JP6432910B2 - Medical image processing apparatus, ultrasonic diagnostic apparatus, and medical image processing method - Google Patents

Description

  The present invention relates to a medical image processing apparatus, an ultrasonic diagnostic apparatus, and a medical image processing method for performing image processing on an ultrasonic image obtained by imaging a subject to which an ultrasonic contrast agent is administered.

Conventionally, as means for diagnosing pancreatic cancer, abdominal ultrasound, computed tomography (CT), nuclear magnetic resonance imaging (MRI), positron emission tomography (PET), Various examination techniques using endoscopic retrograde cholangiopancreatography (ERCP), endoscopic ultrasonography (EUS), and the like are used.
Among them, EUS can bring a high-frequency ultrasonic transducer close to the pancreas by in-vivo operation, and its high resolution enables detailed observation of the pancreatic parenchyma, pancreatic duct, and organs near the pancreas. is there.

In addition, the development of electronic transducers and Doppler functions has developed a technique (EUS-FNA (fine needle aspiration)) in which pancreatic lesions are punctured with a puncture needle under EUS, and aspiration cytology is developed. Improvement of diagnostic accuracy surpassing MRI has been reported, and EUS has become an indispensable examination technique in the diagnosis of pancreatic cancer.
By the way, it has been reported that 20 to 35% of pancreatic cancers are accompanied by chronic pancreatitis, and chronic pancreatitis is considered to be one of the risk factors of pancreatic cancer. As a problem of EUS and EUS-FNA, which are considered to have the highest sensitivity for diagnosis of pancreatic cancer, it has been pointed out that the diagnostic accuracy of pancreatic cancer with chronic pancreatitis is lower than that of non-complicated cases. This is because the inflammatory benign mass due to pancreatitis also shows low echoes that appear dark on the EUS screen in the EUS image as in the case of pancreatic cancer, and even if the cells are collected with EUS-FNA, cell changes due to strong inflammation are distinguished from cancer cells. This is because it may be difficult. In 2005, Varadarajulu et al. Reported that the sensitivity of EUS and EUS-FNA for detecting pancreatic cancer without chronic pancreatitis was 91.3%, whereas the sensitivity of EUS and EUS-FNA for detecting pancreatic cancer with chronic pancreatitis was 73. Reported only 0.9%.

In recent years, second-generation ultrasound contrast agents have been developed, and contrast EUS has become possible due to the development of observation devices that detect ultrasound contrast agents such as microbubbles. A major difference between contrast-enhanced CT or contrast-enhanced MRI and contrast-enhanced EUS is that a lesion contrast observation, that is, a blood vessel structure of the lesion and blood flow dynamics can be observed in real time.
In such an inspection using a contrast agent, an inspection using a TIC (Time Intensity Curve) may be performed. The TIC is a time luminance curve showing the time change of the average luminance in the region of interest (ROI) set on the ultrasonic image. By displaying this TIC, the change in the contrast agent concentration is shown. It can be observed, the state of the circulation of the contrast medium in the body can be quantitatively grasped, and the presence or absence of the disease in the subject can be diagnosed (see, for example, Patent Document 1).
Further, when a diagnosis is performed using TIC, even if a non-observation target site of TIC is included in the ROI in addition to the tissue that is the observation target site of TIC, the time for only the observation target site of TIC A device that creates a luminance curve (see, for example, Patent Document 2) has also been proposed.

JP 2005-95376 A JP 2010-75586 A

  By the way, Fusaroli et al. Reported in 2010 that heterogeneous weak contrast findings are characteristic of pancreatic cancer, and that imaging improves the diagnostic accuracy of pancreatic cancer with chronic pancreatitis in EUS. There are few reports on contrast-enhanced EUS, and verification from various aspects is necessary. However, contrast-enhanced EUS has the potential to overcome the problems of conventional EUS and EUS-FNA. The “weak” contrast effect in contrast-enhanced EUS means that the contrast agent does not enter the lesion so much, that is, the blood vessel and blood flow are scarce and the entire lesion looks dark even over time. Conversely, a “strong” contrast effect means that a lot of contrast agent enters the lesion, that is, the blood vessels and blood flow are abundant and the entire lesion looks white. Further, “non-uniformity” means that a white portion and a black-looking portion are mixed in the lesion, and “uniformity” means that the lesion appears uniformly white or uniformly black.

Diagnosis of non-uniform and weak contrast findings depends solely on the experience and impression of the practitioner, and in daily clinical practice, pancreatic cancer that exhibits a contrast pattern close to uniform strong contrast findings is also experienced.
Therefore, the contrast EUS findings, such as “uniform”, “non-uniform”, “strong”, “weak”, provide objectivity so that doctors of any experience can make the same diagnosis accurately, and contrast is close to normal patterns. A technique for picking up pancreatic cancer showing findings is desired.
Since contrast media, that is, individual bubbles, are captured as white spots in the EUS image, as described above, a region of interest (ROI) is set in the lesion, and changes in luminance over time in the ROI are graphed (numerical values). And used in clinical practice. However, this software is only a numerical value of brightness, and it only gives objectivity to the findings such as “strong” and “weak”, and a better method has been desired.
Therefore, the present invention has been made by paying attention to the above-mentioned conventional unsolved problems, and provides information that helps to appropriately diagnose an imaging finding without being influenced by experience or impression. It is an object of the present invention to provide a medical image processing apparatus, an ultrasonic diagnostic apparatus, and a medical image processing method.

  According to one embodiment of the present invention, a display processing unit that displays an ultrasound image on a display device based on ultrasound image data representing a time-series ultrasound image using an ultrasound contrast agent, and a display on the display device An input operation unit that sets a region of interest for the ultrasonic image that has been set, a classification processing unit that divides the region of interest set by the input operation unit into a plurality of cells, and a cell for each frame based on the ultrasonic image data A luminance level calculation unit that calculates the luminance level of a cell and one of the cells as a target cell, and non-uniformity based on the difference between the luminance level of the surrounding cells of the target cell and the luminance level of the target cell The process of determining the sex is repeated by switching the target cell, and the non-uniformity degree calculation unit that calculates the degree of non-uniformity for each frame, and the non-uniformity for each frame calculated by the non-uniformity degree calculation unit. Performance that outputs the degree of uniformity in time series A result output unit, the medical image processing apparatus comprising a provided.

According to another aspect of the present invention, there is provided an ultrasonic diagnostic apparatus including the medical image processing apparatus according to the above aspect.
According to another aspect of the present invention, a step of displaying an ultrasound image on a display device based on ultrasound image data representing a time-series ultrasound image using an ultrasound contrast agent; Dividing the region of interest set by the input operation unit for the displayed ultrasonic image into a plurality of cells, calculating a luminance level of the cell for each frame based on the ultrasonic image data, One of the cells is the target cell, and the process of determining non-uniformity based on the difference between the luminance level of the surrounding cells of the target cell and the luminance level of the target cell is repeated by switching the target cell. Thus, there is provided a medical image processing method comprising a step of calculating the degree of non-uniformity for each frame and a step of outputting the degree of non-uniformity for each frame in time series.

  According to one aspect of the present invention, the degree of brightness non-uniformity is calculated for each frame of the ultrasound image in the region of interest, and the degree of non-uniformity is output in time series. When there is a difference in the degree of inhomogeneity in the ultrasound image, such as in pancreatitis and chronic pancreatitis, etc., it may be helpful for differentiation by referring to the degree of inhomogeneity obtained from the ultrasound image in the region of interest it can.

It is a schematic block diagram which shows an example of the medical image processing apparatus in this invention. It is a flowchart which shows an example of the process sequence of an analyzer. It is an example of the ultrasonic image which set the region of interest. It is an example of the ultrasonic image which set the grid so that a region of interest may be included. It is explanatory drawing for demonstrating the position of a cell. It is a flowchart which shows an example of the process sequence of a madara index | exponent process. It is explanatory drawing for demonstrating the calculation method of a madara index | exponent. It is an example of the ultrasonic image of a pancreatic cancer case showing the change before and after injecting an ultrasonic contrast agent. It is an example of the display screen showing an ultrasonic image. It is an example of the display screen showing the luminance level of an ultrasonic image and a cell. It is an example of the graph which shows an example of a time-dependent change of a Madara index, and the display screen of an average Madara index. It is an example of the ultrasonic image of a chronic pancreatitis showing the change before and after injecting an ultrasound contrast agent. It is an example of the ultrasonic image of a pancreatic cancer case showing the change before and after injecting an ultrasonic contrast agent.

In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of embodiments of the present invention. However, it will be apparent that one or more embodiments may be practiced without such specific details. In other instances, well-known structures and devices are schematically shown in order to simplify the drawing.
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a configuration diagram showing an example of a schematic configuration of a medical image processing apparatus 10 according to the present invention.
The medical image processing apparatus 10 according to the present invention uses, for example, microbubbles as an ultrasonic contrast agent, and uses ultrasonic image data obtained thereby as a processing target.

As illustrated in FIG. 1, the medical image processing apparatus 10 includes, for example, a personal computer, and includes, for example, an ultrasonic image data storage apparatus 1 that stores ultrasonic image data acquired by the ultrasonic image acquisition apparatus 20 and an ultrasonic wave. An analysis device 2 that performs processing on ultrasonic image data stored in the image data storage device 1, an input device 3 such as a mouse or a keyboard for performing various settings necessary for analysis in the analysis device 2, And a display device 4 for displaying the sonic image data, the analysis result of the analysis device 2, and the like.
The ultrasonic image acquisition apparatus 20 is, for example, an ultrasonic diagnostic apparatus, and includes an ultrasonic probe, transmits ultrasonic waves to the subject, receives reflected waves from the subject, and based on the reflected waves. Generate ultrasound image data.

FIG. 2 is an example of a flowchart illustrating an example of a processing procedure of the analysis apparatus 2.
The analysis device 2 first inputs ultrasonic image data and stores it in the ultrasonic image data storage device 1. The input ultrasonic image data is displayed on the display device 4 (step S1). For example, an ultrasonic image of the first frame is displayed.
The ultrasonic image data may be acquired directly from the ultrasonic image acquisition device 20, or a portable storage such as a USB (Universal Serial Bus) memory in which the ultrasonic image data from the ultrasonic image acquisition device 20 is stored. You may acquire from a medium.

When the user operates the input device 3 to perform a region of interest setting operation (step S2), the analysis device 2 performs a region of interest setting process according to the set region of interest (step S3). ).
For example, as illustrated in FIG. 3, the user performs a region-of-interest setting operation by displaying a plurality of points on a display device 4 such as a mouse so as to surround a region to be analyzed by the user in an ultrasonic image displayed on the display device 4. By selecting with, an area surrounded by line segments connecting the selected points is set. In the analysis device 2, the points selected in response to the selection operation of the plurality of points on the display device 4 are sequentially connected by line segments and displayed, and the closed region connecting the selected points is recognized as the region of interest. The method of setting the region of interest is not limited to this. For example, a region such as a circle, ellipse, or rectangle may be set as the region of interest by adjusting the size so as to surround the region to be analyzed. Any method may be used as long as the region of interest can be set so as to surround the region to be analyzed.

In the analysis apparatus 2, when an operation for enlarging or reducing the ultrasonic image displayed on the display apparatus 4 or an operation instruction for moving the display range of the ultrasonic image is performed, the analysis apparatus 2 responds to the instruction. Process.
When the setting operation of the region of interest is completed, the analysis apparatus 2 proceeds to step S4 and displays a grid, for example, as shown in FIG. For example, a preset number of basic grids are displayed. At this time, for example, the grid interval at the reference scale is determined in advance, and the interval is adjusted according to the scale of the displayed ultrasonic image. In addition, the grid is displayed so as to include the region of interest set by the user.

Then, when an adjustment operation such as the interval or number of grids or the display position is performed by the user's operation of the input device 3 (step S5), the grid is updated according to the performed adjustment operation and displayed. (Step S6).
For example, if the number of grid lines is designated for the vertical and horizontal directions, the designated number of grids are displayed. If the grid interval is specified for each of the vertical and horizontal directions, the grid is displayed at the specified interval. For example, after the basic grid is displayed, display for inquiring about the display number and interval of the grid lines in the vertical and horizontal directions, the display position, and the like may be performed, and the grid may be adjusted according to the input operation on the input device 3.

In this example, the basic grid is displayed first, and then the grid is adjusted. However, without displaying the basic grid, the number of grid lines displayed in the vertical and horizontal directions, the interval, and the display position of the grid The grid may be displayed according to the input value or the input display position.
The basic grid stores the size and scale of the region of interest in association with the appropriate number of grid lines in advance, and based on this, the size of the region of interest and the scale thereof are set. An appropriate number of grid lines may be automatically specified, and the specified grid lines may be displayed as the number of basic grids.

For example, if the region of interest is small, setting many grids may make the cells finer, making it difficult to understand the non-uniformity of the cells in the region of interest. There is also a possibility of becoming. On the other hand, when the number of grids set is small, the number of cells included in the region of interest becomes too small, and it may be difficult to understand luminance non-uniformity between cells. Therefore, by setting an appropriate number of grid lines in accordance with the size of the region of interest, its scale, etc., it is possible to easily understand the non-uniformity of the cells included in the region of interest.
In the analysis apparatus 2, when the adjustment of the grid is completed, the process proceeds to step S7, and the variable F for identifying the frame of the ultrasonic image data is set to F = 1.

Then, the process proceeds to step S8, and the luminance is calculated for each cell delimited by the grid for the frame specified by the variable F. That is, the ultrasonic image data for one frame is read from the ultrasonic image data storage device 1, and in the ultrasonic image represented by the ultrasonic image data, for each cell specified by the grid set in step S4, Calculate brightness. That is, if m is a row of cells separated by a grid and n is a column, the luminance is calculated for all m × n cells.
In each frame, the grid is set so as to satisfy the positional relationship of the grid with respect to the ultrasonic image at the time of grid setting, and the luminance is calculated for each cell specified by the grid. That is, in each frame, images represented by cells at the same position represent images of the same part.

The calculation of the cell brightness is performed according to the following procedure.
Note that the luminance calculation is performed only for cells in the region of interest. For cells outside the region of interest, the brightness is not set and the cell is invalidated. The brightness is set only when all the pixels included in the cell are included in the region of interest. For example, when a region of interest boundary exists in one cell, this cell is invalid.
Then, based on the luminance data of all the pixels included in the cell determined to be valid, the representative value is obtained. For example, the total value of the luminance values of all the pixels may be used as the representative value, and the average value of the luminance values of all the pixels, or the mode value, the maximum value, the minimum value, and the intermediate value may be used as the representative value.
Then, based on the obtained representative value, it is classified into three gradations according to the luminance level. For example, when the brightness is classified into black, white, and others, and the brightness is expressed by 256 gradations of 0 to 255, for example, the brightness level is 0 when the brightness is 0 or more and 50 or less, and the brightness is 51 or more and 200 or less. The luminance level is classified as 1, and the luminance level is classified as 2 when the luminance is between 201 and 255.

Note that here, it is effective only when all the pixels included in the cell are included in the region of interest, but it is effective when any pixel included in the cell is included in the region of interest, and all the pixels are interested. It may be invalid only when it is not included in the area. Alternatively, it may be considered effective only when a certain percentage of the cell is included in the region of interest, such as more than half of the pixels included in the cell, or more than 1/3.
If any pixel included in the cell is valid when included in the region of interest, the representative value is included in the region of interest among the pixels included in the cell when calculating the representative value. A representative value may be obtained only for pixel data. Alternatively, based on pixel data included in the region of interest in the cell, pixel data that is not included in the region of interest in the cell may be interpolated, and a representative value may be obtained based on this interpolation. The representative value may be obtained based on the data of all the pixels included in the cell regarded as valid, including the data of the pixels not included in the region.

In addition, the brightness levels are classified into three gradation levels, but the present invention is not limited to this, and may be classified into an arbitrary number of brightness levels as long as the gradation levels are three or more. In the case of classification into luminance levels of three or more gradations, the threshold value dth for determining mottle is set to “2” in the same manner as described above, and when there is a difference in the luminance levels of two or more gradations, the mottle is determined. You may make it judge. Alternatively, when the luminance level is classified into four or more gradation levels, two or more gradations may be determined to be mottle, and three or more gradations may be determined to be mottle. The threshold value dth can be set to an arbitrary value according to the number of gradations of the luminance level. For example, when the brightness level is classified into five gradation levels and the threshold value dth = 2, the brightness level is 0 and the brightness level is 3 or more, the brightness level is 1 and the brightness level is 4 or more, and the brightness level is 2 and the brightness level is 5. It will be judged mottled.
If the luminance level is set for each cell in this way, the cell position and the luminance level are associated with each other and stored in a predetermined storage area.

The cell position is set as follows, for example.
For example, as shown in FIG. 5, assuming that 5 grid lines are set vertically and horizontally and are divided into cells of 5 rows × 5 rows, the upper left cell is 1 in 5 rows × 5 rows of cells. Let the lower right cell be 25. That is, m rows × n columns of cells divided by the grid are numbered in order from the upper left, with the upper left cell as 1, the lower right cell as the last.
Conditions such as whether to enable or disable a cell that includes the boundary of the region of interest in the cell, how to calculate the representative value, or how many gradations the luminance level is classified into Thus, the representative value of the cell, that is, the luminance is determined, and non-uniformity determination described later is performed based on this. Therefore, regarding the above-mentioned various conditions for determining the luminance of these cells, the luminance variation in the region of interest is reflected in the luminance level of the cell in consideration of the grid interval or scale, the luminance variation in the region of interest, etc. It may be set so that. For example, before or after the calculation of the luminance level, after the calculation, whether the cell including the boundary of the region of interest in the cell is enabled or disabled, how the representative value is calculated, or the luminance A display for inquiring about the level of gradation to be classified is displayed on the display device 4 and is configured to operate under conditions according to an instruction operation by the user on the input device 3. Then, as shown in FIG. 10 to be described later, for example, by displaying the calculated luminance level of each cell in one frame along with the actual ultrasonic image on the display device 4, the luminance level of the cell in the region of interest is displayed. After confirming the variation, the conditions may be changed again to derive the luminance level for each cell.

Note that the luminance level for each cell is derived under the same conditions for all frames.
In this way, the luminance level for each cell is calculated for the frame specified by the variable F.
If the luminance level for each cell is calculated for the frame specified by the variable F, the process proceeds to step S9, and then a madara representing the degree of nonuniformity of the luminance of the cells in one frame specified by the variable F. Calculate the exponent.
The calculation of the madara index is performed according to the processing procedure of the madara index calculation process shown in FIG.

First, a variable i for specifying each cell divided by the grid, that is, a variable for specifying a cell specified by a number as shown in FIG. 6 is set to i = 1 (step S21).
Next, the process proceeds to step S22, where the luminance level d (i) of the cell specified by the variable i is read from a predetermined storage area, and it is determined whether this luminance level d (i) is valid. When the brightness level d (i) is valid, that is, when the cell is not set to be invalid and any one of 0, 1, and 2 is set as the brightness level, the process proceeds to step S23. To do.

In step S23, it is determined whether “i-Icol> 0 and d (i-Icol) is valid”. Here, Icol is the number of rows of cells divided by the grid. For example, in the case of FIG. 5, since it is divided into 5 rows × 5 columns, Icol is 5.
When i-Icol> 0 and d (i-Icol) are valid, the process proceeds to step S24 to determine whether “abs [d (i) -d (i-Icol)]> dth”. . When “abs [d (i) −d (i−Icol)]> dth” is satisfied, the process proceeds to step S25, the madara index Sm is incremented by 1, and the process proceeds to step S26. Proceed to step S26.

In step S24, abs [X] is the absolute value of X. dth is a threshold value that represents the difference between the luminance levels of the cells. Here, since the luminance level of the cell is classified into three gradations of “0, 1, 2”, the threshold value dth is “ 2 ”, it is determined whether or not there is a difference of two or more gradations in the luminance level.
In the cell division shown in FIG. 5, since i-Icol> 0 is from the second row, in step S23 and step S24, when i = α, cell α is placed in the row above cell α. And the brightness level of the cell α is compared with the brightness level of the cell (i (α) −Icol) adjacent to the cell α in the row above the cell α. When the difference is larger than the threshold value dth, it is determined that the difference is uneven.

When it is determined that the spot is mottled, the process proceeds to step S25, and the madara index Sm is incremented by one.
In step S26, it is determined whether “i + Icol <Iall + 1 and d (i + Icol) is valid”. Here, Iall is the total number of cells partitioned by the grid. For example, in the case of FIG. 5, since it is divided into 5 rows × 5 columns, Iall is “25”.
If i + Icol <Iall + 1 and d (i + Icol) is valid, the process proceeds to step S27 to determine whether or not “abs [d (i) −d (i + Icol)]> dth”. When “abs [d (i) −d (i + Icol)]> dth” is satisfied, the process proceeds to step S28, and the madara index Sm is incremented by 1, and the process proceeds to step S29. move on.

That is, in the cell division shown in FIG. 5, since i + Icol <Iall + 1 is the lowest row, in step S26 and step S27, when i = α, the row below cell α is adjacent to cell α. When there is a cell in the region of interest to be compared, the luminance level of the cell α is compared with the luminance level of the cell α in the row below the cell α and the luminance level of the adjacent cell (i (α) + Icol). When the value is larger than the value dth, it is determined to be mottled.
In step S29, it is determined whether “(i−1) mod Icol ≠ 0 and d (i−1) is valid”. Here, “XmodY” represents a remainder obtained by dividing X by Y.

When (i-1) mod Icol ≠ 0 and d (i-1) is valid, the process proceeds to step S30, and is "abs [d (i) -d (i-1)]>dth"? Judging. Then, when “abs [d (i) −d (i−1)]> dth” is satisfied, the process proceeds to step S31, the madara index Sm is incremented by 1, and the process proceeds to step S32. Proceed to S32.
That is, in the cell division shown in FIG. 5, since (i−1) mod Icol ≠ 0 is a column excluding the leftmost column, in step S29 and step S30, when i = α, cell α When there is a cell in the region of interest adjacent to the cell α in the left column, the luminance level of the cell α and the luminance level of the cell (i (α) −1) adjacent to the cell α in the left column of the cell α When the difference is larger than the threshold value dth, it is determined that the spot is mottled.

In step S32, it is determined whether “i mod Icol ≠ 0 and d (i + 1) is valid”.
When i mod Icol ≠ 0 and d (i + 1) is valid, the process proceeds to step S33, and it is determined whether “abs [d (i) −d (i + 1)]> dth”. When “abs [d (i) −d (i + 1)]> dth” is satisfied, the process proceeds to step S34, the madara index Sm is incremented by 1, and the process proceeds to step S35. Proceed to

That is, in the cell division shown in FIG. 5, since i mod Icol ≠ 0 is a column except the rightmost column, in step S32 and step S33, when i = α, When there is a cell in the region that is adjacent to the cell α in the column, the luminance level of the cell α is compared with the luminance level of the cell α in the right column of the cell α and the adjacent cell (i (α) +1). When the difference is larger than the threshold value dth, the mottle is determined.
In step S35, the variable i is updated to i + 1, and when the variable i is Iall + 1, that is, when the madara determination is completed for all the cells (step S36), the process proceeds to step S37, and the madara index Sm is changed to the frame. The variable F is identified and stored in a predetermined storage area.
Thus, the madara index calculation process for the frame F is completed.

Then, returning to FIG. 2, since the madara index calculation process (step S9) for the frame F is completed, the process proceeds to step S10, and the frame variable F is updated to F = F + 1.
If the calculation of the Madara index is not completed for all frames, the process returns to Step S8, the luminance level of the cell is calculated for the next frame (Step S8), and the Madara index calculation process is performed (Step S9).
That is, in this embodiment, an ultrasound image is acquired using an ultrasound contrast agent. Since the ultrasonic contrast agent is injected into the subject and then transmitted to the body and disappears, the obtained ultrasonic image changes every moment after the injection of the ultrasonic contrast agent into the subject. Therefore, after injecting an ultrasound contrast agent into a subject, an ultrasound image within an effective period (for example, about 1 minute) that can obtain an ultrasound image according to the condition of the observation target region is used as an analysis target. The madara index is calculated for each frame. The effective period is, for example, about 1 minute after the injection of an ultrasound contrast agent when differentiating between pancreatic cancer and pancreatitis, but it varies depending on the performance and the case of the ultrasound image acquisition device 20 and is 1 minute. It is not limited to this, and can be set arbitrarily.

When the Madara index is calculated for all the frames, the process proceeds from Step S11 to Step S12, and the average value is calculated based on the Madara index for all the frames to obtain the average Madara index. The madara index in each frame is displayed as a graph, for example, with the horizontal axis representing time and the vertical axis representing the madara index, and the average madara index is displayed (step S12). The process ends as described above.
Here, the average madara index for the effective period is calculated, but the user may set the period for calculating the average madara index. For example, after displaying the madara index in the effective period in a time-series graph, the display device 4 is inquired about the period for calculating the average madara index. For example, when the user refers to a time-series graph of the madara index and inputs a section where a good madara index is obtained by operating the input device 3, the section set by the input device 3 For the madara index, an average madara index may be calculated. Similarly, for the graph showing the madara index in time series, the madara index may be displayed in time series only for the section set by the user setting the input device 3.
Further, the average madara index is not necessarily calculated. In other words, at the stage where the madara index is displayed in time series, it is possible to roughly grasp in which range the madara index is changing. Therefore, it is not always necessary to calculate the average madara index.

FIG. 7 shows an example of the luminance level of the cell in the region of interest. Here, for the sake of simplicity, a case will be described in which a grid region composed of cells partitioned by a grid matches a region of interest, and cells in the region of interest are partitioned into 5 rows × 5 columns. In FIG. 7, white cells indicate that the luminance level is “0”, cells that are roughly hatched indicate that the luminance level is “1”, and cells that are finely hatched are The luminance level is “2”.
Since the grid areas are numbered in order from the upper left as shown in FIG. 5, the cells shown in FIG. 7 are numbered from 1 to 25.

As shown in FIG. 7A, in the frame at the time “Time x1”, when only the cell 13 in the third row and third column has the luminance level “2” and the other cells are “0”, two rows When the cell 8 in the third column is the target cell, the cell 8 satisfies i (= 8) + Icol (= 5) <Iall (= 25) +1 and d (i (= 8) + Icol (= 5) )) Is effective and satisfies abs [d (i = 8) −d (i (= 8) + Icol (= 5))]> dth (= 2). Therefore, the relationship between the cell 8 and the cell 13 is determined to be mottled, and the Madara index is Sm = 1.
When the cell 12 in the third row and the second column is the target cell, the cell 12 has i (= 12) + Icol (= 5) <Iall (= 25) +1 and d (i (= 12) + Icol (= 5). ) And abs [d (i = 12) −d (i (= 12) + Icol (= 5))]> dth (= 2). Therefore, the relationship between the cell 12 and the cell 13 is determined to be mottled, and the Madara index is Sm = 2.

In the same procedure, when the cell 13 in the third row and the third column is the target cell, the cell 13 is divided into the upper, lower, left, and right cells 8, 18, 12, and 14, respectively, in steps S23, 24, 26, 27, 29, and 30. , 32, 33, the Madara index Sm is 6.
Further, when the cell 14 in the third row and the fourth column is the target cell, the cell 14 satisfies the relationship between the cell 13 and steps S29 and S30, and therefore, the Madara index Sm is 7. When the cell 18 in the fourth row and the third column is the target cell, the cell 18 satisfies the relationship between the cell 13 and steps S22 and S23, and therefore, the Madara index Sm is 8. As a result, the Madara index in the frame of the time “Time x1” is Sm = 8.

  Similarly, as shown in FIG. 7B, in the frame at the time “Time x2”, the cell 7 in the second row and the second column and the cell 8 in the second row and the third column have the luminance level “2”, the third row and the third column. When the cell 13 in the third row and the cell 14 in the third row and the fourth column have the luminance level “1” and the other cells are “0”, the cell 3 and the cell 7 below the cell 3 Between the cell 8 below and between the cell 6 and the cell 7 on the right is determined to be mottled. Also, between cell 7 and its upper (cell 2), its left (cell 6) and its lower (cell 12), cell 8 and its upper (cell 3), and its right (cell 9) Between each cell, between the cell 9 and the left cell 8 and between the cell 12 and the cell 7 above it is determined to be mottled. As a result, the madara index in the frame at the time “Time x2” is Sm = 10.

  Similarly, as shown in FIG. 7C, in the frame of time “Time x3”, the cell 2 in the first row and the second column, the cell 7 in the second row and the second column, and the cell 18 in the fourth row and the third column are When the luminance level “2”, the cell 8 in the second row and the third column, and the cell 9 in the second row and the fourth column are the luminance level “1” and the other cells are “0”, the cell 1 and its right Cell 2, cell 2 and its left (cell 1) and right (cell 3) cells, cell 3 and its left cell 2, cell 6 and its right cell 7 Between cell 7 and each cell on the left (cell 6) and below (cell 12), between cell 12 and cell 7 above, cell 13 and cell 18 below Between cell 17 and cell 18 on the right, cell 18 and above (cell 13), left (cell 17), right (cell 19), below (cell 23) Between the cells respectively, between the cell 19 and the cell 18 of the left, between the cell 23 and cell 18 thereon, but is determined to plaques respectively. As a result, the Madara index in the frame at the time “Time x3” is Sm = 16.

  Similarly, as shown in FIG. 7D, in the frame of time “Time x4”, the cell 4 in the first row and the fourth column, the cell 12 in the third row and the second column 12, the cells 22 and 5 in the fifth row and the second column. Cell 24 in row 24 column has luminance level “2”, cell 6 in row 2 and column 1, cell 7 in row 2 and column 2, cell 11 in row 3 and column 1, cell 14 in row 3 and column 4, and When the cell 15 in the third row and the fifth column has the luminance level “1” and the other cells are “0”, the cell 3 and the right cell 4, the cell 4 and the left (cell 3), Between each cell on the right (cell 5) and below (cell 9), between cell 5 and left cell 4, between cell 9 and cell 4 above, cell 12 and its right (cell 13) and below (cell 17), between cell 13 and left cell 12, and between cell 17 and above (cell 12) and below (cell 22) , Between cell 19 and cell 24 below, between cell 21 and cell 22 on the right, cell 22 and above (cell 17), left (cell 21), and right (cell 23) Between cell 23 and its left cell 22 and right cell 24, between cell 24 and its upper cell (cell 19), its left (cell 22) and right cell (cell 25). , The gap between the cell 25 and the left cell 24 is determined to be mottled. As a result, the madara index in the frame at the time “Time x4” is Sm = 22.

Next, the operation of the above embodiment will be described.
FIG. 8 is an example of a time-series ultrasound image obtained using microbubbles as an ultrasound contrast agent, and is an ultrasound image of a typical pancreatic cancer case. A hypoechoic mass is found in the pancreas. In general, pancreatic cancer appears to be lower echo (black) than normal pancreas.
When an ultrasound contrast agent is injected into a subject, the contrast agent flows into the tumor as white dotted bubbles.
In the case of pancreatic cancer in FIG. 8, before the contrast agent was added (FIG. 8 (a)), 15 seconds after the injection (FIG. 8 (b)), 30 seconds (FIG. 8 (c)), and 45 seconds later (FIG. 8). 8 (d)). As can be seen at first glance, the bubbles do not flow uniformly into the tumor, and there is unevenness between where the bubbles flow and where they do not.

Although there are pancreatic cancer cases in which bubbles hardly enter depending on the case, the typical pancreatic cancer case has unevenness in a pattern as shown in FIG.
When ultrasound image data representing such an ultrasound image is read into the analysis device 2 (step S1), for example, the ultrasound image is displayed on the display device 4 as shown in FIG.
Then, when the region of interest is set according to the user's operation of the region of interest (steps S2 and S3) and further the grid is set according to the user's grid adjustment, for example, as shown on the left side of FIG. A grid is displayed (steps S4 to S6).

Further, when the cell brightness calculation process is performed according to the set grid, for example, a calculation result of the brightness level for each cell as shown in the center of FIG. 10 is obtained (step S8).
Then, based on the calculation result of the luminance level for each cell, a madara index calculation process is performed (step S9). By performing this process for each frame, a madara index for each frame is obtained.
The obtained Madara index for each frame is displayed on the display device 4 in time series as shown in FIG. 11, and the average Madara index is further displayed on the display device 4 as shown in the lower right of FIG. 11 ( Step S12).

For example, in the case of FIG. 11, the average madara index is calculated as about 20.67.
In addition, FIG. 11 represents the madara index for 1 minute after contrast medium injection.
On the other hand, FIG. 12 shows an ultrasound image of a typical chronic pancreatic inflammation example. Also in this case, before contrast agent injection (FIG. 12 (a)), 15 seconds after contrast agent injection (FIG. 12 (b)), 30 seconds (FIG. 12 (c)), and 45 seconds later (FIG. 12 (d)). ) Shows an ultrasonic image.
As shown in FIG. 12A, inflammation is also extracted as a low echo in normal ultrasound. Prior to imaging, a hypoechoic mass is observed in the pancreas. However, this alone makes it difficult to distinguish between tumor (cancer) and inflammation.

When contrast is performed, as shown in FIGS. 12B to 12D, bubbles appearing as white dots uniformly flow into the hypoechoic mass, and unevenness is hardly caused.
The average madara index for one minute after the start of contrast is calculated in the same procedure as in the case of the pancreatic cancer case of FIG. 8, and is “6”, which is different from the case of the pancreatic cancer case shown in FIG. 8 (about 20.67). Showed very low values.
FIG. 13 shows an ultrasound image of a pancreatic cancer case. Although a hypoechoic mass is observed in the pancreas, the contrast does not appear as uneven as in the case of FIG.

If a specialized specialist sees this contrast finding, it may be able to make an accurate diagnosis, but there are cases where it is difficult to differentiate from chronic pancreatitis just by the impression that jumps into the eye, as in this case. When the average madara index is calculated for the pancreatic cancer case of FIG. 13, it is “29”, and when it is actually quantified, it can be seen that it is different from chronic pancreatitis, and it can be easily diagnosed that there is a high possibility of pancreatic cancer.
Therefore, for example, by analyzing past cases of pancreatic cancer and pancreatitis, values suitable for differentiating between pancreatic cancer and pancreatitis are detected in advance for various conditions necessary for the calculation of the Madara index. Set as the initial value of the condition.

  Then, by using the analysis apparatus 2 in which various conditions are set as described above, if an ultrasonic image of a pancreatic cancer or pancreatic inflammation example is analyzed, the transition of the Madara index and the average Madara index can be obtained. As described above, the average madara index of pancreatic cancer tends to be larger than the average madara index of pancreatitis, and the range of values that the average madara index can take is compared between pancreatic cancer and pancreatitis. It tends to be divided. Therefore, refer to changes in mean madara index and madara index even in the case of a practitioner who is relatively unfamiliar with differentiating between pancreatic cancer and pancreatitis, or in cases where it is difficult to differentiate between pancreatic cancer and pancreatitis using only ultrasound images. Can help to differentiate between pancreatic cancer and pancreatitis.

  In some cases, even in the case of pancreatic cancer, the ultrasound contrast agent does not enter the site of the pancreatic cancer and the ultrasound image becomes a blackish image and nonuniformity may not be obtained. In such a case, the average madara index becomes a relatively low value, and if differentiation is performed from the average madara index of general pancreatic cancer or pancreatitis, there is a possibility that appropriate differentiation cannot be performed. Therefore, the conventional TIC may be obtained at the same time, and the result may be displayed on the display device 4. In this way, by displaying the analysis results by different analysis methods for the same ultrasound image, it is a case that is not a general pancreatic cancer or pancreatitis, and it is judged that it is not pancreatic cancer by discrimination by the mean madara index Even if it exists, it will be judged as pancreatic cancer from the result of TIC, and more appropriate discrimination can be aimed at by using both results.

In the above embodiment, a case has been described in which a grid is set so as to include a region of interest, and the region of interest is divided into rectangular cells. However, the present invention is not limited to this. For example, any shape may be used as long as the region of interest can be divided by a plurality of cells having the same shape, such as a triangle and a hexagon, without overlapping each other. The cells may be arranged in a staggered manner.
In addition, the luminance level of one cell is compared with that of the upper, lower, left, and right cells, and when the difference in luminance level is equal to or greater than the threshold value dth, the mottle is determined, but this is not restrictive. For example, the luminance level may be compared with a cell having an oblique relationship with respect to one cell, or the luminance levels may be compared for all cells having an oblique relationship with each other. .

  In addition, the present invention is not limited to the case where the luminance level is compared with one cell up, down, left, and right. For example, an adjacent cell of the same gradation is regarded as one cell, and the adjacent cell of the same gradation is used. It is also possible to compare one cell group consisting of and the upper, lower, left and right cells. In this case, each cell included in the cell group may be compared with one cell group in a vertical and horizontal relationship. For example, when P adjacent cells of the same gradation are made into one cell group, the luminance level of each of the P × 2 cells in the upper and lower relations of the P cells included in the cell group The brightness level of one cell group is compared. Further, the leftmost cell of the P cells included in the cell group and the leftmost cell thereof are compared. Similarly, the rightmost cell and the rightmost cell of the P cells included in the cell group are compared. And the number of which the difference is greater than or equal to the threshold value dth may be determined as the mottled number for this cell group. Or, among cells that have a vertical and horizontal relationship with a single cell group, compare only cells that have the same gray level cell except black and have the same number of cells in the cell group. As a target, the plurality of continuous cells may be regarded as one cell, and the luminance level may be compared between the plurality of cells and one cell group. For example, when adjacent cells of the same gradation excluding P pieces of black are set as one cell group, when P cells are continuously at the same gradation among the upper, lower, left and right cells of this cell group, P cells are compared and P cells that can be regarded as comparison objects are defined as one cell, P cells are compared with one cell group, and the difference between them is equal to or greater than a threshold value dth. A certain number may be determined as a mottled number for this cell group.

Further, the number of P is not necessarily limited to the number of continuous cells of the same gradation excluding black, and can be arbitrarily set, for example, two or three, and the mottled determination target cell is one. Whether to compare a plurality of consecutive cells of the same gradation excluding black as one cell, how many cells are considered as one cell, or one cell to be compared It is also possible to arbitrarily set conditions such as whether or not a plurality of consecutive cells of the same gradation are regarded as one cell and set as one cell for comparison.
In addition, for example, as described above, the mottled determination target cell is assumed to be one, or a plurality of cells having the same gradation except black are regarded as one cell for comparison, or how many cells are equal to one. Whether to consider a single cell, or to compare a single cell, or to consider a plurality of consecutive cells of the same gradation as a single cell, or to be compared with a cell to be judged The display device 4 displays a sentence for inquiring about a condition such as whether the cell is an up / down / left / right cell, a cell having an oblique relationship, or both of the cells, and the user operates the input device 3 or the like. Thus, if these conditions are set, the cells may be compared according to the set conditions.

In other words, the number of madara indices obtained depends on which cell is compared with which cell, and because the region of interest is divided by a rectangle, it is determined to be mottle depending on the shape of the image represented by the ultrasound image. It may be difficult to do. Therefore, it is possible to more accurately determine the luminance non-uniformity by setting the cell comparison method in consideration of the number of grid lines and the number of gradation levels of the luminance level. it can.
In the above embodiment, the case where the medical image processing apparatus 10 performs processing on input ultrasonic image data has been described. However, the present invention is not limited to this. For example, the processing shown in FIG. 2 is incorporated in the ultrasonic image acquisition apparatus 20 side such as an ultrasonic diagnostic apparatus for acquiring the ultrasonic image data itself, and the ultrasonic image data is acquired on the ultrasonic image acquisition apparatus 20 side. The average madara index may be calculated and displayed.

In the above embodiment, the case where processing is performed on an ultrasound image of the pancreas has been described, but the present invention is not limited to this. For example, if it is an examination target site such as liver, breast, prostate, kidney, etc., an ultrasound image is acquired using an ultrasound contrast agent, and this ultrasound image is used to distinguish malignant or benign. Any ultrasonic image data acquired using an ultrasonic contrast agent can be applied.
In that case, there is a possibility that the degree of non-uniformity, variation in luminance, and the like vary depending on the type of the region to be inspected or the measurement method of the ultrasonic image. At that time, various conditions that affect the calculation of the above-described madara index may be set in accordance with the region to be examined.

In the above embodiment, the input device 3 corresponds to the input operation unit, the processing in step S1 in FIG. 2 corresponds to the display processing unit, the processing in steps S2 to S6 corresponds to the classification processing unit, and step S8. The process in step S9 corresponds to the luminance level calculation unit, the process in step S9 corresponds to the non-uniformity degree calculation unit, and the process in step S13 corresponds to the calculation result output unit.
In addition, the scope of the present invention is not limited to the illustrated and described exemplary embodiments, and includes all embodiments that provide the same effects as those intended by the present invention.
Furthermore, the scope of the invention is not limited to the combinations of features of the invention defined by the claims, but may be defined by any desired combination of particular features among all the disclosed features.

Identification was performed using the medical image processing apparatus 10.
The identification of inflammatory masses due to pancreatic cancer and pancreatitis becomes a problem empirically having a size (region) of 2 to 3 cm. Therefore, a tumor (region) having a size of about 2 cm, which tends to cause a problem in differentiation, was used as a case to be differentiated. Specifically, a tumor having a size (region) of at least 10 mm and a maximum of 36 mm, an average diameter of 22.4 mm, and a median of 23 mm was used.
In addition, for such a tumor (region) about 2 cm in size, the number of cells of 10 × 10 and the luminance level of the cells are set to 5 levels, the threshold value dth is set to “2”, and the luminance level is 2 gradations When there was the above difference, it was set so that it was judged mottled.
As a result, it was confirmed that pancreatic cancer and chronic pancreatitis can be distinguished with high diagnostic accuracy from the chronological change of the obtained spotted index Sm and the average spotted index.

DESCRIPTION OF SYMBOLS 1 Ultrasonic image data storage device 2 Analysis device 3 Input device 4 Display device 10 Medical image processing device 20 Ultrasonic image acquisition device

Claims (8)

A display processing unit for displaying an ultrasound image on a display device based on ultrasound image data representing a time-series ultrasound image using an ultrasound contrast agent;
An input operation unit for setting a region of interest for the ultrasound image displayed on the display device;
A classification processing unit that divides the region of interest set in the input operation unit into a plurality of cells;
A luminance level calculation unit that calculates the luminance level of the cell for each frame based on the ultrasonic image data;
One of the cells is a target cell, and the process of determining non-uniformity based on a difference between a luminance level of a cell around the target cell and the luminance level of the target cell among the cells, A non-uniformity degree calculation unit that calculates the degree of non-uniformity for each frame by repeatedly switching the target cell,
A medical image processing apparatus comprising: a calculation result output unit that outputs the degree of non-uniformity for each frame calculated by the non-uniformity degree calculation unit in time series. The non-uniformity degree calculation unit compares the luminance level between each cell adjacent to the target cell and the target cell among cells included in the region of interest, and the luminance level for the target cell. Count cells that are more than a threshold as non-uniform,
The medical image processing apparatus according to claim 1, wherein the sum of the count results of each cell included in the region of interest for each frame is calculated as an index representing the degree of non-uniformity.   The medical image processing apparatus according to claim 2, wherein the calculation result output unit displays the index for each frame in a time series graph.   The medical image processing apparatus according to claim 3, wherein the calculation result output unit calculates an average value of the indices for each frame and outputs the average value. Before Symbol Classification processing unit sets a grid line to include the region of interest, the regions divided in the grid lines and the cell,
The medical image processing apparatus according to any one of claims 1 to 4, wherein the grid lines are set in accordance with an interval between the input grid lines.   6. The medical image processing apparatus according to claim 1, wherein the medical image processing apparatus is used for differentiation of pancreatic cancer.   An ultrasonic diagnostic apparatus comprising the medical image processing apparatus according to any one of claims 1 to 6. A method of processing ultrasound image data representing a time-series ultrasound image using an ultrasound contrast agent,
Displaying an ultrasound image on a display device based on the ultrasound image data;
Dividing the region of interest set by the input operation unit for the ultrasonic image displayed on the display device into a plurality of cells;
Calculating a luminance level of the cell for each frame based on the ultrasonic image data;
One of the cells is a target cell, and the process of determining non-uniformity based on a difference between a luminance level of a cell around the target cell and the luminance level of the target cell among the cells, Calculating the degree of non-uniformity for each frame by switching the target cell repeatedly,
And a step of outputting the degree of non-uniformity for each frame in time series.

Images