is compared with the display brightness of the same region on the opposite side. And can be judged based on the size. As a result, it can be expected that the interpretation accuracy will be further improved.

In the above-described embodiment, an example in which the interpretation of a mammography image is performed is shown as the present invention, but the gist of the present invention is an X-ray image obtained by photographing other parts of the patient, such as the lung field and the stomach. It is also possible to develop into a device and a method for interpreting images.

1 Mammography device 3 Interpretation workstation 11 Arithmetic logic unit 11A Area setting unit 11B Pixel identification unit 11C Luminance synthesis unit 11D Display control unit 11E Movement control unit 11F Interlocking control unit 12 Input unit 13 Display R _sp (R _sp-1 ... R) _sp-5 ) Simulated area IM _m mammography image that simulates a lesion NS Breast tissue

Claims (12)

With a monitor
An image display means for displaying a two-dimensional X-ray image of the human breast taken by an X-ray mammography apparatus on the screen of the monitor, and an image display means.
On the screen of the monitor on which the X-ray image is displayed by the image display means, a minimum size set for detecting a lesion portion assumed in advance and a simulated area having a specific shape are displayed on the screen. A simulated area setting means for setting the brightness image of the breast so as to be movable, and
Pixel specifying means for specifying a plurality of pixels forming the simulated area on the screen, and
The plurality of original luminances of at least a part of the plurality of pixels identified by the pixel identifying means, and the shadow of the breast structure drawn inside the simulated area on the at least a part of the plurality of pixels. and luminance combining means for combining at least a shape, brightness subjected shading, or the determination can characterize certain processing conditions spread visually of the shadow on the basis of said original luminance of the simulated area,
A superimposed display means for displaying the simulated area having the brightness synthesized by the luminance combining means by superimposing the screen displayed on the monitor.
A mammography interpretation device characterized by being equipped with. The luminance synthesis means
Wherein the original brightness of a plurality of pixels of said at least part, the synthesis of the by the said at least part of the original brightness of a plurality of pixels and processed in the processing of the plant constant brightness synthesized for each pixel brightness Configured to do,
The mammography image interpretation device according to claim 1. Said predetermined process includes a plurality of kinds of Disposal Law prepared in advance,
The luminance combining means, operable to be composed of a plurality of the luminance synthesizing mode corresponding to the plurality of kinds of Disposal Law,
The interpretation device is
A selection means that allows the user to arbitrarily select the plurality of luminance synthesis modes, and
When an arbitrary luminance synthesis mode is selected by the selection means, the user can move the simulated region according to the selected luminance synthesis mode to an arbitrary position on the screen. Prepare,
The superimposed display means is configured to update and display the simulated area moved by the area moving means on the monitor together with the X-ray image.
The mammography image interpretation device according to claim 2. The plurality of treatment methods are
A processing method based on the first luminance synthesis pattern in which the original luminance of only a plurality of pixels at the edge of the simulated region is replaced with the luminance indicating the edge of the simulated region.
A processing method based on a second luminance synthesis pattern in which the original luminance of each pixel inside the simulated region is alpha-blended with the original luminance by adopting the original luminance as an alpha value, and
A processing method based on a third luminance synthesis pattern that switches the process for the synthesis based on the discrimination result of the original brightness of each pixel inside the simulated region based on a predetermined threshold value, and
Have at least
The mammography image interpretation device according to claim 3. The simulated area is set to the specific shape and the minimum size set by the operator to prevent oversight of lesions on the X-ray image.
The mammography image interpretation device according to any one of claims 1 to 4. The simulated area setting means includes an initial display setting means that initially displays the simulated area on the screen of the monitor .
A simulated area moving means for moving the simulated area on the screen,
To prepare
The mammography image interpretation device according to any one of claims 1 to 5. The simulated area setting means includes an area changing means for giving an instruction capable of changing at least one of the minimum size of the simulated area and the specific shape displayed on the screen of the monitor.
The mammography image interpretation device according to any one of claims 1 to 6, wherein the mammography image interpretation device is characterized. The image display means is configured to display two X-ray images collected by photographing the left and right breasts with the X-ray mammography apparatus symmetrically or juxtaposed in the same direction on the screen of the monitor. ,
The simulated area setting means sets the simulated area in each of the two X-ray images displayed on the screen of the monitor at an initial position having the same positional relationship in the symmetrical or the same direction. Is configured as
In the two X-ray images, the superimposed display means maintains the same positional relationship between the two simulated regions whose brightness is synthesized by the brightness combining means for each image, symmetrically or in the same direction. It was configured to be displayed in conjunction with each other.
The mammography image interpretation device according to any one of claims 1 to 7. A two-dimensional X-ray image of the human breast taken by an X-ray mammography device is displayed on the monitor screen.
On the screen of the monitor on which the X-ray image is displayed , a simulated area having a minimum size set for detecting a lesion portion assumed in advance and a specific shape is displayed on the screen of the breast. Set to be movable to the brightness image,
A plurality of pixels forming the simulated area are identified, and
The shadow of the breast structure and the shadow of the simulated area drawn inside the simulated area on the at least a part of the plurality of original luminances of the specified plurality of pixels and the plurality of pixels of the simulated area. Based on the original brightness, at least the shape, shading, or spreading state of the shadow is combined with the brightness that has been subjected to a process for visually determining the state.
The simulated area having the combined brightness is displayed by superimposing the screen displayed on the monitor.
A mammography interpretation method characterized by this. The simulated area is set to the specific shape and the minimum size set by the operator to prevent oversight of lesions on the X-ray image.
The mammography interpretation method according to claim 9. A mammography interpretation device equipped with a computer that can display a two-dimensional X-ray image of the human breast taken by an X-ray mammography device on the screen of a monitor and give input information from an input device to the screen display. A recording medium capable of pre-recording a mammography interpretation program that can be read and executed by the computer.
When the computer reads and executes the program recorded on the recording medium, the computer:
An image display means for displaying a two-dimensional X-ray image of a human breast taken by the X-ray mammography apparatus on the screen of the monitor, and an image display means.
On the screen of the monitor on which the X-ray image is displayed by the image display means, a minimum size set for detecting a lesion portion assumed in advance and a simulated area having a specific shape are displayed on the screen. A simulated area setting means for setting the brightness image of the breast so as to be movable, and
Pixel specifying means for specifying a plurality of pixels forming the simulated area on the screen, and
The plurality of original luminances of at least a part of the plurality of pixels identified by the pixel identifying means, and the shadow of the breast structure drawn inside the simulated area on the at least a part of the plurality of pixels. A luminance synthesizing means for synthesizing at least the shape, shading, or spreading state of the shadow based on the original luminance of the simulated region, and the luminance having been subjected to a process for visually determining the state of spreading.
A superimposed display means for displaying the simulated area having the brightness synthesized by the luminance combining means by superimposing the screen displayed on the monitor.
A recording medium characterized by functionally providing. A pre-recording medium that can be read and executed by a computer that can display a two-dimensional X-ray image of the human breast taken by an X-ray mammography device on the screen of a monitor and can give input information to the screen display. In the mammography interpretation program recorded in
When the computer reads the program from the recording medium and executes the program, the computer:
An image display means for displaying a two-dimensional X-ray image of a human breast taken by the X-ray mammography apparatus on a monitor screen, and an image display means.
On the screen of the monitor on which the X-ray image is displayed by the image display means,
A simulated area setting means for setting a minimum size set for detecting a lesion portion assumed in advance and a simulated area having a specific shape so as to be movable on a brightness image of the breast displayed on the screen.
Pixel specifying means for specifying a plurality of pixels forming the simulated area on the screen, and
The plurality of original luminances of at least a part of the plurality of pixels identified by the pixel identifying means, and the shadow of the breast structure drawn inside the simulated area on the at least a part of the plurality of pixels. A luminance synthesizing means for synthesizing at least the shape, shading, or spreading state of the shadow based on the original luminance of the simulated region, and the luminance having been subjected to a process for visually determining the state of spreading.
A superimposed display means that superimposes and displays the simulated area having the brightness synthesized by the luminance combining means on the screen displayed on the monitor.
A program characterized by providing functionally.

Images