JP7541352B2 - Diagnostic Equipment - Google Patents

Description

The present invention relates to diagnostic devices.

The number of people suffering from functional dyspepsia is currently increasing. Although detailed tests show no organic abnormalities, patients with functional dyspepsia experience chronic early satiety, heavy stomach, and stomach pain, which can reduce the quality of life.

To improve stomach-related diseases such as functional dyspepsia, it is important to objectively understand the state of the stomach (location of the stomach, size of the stomach, changes in stomach size over time, and pH of the internal stomach). In particular, it is important to be able to visualize the stomach simply and minimally invasively so that patients themselves can understand the state of their stomach.

Electrogastrograms (EGGs) are a common method for minimally invasively measuring the condition of the stomach, but specialized knowledge is required to analyze the data, and patients are unable to grasp the condition of their stomachs by themselves. In addition, electrogastrograms (EGGs) have the problem that they require expensive amplifiers to detect the weak electromyographic values of the stomach. Also, while it is possible to visualize the stomach wall contour ∂Γ using ultrasound echography, patients cannot easily use ultrasound echography themselves.

A simple, inexpensive, and minimally invasive method for visualizing internal organs is the use of a pulmonary electrical impedance tomography (EIT) system.

As a pulmonary EIT system, Patent Document 1 discloses an electrical impedance measurement belt for placing electrodes around a patient's thorax, the electrical impedance measurement belt including an array of a plurality of spaced electrodes and a support structure on which the electrodes are arranged, the support structure having the electrode array includes two angled legs on which the electrodes are arranged, the two legs diverging at an angle from the apex so that when the belt is wrapped around the patient's thorax, the electrode array extends from the patient's back, where the apex of the angled legs is to be located, to the lower portion of the patient's sternum, essentially parallel to the ribs.

Special table 2015-508314 publication

The lung EIT system of Patent Document 1 obtains sufficient images of the chest cross section whether the subject is lying face down or standing up, because the lung contents are gas. However, the stomach contents are liquid, and are highly dependent on the subject's body position. Therefore, even when using a lung EIT system, it is difficult to visualize the state of the stomach with high accuracy.

The present invention has been made in consideration of the above problems, and has an object to provide a diagnostic device that is inexpensive, minimally invasive, and capable of visualizing the stomach of a subject in a simple manner.

In order to solve the above problems, the present invention proposes the following means.
<1> A diagnostic device according to one aspect of the present invention is a diagnostic device using an electrical impedance tomography sensor including a plurality of electrodes arranged three-dimensionally at intervals, and a support that holds the electrodes and allows the electrodes to be placed on at least the abdomen of a person to be measured, the diagnostic device including: a current/voltage application measurement unit that applies a current or a potential difference between the electrodes, and measures a potential difference and a phase based on a current application/voltage measurement pattern when the current is applied, and measures a current and a phase based on a voltage application/current measurement pattern when the potential difference is applied between the electrodes; a Jacobian matrix calculation unit that calculates a Jacobian matrix of the person to be measured based on the predetermined current application/voltage measurement pattern or the voltage application/current measurement pattern, mesh coordinates obtained by dividing the abdominal contour of the person to be measured, and the coordinates of each of the electrodes; and an electrical property distribution calculation unit that calculates an electrical property distribution of the person to be measured from the Jacobian matrix of the person to be measured calculated by the Jacobian matrix calculation unit and the potential difference and phase or the current and phase measured by the current/voltage application measurement unit ,
The apparatus further includes a stomach wall contour calculation unit that obtains stomach wall contour information of the subject by classifying the electrical property distribution obtained from the electrical property distribution calculation unit into clusters corresponding to each configuration of the subject .
<2> In the diagnosis device according to <1> above, the Jacobian matrix calculation unit may calculate the Jacobian matrix using machine learning.
<3> The diagnostic device according to <1> or <2> above may further include a relative pH estimation unit that estimates a relative pH distribution inside the stomach of the subject from the electrical property distribution calculated by the electrical property distribution calculation unit.
<4> In the diagnostic device according to any one of <1> to <3> above, the electrical property distribution may be a conductivity difference distribution.

According to the above aspects of the present invention, it is possible to provide a diagnostic device that is inexpensive, minimally invasive, and capable of easily visualizing the stomach of a subject, and the diagnostic device.

FIG. 1 is a schematic diagram of a diagnostic device according to an embodiment of the present invention. FIG. 2 is a schematic diagram of a current/voltage application measuring unit. 1 is an example of an arrangement pattern of electrodes in an electrical impedance tomography sensor. 1 is an example of an arrangement pattern of electrodes in an electrical impedance tomography sensor. 1 is an example of an arrangement pattern of electrodes in an electrical impedance tomography sensor. FIG. 13 is a diagram for explaining the relationship between electrode positions and electrode numbers. 13 is a diagram for explaining a current application voltage measurement pattern to an electrode in the 3D neighboring method. FIG. 13 is a diagram for explaining a current application voltage measurement pattern to an electrode in the 3D neighboring method. FIG. 13 is a diagram for explaining a current application voltage measurement pattern to an electrode in the 3D neighboring method. FIG. 11A and 11B are diagrams for explaining a current application voltage measurement pattern for electrodes in a full 3D method. 11A and 11B are diagrams for explaining a current application voltage measurement pattern for electrodes in a full 3D method. 11A and 11B are diagrams for explaining a current application voltage measurement pattern for electrodes in a full 3D method. 10 is a diagram for explaining an example of an arrangement of a stretch sensor, a bend sensor, and a support body. FIG. 1 is a diagram for explaining an abdominal contour ∂Ω of a subject; FIG. 1 is a diagram for explaining a data set I of a known abdominal contour ∂Ω and a data set J of a Jacobian matrix of a known abdominal interior Ω. 13 is a flowchart of calculation of the Jacobian matrix J* in the Jacobian matrix calculation unit. 13 is a flowchart of calculation for classifying into clusters in the stomach wall contour calculation unit. 11 is a flowchart for calculating a relative pH in a relative pH estimation unit. FIG. 13 is a diagram showing conditions for simulating the conductivity difference distribution Δσ and the relative pH distribution. FIG. 13 is a diagram showing conditions for simulating the distribution of conductivity difference Δσ and relative pH before a meal. FIG. 13 is a diagram showing conditions for simulating the distribution of conductivity difference Δσ and relative pH during a meal. FIG. 13 shows the distribution of conductivity difference Δσ and the distribution of relative pH after drinking a liquid. FIG. 13 shows the distribution of conductivity difference Δσ and relative pH while eating solid food. FIG. 1 shows the experimental setup for 3D imaging of the stomach. FIG. 1 shows the experimental protocol (experimental conditions) for 3D imaging of the stomach. FIG. 13 is a diagram showing the distribution of electrical conductivity differences Δσ of the internal stomach Γ of subjects 1 to 3. FIG. 13 is a diagram showing the relative pH distribution of Gamma inside the stomach of subjects 1 to 3. FIG. 13 is a graph showing the relationship between the spatial average conductivity difference and relative pH of the inside of the stomach Γ of a subject 1. 1 is a diagram showing the relationship between the spatial average conductivity difference and relative pH of the inside of the stomach Γ of a subject 2. FIG. 1 is a diagram showing the relationship between the spatial average conductivity difference and relative pH of the inside of the stomach Γ of a subject 3. FIG. FIG. 13 is a diagram showing the stomach wall contour ∂Γ of a subject after using the first step of fuzzy clustering. FIG. 13 is a diagram showing the stomach wall contour ∂Γ of a subject after using the second step of fuzzy clustering. 1 is a diagram showing changes over time in the spatial average conductivity difference <Δσ> and the spatial average relative pH <pH*> of subject 1. FIG. 1 is a diagram showing changes over time in the spatial average conductivity difference <Δσ> and the spatial average relative pH <pH*> of subject 2. FIG. 13 is a diagram showing changes over time in the spatial average conductivity difference <Δσ> and the spatial average relative pH <pH*> of a subject 3. FIG. FIG. 1 is a graph showing the change in pH over time in a healthy subject measured by a conventional method. FIG. 1 shows the change over time in pH in a patient with a gastric ulcer measured by a conventional method. FIG. 13 is a diagram showing the spatial average conductivity difference <Δσ> at a specific time for a subject. FIG. 13 is a diagram showing the spatial average relative pH <pH*> of a subject at a specific time. FIG. 13 is a diagram showing the number of voxels indicating the size of the stomach of a subject.

(Diagnostic device)
A diagnostic device 100 according to one embodiment of the present invention will be described below with reference to the drawings. As shown in Fig. 1, the diagnostic device 100 includes a current/voltage application measurement unit 1 and a diagnostic calculation unit 50. The diagnostic calculation unit 50 includes an abdominal contour estimation unit 2, a Jacobian matrix calculation unit 3, an electrical property distribution calculation unit 4, a stomach wall contour calculation unit 5, a relative pH estimation unit 6, and an output unit 7.

The diagnostic calculation unit 50 of the diagnostic device 100 includes, for example, a central processing unit (CPU), a read only memory (ROM), a random access memory (RAM), and a hard disk drive (HDD)/solid state drive (SSD). The abdominal contour estimation unit 2, the Jacobian matrix calculation unit 3, the electrical property distribution calculation unit 4, the stomach wall contour calculation unit 5, the relative pH estimation unit 6, and the output unit 7 are realized by the CPU executing a predetermined program. The program may be acquired via a recording medium or via a network. A dedicated hardware configuration may also be used to realize the configuration of the diagnostic device 100. Each unit is described below.

(Current and voltage application measurement section)
The current/voltage application measuring unit 1 will be described with reference to Fig. 2. The drawings used in the following description may show characteristic parts in an enlarged scale for the sake of convenience in order to make the characteristics easier to understand, and the dimensional ratios of each component may differ from the actual ones. The materials, dimensions, etc. exemplified in the following description are merely examples, and the present invention is not limited thereto, and may be modified as appropriate within the scope of the effects of the present invention.

First, let us define the directions. Here, we will explain an example in which the person being measured stands on floor surface F. The direction parallel to floor surface F is defined as the x direction, and the direction along floor surface F that is perpendicular to the x direction is defined as the y direction. The z direction is perpendicular to floor surface F. The z direction is perpendicular to the x and y directions. Below, the +z direction may be expressed as "up" and the -z direction as "down". Up and down do not necessarily coincide with the direction in which gravity is applied.

As shown in FIG. 2, the current/voltage application measurement unit 1 includes an electrical impedance tomography sensor 10 and a control unit 30. The electrical impedance tomography sensor 10 includes a plurality of electrodes 20 (number of electrodes Q) arranged three-dimensionally at intervals, and a support 25 that holds the electrodes 20 and allows the electrodes 20 to be placed at least on the abdomen of the person being measured. Here, "able to be placed at least on the abdomen of the person being measured" means that the electrodes 20 are placed at least on the abdomen of the person being measured when the person being measured wears the electrical impedance tomography sensor 10. After the person being measured wears the electrical impedance tomography sensor 10, the current/voltage application measurement unit 1 applies a predetermined current or potential difference between the electrodes 20 and measures the potential difference or current. When applying a current, the potential difference is measured based on a predetermined current application/voltage measurement pattern (a pattern in which two electrodes are selected in sequence from a large number of electrodes, a current is applied to each electrode, and the potential difference is measured in sequence). At this time, it is desirable to also measure the phase (the time shift between the applied current and the measured potential difference). When applying a potential difference, the current is measured based on a predetermined voltage application current measurement pattern (a pattern in which two electrodes are selected from a large number of electrodes in sequence, a potential difference is applied, and the current is measured in sequence). At this time, it is preferable to also measure the phase (the time lag between the applied potential difference and the measured current). Hereinafter, the case where a current is applied will be mainly described, and a detailed description of the case where a potential difference is applied may be omitted. Hereinafter, the number Q of electrodes 20 is, for example, 4 or more. By arranging the electrodes 20 three-dimensionally and having 4 or more electrodes 20, it is possible to estimate the stomach wall contour ∂Γ and the relative pH distribution inside the stomach using the calculation results of the abdominal contour estimation unit 2 and the Jacobian matrix calculation unit 3 described later. In order to increase the accuracy of the calculation, it is preferable to have a large number of electrodes. The arrangement position of the electrodes 20 is not particularly limited. It is preferable that the electrodes 20 are arranged so as to surround the stomach of the person to be measured. The arrangement pattern of the electrodes 20 is described below.

Figure 3 shows an example of an electrode arrangement pattern in which the electrodes 20 are arranged at intervals only on the abdomen of the subject. Here, 16 electrodes 20 are arranged, four at a time at different positions in the z-axis direction and four at a time at different positions in the xy plane. The electrodes 20 in Figure 3 are equally spaced apart in the z-axis direction and in the circumferential direction, but the arrangement pattern of the electrodes is not limited to this and may be, for example, a staggered arrangement. However, it is preferable to arrange the electrodes 20 so that the upper end to the lower end of the stomach of the subject is between the electrode 20 located at the top and the electrode 20 located at the bottom in the z-axis direction. It is also preferable to arrange the electrodes 20 so that the left end to the right end of the stomach of the subject is within a circumference with both ends of the electrodes 20 as a diameter in the x-axis direction or y-axis direction. By arranging the electrodes 20 in this way, the electrical property distribution of the subject can be measured more accurately.

Figure 4 shows an example of an electrode arrangement pattern in which electrodes 20 are arranged at intervals on and around the abdomen of the subject. Here, 16 electrodes 20 are arranged on the abdomen, four at a time at different positions in the z-axis direction and four at a time at different positions in the xy plane, and 16 electrodes 20 are arranged around the abdomen of the subject. The electrodes 20 may be spaced equally apart, or may not be spaced equally apart. However, it is preferable to arrange the electrodes 20 so that the entire stomach (from the upper end to the lower end of the stomach) is between the uppermost electrode 20 and the lowermost electrode 20. In addition, in the case of the arrangement in Figure 4, since the electrodes 20 are arranged around the abdomen, the potential difference can be measured with current flowing not only on the abdominal side of the stomach but also on the back side and both flank sides. By arranging the electrodes 20 in this way, the electrical property distribution of the stomach of the subject can be calculated more accurately.

Figure 5 shows an example of an electrode arrangement pattern in which the electrodes 20 are arranged at intervals around the abdomen of the person being measured. Here, 16 electrodes 20 are arranged around the abdomen of the person being measured, with positions changed in the z-axis direction, for a total of 64 electrodes 20. The electrodes 20 may be spaced equally apart, or may not be spaced equally apart. However, it is preferable to arrange the electrodes 20 so that the entire stomach (from the upper end to the lower end of the stomach) is between the electrode 20 located at the highest position in the z-axis direction and the electrode 20 located at the lowest position. In addition, in the case of the arrangement in Figure 5, since the electrodes 20 are arranged around the abdomen, current flows over a wide area not only on the abdominal side of the stomach, but also on the back side and both flank sides. By arranging the electrodes 20 in this way, the distribution of electrical properties of the person being measured can be measured more accurately.

The electrode 20 is electrically connected to the control unit 30. There are no particular limitations on the material or shape of the electrode 20 as long as it can apply a current or potential difference to the subject. Examples of the electrode 20 include metals such as Au, Ag, and Cu, conductive polymers, fibers whose surfaces are coated with metals, and fibers whose surfaces are coated with conductive polymers.

The method of electrically connecting the electrode 20 and the control unit 30 is not particularly limited. For example, the electrode 20 and the control unit 30 may be connected with a lead wire, or the electrode 20 and the control unit 30 may be connected with wiring with woven conductive fibers. The use of woven wiring improves the comfort of the support unit 25.

There are no particular limitations on the support 25, so long as it can hold the electrode 20 and can place the electrode 20 at least on the abdomen of the person being measured. It is preferable that the support 25 can apply a predetermined pressure to at least the abdomen of the person being measured. Applying a predetermined pressure improves the adhesion between the electrode 20 and the person being measured, making it possible to more accurately apply a current or potential difference and measure the potential difference or current. The support 25 also covers at least the abdomen of the person being measured, or the abdomen and the area around the abdomen. The material of the support 25 is preferably a dielectric material such as elastomer, leather, or cloth. The shape of the support 25 is not particularly limited, but examples include a belt-like shape, a belly band-like shape, and a suit-like shape.

(Control Unit)
The control unit 30 includes, for example, a multiplexer for switching between a current application electrode (or a voltage application electrode for applying a potential difference) that applies a current and a voltage measurement electrode (or a current measurement electrode for measuring a current) that measures a potential difference, and an impedance analyzer for performing voltage measurement (or current measurement) and phase measurement. The impedance analyzer is a component that changes the applied frequency and amplitude to measure impedance, that is, the ratio of the measured potential difference (applied potential difference) to the applied current (measured current), and its phase. The control unit 30 executes a predetermined program in, for example, a CPU, and controls the multiplexer and the impedance analyzer to perform impedance measurement (measurement of the ratio of the potential difference to the current, and its phase). The control unit 30 may be controlled only within the current/voltage application measurement unit 1 to perform impedance measurement, or the control unit 30 may be controlled according to a program executed in the diagnosis calculation unit 50 to perform impedance measurement. The result of the impedance measurement is sent to the electrical property distribution calculation unit 4. There is no particular limitation on the method of transmitting information to the electrical property distribution calculation unit 4. The data may be sent from the control unit 30 to the electrical property distribution calculation unit 4 of the diagnostic calculation unit 50 via a wired line, or may be sent to the electrical property distribution calculation unit 4 of the diagnostic calculation unit 50 wirelessly.

The control unit 30 applies a current between the electrodes 20 and measures the potential difference based on a predetermined current application voltage measurement pattern (a pattern of which electrodes the current is applied between and which electrodes the potential difference is measured between). Alternatively, the control unit 30 applies a potential difference between the electrodes 20 and measures the current based on a predetermined voltage application current measurement pattern. When applying a current or a potential difference, there is no particular limitation as to which electrodes 20 the current (potential difference) is applied between and which electrodes the potential difference (current) is measured between, but it is preferable to apply the current (potential difference) "evenly" to the electrodes 20 arranged three-dimensionally and measure the potential difference (current). "Applying a current (potential difference) evenly and measuring the potential difference (current)" means that the current potential difference is applied and measured so that all the electrodes 20 are used at least once to apply or measure the current potential difference. In order to explain the current application voltage measurement pattern to the electrodes 20, an example of how to assign electrode numbers in the electrode arrangement pattern of FIG. 5 will be explained. Hereinafter, with reference to FIG. 6, a method of numbering the electrodes 20 when each layer has eight electrodes 20 and is configured with four layers will be described. The number of each electrode 20 is composed of two numbers, the first number representing the layer number, and the second number representing the position of the electrode 20 in the layer. The numbers representing the positions of the electrodes 20 are numbered counterclockwise from the front of the center of the abdomen, for example. Hereinafter, with reference to FIGS. 7A, 7B, 7C, 7D, 7E, and 7F, a case in which a current is applied "evenly" and a potential difference is measured will be described. In addition, examples of the current application voltage measurement pattern include a three-dimensional (3D) adjacent method and a full three-dimensional (full 3D) method. The number M of the current application voltage measurement pattern differs for each current application voltage measurement pattern. In the following, each current application voltage measurement pattern will be described. In addition, the current application voltage measurement pattern described below can also be applied to a voltage application current measurement pattern. Considering the effect on the living body and the simplicity of the device, it is preferable that the applied current value and its applied frequency be, for example, an alternating current of 1.0 mA or less in the Hz to MHz range.

The 3D adjacent method is a method of applying a current between adjacent electrodes. For example, consider the case where there are 8 electrodes in one layer, and there are 32 electrodes in total in 4 layers, and use FIG. 7A to explain. Apply a current to the 1-1 and 1-2 electrodes on the same surface in the z direction, and measure the potential difference between adjacent electrodes such as the 1-3 and 1-4 electrodes, the 1-4 and 1-5 electrodes, etc. After measuring the potential difference between the last two electrodes of the layer (here, the 1-7 and 1-8 electrodes), measure the potential difference between the last electrode in the layer (here, the electrode 1-8) and the first electrode of the layer above it (here, the electrode 2-1). Thereafter, measure the potential difference between each electrode in the same procedure, and measure the potential difference between adjacent electrodes up to the 4-7 and 4-8 electrodes. Next, apply a current to the 1-2 and 1-3 electrodes on the same surface in the z direction, and measure the potential difference in the same way. As shown in Figure 7B, for the 1-8, 2-8, and 3-8 electrodes, a current is applied between the 2-1, 3-1, and 4-1 electrodes on the layer above, and the potential difference is measured in the same way, as shown in Figure 7C. In this case, with the 3D adjacent method, the number M of current application voltage measurement patterns is 928 in total.

The full 3D method is a method of measuring the potential difference between all combinations of a reference electrode and an electrode other than the reference electrode, or between as many electrodes as possible. For example, referring to FIG. 7D, a current is applied between the 1-8 electrode and the 1-1 electrode, which are adjacent to each other on the same surface in the z direction, and the potential difference is measured not only for adjacent electrodes such as the 1-2 electrode (reference electrode) and the 1-3 electrode, but also for electrodes that are not adjacent to the reference electrode, such as the 1-2 electrode and the 2-1 electrode, and the 1-2 electrode and the 3-1 electrode, and the potential difference between each of the reference electrodes and the electrodes other than the reference electrode is measured in sequence. Also, as shown in FIG. 7E, a current is applied to the 1-8 electrode and the 1-2 electrode, which are not adjacent even though they are on the same surface in the z direction, and the potential difference between each of the reference electrodes and the electrodes other than the reference electrode is measured in sequence in the same manner. Furthermore, as shown in FIG. 7F, a current is applied to the 1-8 electrode and the 2-2 electrode, which are adjacent to each other on a surface that is not on the same z direction, and the potential difference is measured in the same manner. In the full 3D method, the total number of current application and voltage measurement patterns, M, is 189,225.

Below, we will explain an example of measuring the potential difference using the 3D adjacent method with the diagnostic device 100 of this embodiment.

(Abdominal contour estimation section)
As shown in Fig. 2 and Fig. 8A, for example, the abdominal contour estimation unit 2 arranges a stretch sensor 22 and a bend sensor (angle detection sensor) 24 at four or more abdominal contour measurement points 26 on the support part 25, obtains the coordinates (x, y, z) of the abdominal contour measurement points 26 as reference points before the subject wears the support part 25, and estimates the abdominal contour ∂Ω of the subject from the change in the coordinates (x, y, z) of the abdominal contour measurement points 26 after the subject wears the support part 25. The stretch sensor 22 and the bend sensor 24 are preferably positioned in the same z-direction as each electrode, but the x and y positions may be provided at unique positions or may be positioned near the coordinates (x, y, z) of the electrode 20. The stretch sensor 22 and the bend sensor 24 superimpose the x and y abdominal contours ∂Ω at the z-direction positions of each electrode in the z-direction to estimate the three-dimensional abdominal contour ∂Ω of the subject. In order to obtain the abdominal contour ∂Ω of the subject from the coordinate data of the abdominal contour measurement points 26, the abdominal contour estimation unit 2 supplements the intervals between the abdominal contour measurement points 26 (between each coordinate point). Although the abdominal contour ∂Ω of the subject can be accurately estimated by using such stretch sensor 22 and bend sensor 24, the abdominal contour ∂Ω of the subject can be simply estimated by using only one of the stretch sensor 22 and the bend sensor 24, or even if neither of them is used, by using a support part 25 of an average size such as S, M, or L. The abdominal contour estimation unit 2 estimates the abdominal contour ∂Ω of the subject more accurately using the position coordinates of the abdominal contour measurement points 26 and an interpolation curve such as a B-spline curve. Here, the abdominal contour ∂Ω of the subject is estimated by a B-spline curve, but a Bezier curve, a Lamé curve, or the like may also be used. FIG. 8B shows the abdominal contour ∂Ω of the subject estimated by a B-spline curve. As shown in FIG. 8B, for example, the abdominal contour ∂Ω of the subject is divided into each position where an electrode is located (divided into each layer), and the abdominal contour ∂Ω ^l to ∂Ω ⁴ of the subject is estimated for each layer. From the abdominal contour ∂Ω ^l to ∂Ω ⁴ of the subject obtained for each layer, the three-dimensional abdominal contour ∂Ω of the subject is estimated more accurately using an interpolation curve such as a B-spline curve not only in the x and y directions but also in the z direction. The position of the q-th electrode 20 is represented by the length r _q from the electrode position to the center (origin) O of the layer, and the angle θ between the line connecting the position of the electrode 20 and the origin O and the x-axis. The information on the abdominal contour ∂Ω of the subject obtained is sent to the Jacobian matrix calculation unit 3.

The stretch sensor 22 is an elastic strain sensor. The stretch sensor 22 measures the displacement in the stretch direction near the abdominal contour measurement point 26 when the subject measures the support 25.

The bend sensor 24 is a sensor that can measure angular displacement. The bend sensor 24 measures the displacement in the bending direction near the abdominal contour measurement point 26 when the subject measures the support 25.

In the above explanation, an example of estimating the abdominal contour ∂Ω using the stretch sensor 22 and bend sensor 24 has been described, but in the diagnostic device 100 of this embodiment, as long as the coordinates of the abdominal contour measurement points 26 are available, it is not necessary to include sensors such as the stretch sensor 22 and bend sensor 24. When no sensors are used, for example, the coordinates of the abdominal contour measurement points 26 used in the calculations of the abdominal contour estimation unit 2 may be data obtained by measurement separately from the diagnostic device 100.

(Jacobian matrix calculation part 3)
The Jacobian matrix is a sensitivity matrix that indicates how much the measured potential difference when a current is applied (or the measured current when a voltage is applied) changes with respect to the change in the spatially distributed electrical properties (conductivity, dielectric constant) relative to the reference. The Jacobian matrix (sensitivity matrix) of the subject varies depending on the spatial distribution of the subject's electrical properties and body shape, and the electrical property distribution can be calculated by knowing the Jacobian matrix of the subject. The Jacobian matrix calculation unit 3 calculates the Jacobian matrix J* (* is a symbol that means that it is estimated for the subject) of the subject's abdominal interior Ω using the predetermined current application voltage measurement pattern (or voltage application current measurement pattern) of the electrode 20, the subject's abdominal contour ∂Ω (mesh coordinates obtained by dividing the abdominal contour ∂Ω) estimated by the abdominal contour estimation unit 2, and the coordinates of the electrode 20. The Jacobian matrix calculation unit 3 may (1) calculate the Jacobian matrix J* using the following formula (9) based on a stomach X-ray image or MRI image of the subject taken in advance, and create a custom-made Jacobian matrix J*, or (2) use three-dimensional position information of the stomach based on general information such as age, sex, nationality, height, and weight and G abdominal contours ∂Ω of any shape as a first database, create a second database (a data set I of known abdominal contours ∂Ω and a data set J of the Jacobian matrix of known abdominal interior Ω) from the first database, and select the optimal Jacobian matrix J* for the subject from the second database using machine learning or the like for the abdominal contour ∂Ω of the subject estimated by the abdominal contour estimation unit 2.

Below, the first and second databases of (2) will be described with reference to Figure 9. First, in the first database, for example, 3D position information of the stomach of a healthy person of a specific age, nationality, and sex that is publicly available (for example, a publicly available 3D image of the stomach) is prepared, and information of abdominal contours ∂Ω of all shapes (geometry) G pieces is prepared assuming obese and thin people (for example, images that have been processed from a publicly available 3D image of the stomach assuming obese and thin people).

In the first database, the interior Ω ^g of the abdominal contour ∂Ω ^g of the g-th geometry is divided into three-dimensional meshes so that an appropriate resolution can be obtained according to the Q number of electrodes 20. For example, when the number of electrodes 20 is Q=32, the region including the known abdominal contour ∂Ω ^g may be divided into approximately 1024 in the xy plane and approximately 32 in the z direction, resulting in a total of approximately 32768 points, to create mesh n (1≦n≦N). In this case, N is 32768 points. The number and shape of the meshes can be appropriately set according to the number of electrodes 20 and the required resolution. This operation may be performed in the first database or in the next second database.

Next, the data set I of the known abdominal contour ∂Ω for the information of the geometry G of abdominal contour ∂Ω in the second database will be described. The data set I of the known abdominal contour ∂Ω is a matrix consisting of data of the known abdominal contour ∂Ω having elements of (Q+3)×N (number of spatial meshes)×G (number of geometries in the first database). Here, strictly speaking, Q is the number of abdominal contour measurement points, and may have a value different from the number of electrodes, but for convenience, the number Q of abdominal contour measurement points and the number Q of electrodes are set to the same value. The data set I of the abdominal contour ∂Ω is composed of the abdominal contour ∂Ω of the subject estimated by the abdominal contour estimation unit 2, N mesh coordinates (x _n , _yn , z _n ) obtained by dividing the abdominal interior Ω from the abdominal contour ∂Ω, and the distance r from the origin O to each electrode 20. The 3 in Q+3 refers to the coordinate position (x _n , _yn , z _n ) in the nth mesh of the gth geometry, and Q refers to the radius r of the abdominal contour measurement point Q in the gth geometry.

The data set I is expressed by the following formula (1). I ^g is an input variable in the known abdominal contour ∂Ω ^g , and is expressed by the following formula (2). I ^g _n in formula (2) is an input variable of mesh n in the known abdominal contour ∂Ω ^g , and is expressed by the following formula (3). X ^g _n in formula (3) indicates the Cartesian coordinates (x _n , _yn , z _n ) of mesh n in the known abdominal contour ∂Ω ^g , and is expressed by the following formula (4). r ^g in formula (3) is the distance from the origin of the electrode 20 placed on the known abdominal contour ∂Ω ^g , and is expressed by the following formula (5). Q in formula (5) indicates the number of abdominal contour measurement points (which may be the same as the number of electrodes). Note that T in the formula indicates the transpose of a matrix element, and the right side R indicates a set of real numbers, the superscript indicates an element of a matrix or an element of a column vector, and the subscript indicates an element of a row vector.

The data set J of the Jacobian matrix of the known abdominal interior Ω may be created by, for example, calculating the Jacobian matrix J using the finite element method from the first database and any shape of abdominal contour ∂Ω described above. Since the Jacobian matrix J varies depending on the abdominal contour ∂Ω of the subject, it is preferable to prepare a large number of known abdominal contours ∂Ω and Jacobian matrices J. Here, the known abdominal contour ∂Ω is ∂Ω ^g (1≦g≦G). G is the number of known abdominal contours ∂Ω in the data set, and is, for example, 100 to 10,000. The types of ∂Ω ^g and the number of G can be modified as appropriate. The data set J of the Jacobian matrix of the known abdominal interior Ω is a known sensitivity matrix having elements of M (number of current application voltage measurement patterns)×N (number of spatial meshes)×G (number of first databases). This refers to data on multiple Jacobian matrices J obtained by the finite element method or the like from G three-dimensional shapes represented by X vectors and multiple abdominal contours ∂Ω of known specimens. The Jacobian matrix data set J is the sensitivity matrix of a predetermined current application voltage measurement pattern (or voltage application current measurement pattern).

The Jacobian matrix J of the abdominal contour ∂Ω is expressed by the following formula (6). In formula (6), M indicates the number of current application voltage measurement patterns, N indicates the number of meshes, and G indicates the number of geometries. The Jacobian matrix J ^g of geometry g (1≦g≦G) is expressed by formula (7), and the Jacobian matrix J ^g _n of mesh n (1≦n≦N) of geometry g (1≦g≦G) is expressed by formula (8). The Jacobian matrix elements J ^g _nm of current application voltage measurement patterns m (1≦m≦M) and m of mesh n (1≦n<N) of geometry g (1≦g≦G) are calculated using the following formula (9). Here, σ _n indicates the conductivity in mesh n as an example of an electrical property distribution, but may be other electrical property distributions (conductivity difference distribution Δσ, permittivity distribution, permittivity difference distribution, phase distribution, phase difference distribution). _An indicates the volume of the nth mesh, but if you want to perform a simple calculation due to calculation costs, you can use the mesh area in the xy directions to approximate the z direction. Vm(e,d) indicates the measured potential difference V in current application voltage measurement pattern m. e indicates the current application electrode pair in current application voltage measurement pattern m, and d means the voltage measurement electrode pair in current application voltage measurement pattern m. V(i ^e ) indicates the potential difference between voltage measurement electrode pair d induced by current application to current application electrode pair e. V(i ^d ) is the potential difference between current application electrode pair e induced by current application to voltage measurement electrode pair d. ∇ is the nabla symbol, which is a differential operator.

Hereinafter, a method for calculating the Jacobian matrix J* (where * denotes that the matrix is estimated for the subject) of the subject based on a predetermined current application voltage measurement pattern m (1≦m≦M) from the subject's abdominal contour ∂Ω estimated by the abdominal contour estimation unit 2 using a data set J of the Jacobian matrix of the known abdominal interior Ω output from the second database in FIG. 9 and a data set of the known abdominal contour ∂Ω will be described. When calculating the Jacobian matrix J*, the data set I of the abdominal contour ∂Ω and the data set J of the Jacobian matrix of the abdominal interior Ω are used as input variables, and the Jacobian matrix J* of the subject is calculated using, for example, a nearest neighbor search method or machine learning such as a neural network. Examples of nearest neighbor search methods include, but are not limited to, the K-nearest neighbor method, approximate nearest neighbor search, locality sensitive hashing, and kd tree. Here, the K-nearest neighbor method will be used as an example.

Next, a method of calculating the Jacobian matrix J* of the subject using the K-nearest neighbor method will be described with reference to FIG. 10. The Jacobian matrix calculation unit 3 divides the abdominal contour ∂Ω* of the subject sent from the abdominal contour estimation unit 2 into the same number of meshes N as the data set of the known abdominal contour ^∂Ωg . Then, an input variable I ^* is created from mesh n based on a predetermined current application voltage measurement pattern m and the coordinates of the electrode 20 sent from the current/voltage application measurement unit 1 (S12). The input variable I ^* is an input variable in the abdominal contour ∂Ω* of the subject, and is expressed by the following formula (10). I ^* _n in formula (10) is an input variable of mesh n in the abdominal contour ∂Ω of the subject, and is expressed by the following formula (11). X* _n in formula (11) indicates the Cartesian coordinates (x* _n , y* _n , z* _n ) of mesh n in the abdominal contour ∂Ω of the subject, and is expressed by the following formula (12). T indicates the transposition of matrix elements. r ^* in formula (11) is the distance from the origin O of the electrodes 20 (more precisely, the abdominal contour measurement points) placed on the abdominal contour ∂Ω of the subject, and is expressed by the following formula (13). Q in formula (13) indicates the number of electrodes 20 (more precisely, the abdominal contour measurement points).

Next, the flow of calculation of the Jacobian matrix J ^* of the subject by the K-nearest neighbor algorithm (K-NN) will be described. The Jacobian matrix calculation unit 3 inputs initial values (e.g., n=1, m=1) (S13). Next, the Jacobian matrix calculation unit 3 first calculates a Euclidean distance matrix C between the measured I ^* of the subject and the data set I of the known abdominal contour ∂Ω output from the second database (S14). The Euclidean distance matrix C is expressed by the following formula (14), and indicates I ^* _nm and input variables ^Ignm of K clusters with small Euclidean distance. The number of clusters K is not particularly limited, and is, for example, 5. The Euclidean distance matrix C is composed of elements K× _N of the number of clusters K and the number of meshes N that are determined independently. The Euclidean distance matrix Cn in mesh n (1≦n≦N) is shown by the following formula (15), which indicates the Euclidean distance between I ^g _n and I ^* _n , and is determined so that the measured I ^* of the subject and the data set I ^g of the known abdominal contour ∂Ω of the geometry g (1≦g≦G) are minimized. Next, using the Euclidean distance matrix C and the data set J of the Jacobian matrix of the known abdominal interior Ω, the Jacobian matrix J ^* _n (the number of elements is M) of the subject at m when n is fixed is calculated (S15). It is calculated by the following formula (16), and J ^g _n indicates the Jacobian matrix of the current application voltage measurement pattern m in mesh n at the known abdominal contour ∂Ω ^g . In FIG. 10, for example, an example of I ₅ ^* is shown, taking the position of mesh n=5 as an example.

After the calculation of J _nm ^* is completed, the Jacobian matrix calculation unit 3 judges whether the number of n is equal to the number of meshes N (S15). If n and N are not equal, the number of n is incremented by one, and the process returns to S15 again (S16). If n and N are equal, the process next judges whether m is equal to the number M of current application voltage measurement patterns (S17). If m and M are not equal, the number of m is incremented by one, and the process returns to S14 again (S18). When n and m are equal to N and M, respectively, the Jacobian matrix calculation unit 3 ends the calculation of the Jacobian matrix J ^* of the subject, and sends the Jacobian matrix J ^* of the subject to the electrical property distribution calculation unit 4. Normally, if the Jacobian matrix J is calculated using the above formula (3) on a normal PC, it takes more than five minutes to calculate. In the diagnostic device 100 according to this embodiment, a known data set I and a data set J of the Jacobian matrix of the abdominal contour ∂Ω ^g are prepared, and the Jacobian matrix J ^* of the subject can be calculated with high accuracy in a short time by using machine learning such as the K-nearest neighbor method based on the subject's abdominal contour ∂Ω and the current application and voltage measurement pattern.

(Electrical property distribution calculation part)
The electrical property distribution calculation unit 4 calculates the electrical property distribution of the subject from the Jacobian matrix J ^* of the subject sent from the Jacobian matrix calculation unit 3 and the potential difference and phase (or current and phase) measured by the current and voltage application measurement unit 1. Here, the electrical property distribution is, for example, the electrical conductivity distribution σ, the electrical conductivity difference distribution Δσ, the dielectric constant distribution, the dielectric constant difference distribution, the phase distribution, and the phase difference distribution. Hereinafter, the electrical conductivity and the electrical conductivity difference (the electrical conductivity at time t relative to the reference at time _t0 ) may be described separately, and the symbol Δ is used to represent the difference.
The following description focuses on the conductivity difference distribution Δσ. The problem of calculating the conductivity difference distribution Δσ from the known Jacobian matrix J ^* of the measured person and the known measured potential difference ΔV (potential difference at time t relative to the reference at time t ₀ ) is called an ill-posed inverse problem, and can be calculated, for example, using iterative calculation. The number of iterations is represented by the number on the right. The initial conductivity difference distribution Δσ ⁰ at the 0th iteration (the number on the upper right is the number of iterations) is calculated from the following formula (17) using the Jacobian matrix J ^* of the measured person. T indicates a transposed matrix. ΔV in formula (17) is a column vector having M elements of a predetermined current application voltage measurement pattern (or voltage application current measurement pattern) as shown in the following formula (18). There are two ways to process ΔVm in formula (18), that is, a method using a measurement time difference with a constant applied current frequency, and a method using several applied current frequency differences with a constant measurement time. Here, the method using the measurement time difference will be described. Using the measured potential ΔV _m (t ₀ ) at time t ₀ of current application voltage measurement pattern m (0≦m≦M) as a reference, the measured potential difference ΔV of the time difference from the measured potential ΔV _m (t) at time t is expressed by the following equation (19). This equation may also be divided by V _m (t ₀ ). In equation (19), m is the current application voltage measurement pattern.
The conductivity difference distribution Δσ of the subject is calculated using the following formula (20) with the initial conductivity difference distribution Δσ ⁰ as the start of the number of repetitions. In formula (20), i represents the number of repetitive calculations. In formula (20), R represents a regularization matrix, and λ represents an arbitrary parameter for converging the calculation, and is, for example, 0.01. R is represented, for example, by the following formula (21), and is a function of the known Jacobian matrix J ^* of the subject. The calculated conductivity difference distribution Δσ of the subject is sent to the stomach wall contour calculation unit 5 or the relative pH estimation unit 6.

(Stomach wall contour calculation unit)
The stomach wall contour calculation unit 5 obtains stomach wall contour ∂Γ information of the subject by classifying the electrical property distribution (conductivity distribution σ, conductivity difference distribution Δσ, permittivity distribution, permittivity difference distribution, phase distribution, and phase difference distribution) of the subject obtained from the electrical property distribution calculation unit 4 into clusters corresponding to the components of the subject's stomach inside Γ, stomach wall contour ∂Γ, and outside the stomach. In addition, electrical property distribution information of the subject's stomach inside Γ is also obtained. The stomach wall contour calculation unit 5 sends the obtained stomach wall contour ∂Γ information and electrical property distribution information of the stomach inside Γ to the relative pH estimation unit 6 and the output unit 7. Compared to other tissues such as the liver and spleen surrounding it, the stomach wall has a particularly high conductivity difference among electrical property distributions, and by clustering the electrical property difference distribution Δσ, the stomach wall contour ∂Γ can be extracted in distinction from other organs. The inside of the stomach also has a high conductivity difference due to the secretion of gastric acid, and therefore, by obtaining the stomach wall contour ∂Γ, the conductivity difference distribution Δσ of the stomach inside Γ can be obtained with high accuracy. Furthermore, in addition to the conductivity difference, the gastric wall has a small dielectric constant compared to other tissues such as the liver and spleen, and as a result, there is a large difference in phase, so electrical properties such as dielectric constant and phase may be used in addition to the conductivity difference. The method of classifying the electrical property distribution of the subject into clusters corresponding to the components of the subject (outside of the stomach, stomach wall contour ∂Γ, inside of the stomach Γ) is not particularly limited, but examples include fuzzy clustering, k-means, EM algorithm, ISODATA method, and longest distance method. Here, an example of two-step fuzzy clustering in which fuzzy clustering is repeated twice will be described.

FIG. 11 is a flow chart showing how to classify the electrical property distribution (conductivity distribution σ, conductivity difference distribution Δσ, permittivity distribution, permittivity difference distribution, phase distribution, and phase difference distribution, etc.) of a subject into clusters corresponding to the subject's constituent components using a 2-step fuzzy clustering (2-step fuzzy c-means algorithm). Hereinafter, the flow of classifying the electrical property distribution into clusters corresponding to the subject's constituent components will be described using FIG. 11, taking the example of electrical property distribution σ as an example. The electrical property distribution σ sent from the electrical property distribution calculation unit 4 is divided into three types of clusters K=3, ^k=1 σ ⁰ for the stomach inside Γ, ^k=2 σ ⁰ for the stomach wall contour ∂Γ, and ^k=3 σ ⁰ for the stomach outside, and the clusters are initialized (S21). At this time, the number of repetitions j is set to 0. Next, the cluster center position ^kr _c ^j is calculated from the coordinates ^kr _n = (x _n , _yn , z _n ) of the conductivity ^k σ _n of the k-th clustered mesh n using the following formula (22) (S22). Next, the fuzzy weight ^k μ ^j n is calculated using formula (23) (S23). Next, the obtained fuzzy center of gravity ^k f _c ^j is calculated using formula (24) (S24). Cluster reassignment is performed using formula (25). After calculating each cluster center of gravity ^kr _c ^j and the fuzzy center of gravity ^k f _c ^j , the conductivity σ _n in mesh n is assigned to cluster _k as the product of the center of gravity ^kr _c with the smallest Euclidean distance between the coordinate rn of mesh n and the fuzzy center of gravity ^k f _c ^j calculated using the following formula (23 ₎ and the Fiji ^weight (S25). After each cluster rearrangement, if the latest j-th calculation of each fuzzy center of gravity ^k f _c ^j does not match the center of gravity ^k f _c ^j-1 of each fuzzy center of gravity k f c j in the calculation one time before (j-1), the process returns to S22 (S26). If the latest fuzzy center of gravity k f c j-1 of each calculation matches the fuzzy center of gravity k f c j in the calculation one time before, the process proceeds to step 2. Step 2 is the same as step 1, but the number of clusters K is 2, that is, calculation is performed only for k=1 (inside the stomach) and k=2 (stomach wall), and the cluster center of gravity ^k r _c and the fuzzy center of gravity ^k f _c ^j are calculated again for each cluster. Step 2 is the same as step 1, in which cluster initialization (S21a), calculation of cluster center of gravity (S22a), calculation of fuzzy weighting (S23a), calculation of fuzzy center of gravity (S24a), cluster resetting (S25a), and judgment (S26a) are performed. By performing the same process again, an electrical property distribution clustered into the components of the subject can be obtained. By classifying the electrical conductivity distribution σ of the subject into clusters corresponding to the components, such as by fuzzy clustering, the stomach wall contour ∂Γ can be clarified with higher accuracy than by dividing it by a threshold value of electrical conductivity. In addition, since the stomach wall contour calculation unit 5 can calculate the stomach wall contour ∂Γ in a short time, the shape change of the stomach can be obtained in real time, and further, the time change of the stomach size can be obtained by calculating the number of voxels of the stomach wall contour ∂Γ and the stomach interior Γ. By obtaining the shape change of the stomach wall contour ∂Γ in real time, it is possible to grasp the movement of the entire stomach, the movement of the upper stomach, and the movement of the lower stomach, which are important for stomach diagnosis.

The stomach wall contour calculation unit 5 may use a conventional method to emphasize the boundary between the subject's stomach wall contour ∂Γ and the inside of the stomach Γ, and the boundary between the stomach wall contour ∂Γ and the outside of the stomach. The method of emphasizing the boundary is not particularly limited, and examples include the Canny method, the Sobel method, and the Gaussian Laplacian method.

(Relative pH Estimation Unit)
FIG. 12 is an example of an algorithm for determining the relative pH in the relative pH estimation unit 6. Hereinafter, an example of the electrical conductivity distribution σ will be taken as the electrical property distribution, and the algorithm will be described. The relative pH estimation unit 6 estimates the distribution of the relative pH inside the stomach Γ based on the electrical property distribution (here, the electrical conductivity distribution σ) of the subject sent from the electrical property distribution calculation unit 4 and the stomach wall contour ∂Γ information of the subject sent from the stomach wall contour calculation unit 5. The estimated relative pH distribution is sent to the output unit 7. The relative pH estimation unit 6 may estimate the relative pH distribution inside the stomach of the subject from the electrical property distribution of the subject. Here, the relative pH does not indicate the strict pH, but indicates the pH relative to the following standard, and is represented by the symbols pH* and *. When a drink or meal enters the stomach Γ, gastric acid (mainly hydrochloric acid HCl) is secreted from the stomach wall contour ∂Γ, and the pH of the stomach Γ changes. The exact relative pH of Γ inside the stomach is expressed as pH = -log [H ⁺ ], where [H ⁺ ] is the concentration of H ⁺ ions in the solution per liter in moles. First, the relative pH before a meal (drinking) is estimated based on, for example, the following formula (26) which is based on Kohlrausch's square root law (Onsager's conductivity formula). In the following formula (26), σ indicates the conductivity calculated by the electrical property distribution calculation unit 4, and c is the gastric acid concentration to be obtained from this formula. Λ ^∞ indicates the known limiting molar conductivity of gastric acid, which is 273.5 x 10 ⁴ Sm ² mol ^-1 according to Kohlrausch's law of independent ion migration if gastric acid is assumed to be hydrochloric acid HCl. In addition, α and β are constants calculated from the temperature, dielectric constant, and viscosity coefficient of the solvent of Γ inside the stomach. If we assume that the solvent is water, the temperature of Γ inside the stomach is 36.5°C (=309.5K), the dielectric constant is 80.4, and the viscosity coefficient η=0.8903×10 ^-2 P Boise (1P=10 ^-1 Pas), then the constants such as Λ ^∞ , α, and β can be obtained as shown in Fig. 12, and c can be calculated using formula (26). There are three points to note here: 1) c is gastric acid (electrolyte concentration) rather than the concentration of H ⁺ , and the ionization constant of HCl may be taken into consideration to calculate an accurate [H ⁺ ], but since gastric acid is a strong gastric acid, the ionization constant may not be taken into consideration for simplicity, and c may be regarded as [H ⁺ ]. In addition, 2) the conductivity σ ₀ calculated by the electrical property distribution calculation unit 4 is not an absolute value with a unit but a relative value, so c calculated by formula (26) may also be considered as a relative value. Furthermore, 3) since c is an electrolyte concentration per unit volume, a more accurate comparison is possible if the volume V of the ingested solvent such as water is known. If the ingested water is known, c calculated by formula (26) may be divided by the water volume Vw, the stomach volume V0 on an empty stomach, and the stomach volume Vt at time t to obtain a value per unit volume, or, since it is based on relative values, it is not necessary to divide by Vw, V ₀ , Vt, etc. Below, a description will be given of the case where a liquid is ingested and the case where a solid food is eaten according to a simulation.

First, electromagnetic simulation was performed under the simulation conditions shown in Table 1, FIG. 13, FIG. 14A, and FIG. 14B, that is, under the conditions shown in Table 1 for the geometry of the abdomen and stomach shown in FIG. 13, under the conditions of FIG. 14A before meal (during drinking) (BM) and FIG. 14B during meal (DM) conditions. First, the case of drinking liquid will be described. In general, it is said that when about 200 mL of liquid enters the stomach of a healthy person, it passes through the gastric tract (stomach tract) and is discharged into the duodenum in about several minutes to several tens of minutes, and the expansion of the stomach volume _V0 and the peristaltic movement of the stomach when fasting are small. When these liquids enter the stomach, it can be assumed that there is almost no change in the stomach volume _V0 when fasting. In the case of a liquid with high viscosity, the viscous friction force becomes large and is discharged by the peristaltic movement of the stomach, and there is a volume change in the stomach interior Γ, so this will be described in the case of a solid matter described later. The relative pH of the stomach when a beverage is consumed can be estimated by the following method. The relative pH after drinking a beverage is calculated, for example, from the following formula (27). In the following formula (27), σ ^BM ₀ indicates the fasting conductivity calculated before drinking the liquid, σ ^BM _t indicates the conductivity at time t after drinking the liquid, Λ ^∞ indicates the limiting molar conductivity difference, α and β are constants, c ^BM _0H+ indicates the fasting gastric acid concentration before drinking the liquid, and c ^BM _tH+ indicates the gastric acid concentration c ^BM _tH+ at time t after drinking the liquid. The sum of c ^BM _0H+ and c ^BM _tH+ is expressed by the following formula (28). In formula (28), m ₀ indicates the mass of gastric acid (H ⁺ ), and Δm indicates the change in the mass of gastric acid after drinking the liquid. V ₀ is the fasting stomach volume before drinking the liquid. c ^BM _0H+ can be calculated using equation (26) from the conductivity distribution σ obtained from the results obtained by measurement while drinking the liquid, and V ₀ can also be calculated from the stomach wall contour ∂Γ obtained from the results obtained by measurement before drinking the liquid. [H ⁺ ] is expressed by the following equation (29). The relative pH can be calculated from equations (26) and (29). In this way, the relative pH can be calculated using equation (29) from the conductivity distribution σ obtained after drinking the liquid. This allows the conductivity distribution σ and relative pH distribution before the meal (BM in beverage) to be obtained as shown in FIG. 15.

Next, the case of eating solid food will be described. When solid food is eaten or highly viscous liquid is drunk, the contents that enter the stomach are discharged by the peristaltic movement of the stomach. In general, the stomach of a healthy person takes a longer time to discharge solid food into the duodenum than liquid, and the expansion from the stomach volume _V0 when fasting and the peristaltic movement of the stomach are large, while it is said that patients with stomach diseases such as functional dyspepsia have small expansion from the stomach volume _V0 when fasting and small peristaltic movement of the stomach. Therefore, the state of the stomach can be examined by comparing the volume Vt of the stomach interior Γ at time t with _V0 . The relative pH of the stomach when eating solid food or drinking highly viscous liquid can be estimated by the following method. The relative pH after eating solid food or drinking highly viscous liquid is calculated, for example, from the following formula (30). In the following formula (30), σ ^DM ₀ denotes the conductivity distribution σ calculated before eating a solid food, σ ^DM _t denotes the conductivity distribution calculated at time t while eating a solid food, Λ ^∞ denotes the limiting molar conductivity difference, α and β are constants, c ^DM _0H+ denotes the gastric acid concentration before eating a solid food, and c ^DM _tH+ denotes the gastric acid concentration at time t while eating a solid food. The sum of c ^DM _0H+ and c ^DMtH ₊ is expressed by the following formula (31). In formula (31), m ^DM _0H+ denotes the mass of gastric acid (H ⁺ ) before eating a solid food, Δm ^DM _tH+ denotes the change in the mass of gastric acid while eating a solid food, V ^DM _0H+ denotes the stomach volume before eating a solid food, and V ^DM _tH+ denotes the change in stomach volume while eating a solid food. c ^DM _0H+ can be calculated using equation (26) from the conductivity distribution σ estimated from the measurement before eating solid food. V ^DM _0H+ can be obtained from the stomach wall contour ∂Γ estimated from the measurement before eating solid food, and V ^DM _tH+ can be obtained from the stomach wall contour ∂Γ estimated from the measurement while eating solid food. [H ⁺ ] is expressed by the following equation (32). The relative pH can be obtained from equations (26) and (32). As a result, the conductivity distribution σ and the relative pH distribution during a meal (DM) can be obtained as shown in FIG. 16.
The diagnostic device 100 according to this embodiment can measure the current potential difference and display the image in a short time (approximately 0.1 to 1 second), and can also perform calculations in a short time, so that it is possible to measure changes in the shape and relative pH of the stomach during a meal in real time.

(Output section)
The output unit 7 outputs the electrical conductivity distribution σ, electrical conductivity difference distribution Δσ, dielectric constant distribution, dielectric constant difference distribution, phase distribution, and phase difference distribution within the stomach wall contour ∂Γ and the stomach wall contour ∂Γ sent from the stomach wall contour calculation unit 5, as well as the relative pH distribution sent from the relative pH estimation unit 6. The output unit 7 may also output the electrical conductivity distribution σ, electrical conductivity difference distribution Δσ, dielectric constant distribution, dielectric constant difference distribution, phase distribution, and phase difference distribution sent from the electrical property distribution calculation unit 4 preceding the stomach wall contour calculation unit 5. These electrical property distributions and relative pH distributions may be converted into 4D (four-dimensional) images of 3D (three-dimensional) space and time, 3D images with fixed time, 2D images or 1D (one-dimensional) values obtained by spatially averaging the 3D images, and time average values obtained by temporally averaging, and the outputs may be displayed. Hereinafter, <> may be used as a symbol for spatial average.

There are two types of gastric motility: adaptive relaxation (the motility of the upper stomach that puts food into the stomach) and gastric emptying (the motility of the lower stomach that emptys digested material into the duodenum). Symptoms of gastric disease include "early satiety" caused by abnormal adaptive relaxation due to reduced motility of the upper stomach, and "heavy stomach" caused by abnormal gastric emptying due to reduced motility of the lower stomach. Therefore, the output unit 7 may display the change in the volume V0 of the entire stomach over time, or may display the upper stomach (fundus), middle stomach (body), and lower stomach (pylorus) separately, rather than the entire stomach volume. Furthermore, one of the pathological conditions of gastric disease is, for example, stomach pain (epigastric pain, epigastric burning sensation) caused by gastric hypersensitivity, and in some cases, the secretion of gastric acid from the middle stomach (body) is important. Therefore, the relative pH may be displayed as the relative pH of the entire stomach, or may be displayed separately for the upper stomach (fundus), middle stomach (body), and lower stomach (pylorus).

The output destination of the electrical property distribution clustered into the constituents of the subject is not particularly limited. The output destination may be a display unit such as an LCD display, or a storage device such as a HDD.

The above is a detailed explanation of the diagnostic device 100 according to this embodiment. Note that the technical scope of the present invention is not limited to the above embodiment, and various modifications can be made without departing from the spirit of the present invention. In the above example, the relative pH was calculated by the relative pH estimation unit 6 after the calculation by the stomach wall contour calculation unit 5, but the relative pH may be calculated by the relative pH estimation unit 6 first, and then the stomach wall contour calculation unit 5 may perform the calculation.

In addition, the components in the above embodiment may be replaced with well-known components as appropriate without departing from the spirit of the present invention, and the above-mentioned modifications may be combined as appropriate.

The present invention will be described below with reference to examples, but the present invention is not limited to the following examples.

Example 1
Next, an example of an experiment conducted to verify the effectiveness of the diagnostic device 100 according to this embodiment will be described. This experiment was conducted with the approval of the Chiba University Graduate School of Engineering Bioethics Committee. The subjects were three people (a 22-year-old male (subject 1), a 30-year-old male (subject 2), and a 32-year-old male (subject 3)), two of whom (the 22-year-old and 30-year-old males) were healthy, but the 32-year-old male suffered from mild functional dyspepsia and had a condition in which he had a bad stomach and could not eat much breakfast. The subjects wore a suit-like support 25 equipped with 8 (x, y directions) x 4 (z direction) = 32 electrodes 20 as shown in FIG. 17. The predetermined current application voltage measurement pattern was the 3D adjacent method shown in FIGS. 7A to C. The application frequency between the electrodes of the current/voltage application measurement unit 1 was 1450 Hz, the application current was 1 mA, and a current was applied between each electrode to measure the potential difference. In the abdominal contour estimation unit 2, since all three subjects have the same body structure, the coordinates of the contour measurement position for one representative subject's abdomen were measured using a bend sensor (angle detection sensor) and a stretch sensor, and the abdominal contours ∂Ω ^l to ∂Ω ⁴ , which are the shapes from layer 1 to layer 4, were obtained, and the three-dimensional abdominal contour ∂Ω of the subject was estimated using a B-spline curve. In the Jacobian matrix calculation unit 3, a Jacobian matrix J* was calculated based on the body structure of the subject based on FIG. 10. In the electrical property distribution calculation unit 4, only the conductivity difference distribution was imaged, without using the permittivity distribution and phase distribution. In the stomach wall contour calculation unit 5 of the diagnostic device of this embodiment, two-step fuzzy clustering was used. The experiment was started under the above experimental conditions.

The relationship between the measurement time and the eating and drinking behavior will be described below with reference to FIG. 18. The measurement time is every 3 minutes, and each time is indicated by a symbol t ₀ to t _40. The subject did not eat or drink for more than 6 hours before the start of the experiment, and the fasting state was set as the initial state. From 30 minutes before the start of the experiment, the subject was in a sitting position and resting state to prepare the physiological condition. After the start of the experiment, the subject drank 200 mL of water every 30 minutes, divided into three times, totaling 600 mL of water, and 90 minutes later, ate three inari sushi. The three inari sushi had a mass of 300 g, and when liquefied in a shaker to simulate the inside of the stomach, it was equivalent to approximately 100 mL of viscous liquid. Measurements were performed 41 times every 3 minutes (including measurements at 0 minutes at the start of the experiment) for a total of 120 minutes, and the measured potential difference V0 in the fasting state at the start of the experiment t ₀ was used as the reference. The first 200 mL of water was promptly drunk in about 30 seconds immediately after the measurement at the start of the experiment (t ₀ ). Then, the second 200 mL of water was promptly drunk in about 30 seconds immediately after the measurement 30 minutes later (t ₁₀ ). Then, the third 200 mL of water was promptly drunk in about 30 seconds immediately after the measurement 60 minutes later (t ₂₀ ). Then, three 300 g pieces of inari sushi were promptly eaten in about 4 minutes immediately after the measurement 90 minutes later (t ₃₀ ), and the measurement was continued.

Figure 19 shows output images of 3D conductivity difference distribution Δσ of three subjects (two healthy subjects and one mild dyspepsia patient) output by the electrical property distribution calculation unit 4, and the images are taken before the two-step fuzzy clustering process, at t ₁ (immediately after drinking 200 mL of water), t ₁₁ (immediately after drinking 200 mL of water for the second time), t ₂₁ (immediately after drinking 200 mL of water for the third time), and t ₃₁ (during a meal), 3 minutes after the start of the experiment. In the diagram of t ₁ after the start of the experiment in Figure 19, the area surrounded by a dotted line indicates the 3D position of the stomach. The diagram at the top of Figure 19 shows the electrode arrangement pattern.

FIG. 20 shows the relative pH [-] output by the relative pH estimation unit 6. Here, the relative pH estimation unit 6 is executed before the output of the stomach wall contour calculation unit 5, and the result of the relative pH estimation unit 6 is input to the stomach wall contour calculation unit 5. FIG. 20 shows the relative pH distribution of subjects 1 to 3 at t ₁ (immediately after drinking 200 mL of water) 33 minutes after the start of the experiment, t ₁₁ (immediately after drinking 200 mL of water for the second time), t ₂₁ (immediately after drinking 200 mL of water for the third time), and t ₃₁ (during a meal) 93 minutes after the start of the experiment. The area surrounded by the dotted line in the figure of FIG. 20 3 minutes after the start of the experiment indicates the 3D position of the stomach. The upper figure of FIG. 20 shows the electrode arrangement pattern.

Figures 21A, 21B, and 21C are graphs showing the spatial average conductivity difference <Δσ> and the spatial average relative pH <pH*> according to the Onsager conductivity formula in the relative pH estimation unit 6, and show the results of determining the spatial average relative pH <pH*> from the spatial average conductivity difference at 40 points at each time (the conductivity difference at the start of the experiment t0 is used as the reference, and the spatial average conductivity difference <Δσ> at time t). Figure 21A shows the results for subject 1 (healthy subject), Figure 21B shows the results for subject 2 (healthy subject), and Figure 21C shows the results for subject 3 (a patient with mild functional dyspepsia).

22A and 22B are 3D images showing the spatial positions of the stomach interior Γ (k=1) and the stomach wall contour ∂Γ (k=2) after two-step fuzzy clustering (a) first-step fuzzy clustering ( FIG. 22A ) and (b) second-step fuzzy clustering ( FIG. 22B ) output by the gastric wall contour calculation unit 5, based on the 3D image of the conductivity difference Δσ of one of the subjects (subject 1: healthy subject) output by the electrical property distribution calculation unit 4 in FIG. 19 . In the first step in Figure 22A, the stomach was divided into three clusters ( ^{k = 1} Δσ for inside Γ, ^{k = 2} Δσ for stomach wall contour ∂Γ, and ^{k = 3} Δσ for outside the stomach), and in the second step in Figure 22B, the stomach was divided into two clusters ( ^{k = 1} Δσ for inside Γ and ^{k = 2} Δσ for stomach wall contour ∂Γ), and the stomach wall contour ∂Γ and inside the stomach Γ were extracted. Representative images are shown at t ₁ (immediately after drinking 200 mL of water) 3 minutes after the start of the experiment and t ₃₁ (immediately after eating a meal).

23A, 23B, and 23C show the spatial average conductivity difference <Δσ> in each voxel located in two types of clusters ( ^k=1 Δσ of stomach interior Γ and ^k=2 Δσ of stomach wall contour ∂Γ) for three subjects on the left vertical axis. Furthermore, the time change of relative pH (relative pH <pH*> on the right vertical axis) which is the output of the diagnostic device of this embodiment is shown. The horizontal axis shows time (t1 to t40). FIG. 23A shows the result of subject 1 (healthy subject). FIG. 23B shows the result of subject 2 (healthy subject). FIG. 23C shows the result of subject 3 (patient with mild functional dyspepsia). Next, the time change of <Δσ> will be explained for each time t1 to t40. As shown in the subjects (healthy subjects) in Figures 23A and 23B, the spatial average conductivity difference <Δσ> immediately after drinking 200 mL of water, i.e., at t ₁ , t ₁₁ , and t _21, shows a rapid decrease in the conductivity difference due to the dilution of gastric acid by water, and at the measurement time thereafter, gastric acid is secreted, and the decrease in <Δσ> is gradual or increases, and then gastric acid secretion stops and liquid is discharged from the stomach, so <Δσ> shows a tendency to decrease. At t ₃₁ immediately after eating inari sushi, there is a similar tendency in <Δσ>, but since meals are solids that take time to digest and are discharged from the stomach for a long time, and gastric acid secretion continues for a long time, it can be seen that the decrease in <Δσ> is not as significant as in the case of water. However, the subject 3 (a patient with mild functional dyspepsia) in Figure 23C shows a significantly different tendency. For example, at time _t31 immediately after eating inari sushi, <Δσ> hardly changes.

Next, the relative pH will be described. As shown in the subject (healthy person) in Figs. 23A and 23B, immediately after drinking 200 mL of water, i.e., the spatial average relative pH of the relative pH at t ₁ , t ₁₁ , and t ₂₁ , rises rapidly due to the secretion of gastric acid, and then, as the gastric acid is diluted, there is a gradual rise in the relative pH <pH*>, and at the measurement time thereafter, gastric acid is secreted and the relative pH shows a tendency to increase. It can be seen that even at t ₃₁ immediately after eating inari sushi, there is a tendency for <pH*> to increase. However, the subject (mild functional dyspepsia patient) in Fig. 23(c) shows a significantly different tendency. For example, at t ₃₁ immediately after eating inari sushi, <pH*> hardly changes. From the above, it was confirmed that the diagnostic device according to this embodiment can visualize the relative pH inside the stomach Γ before and after eating.

Incidentally, Figures 24A and 24B are examples of measurements taken using a highly invasive gastric relative pH measuring device of the prior art (a device that measures the relative pH inside the stomach by inserting a sensor directly into the stomach's Gamma through the nasal cavity) (cited in Ashida Kiyoshi: 24-Hour Relative pH Monitoring, edited by Takemoto Tadayoshi: Japanese Society of Digestive Ulcers, Medical Science Publishing, 1995), and show similar trends. Figure 24A shows the results for a healthy subject, and Figure 24B shows the results for a patient with a gastric ulcer.

Fig. 25A shows the time change [-] of the spatial average conductivity difference <Δσ> between the stomach interior Γ (k=1) and the stomach wall contour ∂Γ (k=2) after 2-clustering, at times of particular interest: t ₁ (immediately after drinking 200 mL of water after measurement at t ₀ ), t ₁₁ (immediately after drinking 200 mL of water after measurement at t ₁₀ ), t ₂₁ (immediately after drinking 200 mL of water after measurement at t ₂₀ ), and t ₃₁ (immediately after eating inari sushi after measurement at t ₃₀ ), and Fig. 25B shows the relative ph [-] at those times. As shown in these figures, it can be seen that the trends are significantly different between healthy subjects and patients suffering from functional dyspepsia.

Figure 26 shows the number of voxels of two types of clusters ( ^k=1 Δσ of stomach interior Γ and ^k=2 Δσ of stomach wall contour ∂Γ) calculated from Figure 19. The average values of t1 _, t11 _{, and t21} (immediately after drinking 200 mL of water) 3 minutes after the start of the experiment and _t31 (immediately after eating) 93 minutes after the start of the experiment were compared between a healthy subject and a functional dyspepsia patient. As a result, the stomach volume immediately after eating compared to immediately after drinking water increased significantly by _ΔV31 = 23% in the healthy subject, but only increased by _ΔV31 = 6% in the functional dyspepsia patient. This confirmed that the shape of motor function could be extracted.

As described above in detail, the diagnostic device 100 according to this embodiment can diagnose the stomach condition of the subject in a very minimally invasive manner with high accuracy in a short time by simply applying a weak current (weak potential difference).

1 Current and voltage application measurement unit 2 Abdominal contour estimation unit 3 Jacobian matrix calculation unit 4 Electrical property distribution calculation unit 5 Stomach wall contour calculation unit 5
6 Relative pH estimation unit 7 Output unit 10 Electrical impedance tomography sensor 20 Electrode 25 Support 30 Control unit 100 Diagnostic device

Claims (4)

A diagnostic device using an electrical impedance tomography sensor, comprising: a plurality of electrodes arranged three-dimensionally at intervals; and a support for holding the electrodes and enabling the electrodes to be placed on at least an abdomen of a subject,
a current/voltage application measurement unit that applies a current or a potential difference between the electrodes, and measures a potential difference and a phase based on a current application/voltage measurement pattern when the current is applied, and measures a current and a phase based on a voltage application/current measurement pattern when the potential difference is applied between the electrodes;
a Jacobian matrix calculation unit that calculates a Jacobian matrix of the subject based on the predetermined current application voltage measurement pattern or the voltage application current measurement pattern, mesh coordinates obtained by dividing the abdominal contour of the subject, and coordinates of each of the electrodes;
an electrical property distribution calculation unit that calculates an electrical property distribution of the subject from the Jacobian matrix of the subject calculated by the Jacobian matrix calculation unit and the potential difference and phase or the current and phase measured by the current/voltage application measurement unit;
Equipped with
The diagnostic device further includes a stomach wall contour calculation unit that obtains stomach wall contour information of the subject by classifying the electrical property distribution obtained from the electrical property distribution calculation unit into clusters corresponding to each configuration of the subject . The diagnostic device according to claim 1 , wherein the Jacobian matrix calculation unit calculates the Jacobian matrix using machine learning. The diagnostic device according to claim 1 , further comprising a relative pH estimation unit that estimates a relative pH distribution inside the stomach of the subject from the electrical property distribution calculated by the electrical property distribution calculation unit. The diagnostic device according to claim 1 , wherein the electrical property distribution is a conductivity difference distribution.

Images