WO2015141668A1

WO2015141668A1 - Method for non-invasively determining angle formed by structure inside organism

Info

Publication number: WO2015141668A1
Application number: PCT/JP2015/057880
Authority: WO
Inventors: 達則坂本; 壽一伊藤; 知宏黒田; トゥッカマティアスカルボネン、
Original assignee: 国立大学法人京都大学
Priority date: 2014-03-17
Filing date: 2015-03-17
Publication date: 2015-09-24
Also published as: JP2015175785A

Abstract

The present invention provides a computer program, a device, a method and the like which make it possible to evaluate endolymphatic hydrops from a live cochlea without requiring extraction and slicing of the cochlea or a solvent treatment thereof for dyeing. The present invention uses a cross-sectional image group including two or more cross-sectional images in which the cochlea inside surface and the Reissner's membrane appear, and determines a plane approximating one section of the surface shape of each of the two tissues inside a three-dimensional space. It is possible to use an optical flow method in order to determine said plane.

Description

Method for noninvasively determining an angle formed by a structure inside a living body

The present invention generally relates to analysis of tomographic images, and more particularly, based on the tomographic images, determines the relative orientation (eg, relative angle) of two structures within an object such as a living body. The present invention relates to a computer program, an apparatus, a method, and the like.

聴覚 The cochlea, the organ that controls hearing, is an organ that exists in the inner ear, is surrounded by hard bones, and is divided into two compartments with different ionic compositions. A three-story tube is wound like a snail, and the central part is called a membrane maze (or the central floor) in one compartment, and a liquid with a high potassium concentration called endolymph is inside. In. The three-story upper (vestibule floor) and lower (tymroom floor) are continuous compartments connected by the tip of the cochlea, filled with a liquid with a low potassium concentration equivalent to normal extracellular fluid. The boundary between the central floor and the vestibular floor is the Ricenell membrane, which is normally stretched straight.

Meniere's disease is an intractable disease that repeats dizziness and hearing loss, and in the most severe cases, hearing may be abolished. The cause has not yet been determined, but the endolymphatic edema has been observed as a result of a marked enlargement of the endolymph when preparing a postmortem temporal bone specimen from a patient with Meniere's disease. It is assumed to be deeply involved in the pathology. However, it has not been proved because there is no direct diagnosis of endolymphatic hydrops in living humans or animals. So far, diseases associated with endolymphatic edema have been considered, such as low-frequency sensorineural hearing loss, Lermoye syndrome, delayed endolymphatic edema.

Research on endolymphatic edema is necessary. To do so, it is necessary to observe endolymphedema in humans and model animals. However, the inner ear is in a hard bone, and it is difficult to evaluate endolymphatic edema in the form of a cochlea even if it is removed after death as well as being alive in humans and animals. There was no method that could evaluate the form of endolymphatic edema other than thinning the excised specimen (Non-patent Document 1).

Optical coherence tomography (Optical Coherence Tomography: OCT) is a technology that uses the fact that near-infrared rays are not easily absorbed by living tissue, and detects tomographic images with high-sensitivity detection of the reflected infrared rays inside the tissue. It is. The present inventors have shown that the internal structure including endolymphatic edema can be depicted without destroying the cochlea of a living mouse (Non-patent Document 2). Moreover, a dental diagnostic apparatus using OCT is known (Patent Document 1).

Utility Model Registration No. 3167523

As described above, conventionally, in order to evaluate endolymphatic edema, it was necessary to prepare a section. However, since the evaluation by the section is an evaluation in a state different from the state having the function, there is a possibility that the original disease state is not reflected. In addition, since the evaluation is focused on a specific cross section where the section can actually be made, the state of endolymphatic edema may not be evaluated correctly depending on how the section is made, such as cutting at an angle. there were. In order to avoid that, in the conventional method, it was necessary to take great care to make a section through the cochlear shaft. Furthermore, according to the prior art, it is difficult to correctly evaluate endolymphatic edema as an area ratio between the vestibular floor and the central floor unless at least one entire cross section of the osseous cochlear duct is observed.

In contrast to the above-mentioned problems of the prior art, the present invention makes it possible to evaluate endolymphatic edema from a living cochlea without removing the cochlea and performing solvent treatment for sectioning and staining. It is an object to provide a computer program, an apparatus, a method, and the like.

In order to solve the above-mentioned problems, the present inventors diligently studied the possibility of evaluating endolymphatic edema based on cochlear tomographic images in live animals. In that process, we paid particular attention to OCT that can acquire tomographic images from living bodies with minimally invasive, real-time and high resolution. However, it is necessary to observe a sufficient depth from the inner ear surface (especially the depth to the cochlea axis) to ensure that the evaluation is at an appropriate angle cut, whereas OCT Then there was a problem that the depth of penetration was limited. It was also difficult to draw an image from the surface of the cochlear bone capsule in a cross section that accurately passes the cochlear axis. Therefore, the present inventors have further studied, and based on a plurality of tomographic images obtained by using OCT, the outer shapes of both tissues in a virtual space in the vicinity of a certain position where the inner surface of the cochlea and the Ricenell membrane are in contact with each other. We came up with a methodology for evaluating the deviation of the Risnel film by approximating each plane and obtaining the angle between these two planes as the relative angle between the two tissues.
Furthermore, the inventors have found that the optical flow method used for mobile tracking can be applied for this methodology. That is, in general, in moving object tracking, it is assumed that a moving object always appears in the vicinity in temporally continuous images in a moving image. Similarly, in the group of tomographic images used in the medical field, in the adjacent tomographic image (spatial continuous image), one continuous object should always appear in the vicinity of the image. A point in the vicinity of the outline of each tissue in one image can be associated with a specific point in the image adjacent to the tissue. By approximating the plane passing through the series of points thus obtained, the target plane can be determined.
Based on such findings, the present inventors have conducted further research and completed the present invention.

That is, the present invention is as follows.
[1] A computer program for determining the relative orientation of two structures in the vicinity of each other in an object,
Computer
(M1) means for receiving a tomographic image group including at least two tomographic images in which position information in the object is associated and the two structures appear; and
(M2) The computer program for causing a computer to function as means for determining a plane that approximates a part of the surface shape of each of the two structures in a three-dimensional space based on the tomographic image group.
[2] The means (M2) calculates an optical flow between at least one set of tomographic images included in the tomographic image group, and the positional information associated with each of the at least one set of tomographic images and The computer program according to [1], wherein the computer program is configured to determine the plane using the calculated optical flow.
[3] The means (M2)
Displaying at least one tomographic image included in the tomographic image group on an image display means;
Using the displayed tomographic image or a plurality of tomographic images, for each of the two structures, the user selects at least two points in the vicinity of the contour line, receives the selected points,
For each of the selected points, identify a point corresponding to the selected point in a tomographic image close to the selected tomographic image; and
Configured to determine, for each of the two structures, a plane that approximates a portion of the surface shape based on the selected point and the identified point;
The computer program according to [1] or [2] above.
[4] Any of the above [1] to [3], wherein the means (M1) includes means for determining the tomographic image group from a plurality of given tomographic images based at least in part on a user input. A computer program according to the above.
[5] Further computer
(M3) determining a straight line intersecting the two planes determined by the means (M2) in the three-dimensional space, calculating a rotation angle from one plane to the other plane with respect to the determined straight line; and The computer program according to any one of the above [1] to [4], which functions as means for outputting the calculated rotation angle as a relative angle between the two structures.
[6] A device for determining the relative orientation of two structures in the vicinity of each other in an object, and the computer program according to any one of [1] to [5] functions An apparatus comprising: a computer configured as described above.
[7] (M4) The apparatus according to [6], further including means for acquiring a tomographic image in the object.
[8] The apparatus according to [7], wherein the means (M4) is an optical coherence tomography device.
[9] The apparatus according to [8], wherein the optical coherence tomography device is configured to acquire a tomographic image of the inner surface of the cochlea and the Risnel membrane.
[10] A method for determining the relative orientation of two structures in the vicinity of each other in an object,
(S1) preparing a tomographic image group including at least two tomographic images in which positional information in the object is associated and the two structures appear; and
(S2) including a step of determining a plane that approximates a part of the surface shape of each of the two structures in a three-dimensional space based on the tomographic image group, and based on the determined two planes, Said method wherein the relative orientation is determined.
[11] The step (S2) calculates an optical flow between at least one set of tomographic images included in the tomographic image group, and the positional information associated with each of the at least one set of tomographic images; The method according to [10] above, including determining the plane using the calculated optical flow.
[12] The step (S2)
Displaying at least one tomographic image included in the tomographic image group on an image display means;
Using the displayed tomographic image or a plurality of tomographic images, for each of the two structures, select at least two points near the contour line;
For each of the selected points, identify a point corresponding to the selected point in a tomographic image close to the selected tomographic image; and
Determining, for each of the two structures, a plane that approximates a portion of a surface shape based on the selected point and the identified point;
The method according to [10] or [11] above.
[13] The method according to any one of [10] to [12] above, wherein the tomographic image is acquired using an optical coherence tomography device.
[14] (S3) A straight line intersecting the two planes determined in the step (S2) in the three-dimensional space is determined, and a rotation angle from one plane to the other plane with respect to the determined straight line is calculated. The method according to any one of [10] to [13], further including the step of determining the calculated rotation angle as a relative angle between the two structures.
[15] The method according to any one of [10] to [14] above, wherein the two structures are a cochlear inner surface and a Ricenell membrane.
[16] including a step of determining a relative angle between the inner surface of the cochlea and the Ricenel membrane by using the method according to [14] above with the inner surface of the cochlea and the Ricenel membrane as the two structures, A method for assisting in the diagnosis of endolymphatic hydrops in a mammal, wherein endolymphatic hydrops are diagnosed.

According to the present invention, it is not necessary to acquire data parallel to the cochlear axis, and the inner surface of the cochlea and the Ricenell membrane are distorted on the cut surface in any direction. Even if it is visible at a certain angle, a correct relative angle can be calculated by approximating a part of the surface shape of both tissues three-dimensionally. In addition, according to the present invention, even if there is no tomographic image having a depth up to the cochlea axis, it is possible to evaluate the shift of the Ricenell film without any problem using a partial image outside the cochlea. Therefore, by using the present invention, it is possible to evaluate endolymphatic edema based on the cochlea of a living animal. Furthermore, the procedure was complicated in the prior art (for example, it takes several hours for fixation in mice, 7 days for decalcification, several days for sectioning and staining, and more than one month for fixation in humans) In the present invention, since it can be easily digitized after image acquisition and analyzed on a computer, it is excellent in simplicity, rapidity, and economy. The analysis program can be configured using an existing library that is generally available.
The use of the present invention makes it possible to diagnose endolymphatic edema in vivo, so that medication or stimuli that cause inner ear damage (for example, loud sounds or barotrauma) cause endolymphatic edema. Observe in real time. By using the present invention, it becomes possible to observe the state of temporary endolymphatic edema that could not be diagnosed until now, and it becomes a tool for clarifying the pathophysiology of endolymphatic edema that has not been known so far. . The present invention can also be used to evaluate animal experimental models for the development of treatments for endolymphatic edema.
The present invention can be applied not only to a cochlea but also to any tissue or structure in a target (eg, a living body) from which a tomographic image can be acquired.

It is a figure for demonstrating "relative orientation" and "relative angle" in this invention. (B) is an enlarged view of the vicinity of the point A in (a). FIG. 2 is a diagram illustrating a configuration of an apparatus of the present invention according to an exemplary embodiment. FIG. 6 shows a flowchart of a procedure for determining relative orientation according to the present invention according to an exemplary embodiment. It is an example of the image which applied the computer program of this invention to the OCT tomographic image of a mouse cochlea. It is a graph which shows the angle of a Ricenell membrane and the cochlea inner side surface computed by the method of this invention.

Hereinafter, the present invention will be described in more detail.

(Definition)

In this specification, examples of the “object” include a living body, but are not limited thereto. The living body includes, for example, animals (for example, mammals such as humans, dogs, cats, monkeys, horses, sheep, goats, mice, rats, rabbits, and other vertebrates) and plant bodies. The target object has at least two structures that are targets for determining the relative orientation, and may have other components.

In the present specification, the “structure” is not particularly limited as long as it is a physical entity having a specific shape. The structure is, for example, a living tissue. The structure may be inside the object or may form a part of the surface of the object. In the present invention, “two structures close to each other”, which are targets for determining the relative orientation, are close enough that at least a part of each structure appears in one tomographic image. If it is, it will not specifically limit. The two structures may be in contact with each other.

More specifically, examples of the two structures include a cochlea inner surface and a Ricenell membrane. As mentioned above, it has been suggested that Meniere's disease, bass-sensitive sensorineural hearing loss, Lermoye syndrome, delayed endolymphatic hydrops, etc. are associated with endolymphatic hydrops. The relative angle change between the cochlear inner surface and the Ricenell's membrane is observed due to significant expansion. By determining the relative orientation between the cochlear inner surface and the Ricenell's membrane, endolymphatic edema, and the above mentioned Evaluation and medical insights about such diseases can be obtained. Therefore, the inner surface of the cochlea and the Ricenell membrane are preferred objects as the two structures.
Alternatively, for example, when a tube is arranged inside the object and a valve is fixed to the inner wall of the tube, the object is destroyed by obtaining a tomographic image around the fixed part of the inner wall of the tube and the valve. The angle between the inner wall of the tube and the valve can be determined non-invasively.

As used herein, “relative orientation” refers to the relationship between the orientation of one structure and the orientation of the other structure. Usually, in the space around the position where both structures are close to each other, It can be grasped as the direction in which the surface formed by the outer shape of one structure extends with respect to the surface formed by the outer shape of the other structure. The relative orientation may be given as a quantitative one (eg, a relative angle defined as described below), a qualitative one, or a visual one. It may be given as information (for example, display on a display of a three-dimensional model serving as an indicator of the relative orientation of both structures).

The relative angle between the two structures can be defined as follows.
For example, consider a case where two structures T ₁ and T ₂ are in contact with each other, and therefore a line 1 is formed where the outer shapes of both structures are in contact with each other (FIG. 1). Focusing on one point A on the line l, in the vicinity of the point A can be approximated to a respective outer planar S ₁ and S ₂ of the two structures, a line profile come into contact with each other can be approximated as a straight line l '. The planes S ₁ and S ₂ are preferably tangent planes of the structures T ₁ and T ₂ respectively passing through the point A. A rotation angle θ from the plane S ₁ to the plane S ₂ with respect to the straight line l ′ can be defined as a relative angle between both structures near the point A. The planes S ₁ and S ₂ and the rotation angle θ can be determined according to the invention.
Note that although the relative angle obtained according to the selection of the point A on the line l varies, those skilled in the art can appropriately determine the position to which attention should be paid according to the object and purpose of analysis. Moreover, since the shape change of the structure in the living body is generally smooth, a constant relative angle can be obtained in a minute space regardless of the position of interest. Therefore, instead of selecting the point A on the line 1 where the outer shapes of the two structures are in contact with each other, as shown in FIG. 1, the outer shape of the structure T _{1 in} the vicinity of the point A (or near it). the point _{a 1,} select a point _{a 2} of the outer shape of the structure _{T 2} in the vicinity of the point a (or near) the structure _{T 1} and T in the vicinity of each point _a 1, _{a 2} _The planes S ₁ and S ₂ that approximate the outer shape of ₂ may be determined.
Alternatively, if necessary, the relative angles of the two structures may be determined by selecting a plurality of points and obtaining the relative angles, and then averaging the obtained relative angles.
In addition, even when the two structures T ₁ and T ₂ are close to each other but are not in contact with each other, for example, a point A near the two structures is determined. select one point a ₁ on the outer shape of the structure T ₁ (or near) in the vicinity, the outer shape of the structure T ₂ in the vicinity of the point a (or near) 1 point a ₂ of each Determine planes S ₁ and S ₂ that approximate the outer shapes of the structures T ₁ and T ₂ in the vicinity of the points A ₁ and A ₂ , and further extend both planes so that the two generated planes intersect each other Thus, the relative angle can be defined in the same manner as in the above case where both structures are in contact.

In the present specification, the “tomographic image” may be obtained by any means as long as the analysis according to the present invention is possible. As means for acquiring a tomographic image, for example, optical coherence tomography (OCT), nuclear magnetic resonance imaging (MRI), X-ray computed tomography (X-ray CT), or ultrasonic imaging (echo) is used. Things.
Among these, a means using OCT is preferable because it has low invasiveness to a living body and can acquire a high-resolution tomographic image. OCT may be time-domain OCT (time-domain OCT, TD-OCT) or Fourier-domain OCT (Fourier-domain OCT, FD-OCT). FD-OCT is preferable because data can be acquired. Examples of FD-OCT include spectral domain OCT (Spectral-domain OCT, SD-OCT) and wavelength sweep type OCT (Swept source OCT, SS-OCT). SS-OCT is exemplified as a preferred one. A commercially available device can be used as the OCT device or a device for the various means described above.

The tomographic image can be acquired by a normal procedure according to a standard textbook, a manual of the equipment to be used, or the like. For example, the procedure for obtaining a cochlear tomographic image using an OCT device is described in the examples below and Non-Patent Document 2 above, and the following settings can be used, for example: 800-1400nm (1300nm in the example), scan frequency up to 200KHz (16kHz in the example, but slower if the target doesn't move). A method of irradiating an object with laser light using an objective lens (used in the examples), or a double-structured fiber that rotates only the inner one, and the light that passes through the inner fiber is reflected by the tip lens. For example, a method of obtaining an OCT image by irradiating from an outer window can be used.
In the case where invasiveness becomes a more serious problem, such as when the subject is a human, for example, an OCT probe may be inserted through the middle ear after the eardrum has been excised to obtain an image.

The “tomographic image group” for use in the present invention includes at least two tomographic images in which positional information in the object is associated and the two structures appear.
The above-mentioned “positional information in the object” is information indicating which position of the object the tomographic image represents the tomographic plane, and therefore each point in the tomographic image is in (or outside of) the object. It is known which position corresponds to. The position may be relative, that is, the relative positional relationship between any one point in one image included in the group of tomographic images and any one point in the other image (ie, the length of the vector). And the direction) may be known.
“Two structures appear” means that at least a part of each structure is seen in the tomographic image.
The number of tomographic images included in the tomographic image group varies depending on a specific analysis target, an image set, and the like, but is at least two, preferably at least three. The upper limit of the number of images is not limited as long as different images can be actually acquired and can be used for processing according to the present invention, but is 512, for example. In addition, it is preferable that the plurality of images included in the tomographic image group have a good signal-noise ratio and that there is no large positional shift between the images.

定義 Definitions or explanations of other terms are given below as required.

(Computer program and device)

Next, a computer program (hereinafter also referred to as “the computer program of the present invention”) and an apparatus (hereinafter referred to as “the apparatus of the present invention”) for determining the relative orientation of the two structures according to the present invention. Will be explained. In the following, the apparatus of the present invention will be described in particular with reference to the drawings. However, the computer program of the present invention is a computer program for executing the functions of the apparatus of the present invention, and is configured according to the description of the apparatus of the present invention. can do.

FIG. 2 is a diagram illustrating the configuration of the apparatus of the present invention according to an exemplary embodiment. The apparatus of the present invention comprises (M1) means for receiving a tomographic image group including at least two tomographic images in which positional information in the object is associated and the two structures appear, and (M2) Based on the group of tomographic images, the apparatus has at least means for determining a plane that approximates a part of the surface shape of each of the two structures in a three-dimensional space. In a preferred embodiment, it is shown in FIG. As described above, based on the determination of the plane in the three-dimensional space by the means (M2), the relative orientation of the two structures is calculated as the relative angle between the two structures, and the above-described output for outputting it (M3) may be further included. The means (M1), (M2) and (M3) may be implemented in one computer as shown in FIG. 2, or are configured to be executed using a plurality of computers. May be.
Similarly, as shown in FIG. 2, (M4) means for acquiring a tomographic image in the object (tomographic image acquiring means) may be included in the configuration of the apparatus of the present invention. As the means (M4), those described above in this specification can be used. Furthermore, various input means and output means (for example, a display, a printer, etc.) not shown may also be components of the apparatus of the present invention.

The above means (M1) is a means for receiving a tomographic image group to be subjected to analysis by the means (M2), and the definition of each term is as described above. The tomographic image group may be received by reading data of a plurality of previously acquired tomographic images recorded in an auxiliary storage device such as a hard disk drive or an external storage device (eg, CD-ROM). Alternatively, the tomographic image acquired by the tomographic image acquiring means (M4) may be received in real time.

The means (M1) selects the tomographic image group from a plurality of given tomographic images based at least in part on a user input so that an appropriate tomographic image group is selected for analysis by the means (M2). It is preferred to have a means for determining. For example, such means displays a plurality of candidate tomographic images on the image display means, and selects at least one image to be used for analysis by the means (M2) from the displayed images. Can be configured. The means may be configured such that all images used by the means (M2) are selected by the user, or based on an image selected by the user, a preset reference (for example, already Another image may be selected automatically or semi-automatically (that is, presentation and confirmation to the user) according to a criterion in terms of the positional relationship of the selected image with respect to the tomographic plane).

The means (M2) is a means for determining a plane that approximates a part of the surface shape of each of the two structures in the three-dimensional space based on the tomographic image group determined by the means (M1). is there. The planes may be the planes S ₁ and S ₂ described above with reference to FIG. These planes are planes that approximate the outer shapes of the structures T ₁ and T ₂ in the vicinity of the point of interest A (or two points A ₁ and A ₂ ), respectively. Such plane determination based on a plurality of tomographic images can be performed using a technique known per se. Examples of such a technique include, but are not limited to, a method described later based on an optical flow method, a method of creating three-dimensional volume data from a tomographic image, extracting a continuous object therefrom, and performing plane approximation, and the like. It is done. In particular, the method described later based on the optical flow method is preferable from the viewpoint of high-speed processing and high reliability of results.
Further, the means (M2) reduces image noise by applying generally used filtering such as a bilateral filter and a column filter to the tomographic image group used for analysis before performing the process of determining the plane. It is preferable that it is comprised.

The determination of the plane based on the optical flow method is performed by the following general procedure: calculating an optical flow between at least one set of tomographic images included in the tomographic image group, and determining the at least one set of tomographic images. The plane can be determined using the position information associated with each of the positions and the calculated optical flow.
As described above, the method is based on the assumption that in the tomographic images close to each other, each structure should always appear in the vicinity of the image. Therefore, the set of tomographic images is an image representing the tomographic planes close to each other. It is preferable that In a particularly preferred embodiment, the set of tomographic images are adjacent tomographic images parallel to each other. In this specification, tomographic images that are close to or adjacent to each other means that the tomographic planes represented by the tomographic images are close to or adjacent to each other.
When the above assumption holds, each point of each structure in one of the set of tomographic images may be associated as being continuous with any point in the structure in the other tomographic image. it can. An optical flow (that is, a distribution of flow vectors) can be calculated by representing a position where each point in one image has moved in the other image by a series of vectors (flow vectors). In this specification, calculating the optical flow between a set of tomographic images means calculating a flow vector map over the entire image by calculating the above flow vector for all points in one image. It does not necessarily need to be generated and means to determine at least one flow vector.

The calculation of the optical flow can be performed using, for example, various methods used in the field of moving object tracking in a moving image. Examples of such a method include a pattern recognition-based method, Particle Filter, Mean Shift tracking, gradient method, and the like.

In order to determine a plane that approximates a part of the surface shape of each structure, a flow relating to a point in the tomographic image near the outline of each structure (that is, a point on the outline or a point in the structure near the outline) It is necessary to determine the vector. In general, since more than two points are needed to determine the plane, at least one point (in this case, in addition to the point and one point identified as corresponding to the point in a different image) Need another point in the vicinity of the line), more preferably the flow vector needs to be determined at at least two points (thus identifying the at least two points and corresponding those points in different images) At least 2 points, and thus a total of at least 4 points). The plurality of points selected for determining the plane may be points in one image or points in different images.

Such a point can be acquired, for example, by allowing the user to select a corresponding point in the tomographic image displayed on the image display means, but a more automated method may be used. Accordingly, the means (M2) causes the image display means to display at least one tomographic image included in the tomographic image group, and in the displayed image, at least two points in the vicinity of the contour line for each structure of interest are displayed by the user. The selected point may be received and the selected point may be received. The means (M2) applies the above-described method to each of the points selected in this way, and corresponds the selected point in the tomographic image close to the selected tomographic image. May be configured to identify points to do. The adjacent tomographic images may be selected by the user based on position information associated with each tomographic image, or automatically or semi-automatically (ie, presented to the user) by means (M2) based on the position information. And confirmation).

The means (M2) is, for example, selected as described above, at least two points in the vicinity of the outline of the structure of interest in one tomographic image, and at least two points specified to correspond to them in another tomographic image Based on the above, a plane that approximates a part of the surface shape of the structure is determined. The set of points used for the determination of the plane is preferably at least mostly in the vicinity of the outline of the structure. The plane is determined by approximating one plane passing through all of these points (the positions of all the points are known in a three-dimensional space) using, for example, the least square method. Can do. By performing the same procedure for each of the two structures, two target planes are determined.

The above means (M3) is a means for calculating the relative angle between the two planes determined by the means (M2) and outputting it. The definition of relative angle is given above, and the skilled person can determine the angle based on the two planes.

(Method)
The present invention also provides a method for determining the relative orientation of two structures in the vicinity of each other in an object (hereinafter also referred to as the method (I) of the present invention). FIG. 3 shows a flowchart of an exemplary embodiment of the method. The method (I) of the present invention includes at least the steps (S1) and (S2) described above, and as shown in FIG. 3, in a preferred embodiment, the structure is based on the two planes determined by the step (S2). The step (S3) for determining the relative angle between the bodies is included. Further, (S4) a step of acquiring a tomographic image in the object may be included in the method of the present invention.

The above step (S1) is a step of preparing a tomographic image group to be subjected to the analysis in the step (S2), and the definition of each term is as described above. The preparation of the tomographic image group may be to read out data of a plurality of tomographic images acquired in advance and recorded in an auxiliary storage device such as a hard disk drive or an external storage device (eg, CD-ROM). Alternatively, the tomographic image acquired in step (S4) may be received in real time.

In step (S1), in order to select an appropriate tomographic image group for analysis in step (S2), the tomographic image group is obtained from a plurality of given tomographic images based at least in part on a user input. It is preferable to have a step of determining. Such a process can be carried out according to the above description relating to the means (M1) of the apparatus of the present invention.

The step (S2) is a step of determining a plane that approximates a part of the surface shape of each of the two structures in the three-dimensional space based on the tomographic image group determined in the step (S1). is there. The planes may be the planes S ₁ and S ₂ described above with reference to FIG. These planes are planes that approximate the outer shapes of the structures T ₁ and T ₂ in the vicinity of the point of interest A (or two points A ₁ and A ₂ ), respectively. Such plane determination based on a plurality of tomographic images can be performed using a technique known per se, for example, using the means (M2) of the apparatus of the present invention.

The above step (S3) is a step of calculating the relative angle between the two planes determined in step (S2). The definition of relative angle is given above, and the skilled person can determine the angle based on the two planes.

The present invention also provides a method for assisting diagnosis of endolymphatic hydrops in a mammal (hereinafter also referred to as the method (II) of the present invention). The method evaluates endolymphatic edema based on the relative orientation obtained by applying the method (I) of the present invention for the determination of the relative orientation of the cochlear inner surface and the Ricenell membrane. Can be done. Endolymphatic edema is considered to be present in various diseases and to various extents, and therefore quantitative evaluation is preferable. Therefore, the relative orientation is preferably a relative angle between both tissues. Here, the evaluation of endolymphatic edema based on the relative angle is, for example, the relative angle calculated by the method (I) of the present invention based on the tomographic image of the cochlea from an individual that is known not to have endolymphedema. This can be done by comparing the relative angle calculated by the method (I) of the present invention based on the tomographic image of the cochlea from the individual to be tested. In addition, based on tomographic images of the cochlea of a large number of normal individuals and a large number of individuals with endolymphatic edema, the threshold value or normal range of the relative angle that serves as an index of endolymphatic hydrops is determined, and the individual to be tested is determined. The relative angle may be evaluated by comparing it with the threshold or the normal range. In that case, the degree of endolymphatic edema may be determined using the threshold or the magnitude of deviation from the normal range.

Hereinafter, the present invention will be described more specifically with reference to examples. However, the present invention is not limited to the following examples.

Example 1: Determination of the relative angle between the inner surface of the cochlea and the Ricenel membrane in normal mice and endolymphatic hydrops model mice from OCT images

animal:
Normal mice and Slc26a4 knockout mice were used. Slc26a4 knockout mice are known to cause severe endolymphatic edema.
Method:
(1) Acquisition of OCT image Analysis was performed on living organisms, excised, and excised decalcified cochlea.
[In the case of a living cochlea] General anesthesia is performed by intraperitoneal administration of an anesthetic, and when a thin bone called a bra covering the cochlea is removed by incision behind the ear, the cochlea bone is exposed. The blood was wiped off and a small amount of water or glycerol was dropped to leave the surface moist and smooth and left under the OCT probe to obtain an image.
[In the case of an isolated cochlea] After fixation with open heart return, the inner ear was removed and placed under the OCT probe to obtain an OCT image.
[In the case of demineralized cochlea] In addition to fixation, the inner ear was decalcified by shaking in 10% EDTA at 4 ° C for 7 days, and then placed under the OCT probe to obtain an image.
As the OCT system, an OCS1300SS OCT system (manufactured by Thorlabs, USA) was used, and measurement was performed under the conditions of a center wavelength of 1300 nm, a spectral bandwidth of 100 nm, and an average output of 10 mW.
(2) Angle measurement Data was read with a special program, noise reduction processing was performed with a bilateral filter and column filter, and images that were significantly displaced were removed, and then displayed as a series.
Select one cross-section, and point 1 on the inner surface of the cochlea in the second rotation of the cochlea (referred to as point A), two points on the inner surface of the cochlea near point A, and two points on the Risnel film near point A Specified and used the optical flow method (more specifically, optical flow estimation based on template matching, which uses a small area in an image itself as a feature quantity and searches for a similar small area between different images). From the images, the surface in contact with the inner surface of the cochlea at point A and the surface in contact with the Ricenell membrane near point A through point A were identified by the least square method, and the angle between the two surfaces was calculated. The angle was obtained by the above calculation at three points randomly selected for one cochlea and the average was calculated.
FIG. 4 shows an image displaying the intersection line between the two planes identified in the selected tomographic image. (a) is an image of a normal mouse and (b) is an image of a knockout mouse.

result:
The angles calculated for normal mice (n = 4) and Slc26a4 knockout mice (n = 3) are shown in Table 1 below.

Furthermore, the average of the calculated angles for each of normal mice (n = 4) and Slc26a4 knockout mice (n = 3) is shown in FIG. The obtained angle was significantly smaller in normal mice (n = 4) and Slc26a4 knockout mice (n = 3) in knockout mice (p = 0.0002).

This application is based on Japanese Patent Application No. 2014-053790 filed in Japan (filing date: March 17, 2014), the contents of which are incorporated in full herein.

Claims

A computer program for determining the relative orientation of two structures in the vicinity of each other in an object,
Computer
(M1) means for receiving a tomographic image group including at least two tomographic images in which position information in the object is associated and the two structures appear; and
(M2) The computer program for causing a computer to function as means for determining a plane that approximates a part of the surface shape of each of the two structures in a three-dimensional space based on the tomographic image group.
The means (M2) calculates an optical flow between at least one set of tomographic images included in the tomographic image group, and calculates the positional information associated with each of the at least one set of tomographic images. The computer program according to claim 1, wherein the computer program is configured to determine the plane using the optical flow.
The means (M2)
Displaying at least one tomographic image included in the tomographic image group on an image display means;
Using the displayed tomographic image or a plurality of tomographic images, for each of the two structures, the user selects at least two points in the vicinity of the contour line, receives the selected points,
For each of the selected points, identify a point corresponding to the selected point in a tomographic image close to the selected tomographic image; and
Configured to determine, for each of the two structures, a plane that approximates a portion of the surface shape based on the selected point and the identified point;
The computer program according to claim 1 or 2.
The means (M1) according to any one of claims 1 to 3, wherein the means (M1) comprises means for determining the tomographic image group from a plurality of given tomographic images based at least in part on user input. Computer program.
Computer
(M3) determining a straight line intersecting the two planes determined by the means (M2) in the three-dimensional space, calculating a rotation angle from one plane to the other plane with respect to the determined straight line; and The computer program according to any one of claims 1 to 4, which functions as means for outputting the calculated rotation angle as a relative angle between the two structures.
An apparatus for determining the relative orientation of two structures in the vicinity of each other in an object, wherein the computer program according to any one of claims 1 to 5 is configured to function. Said apparatus comprising a computer.
(M4) The apparatus of Claim 6 which further has a means for acquiring the tomographic image in a target object.
The apparatus according to claim 7, wherein the means (M4) is an optical coherence tomography device.
The apparatus according to claim 8, wherein the optical coherence tomography device is configured to acquire tomographic images of the inner surface of the cochlea and the Risnel membrane.
A method for determining the relative orientation of two structures in the vicinity of each other in the object,
(S1) preparing a tomographic image group including at least two tomographic images in which positional information in the object is associated and the two structures appear; and
(S2) including a step of determining a plane that approximates a part of the surface shape of each of the two structures in a three-dimensional space based on the tomographic image group, and based on the determined two planes, Said method wherein the relative orientation is determined.
The step (S2) calculates an optical flow between at least one set of tomographic images included in the group of tomographic images, and calculates the position information associated with each of the at least one set of tomographic images. The method of claim 10, comprising determining the plane using an optical flow.
The step (S2)
Displaying at least one tomographic image included in the tomographic image group on an image display means;
Using the displayed tomographic image or a plurality of tomographic images, for each of the two structures, select at least two points near the contour line;
For each of the selected points, identify a point corresponding to the selected point in a tomographic image close to the selected tomographic image; and
Determining, for each of the two structures, a plane that approximates a portion of a surface shape based on the selected point and the identified point;
The method according to claim 10 or 11.
The method according to any one of claims 10 to 12, wherein the tomographic image is acquired using an optical coherence tomography device.
(S3) determining a straight line where two planes determined by the step (S2) intersect in the three-dimensional space, calculating a rotation angle from one plane to the other plane with respect to the determined straight line; and The method according to any one of claims 10 to 13, further comprising the step of determining the calculated rotation angle as a relative angle between the two structures.
The method according to any one of claims 10 to 14, wherein the two structures are a cochlea inner surface and a Ricenell membrane.
15. Using the method of claim 14 with the cochlear inner surface and the Ricenell membrane as the two structures, determining the relative angle between the cochlear inner surface and the Ricenell membrane, based on the relative angle, the endolymph A method of assisting in the diagnosis of endolymphatic hydrops in a mammal, wherein edema is diagnosed.