JP4907798B2 - Ultrasonic diagnostic equipment - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an ultrasound diagnostic apparatus that obtains an ultrasound image in a subject based on an echo signal due to ultrasound applied to the subject, and more particularly, from an intensity distribution of the echo signal in a living organ in the subject. The present invention relates to an ultrasonic diagnostic apparatus having a function of extracting a minute structure.
[0002]
[Prior art]
There are a wide variety of medical applications of ultrasonic signals, and an ultrasonic diagnostic apparatus is one of them.
[0003]
The mainstream of this ultrasonic diagnostic apparatus is a type that obtains a tomographic image of a soft tissue of a living body using an ultrasonic pulse reflection method. This imaging method can obtain a tomographic image of a tissue non-invasively and can be displayed in real time compared to other medical modalities such as an X-ray diagnostic apparatus, an X-ray CT scanner, an MRI apparatus, and a nuclear medicine diagnostic apparatus. The apparatus is small and relatively inexpensive, has no exposure to X-rays, and has many advantages such as blood flow imaging based on the ultrasonic Doppler method. For this reason, it is widely used in the diagnosis of the circulatory organ (heart), abdomen (liver, kidney, etc.), mammary gland, thyroid gland, urology, and gynecology. In particular, it is possible to observe heart pulsations and fetal movements in real time with a simple operation by simply placing an ultrasonic probe on the body surface, and there are no worries about X-ray exposure. Advantages such as being able to easily perform inspection by moving the ultrasonic diagnostic apparatus to the bedside are preferred.
[0004]
Also, currently used ultrasonic diagnostic apparatuses usually have various measurement functions. The term “measurement” used herein refers to quantifying a physical event in a subject, and the measurement result is converted into a numerical value itself and / or converted into an amount such as color or luminance corresponding to the numerical value and presented. The
[0005]
The main measurement functions installed in conventional ultrasonic diagnostic apparatuses are listed below.
1. Shape measurement: With this shape measurement function, for example, the size of a liver tumor, the wall thickness of a myocardium, the size of a fetus, and the like are measured.
2. Velocity measurement: This velocity measurement function includes, for example, arterial blood flow velocity using the Doppler method, and blood flow velocity mapping of the intravascular liver using the color Doppler method.
3. Measurement of volume, flow rate, etc .: With this measurement function, for example, the volume of the heart cavity can be estimated based on several lengths in the heart chamber, and the blood flow can be measured from the change over time in the signal strength of the contrast agent. Done.
[0006]
The measurement values obtained by such measurement are naturally useful information for evaluating the severity of the disease. For example, information such as tumor size and the degree of regurgitation in blood vessels immediately indicates the degree of need for treatment.
[0007]
On the other hand, there is a lot of measurement information that is indirectly useful for diagnosing the health condition of a subject even if it is not for directly evaluating a disease. Rather, this is more common for everyday measurements. For example, the subject's height, weight, blood pressure, or various numerical values obtained by blood tests fall into this category.
[0008]
Furthermore, as a matter that is different from such various measurement functions, there is a diagnosis close to quantification based on empirical judgment of a doctor. This valuable diagnosis can be found throughout the medical field. For example, one such diagnosis is a diagnosis of liver cirrhosis.
[0009]
Cirrhosis means that fibrotic tissue increases in the liver due to repeated destruction and regeneration of hepatocytes, the number of hepatocytes gradually decreases, and the liver becomes hard and contracted. At the initial stage of cirrhosis, there is no subjective symptom of the patient, and it is difficult to visually recognize the minute fiberized structure in the ultrasonic diagnostic image. However, as the degree of cirrhosis increases, the speckle pattern non-uniformity in the liver parenchyma can be visually recognized. Therefore, in medical practice, this non-uniformity is visually observed and used as a standard for judging the degree of cirrhosis. .
[0010]
The speckle patterns that appear in this ultrasound diagnostic image are the parts where the echo signal intensity is high and low due to countless superposition of scattered waves when countless scatterers are distributed with fineness below the resolution of the ultrasound. This is a phenomenon that occurs. This is a physical phenomenon close to so-called interference fringes, and it is well known that the pattern itself does not directly reflect the structure of the organ. In the observation of cirrhosis, the speckle pattern does not directly reflect the structure of the fiberized tissue. However, as the severity of cirrhosis increases, this speckle pattern exhibits a characteristic visual pattern, which is used for diagnosis.
[0011]
For example, FIGS. 15A and 15B schematically show a tomographic image of the liver that is referred to when observing the above cirrhosis. FIG. 6A is a tomographic image of a normal person having no abnormality in the liver, and a pattern called a speckle pattern of the liver appears relatively uniform. On the other hand, FIG. 7B schematically shows a tomogram of an abnormal liver having a disease, and the speckle pattern becomes non-uniform compared to the image of FIG. It can be confirmed.
[0012]
Therefore, when an ultrasound diagnostic image of the liver is presented, the “uniformity” of the speckle pattern is visually observed. If the unevenness is strong, it is diagnosed that there is a suspicion of an abnormal liver with cirrhosis. That is why.
[0013]
However, until now, there has been no case where the “speckle pattern non-uniformity” in this example, that is, the “abnormality” has been quantified, and the diagnosis has been based solely on the empirical judgment of the doctor.
[0014]
And, in recent years, the objective is to elucidate objectively and scientifically the question of what kind of human recognition patterns are used for diagnosis based on the above-mentioned doctor's empirical judgment. Research has been done.
For example:
1, Yamaguchi T, Hachiya H, “Modeling of the Circulatory Liver Consendering the Liver Robot Structure”, Jpn, J. et al. App; Phys. Vol. 38 (1999) pp. 38. 3382-3392;
2, Otsuka, Yamaguchi, Hachiya: “Study by simulation of ultrasonic B-mode image of lesioned liver”, Shingaku Giho, US96-16 (1996-06), pp. 15-22:
3. Tsuneo Kikuchi, Toshihiro Nakazawa et al., “Development of quantitative diagnosis technique for inter-disorders based on echo waveform spectrum shape of ultrasonic diagnostic apparatus”, Nichio Medical Basic Technology Study Group, BT-2000-31, pp. 9-15 (2001);
There are papers such as.
[0015]
According to these documents, the speckle pattern of the tomographic image of the liver changes with the progress of cirrhosis (see FIGS. 15 (a) and 15 (b)). As the fibrotic tissue increases in size as it progresses, it is recognized as a structure for the ultrasonic pulse, and the structural information gradually appears and increases in the speckle pattern. Therefore, it is considered that the aspect gradually changes with this.
[0016]
In the prior art disclosed so far, several attempts have been made to quantify the degree of progression of cirrhosis. For example, in Japanese Patent Application No. 2000-0542201, there is an invention called “ultrasonic diagnostic apparatus and quantitative analysis method of normal tissue using ultrasonic waves”.
[0017]
The invention described in Japanese Patent Application No. 2000-054201 is based on the statistical properties of speckle patterns as described below.
[0018]
A curve 51 in FIG. 16A shows a probability density distribution of luminance values of echo signals reflected from a normal liver. From a stochastic / statistical point of view, if the scatterers are randomly distributed, the probability density distribution P (x) of the amplitude value which is the intensity of the echo signal reflected from the scatterers is P ( x) = (x / σ²) Exp (-x²/ 2σ²). Where σ²Represents the variance and is normalized to an average of 0.
[0019]
When the liver is normal, it can be assumed that many scatterers (except for obvious structures such as blood vessels) are present in the liver at random, so the probability density of echo signal intensity (amplitude) representing the liver The function exhibits a Rayleigh distribution as shown by a curve 51 shown in FIG. However, as the above-mentioned fiberized structure increases in the liver, the speckle pattern reflects the structure and cannot be said to be random. As a result, the probability density function of luminance deviates from the Rayleigh distribution as shown by a curve 52 in FIG.
[0020]
As described above, it is possible to determine whether the liver is normal or abnormal by observing the outline of the probability density distribution curve of the echo signal intensity. That is, an error between the probability density distribution obtained by actual measurement and the Rayleigh distribution as a theoretical value is used as a criterion for the evaluation.
[0021]
However, the resolution of the ultrasonic diagnostic image is determined by the transmission frequency, transmission wave number, transmission aperture, etc., and the initial fibrosis structure in liver cirrhosis as described above, or a minute lesion existing in the tissue (general The lesions close to or less than the limit of resolution at the time of diagnosis are buried in the speckle pattern or cannot be seen, or have been visualized in a state that is difficult to distinguish from the speckle pattern. Ideally, liver cirrhosis can be diagnosed based on images at an early stage when the patient's subjective symptoms do not appear so much. It was very difficult to find and further quantify the microstructure in it.
[0022]
[Problems to be solved by the invention]
The present invention has been made in view of the above circumstances, and an object of the present invention is to smooth the image of the speckle part and extract the microstructure using the statistical properties of the speckle pattern. Thus, an object of the present invention is to provide an ultrasonic diagnostic apparatus including an analysis algorithm capable of observing minute abnormal lesions in a homogeneous tissue structure including the degree of progression of cirrhosis.
[0023]
  In order to solve the above-mentioned problem, the invention according to claim 1 is the first signal of the received signal based on the subject region in the ultrasonic diagnostic apparatus for obtaining a tomographic image by irradiating the subject with an ultrasonic pulse.Intensity or amplitudeValue and second signalIntensity or amplitudeValue similarityBy testDetermination means for determining, and the first and second signalsIntensity or amplitudeIf the value similarity is high, the second signalIntensity or amplitudeA large weighting factor for the valueMultiplicationIf the similarity is low, the second signalIntensity or amplitudeA small weighting factor for the valueMultiplicationAnd the first and second signals based on the set weighting factor.Intensity or amplitudeAn averaging means for weighted averaging of values and a display means for displaying an ultrasonic image generated based on the weighted average value are provided.
[0024]
  In order to solve the above-mentioned problem, the invention according to claim 2 is the first reception based on the first subject region in the ultrasonic diagnostic apparatus that obtains the ultrasonic image by irradiating the subject with the ultrasonic pulse. signalIntensity or amplitudeValue and the second received signal based on the second subject regionWith intensity or amplitude valueSimilarity based on statistical propertiesBy testA determination means for determining, and the first received signal based on the similarityIntensity or amplitudeValue and second received signalIntensity or amplitudeCombining means for obtaining a third received signal by combining the values, and the third receptionsignalAnd a display means for generating an ultrasonic image based on the above.
[0025]
  In order to solve the above-described problem, the invention according to claim 3 is the first signal of the received signal based on the subject region in the ultrasonic diagnostic apparatus that obtains an ultrasonic image by irradiating the subject with an ultrasonic pulse.Intensity or amplitudeValue and second signalIntensity or amplitudeValue similarityBy testDetermination means for determining and the first signalIntensity or amplitudeValue and second signalIntensity or amplitudeAn averaging means for weighted averaging of values, and a display means for displaying an ultrasonic image generated based on the averaged values, wherein the averaging means includes the first signal.Intensity or amplitudeValue and second signalIntensity or amplitudeIf the value similarity is high, the second signalIntensity or amplitudeMultiplying the value by a large weighting factor, the first signalIntensity or amplitudeValue and second signalIntensity or amplitudeIf the similarity of the values is low, the second signalIntensity or amplitudeMultiplying the value by a small weighting factor and weighting the first and second signalsIntensity or amplitudeIt is characterized by a weighted average of values.
[0026]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram illustrating a control configuration of the ultrasonic diagnostic apparatus according to the present embodiment.
[0027]
Although the present invention can be applied to various diagnostic apparatuses, this embodiment will be described with respect to an ultrasonic diagnostic apparatus. In addition, the diagnosis site can be applied to the liver, pancreas, myocardium, etc., which have a relatively homogeneous tissue structure under normal conditions. In this example, the case of diagnosing the severity of liver cirrhosis will be described.
[0028]
[Description of structure]
As shown in FIG. 1, the ultrasonic diagnostic apparatus according to this embodiment includes an ultrasonic probe 12 that is responsible for transmitting and receiving ultrasonic signals to and from a subject, and driving the ultrasonic probe 12 and receiving the ultrasonic probe 12. The apparatus main body 11 which processes a signal, the input device 13 which is connected to this apparatus main body 11 and can input the instruction information from an operator into the apparatus main body 11 and the monitor 14 are provided. The input device 13 includes buttons, a keyboard, a trackball, and the like that can control the diagnostic device and set various image quality conditions.
[0029]
The apparatus main body 11 includes an ultrasonic transmission unit 21, an ultrasonic consultation unit 22, a B-mode processing unit 23, a Doppler processing unit 24, an image generation circuit 25, a signal analysis unit 26 (main in the present invention), a control processor (CPU). 27, a storage medium 28, and other interfaces 29. These may be configured by hardware such as an integrated circuit, but may be software programs modularized in software.
[0030]
The ultrasonic transmission unit 21 includes transmission circuits such as a delay circuit and a pulsar circuit (not shown), and the ultrasonic diagnosis unit 22 includes reception circuits such as an A / D converter and an adder. A sound wave is generated and sent to the vibration element of the probe 12, and an echo signal scattered by the tissue in the subject is received again by the probe 12 to obtain a received signal.
[0031]
The output from the ultrasound consultation unit 22 is sent to the B mode processing unit 23. Here, logarithmic amplification of the echo signal, envelope detection processing, and the like are performed, and the signal intensity is data expressed by brightness. The Doppler processing unit 24 performs frequency analysis of velocity information from the echo signal and sends the analysis result to the image generation circuit 25.
[0032]
In the image generation circuit 25, a scanning line signal sequence of ultrasonic scanning is converted into a scanning line signal sequence of a general video format typified by a television or the like, and combined with character information and scales of various setting parameters. And output to the monitor 14 as a video signal. Thus, a tomographic image representing the subject tissue shape is displayed on the monitor 14. Further, the image generation circuit 25 is equipped with a storage memory for storing image data and can be called by an operator after diagnosis, for example.
[0033]
The control processor 27 has a function as an information processing apparatus (computer) and is a control means for controlling the operation of the main body of the ultrasonic diagnostic apparatus. Also in the signal analysis of the present invention, a command to transfer necessary programs and data from the storage medium 28 to the signal analysis unit 26 is sent.
[0034]
The storage medium 28 stores the above-described various analysis software programs in addition to storing the diagnostic image (details will be described later).
[0035]
The signal analysis unit 26 reads an output signal immediately after the ultrasonic reception unit 22 (referred to as a radio frequency (RF) signal) or an image luminance signal after passing through the B-mode processing unit 23, and performs analysis processing of the present invention described later. The result is displayed on the display unit via the image processing unit 25, stored in the storage medium 28, or transferred to an external PC or printer via the network interface 29.
[0036]
[Description of analysis method]
Next, an analysis method in the signal analysis unit 26 will be described with reference to FIG.
[0037]
First, an echo signal to be analyzed is selected by the operator. Since this signal is preferably an echo signal obtained from the liver parenchyma even when an RF signal is used, a region 41 (hereinafter referred to as ROI 41) is designated in the image as shown in FIG. Thus, spatially corresponding echo data is taken into the signal analysis unit 26. In this example, the shape of the ROI 41 is a square, but a circle (ellipse) or a free closed curve can also be specified. It is also possible to specify a plurality of ROIs 41.
[0038]
Next, the following processing is performed on the signal in the ROI 41. Here, it is assumed that the RF signal or the image luminance signal in the ROI 41 is numbered in a two-dimensional array spatially corresponding to the diagnostic image. For example, as shown in FIG. 3, it is assumed that there are Nx × Ny data in the ROI 41, Nx data in the x direction and Ny data in the y direction. Further, the coordinates of the point P in the ROI 41 are (x, y) (where 1 ≦ x ≦ Nx, 1 ≦ y ≦ Ny).
[0039]
The arithmetic processing described below is performed on each point P (x, y) in the ROI 41.
[0040]
First, as shown in FIG. 4, a region near P (x, y) is secured. This neighboring region is more ideal when it is a circle centered on the point P. However, for simplicity, it is ± a in the x direction and ± b in the y direction as shown in FIG. Let's think in a square area. (Note: The neighboring region in FIG. 4 is included in the ROI 41 in FIG. 3, and is generally sufficiently smaller than the ROI 41 shown in FIG. 3.) Here, a point in this neighboring region is represented by Q (i , J) and signal strength values at points P and Q are Ip and Iq, respectively.
[0041]
Next, smoothing processing is performed on the above-mentioned neighboring area from the viewpoint of “similarity”. Hereinafter, the smoothing process will be described in general, and then the smoothing process according to the present invention will be described in detail.
[0042]
<A: General explanation>
The smoothing process is a process for obtaining a so-called “blurring” effect by somewhat weighting the value (value) of a nearby point to the value of a certain point. In the conventional general smoothing process, weighting correlated with the distance between two points is often performed (that is, the weighting coefficient of a near point is large and the weighting number of a far point is small).
[0043]
<B: Specific Explanation of Smoothing Processing According to the Present Invention>
On the other hand, the weighting coefficient according to the method of the present invention is determined from the viewpoint of “similarity” whether the two points are statistically similar regardless of the distance between the two points. When the similarity of the nearby point Q is high with respect to the point P, for example, the value of the point Q is weighted to the point P by a coefficient close to 1, for example, and when the similarity is low, for example, a coefficient close to 0 , The value of the point Q is weighted to the point P. The process that is smoothed according to the similarity is called “coherent filter” process.
[0044]
Hereinafter, an example of coherent filter processing will be described.
[0045]
First, the following evaluation function W is defined:
W = D− | Iq−Ip | / σ (1)
Here, σ is a standard deviation obtained from the probability density distribution of the echo signal intensity in the ROI 41, and D is a threshold set separately (see FIG. 5). If W <0 (that is, if the second term on the right side is greater than the threshold value D), the point Q is determined to be “not similar” to the point P and is excluded from the weighting target (this is This corresponds to the case where the rejection area is selected as D in the statistical test method). If W> 0, the amplitude value Iq at the point Q is weighted to Ip. However, the weighting coefficient Cw (i, j) at that time is a function of the following intensity difference:
Cw (i, j) = [1-{(Iq-Ip) / σD}²]².
[0046]
This weighting factor is obtained for all points in the neighborhood and added to the point P to obtain the value Ip ′ of the point P after the calculation:
Ip ′ = Ip + {Σ (Cw (i, j) × Iq (i, j))} / Ctot
Ctot is the total amount of Cw.
[0047]
The rejection area D can be specified and changed by the operator. Needless to say, the optimum conditions are stored in advance in the ultrasonic diagnostic apparatus. As described above, since the amplitude of the echo signal has a statistical property such that it follows a Rayleigh distribution, this rejection area D is determined from the probability density function of the Rayleigh distribution.
[0048]
The result of the image processing obtained by this method is shown in FIG. However, (A) is an original image, and the target is a normal liver, but a structure such as a boundary in the upper part of the liver and a relatively large blood vessel cross section in the liver is confirmed. And (B) is the image after a calculation. Here, as the standard deviation σ, a value calculated based on a statistical amount of a relatively homogeneous portion in the liver parenchyma in this data was used.
[0049]
In general, when a simple smoothing process is performed, so-called “edge blur” occurs, the spatial resolution of the image is impaired, and the entire image is blurred. However, when looking at the result (B) obtained by this method, since the luminance of the substantial part is a speckle pattern according to the Rayleigh distribution, similarities are recognized, and as a result, a very large smoothing process is performed. Yes. On the other hand, the structure of the liver boundary wall and the blood vessel wall does not follow the statistical distribution of the liver parenchyma, and therefore is not subjected to smoothing processing and is depicted as a structure as it is. As described above, the image processing result obtained by this method is characterized in that the boundary of the structure is very steep compared to the normal smoothing process.
[0050]
Next, the procedure flow of the operator will be described with reference to FIG.
[0051]
First, the operator scans the liver of the subject and selects a cross section to be analyzed (S71). Next, the ROI to be analyzed is designated (S72). Next, a standard deviation σ necessary for the evaluation function is obtained. At this time, a method of separately specifying the ROI for obtaining the standard deviation σ by the operator (S73), and calculating the standard deviation σ in advance from the echo signal of the entire image at that time and from the data in the ROI to be analyzed, It is possible to select the method (S74) by which this value is called. Next, actual analysis is performed (S75), and the analysis result is displayed on the display unit (S76). At this time, the threshold value D takes an arbitrary value, but by changing this threshold value D, it is not similar to the speckle pattern, that is, the degree of recognition as a structure changes. The value of the threshold value D is changed according to the analysis result, and recalculation is performed (S77). The analysis ends when a desired image is obtained as a result of repeated calculations (S78).
[0052]
<Various ideas for improving accuracy ... 1>
Next, a first method for improving the accuracy of the analysis calculation will be described.
[0053]
When performing this analysis, the operator inputs to the system that the analysis is started, so that the ultrasonic diagnostic apparatus of the present invention changes to dedicated transmission / reception conditions. This is to achieve the following objectives:
[1] Improvement in analysis accuracy by increasing the number of samples of acquired data: Since analysis uses statistical properties, it is better that the number of samples of echo data is large. However, simply increasing the number of transmissions / receptions only takes an echo signal having the same information, and does not increase the substantial amount of information. In order to achieve this object, the scanning line density for transmission / reception is larger than that during normal diagnosis, for example, twice or four times. Alternatively, ultrasonic transmission / reception is performed a plurality of times on the same scanning line under transmission conditions having different frequencies.
[0054]
Since the above processing leads to a decrease in real-time observation capability due to a decrease in frame rate, until the time immediately before analysis, the system operates under normal scanning line conditions, and this transmission / reception is performed at the timing of analysis start (such as an instruction from the operator) It changes to the conditions.
[0055]
[2] Improvement in S / N ratio in the high frequency band: It can be said that transmission and reception in the high frequency band has higher resolution and more spatial information because of the basic properties of ultrasonic waves. On the other hand, the sound wave is greatly attenuated at a high frequency, and there arises a problem that the received signal cannot be acquired up to the deep region. In order to solve this problem, the same number of transmissions / receptions on the same scanning line is increased, and the averaging process is performed at the RF signal level. For example, when the averaging process is performed twice on the same received RF data, the random noise is reduced and the steady echo signal amplitude level is increased by about 6 dB. In this method, as in the above [1], an increase in scanning line density leads to a decrease in real-time observation capability due to a decrease in frame rate. Therefore, until just before the analysis, the system operates under normal scanning line conditions, and the analysis start timing. The transmission / reception conditions are changed according to instructions from the operator.
[0056]
[3] Avoiding signal saturation: If there is a medium having a large scattering coefficient, the received signal may be saturated. Further, if the operator makes a mistake in the gain setting on the apparatus, the reception signal saturation similarly occurs. When the signal is saturated, there is a risk that the statistics of the signal change and an incorrect analysis result is presented. In this system, the signal analysis unit 26 shown in FIG. 1 monitors the signal level of the ultrasonic reception unit 22 based on the information on the maximum value that the received signal can take, and a signal that reaches the maximum value is generated ( Alternatively, if a value close to that occurs, the analysis is stopped and a message prompting the operator to remeasure is displayed.
[0057]
[4] Analysis for a plurality of frames The signal analysis unit 26 is provided with a memory for storing RF data for a plurality of frames. As shown in FIG. In consideration of the above, it is also possible to perform the above-described analysis processing (coherent filter processing) on a three-dimensional neighborhood. In this case, in order to increase the amount of signal information in the z direction, it is desirable to change the scan surface in the living body over time by moving the probe. However, in our study, the probe is not moved intentionally. In both cases, it is confirmed that the information of the echo signal changes with time due to the minute movement of the operator or the heartbeat and breathing of the subject, and the analysis accuracy of the present technique is improved by the analysis by the plurality of frames.
[0058]
<Other functions>
[5] Display / non-display of analysis area: The analysis result image is intended to be used as a new information for diagnosis by extracting the feature quantity of the structure from the speckle pattern. There is also a situation where it is desired to reconfirm the echo image. The information of the diagnostic image before analysis is stored by the signal analysis unit 26 of FIG. 1 or once recorded in the storage medium 28, and is retained by the operator using the input device even after analysis. It is possible to display on the monitor 14 by the call command. As a display form at that time, it is possible to perform a display such as a parallel display, a superimposed display, or a display that is alternately switched by a button input.
[0059]
[6] Expansion of analysis area: In the ultrasonic diagnostic apparatus of the present invention, the ROI portion designated as the analysis area can be enlarged and displayed before analysis, during analysis, or after analysis. In general, the information of the diagnostic image processed by the image generation circuit is larger than the number of pixels displayed as a television format on the display unit. Therefore, the enlarged display in this case is different from simply enlarging a photograph or the like, resulting in an increase in the amount of information displayed in the ROI itself. Further, as described above, in the transmission / reception control method (idea [1]) according to the present invention, the scanning line density is denser than that of the conventional method. It will be advantageous.
[0060]
<Various ideas for improving accuracy ... 2>
Next, a second method for improving the accuracy of the analysis calculation will be described.
[0061]
When attempting to quantify tissue properties from a diagnostic image, it is common to analyze the local region of the image as a “sample”. This is because the diagnostic image includes blood vessels, organ boundaries, gallbladder, and the like in addition to the tissue region, and analysis with these included results in increasing the error.
[0062]
As described above, a statistical method called “test” is well known as a method for estimating a population from a limited part (sample) of an image. This is a method of determining whether to reject this hypothesis by making one hypothesis on the property of the population and examining the property of the sample.
[0063]
Taking liver tissue diagnosis as an example, it is as follows.
[0064]
First, a hypothesis is made that the liver (population) is normal. A set of amplitudes of echo signals obtained from a normal liver is known to follow a Rayleigh distribution as described above. Therefore, it is a technique called a test to determine whether or not this hypothesis applies to a sample that has been taken out.
[0065]
In general, χ²There are various methods such as a test, a t-test, and an F-test, but since the method itself is already widely known, the details thereof are omitted here.
[0066]
However, when trying to use these tests for the normal tissue diagnosis, the following problems arise.
[0067]
That is, in the normal tissue diagnosis, the statistic of “normal tissue” corresponding to the population cannot be obtained directly. This statistic means “average value or variance when patients with cirrhosis are normal liver”, and it is not possible to obtain this value from patients with cirrhosis (patients with suspected cirrhosis). Is possible. In addition, this value does not make sense with an average value or a variance value obtained from an echo signal of another subject (human body) having a normal liver. This is because these statistics vary depending on the ultrasonic irradiation sound pressure and gain setting. This problem cannot be solved even if the setting of the diagnostic apparatus is the same because the value changes due to a difference in biological attenuation of the subject.
[0068]
Therefore, in the present invention, a local region suitable as a statistic of “normal tissue” is extracted from the subject currently being diagnosed, and this is extracted as a statistic of the population (variance value σ², Standard deviation σ).
[0069]
Hereinafter, this method will be described.
[0070]
FIG. 9 shows a variance value σ calculated from a spatial arrangement similar to a diagnostic image (see FIG. 8) based on the echo signal of the liver diagnosed as cirrhosis.²It is the analysis result of the said liver which shows. A method for analyzing the curves A and B in the figure will be described below.
[0071]
A: First, as shown in FIG. 10, a small region R01 of a certain size for taking a sample is set, and the average value μ and variance σ are shifted while shifting the position little by little.₁ ²And this variance value σ₁ ²Is a curve A shown in FIG.
B: Next, the above average value μ is substituted into the following equation, and the variance σ₂ ²The curve B is obtained.
σ₂ ²= (4 / π-1) μ (1)
The above equation (1) is valid only when the assumption that the probability density distribution of the sample “follows the Rayleigh distribution” holds. Therefore, if the sample has a non-Rayleigh distribution, this equation (1) does not hold.
[0072]
As is clear from FIG. 9, the values in the section (1) in FIG. From this, in the section (1), the variance σ calculated using the formula (1)₂ ²And the actual dispersion value σ₁ ²Can be predicted to be nearly identical. Therefore, in this section (1), it can be determined that the probability density distribution of the sample substantially follows the Rayleigh distribution.
[0073]
On the other hand, in the range of the section (2), both values are greatly different. This indicates that Expression (1) does not hold. Therefore, in this section (2), it can be determined that there is a strong possibility that the probability density distribution of the sample is a non-Rayleigh distribution.
[0074]
What is important here is the fact that a range such as section (1) is found even locally. Thus, even within a tissue region diagnosed as cirrhosis, it is possible to find a small region according to the Rayleigh distribution by searching while changing the position of the sample.
[0075]
In the present invention, a sample having a variance value similar to the Rayleigh distribution is searched by the above-described method (see FIG. 11), and this variance value is used as the variance σ of the population.₀ ²And the variance value in the analysis region in the vicinity is σ₁ ²Then, test from both values. Thus, it is a feature of the present invention that “pseudo population distribution” is obtained by searching.
[0076]
As described above, there are several types of test methods themselves, and since these are already widely used, the description thereof will be omitted here. Further, the “test” here is a judgment method for judging “whether or not the sample shows the Rayleigh distribution”, if interpreted broadly, and it is not necessary to use a particularly strict test method. For example, the above σ₁ ²And σ₂ ²If the ratio is 2 or more, a method of rejecting may be used.
[0077]
In any case, if the hypothesis is rejected as a result of the test, the region is judged to be “non-Rayleigh”.
[0078]
<Application to CFAR>
The method of the present invention can also be applied to so-called Constant False Alarm Rate processing (CFAR processing). This CFAR is a technique well known in the radar technology whose principle is similar to that of an ultrasonic diagnostic apparatus.
[0079]
The principle will be briefly described below.
[0080]
FIG. 12 shows an example of a signal displayed on the radar (the CFAR processing performed on the video luminance signal is called LOG / CFAR processing, but is simply called CFAR processing here). In this example, three signals typified by airplanes are included in scatterers typified by clouds.
[0081]
In this example, the point to be extracted is clear, but when it is displayed as luminance information, the extract may be difficult to see due to the influence of the scatterers. Therefore, in order to remove the scattered matter, so-called gain adjustment that does not display a threshold value or less is performed.
[0082]
However, in the case of this example, if T1 shown in the figure is used as a threshold value, the point C is not extracted. Further, when T2 shown in the figure is a threshold value, a point C is displayed, but a scattered object near the point A is visually recognized instead. Therefore, in such a case, further CFAR processing is performed.
[0083]
The CFAR processing is processing for subtracting an average value of signals in the vicinity of a certain point X excluding the point from the point X, and then setting a display threshold again.
[0084]
When this CFAR process is performed on the signal displayed on the radar shown in FIG. 12, the result shown in FIG. 13 is obtained. As can be seen from the figure, by applying this CFAR process, the entire slope of the scatterer is removed, and the threshold value T3 for extracting the points A to C can be easily set.
[0085]
What has been described above is so-called CFAR processing.
[0086]
There is already a paper report that the lesion of the liver could be extracted relatively well using this CFAR, but the CFAR can be performed well when there are sparse points to be extracted. It is understood that if dots are densely present, that is, if advanced cirrhosis is assumed, it will not work theoretically. This is because the “average value in the vicinity” includes not only scatterers but also adjacent extraction points, so that it no longer shows non-Rayleigh properties (see FIG. 14). Therefore, in such a case, the subtraction result is underestimated.
[0087]
Therefore, in such a case, by using this method to search for a region exhibiting a nearby Rayleigh distribution and using the average value obtained therefrom, the same degree of accuracy as in the case where the aforementioned extract is sparse can be obtained. Can keep.
[0088]
<Presentation method of judgment results>
There is a high possibility that the portion rejected by the above judgment method is not a normal tissue. Therefore, in this example, the tomographic image of this portion is displayed, and the portion that has not been rejected is displayed in black with a luminance value of 0, for example. By such a technique, it is possible to highlight and display a portion that may be a disease site.
[0089]
As a modified example, for example, an emphasis display method in which the rejected area is displayed in red or the like in a B-mode monochrome luminance image can be considered.
[0090]
Further, if the original B-mode tomographic image can be observed even after the above-described disease highlight display image is obtained, it is convenient for confirming the analysis result and confirming the original tissue properties.
[0091]
Therefore, in the present invention, the highlighted display image of the analysis result and the original B-mode tomographic image are configured to be switchable and displayed by an operator's instruction (for example, button operation). In addition, it is possible to display simultaneously on one screen in parallel.
[0092]
<Display statistics>
In the present invention, the variance σ of the population obtained by the above method₀ ²Or the variance σ of the region of interest₁ ²Alternatively, the square root (standard deviation) and the average value μ thereof can be displayed on the display unit.
[0093]
<Display of probability density curve>
In the present invention, it is assumed that the sample data of each region such as the average value and the variance value obtained by the above-described analysis is displayed as a probability density distribution as shown in FIG. Furthermore, in the present invention, it is also possible to display a probability density curve for one point or local region on the image designated by the operator after analysis on a separate screen.
[0094]
【The invention's effect】
  As described above, the ultrasonic diagnostic apparatus according to the present invention.In placeTherefore, at the time of ultrasonic diagnosis, the existence of structures close to the limit of the resolution of ultrasonic pulses, which is difficult to distinguish from speckle patterns by visual inspection, is extracted using statistical properties and is easy to visually recognize. By generating an image, it is possible to more easily diagnose the severity of cirrhosis.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a control configuration of an ultrasonic diagnostic apparatus according to the present invention.
FIG. 2 is an explanatory diagram for designating an analysis region in the ultrasonic diagnostic apparatus according to the present invention.
FIG. 3 is a conceptual diagram showing the arrangement of data in the analysis region in the ultrasonic diagnostic apparatus according to the present invention.
FIG. 4 is a diagram showing a region used for smoothing processing for one point in the analysis region in the ultrasonic diagnostic apparatus according to the present invention.
FIG. 5 is a conceptual diagram for explaining the relationship between signal strengths at two points on which smoothing processing is performed.
FIG. 6 is an example of an image obtained by the ultrasonic diagnostic apparatus according to the present invention and its analysis technique.
FIG. 7 is a flowchart showing an operator's procedure when performing analysis according to the present invention.
FIG. 8 is a conceptual diagram when smoothing is performed using a plurality of frames for one point in the analysis region of the present invention.
9 is a variance value σ calculated from a spatial arrangement similar to that of the diagnostic image shown in FIG. 8 based on the echo signal of the liver diagnosed as cirrhosis.²It is the analysis result of the said liver which shows.
FIG. 10 is an explanatory diagram for explaining a sample acquisition method for drawing the curve A shown in FIG. 9;
11 is a conceptual diagram showing a process of searching for a sample having a variance value similar to the Rayleigh distribution by the method shown in FIG.
FIG. 12 is an explanatory diagram illustrating a signal displayed on a radar as an example;
13 is a diagram illustrating a result of performing CFAR processing on the explanatory diagram illustrated in FIG. 12; FIG.
FIG. 14 is a diagram showing a difference between an average value of a conventional method and an average value after searching for a Rayleigh part when points to be extracted are densely present, that is, when advanced cirrhosis is assumed.
FIG. 15 is a schematic diagram showing a difference in appearance of diagnostic images of normal liver and cirrhosis liver.
FIG. 16 is a schematic diagram showing a difference in probability density distribution of signal intensity between normal liver and cirrhosis liver.
[Explanation of symbols]
11 ... Device body
12 ... Ultrasonic probe
13 ... Input device
14 ... Monitor
21 ... Ultrasonic transmission unit
22 ... Ultrasonic wave receiving unit
23 ... B-mode processing unit
24 ... Doppler processing unit
25. Image reproduction circuit image memory
26 ... Signal analysis unit
27. Control processor (CPU)
28 ... Storage medium
29 ... Other interfaces
41 ... area
51 ... Probability density distribution of luminance value of echo signal reflected from normal liver
52 ... Luminance value of echo signal reflected from liver with increased fiber structure
Probability density distribution
P ... Subject
T1 ... Threshold
T2 ... Threshold
T3: Threshold value for extracting points A to C

Claims (10)

In an ultrasonic diagnostic apparatus that obtains a tomographic image by irradiating a subject with an ultrasonic pulse,
A determination means for determining the similarity between the intensity or amplitude value of the first signal and the intensity or amplitude value of the second signal of the received signal based on the subject site by testing ;
If the degree of similarity is high intensity or amplitude value of the first and second signals are multiplied by a large weighting factor to the strength or amplitude value of the second signal, wherein if the similarity is low of the second signal Averaging means for multiplying the intensity or amplitude value by a small weighting factor, and weighted averaging the intensity or amplitude values of the first and second signals based on the set weighting factor;
Display means for displaying an ultrasonic image generated based on the weighted average value;
An ultrasonic diagnostic apparatus comprising: In an ultrasonic diagnostic apparatus that obtains an ultrasonic image by irradiating a subject with an ultrasonic pulse,
The similarity between the intensity or amplitude value of the first received signal based on the first subject part and the strength or amplitude value of the second received signal based on the second subject part is tested based on statistical properties Judging means for judging by :
Combining means for combining the intensity or amplitude value of the first received signal and the intensity or amplitude value of the second received signal based on the similarity to obtain a third received signal;
Display means for generating an ultrasonic image based on the third received signal ;
An ultrasonic diagnostic apparatus comprising: In an ultrasonic diagnostic apparatus that obtains an ultrasonic image by irradiating a subject with an ultrasonic pulse,
A determination means for determining the similarity between the intensity or amplitude value of the first signal and the intensity or amplitude value of the second signal of the received signal based on the subject site by testing ;
And averaging means for weighting averaging the intensity or amplitude value of the intensity or amplitude value and the second signal of the first signal,
Display means for displaying an ultrasonic image generated based on the averaged value,
Said averaging means, said intensity or amplitude value of the first signal and in the case similarity of the intensity or amplitude value of the second signal is higher multiplies a large weighting factor to the strength or amplitude value of the second signal, wherein a strength or amplitude value of the first signal if the second signal strength or the degree of similarity is low amplitude value multiplied by a small weighting factor to the strength or amplitude value of the second signal, said first and second weighted An ultrasonic diagnostic apparatus characterized by performing weighted averaging of the intensity or amplitude value of the two signals.   The ultrasonic diagnosis according to claim 1 or 3, further comprising display means for reconstructing and displaying the result of the weighted average so as to spatially correspond to the tomographic plane of the subject. apparatus.   The display means includes means for displaying the result of the weighted average reconstructed so as to spatially correspond to the tomographic plane of the subject in parallel or superimposed with a diagnostic image before analysis. The ultrasonic diagnostic apparatus according to claim 4.   The determination of the similarity is performed based on a hypothesis according to a probability density distribution in which a probability density distribution of signal values is a theoretical value according to a Rayleigh distribution. Ultrasound diagnostic equipment. The ultrasound according to any one of claims 1 to 5, further comprising a data capturing unit that captures data before the received signal is converted into image data and uses the data for the determination. Diagnostic device.   The ultrasonic diagnostic apparatus according to claim 6, further comprising means capable of setting a rejection area for the determination by an operator. The determination means performs the determination on a region corresponding to a predetermined range of the subject in the received signal,
9. The ultrasonic diagnostic apparatus according to claim 1, wherein the irradiation condition of the ultrasonic pulse and the reception condition of the reception signal are different between the predetermined region and the other region . 10 . The ultrasonic diagnostic apparatus as described. In an ultrasonic diagnostic apparatus that obtains a tomographic image by irradiating a subject with an ultrasonic pulse,
Analysis operation means for extracting a specific signal using the statistical properties of echo signal values or pixel values generated from the subject site;
Display means for displaying the result extracted from the analysis calculation means,
The analysis operation means includes determination means for determining the similarity between the intensity or amplitude value of the first signal and the intensity or amplitude value of the second signal in the echo signal or reception signal to be analyzed, and the first Based on the similarity between the intensity or amplitude value of one signal and the intensity or amplitude value of the second signal , a weight setting for multiplying a large weighting factor when the similarity is high and a small weighting factor when the similarity is low to the unit, the strength or amplitude value of the first signal, and averaging means for performing weighted average the extracted and strength or amplitude value of the second signal multiplied by the weighting factor, characterized in that it comprises a Ultrasonic diagnostic equipment.

Images