JP7414236B2 - Lung shadow measurement program, lung shadow measurement method, and lung shadow measurement system - Google Patents

Description

The present invention relates to a lung shadow measurement program, a lung shadow measurement method, and a lung shadow measurement system.

2. Description of the Related Art Conventionally, there is a known technique for presenting a CT image of the lungs of a patient with diffuse lung disease to a doctor to assist in diagnosis.

Fujifilm, "Fujifilm and Kyoto University successfully jointly develop diagnostic support technology for interstitial pneumonia using AI technology" (https://www.fujifilm.co.jp/corporate/news/articleffnr_1417.html)

However, the conventional techniques have a problem in that the diagnostic accuracy of lung diseases may decrease. For example, it is known that as diffuse lung disease worsens, patients are unable to breathe in enough air. Therefore, if the imaging date and time differ, the volume of the lung estimated from the image may differ. As a result, the doctor must perform a diagnosis using images while taking into account changes in volume, and it is thought that the diagnostic accuracy may deteriorate in some cases.

One aspect of the present invention is to prevent a decline in the diagnostic accuracy of lung diseases.

In one embodiment, the lung shadow measurement program causes a computer to perform a process of calculating a first volume, which is the volume of the lungs, based on images taken of the patient's lungs. The lung shadow measurement program causes a computer to execute a process of calculating a second volume, which is the volume of a region in the lung where shadows are distributed, based on the image. The lung shadow measurement program causes the computer to perform a process of normalizing the second volume based on the first volume. The lung shadow measurement program causes the computer to execute a process of displaying information regarding the normalized second volume on the display device.

In one aspect, it is possible to suppress a decline in diagnostic accuracy of lung diseases.

FIG. 1 is a diagram showing an example of the configuration of a lung shadow measurement system. FIG. 2 is a diagram showing an example of the data structure of the patient DB. FIG. 3 is a diagram showing an example of the data structure of the case DB. FIG. 4 is a diagram showing an example of the configuration of the calculation section. FIG. 5 is a diagram showing an example of a screen displaying a non-normalized volume. FIG. 6 is a diagram showing an example of a screen displaying normalized volumes. FIG. 7 is a flowchart showing the processing flow of the lung shadow measuring device. FIG. 8 is a flowchart showing the flow of volume calculation processing. FIG. 9 is a diagram illustrating an example of a hardware configuration.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Examples of a lung shadow measurement program, a lung shadow measurement method, and a lung shadow measurement system according to the present invention will be described in detail below with reference to the drawings. Note that the present invention is not limited to this example. Moreover, each embodiment can be combined as appropriate within a consistent range.

The lung shadow measurement system is a system that supports a doctor's examination by displaying patient images (CT images, etc.) stored in a DB. For example, the lung shadow measurement system 1 can present to a doctor an image of a patient under examination and information regarding a disease associated with the acquired image. This allows the doctor to perform a medical examination with reference to the presented images and information regarding the disease. The user of the lung shadow measurement system is, for example, a doctor.

The configuration of a lung shadow measurement system according to an embodiment will be described using FIG. 1. FIG. 1 is a diagram for explaining the configuration of a lung shadow measurement system. As shown in FIG. 1, the lung shadow measurement system 1 includes a lung shadow measurement device 10 and a display device 20. The lung shadow measuring device 10 also includes an interface section 11, a storage section 12, and a control section 13.

The interface unit 11 is an interface for data input/output and data communication with other devices. Further, for example, the interface unit 11 outputs image data to the display device 20. At this time, the display device 20 displays an image based on the image data received via the interface section 11. The display device 20 is, for example, a display or a touch panel.

The storage unit 12 is an example of a storage device that stores programs executed by the control unit 13 and data necessary for executing the programs. The storage unit 12 stores a patient DB 121, a case DB 122, an image DB 123, and normalization function information 124.

The patient DB 121 stores information for each patient. FIG. 2 is a diagram showing an example of the data structure of the patient DB. As shown in FIG. 2, the patient DB 121 stores the patient ID, name, gender, and date of birth for each patient. In the example in Figure 2, the name of the patient whose patient ID is "0000000074" is "Patient Taro," his gender is male, and his date of birth is "1/1/1950." .

The case DB 122 stores test results. FIG. 3 is a diagram showing an example of the data structure of the case DB. As shown in FIG. 3, the case DB 122 stores image paths for each patient's examination date and the types of shadows detected in the images. In the example in Figure 3, in the image taken by the patient whose patient ID is "0000000074" in an examination on "2019/11/15", the shadow type of patch 001 is emphysema, and the shadow type of patch 002 is GGO. It has been shown that

Here, the patches are a plurality of regions obtained by dividing a CT image. For example, a patch is a 16×16 pixel image obtained by dividing a CT image. Further, the shade type is obtained based on the degree of similarity between the feature of each patch and the feature of each predetermined shade. For example, the lung shadow measurement system 1 can detect the shadow type corresponding to each patch by using a machine learning model that calculates the similarity between the feature amount extracted from a patch and the feature amount of each shadow type. can.

The image DB 123 stores images obtained in examinations. Furthermore, the path for acquiring images in the image DB 123 is stored in the case DB 122.

The normalization function information 124 is information on a function used in normalization processing by the calculation unit 133, which will be described later. Further, the normalization function may be different depending on the type of disease. In that case, the normalization function information 124 is information on a function associated with the type of disease. For example, normalization function information 124 may be predetermined coefficients and constants used in the function.

The control unit 13 executes programs stored in an internal storage device by, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), etc., using the RAM as a work area. This is realized by Further, the control unit 13 may be realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The control unit 13 includes a display control unit 131, an extraction unit 132, and a calculation unit 133.

The display control unit 131 causes the display device 20 to display an image generated based on the results of processing by each processing unit of the control unit 13. For example, the display control unit 131 outputs image data to the display device 20 via the interface unit 11.

The extraction unit 132 extracts feature amounts from the image. The extraction unit 132 may extract the feature amount of the entire image, or may extract the feature amount of each patch.

The calculation unit 133 calculates numerical values to be presented to the user. The calculation unit 133 calculates the estimated volume of the region in which organs and shadows are distributed from the image. Further, the calculation unit 133 can normalize the calculated volume. Here, the image used by the calculation unit 133 to calculate each volume may be referred to as a medical image.

Here, the necessity of normalization will be explained. For example, one organ, the lungs, changes in size depending on the amount of air a patient inhales. Therefore, even for the same patient, the size of the lungs in the image may differ depending on the timing of imaging. Additionally, as diffuse lung disease worsens, a person may not be able to take a deep breath. Therefore, the degree of progression of the disease may affect the size of the organ shown in the image.

For this reason, when comparing and evaluating the volumes of regions where shadows are distributed for a plurality of images taken at different times and dates, it is desirable that the volumes of the lungs be the same for each time and date of imaging. Therefore, the calculation unit 133 normalizes the volume of the region where shadows are distributed to reduce the influence of changes in lung volume.

FIG. 4 is a diagram showing an example of the configuration of the calculation section. As shown in FIG. 4, the calculation section 133 includes an organ detection section 133a, an organ volume calculation section 133b, a shadow detection section 133c, a shadow region volume calculation section 133d, and a normalization section 133e.

The organ detection unit 133a detects a predetermined organ from the image. For example, the organ detection unit 133a detects a region in the image that is similar to a predetermined feature amount of the lungs as a lung.

The organ volume calculation unit 133b calculates a first volume, which is the volume of the lungs, based on an image of the patient's lungs. It is assumed that it is known from the detection result of the organ detection unit 133a that the image for which the volume is calculated by the organ volume calculation unit 133b is a photograph of the patient's lungs. The organ volume calculation section 133b is an example of a first volume calculation section.

Here, as shown in FIG. 4, the medical image may be a plurality of slice images of the lungs obtained by a CT scan or the like. In this case, for example, the organ volume calculation unit 133b calculates the first volume by integrating the areas of regions in which the lungs are captured in a plurality of images obtained by linearly interpolating a plurality of slice images of the lungs. According to linear interpolation, a slice image between two slice images can be virtually generated.

The shadow detection unit 133c detects an area where shadows are distributed from the image. For example, the shadow detection unit 133c detects a region in the image that is similar to a predetermined feature amount of a shadow. For example, the shade detection unit 133c detects which shade each patch is most similar to.

Here, the shadow appearing in the image can be said to have predetermined characteristics. Furthermore, for example, shadows in lung images may appear not only due to specific lesions and diseases, but also due to changes due to aging, bleeding, changes due to smoking, and the like. Here, it is assumed that some of the shadows are abnormal shadows, and furthermore, some of the abnormal shadows are determined to be lesions. For example, the shadow detection unit 133c may simply detect the type of shadow for each patch, and the user may specify the types of lesions and diseases based on the detection results.

The shadow region volume calculation unit 133d calculates a second volume, which is the volume of a region in the lung where shadows are distributed, based on the image. For example, the organ volume calculation unit 133b calculates the first volume by integrating the area of a region in which the lungs are captured in a plurality of images obtained by linearly interpolating a plurality of slice images of the lungs. For example, the shadow region volume calculation unit 133d calculates the second volume by integrating the area of a region where shadows are distributed in a plurality of images obtained by linearly interpolating a plurality of slice images of the lung. The shadow area volume calculation unit 133d is an example of a second volume calculation unit.

The normalization unit 133e normalizes the second volume based on the first volume. For example, the normalization unit 133e can normalize the second volume using a function that is associated in advance with each disease or lesion specified by the user. As an example, if the reference lung volume is V _A , the first volume is V ₁ , the non-normalized second volume is V ₂ , and the coefficient determined for each disease is α, the normalization unit 133e , the normalized second volume V ₂ ′ can be obtained as shown in equation (1).
V ₂ ′=(V ₁ /V _A )^α×V ₂ …(1)

For example, in the case of a disease in which the volume of the shadow region is directly proportional to the volume of the lung, α is 1.0. On the other hand, in the case of a disease in which the volume of the shadow region does not change significantly with respect to the volume of the lung, α will be closer to 0 than 1.0. For example, it is known that the lesions of diffuse lung disease are widely distributed throughout the lungs and therefore vary greatly with respect to the volume of the lungs. Therefore, in the case of diffuse lung disease, α is often close to 1.0.

Furthermore, the amount of change in the volume of the region is considered to vary depending on the type of shadow. Therefore, the normalization unit 133e may normalize the volume using a different normalization function for each patch according to the type of detected shadow.

The normalization unit 133e can normalize the second volume based on a second image of the lungs taken at a different date and time than the first image. For example, if there are multiple images with different shooting dates and times, the normalization unit 133e performs normalization based on the lung volume calculated from one designated image.

The display control unit 131 causes the display device 20 to display information regarding the normalized second volume. Further, the display control unit 131 can display information regarding either the normalized second volume or the second volume before normalization, according to the user's operation. Further, the display control unit 131 can display a table listing the normalized second volume and a graph representing the normalized second volume.

Here, the first volume may be replaced by volume, especially when the target organ is the lungs. Lung volume is the space inside the lungs and is considered to change in response to changes in lung volume. Therefore, the normalization unit 133e may normalize the second volume based on the volume of the lungs, or may normalize the second volume based on the volume of the lungs. Further, the organ volume calculation unit 133b may directly calculate the lung volume, or may calculate an estimated value of the lung volume based on the lung volume.

The display control unit 131 causes the display device 20 to display a shadow amount progress comparison screen. FIG. 5 is a diagram showing an example of a screen displaying a non-normalized volume. As shown in FIG. 5, the shadow amount progress comparison screen 20a that displays non-normalized volumes includes a patient information display field 201, a volume table 211, a volume graph 212, a check box list 221, radio buttons 222, radio buttons 223, It has a shading histogram 231 and an image display area 241.

The patient information display field 201 displays information regarding the patient. For example, information acquired from the patient DB 121 is displayed in the patient information display column 201. The checkbox list 221 is a group of checkboxes for selecting parts of the lung to display. The shadow histogram 231 is a histogram that displays the appearance frequency of shadows for each position of the lung.

The image display area 241 displays images of the lungs at each imaging date and time. Furthermore, in the image display area 241, patches are represented by squares. Further, the color of the square frame line representing the patch may be different depending on the type of detected shadow.

In the shadow amount progress comparison screen 20a, the volume table 211 displays the non-normalized second volume for each shadow. Volume graph 212 is a bar graph representing the non-normalized second volume. The volume graph 212 is color-coded for each type of shadow.

The radio button 222 and the radio button 223 are radio buttons for selecting whether to display a non-normalized volume, that is, an absolute value, or a normalized volume. On the shadow amount progress comparison screen 20a, since the radio button 222 is selected, information regarding the unnormalized volume is displayed.

FIG. 6 is a diagram showing an example of a screen displaying normalized volumes. The same information as the shadow amount progress comparison screen 20a is displayed in the patient information display column 201, shadow histogram 231, and image display area 241 of the shadow amount progress comparison screen 20b. On the other hand, on the shadow amount progress comparison screen 20b, since the radio button 223 is selected, the volume table 211 and the volume graph 212 display information regarding the normalized volume.

In the example in Figure 5, the unnormalized volume of the area where the shadow type is "emphysema" is 1175.9 cm ³ on 2019/11/15, 2232.4 cm ³ on 2019/10/15, and 2232.4 cm 3 on 2019/9/2019. 15 shows that it was ^2736.0cm3 . On the other hand, in the example in Figure 6, the normalized volume of the area where the shadow type is "emphysema" is 1175.9 cm ³ on November 15, 2019, 1409.5 cm 3 on October 15, 2019, and 1409.5 cm ³ on October 15, 2019. /15 shows that it was ^1574.4cm3 . Here, the volume on November 15, 2019 is the standard for normalization.

In this example, the volume of 2019/10/15 corresponding to each shadow is approximately 0.63 times larger due to normalization. Additionally, the volume of 2019/9/15 corresponding to each shadow is approximately 0.57 times larger due to normalization.

FIG. 7 is a flowchart showing the processing flow of the lung shadow measuring device. As shown in FIG. 7, first, the display control unit 131 causes the display device 20 to display an image corresponding to the patient (step S11). Here, it is assumed that either absolute value or normalization is specified as the volume calculation method. For example, the volume calculation method is specified by radio buttons on the screen, as explained using FIGS. 5 and 6.

If the absolute value is specified as the volume calculation method (step S12, absolute value), the calculation unit 133 calculates the absolute value of the volume (step S13). In this case, the calculation unit 133 may output the calculation result of the shadow area volume calculation unit 133d without executing the processing of the normalization unit 133e. On the other hand, if normalization is specified as the volume calculation method (step S12, normalization), the calculation unit 133 calculates the normalized volume (step S14). Then, the display control unit 131 displays the calculated volume (step S15).

FIG. 8 is a flowchart showing the flow of volume calculation processing. The process in FIG. 8 corresponds to step S15 and step S16 in FIG. As shown in FIG. 8, first, the organ detection unit 133a detects an organ from a slice image (step S101). Next, the organ volume calculation unit 133b calculates the volume of the organ (step S102).

Further, the shadow detection unit 133c detects a shadow from the slice image (step S103). Then, the shadow region volume calculation unit 133d calculates the volume of the shadow distribution region (step S104).

Here, if the volume calculation method is an absolute value (step S105, absolute value), the process ends. On the other hand, if the volume calculation method is normalization (step S105, normalization), the normalization unit 133e normalizes the volume of the shadow distribution region using the normalization function determined from the organ type and shadow type. (Step S106).

As described above, the organ volume calculation unit 133b calculates the first volume, which is the volume of the lungs, based on images of the patient's lungs. The shadow region volume calculation unit 133d calculates a second volume, which is the volume of a region in the lung where shadows are distributed, based on the image. The normalization unit 133e normalizes the second volume based on the first volume. The display control unit 131 causes the display device 20 to display information regarding the normalized second volume. For example, the volume of the lungs may change due to various factors including recovery and deterioration of the disease. In contrast, the lung shadow measuring device 10 can normalize the volume of the shadow region in accordance with changes in the volume of the entire organ. As a result, according to this embodiment, a normalized volume can be presented, so that a decrease in the diagnostic accuracy of lung diseases can be suppressed.

The organ volume calculation unit 133b calculates the first volume based on the first image taken of the lungs.
The shadow area volume calculation unit 133d calculates the second volume based on the first image. The normalization unit 133e normalizes the second volume based on a second image of the lungs taken at a different date and time than the first image. The volume of the lungs may vary depending on the timing of imaging. According to this embodiment, normalization can be performed in consideration of such volume differences.

The organ volume calculation unit 133b calculates the first volume by integrating the area of the lungs in a plurality of images obtained by linearly interpolating a plurality of slice images of the lungs. The shadow region volume calculation unit 133d calculates the second volume by integrating the areas of regions in which shadows are distributed in a plurality of images. In CT imaging, slice images are obtained. In this embodiment, by using a plurality of slice images, it is possible to more accurately calculate a volume estimate.

The normalization unit 133e normalizes the second volume using a function that is associated in advance with each disease or lesion specified by the user. It is conceivable that the amount of change in the volume of the shadow region relative to the amount of change in the volume of the entire lung differs depending on the disease. In this embodiment, by considering such differences between diseases, normalization can be performed more accurately.

The display control unit 131 is characterized by displaying information regarding either the normalized second volume or the second volume before normalization, in response to a user's operation. In this manner, in this embodiment, the volume to be displayed can be switched according to the user's wishes.

The display control unit 131 displays a table listing the normalized second volume and a graph representing the normalized second volume. In this manner, in this embodiment, the normalized volume can be displayed in a manner that is easy for the user to intuitively understand.

Information including processing procedures, control procedures, specific names, and various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. Furthermore, the specific examples, distributions, numerical values, etc. described in the examples are merely examples, and can be changed arbitrarily.

Furthermore, each component of each device shown in the drawings is functionally conceptual, and does not necessarily need to be physically configured as shown in the drawings. That is, the specific form of distributing and integrating each device is not limited to what is shown in the drawings. In other words, all or part of them can be functionally or physically distributed and integrated into arbitrary units depending on various loads, usage conditions, and the like. Furthermore, all or any part of each processing function performed by each device can be realized by a CPU and a program that is analyzed and executed by the CPU, or can be realized as hardware using wired logic.

FIG. 9 is a diagram illustrating an example of a hardware configuration. As shown in FIG. 11, the lung shadow measuring device 10 includes a communication interface 10a, an HDD (Hard Disk Drive) 10b, a memory 10c, and a processor 10d. Furthermore, the parts shown in FIG. 13 are interconnected by a bus or the like.

The communication interface 10a is a network interface card or the like, and communicates with other servers. The HDD 10b stores programs and DB that operate the functions shown in FIG.

The processor 10d reads a program that executes the same processing as each processing unit shown in FIG. It is a hardware circuit. That is, this process executes the same functions as each processing unit included in the lung shadow measurement device 10. Specifically, the processor 10d reads a program having the same functions as the display control section 131, the extraction section 132, and the calculation section 133 from the HDD 10b or the like. The processor 10d then executes a process that executes the same processing as the display control unit 131, extraction unit 132, calculation unit 133, and the like.

In this way, the lung shadow measuring device 10 operates as an information processing device that executes a learning method by reading and executing a program. Further, the lung shadow measuring device 10 can also realize the same functions as the above-described embodiments by reading out the program from the recording medium using the medium reading device and executing the read program. Note that the programs in other embodiments are not limited to being executed by the lung shadow measuring device 10. For example, the present invention can be similarly applied to cases where another computer or server executes a program, or where these computers or servers cooperate to execute a program.

This program can be distributed via a network such as the Internet. In addition, this program is recorded on a computer-readable recording medium such as a hard disk, flexible disk (FD), CD-ROM, MO (Magneto-Optical disk), or DVD (Digital Versatile Disc), and is read from the recording medium by the computer. It can be executed by being read.

1 Lung shadow measurement system 10 Lung shadow measurement device 11 Interface unit 12 Storage unit 13 Control unit 20 Display device 20a, 20b Shadow amount progress comparison screen 121 Patient DB
122 Case DB
123 Image DB
124 Normalization function information 131 Display control unit 132 Extraction unit 133 Calculation unit 133a Organ detection unit 133b Organ volume calculation unit 133c Shadow detection unit 133d Shadow region volume calculation unit 133e Normalization unit 201 Patient information display field 211 Volume table 212 Volume graph 221 Checkbox list 222, 223 Radio button 231 Shading histogram 241 Image display area

Claims (7)

Calculating a first volume, which is the volume of the lung, based on an image taken of the patient's lungs;
Based on the image, calculate a second volume that is the volume of a region in the lung where shadows are distributed;
normalizing the second volume based on the first volume using a function previously associated with each disease or shadow specified by the user ;
displaying information regarding the normalized second volume on a display device;
A lung shadow measurement program characterized by having a computer execute the processing. The process of calculating the first volume calculates the first volume based on a first image taken of the lung,
The process of calculating the second volume calculates the second volume based on the first image,
The lung according to claim 1, wherein the normalization process normalizes the second volume based on a second image of the lung taken at a different date and time than the first image. Shadow measurement program. In the process of calculating the first volume, the first volume is calculated by integrating the area of a region in which the lungs are captured in a plurality of images obtained by linearly interpolating a plurality of slice images of the lungs. ,
The lung according to claim 1 or 2, wherein the process of calculating the second volume calculates the second volume by integrating the area of a region where shadows are distributed in the plurality of images. Shadow measurement program. 4. The displaying process is characterized in that information regarding either the normalized second volume or the unnormalized second volume is displayed in response to a user's operation. The lung shadow measurement program according to any one of the above. 5. The displaying process includes displaying a table listing the normalized second volume and a graph representing the normalized second volume. The lung shadow measurement program described. Calculating a first volume, which is the volume of the lung, based on an image taken of the patient's lungs;
Based on the image, calculate a second volume that is the volume of a region in the lung where shadows are distributed;
normalizing the second volume based on the first volume using a function previously associated with each disease or shadow specified by the user ;
displaying information regarding the normalized second volume on a display device;
A lung shadow measurement method characterized in that processing is performed by a computer. A lung shadow measurement system comprising a lung shadow measurement device and a display device,
The lung shadow measuring device includes:
a first volume calculation unit that calculates a first volume that is the volume of the lungs based on an image taken of the patient's lungs;
a second volume calculation unit that calculates a second volume that is a volume of a region in the lung where shadows are distributed based on the image;
a normalization unit that normalizes the second volume based on the first volume using a function that is associated in advance for each disease or shadow specified by the user ;
a display control unit that causes a display device to display information regarding the normalized second volume;
has
The display device includes:
A lung shadow measurement system, characterized in that information regarding the normalized second volume is displayed.

Images