JP6836250B2 - Bone quality assessment device, method and program - Google Patents

Description

The present invention relates to a bone quality evaluation device, a method and a program, and a fracture risk evaluation device, a method and a program, and in particular, based on an X-ray image (or also referred to as an “X-ray image”) of the subject's lower jaw bone. Bone quality assessment devices, methods and programs for assessing bone quality and fracture risk assessment devices, methods and programs for assessing the risk of osteoporotic fractures (also referred to as "fragile fractures") in the subject.

Osteoporosis was once defined as a disease in which bone density decreases and the risk of fracture increases. For this reason, bone density has been the main consideration in the prevention and treatment of osteoporosis. However, despite the fact that the bone density is within the normal range, there are many cases of fractures, and subsequent studies have shown that the risk of fracture is related not only to bone density but also to bone quality. Along with this, the current definition of osteoporosis is that it is a disease in which bone strength decreases and the risk of fracture increases. It is said that bone strength is related to the ratio of bone density of about 70% and bone quality of about 30%. Furthermore, bone quality is roughly divided into structural characteristics and material characteristics. Of these, structural characteristics include bone size, bone morphology, bone microstructure, etc., and material characteristics include bone matrix, bone turnover, and microinjury. It is said that factors such as (microfracture) and the degree of calcification of bone tissue are factors.

Here, as a technique for evaluating (inspecting) bone density, for example, a dual-energy X-ray Absorptiometry (hereinafter referred to as "DXA") method is known, and "treatment of osteoporosis". According to the Prevention Guideline 2015, it is desirable to evaluate each of the lumbar spine and the proximal femoral bone by the DXA method.

On the other hand, as a technique for evaluating bone quality, for example, one disclosed in Non-Patent Document 1 is known. According to the technique disclosed in Non-Patent Document 1, the variation in the density of each pixel in each X-ray image of the lumbar spine and the femoral neck by the DXA method is obtained by a function called an experimental variogram, and further, this A value called TBS (Trabecular Bone Score) can be obtained from the calculation result of the variation. This TBS correlates with the bone quality and, strictly speaking, with the cancellous bone microstructure which is one of the elements of the bone quality. Therefore, the cancellous bone microstructure can be indirectly evaluated from this TBS. Specifically, it is evaluated that the larger the TBS, the tougher the cancellous bone microstructure is, and the smaller the TBS, the more fragile the cancellous bone microstructure is.

Further, in Non-Patent Document 2, it is not possible to accurately evaluate the fracture risk only by the bone density, and in order to accurately evaluate the fracture risk, it is desirable to be based on the combination of the bone density and TBS. That is disclosed. For example, even if the bone density is within the normal range, if the TBS is small, the risk of fracture is not small and may be, for example, osteoporosis or at least one step before osteopenia. In other words, even if the bone density is the same, the risk of fracture may not be the same, and TBS is said to be useful in identifying this.

Furthermore, Non-Patent Document 3 discloses that the ability of TBS to discriminate osteoporotic vertebral body fractures is also suitable for elderly Japanese women. As a conclusion, it is said that TBS showed a fracture discriminating ability independent of bone mineral density, and its usefulness was suggested especially in cases where the bone mineral density of the lumbar spine was relatively high.

Laurent Pothuaud, 2 others, "Correlations between gray-level variations in 2D projection images (TBS) and 3D microarchitecture: Applications in the study of human trabecular bone microarchitecture", Bone 42 (2008), Elsevier BV, January 29, 2008 Sun, p. 775-787 Didier Hans, "The ABC of Trabecular Bone Score (TBS): From Theory to Clinical Practice", Lausanne University Hospital Kensuke Tanaka, "Ability to Discriminate Osteoporotic Vertebral Fractures by Cancellous Bone Score: Examination in Elderly Japanese Women," Kawasaki Medical School, 2015, No. 41 (2), p. 129-141

However, the prior art disclosed in Non-Patent Documents 1 to 3 described above obtains a value of TBS from an X-ray image obtained by the DXA method and evaluates bone quality from this TBS. That is, these conventional techniques are based on the premise that a relatively large-scale inspection called the DXA method is performed. Therefore, the emergence of a new technique capable of evaluating bone quality by a simpler test and, by extension, evaluating the risk of fracture is desired.

Therefore, an object of the present invention is to provide a novel technique capable of evaluating bone quality by a simpler examination than before and, by extension, evaluating the risk of fracture.

The inventor of the present invention is also the inventor of the invention according to Japanese Patent No. 4077430, and the invention according to the same publication that evaluates the bone density of a subject (human) based on an X-ray image of the lower jaw bone of the subject. Based on the above, it was considered that the bone quality of the subject could be evaluated based on the X-ray image, and it was confirmed by clinical experiments that this idea was valid, and then the present invention was developed. completed. Specifically, the inventor of the present invention determines that the relative ratio between the variation in the shading of a specific region corresponding to a specific portion of the mandibular bone in the X-ray image of the subject's mandibular bone and the average value of the shading is the same. It was confirmed by clinical experiments that it correlates with the bone quality of the subject as described later, and then a technique for evaluating the bone quality of the subject based on this bone quality correlation value was created, and the present invention was completed.

First, the present invention is a bone quality evaluation device that evaluates the bone quality of the subject based on an X-ray image of the mandible of the subject, and includes a bone quality calculation means and a bone quality evaluation means. Bone quality calculating means, bony the relative ratio of the average value of the gray level of the standard deviation and the specific area of the shading of a particular region of a specific region corresponding to a specific portion of the mandible in the X-ray image of the mandible of the subject Obtained as a correlation value. In this case, the bone quality evaluation means evaluates the bone quality based on the relative ratio and the bone quality threshold. As the relative ratio referred to here, for example, a coefficient of variation (or also referred to as "relative standard deviation") obtained by dividing the standard deviation of the shading of a specific region by the average value of the shading of the specific region is appropriate. However, it may be the inverse of this coefficient of variation. It has been confirmed by clinical experiments that this relative ratio correlates with the bone quality of the subject as described later. By using the relative ratio, it is possible to accurately evaluate the bone quality.

The average value of the gray level of a specific region described above in the X-ray image is bone density correlation value, the comparison result between the bone density correlation value and bone density threshold, the comparison result between the bony correlation value and bone quality threshold Based on, the bone mineral density evaluation means evaluates the fracture risk of the subject. By evaluating the fracture risk based on the evaluation result of the bone density, the fracture risk can be evaluated more accurately.

The shading of the entire X-ray image including the specific region changes depending on various conditions such as when the X-ray image is taken. Therefore, the bone density may not be accurately evaluated from the shading of a specific region in such an X-ray image. Therefore, it is desirable that the shading of the X-ray image is appropriately corrected in an appropriate manner before the bone density is evaluated. Therefore, for example, the X-ray image can include an image of a predetermined specimen arranged side by side with the mandible. As this sample, for example, there is a solid substance having a constant and known transmittance for X-rays, and in particular, the aluminum block disclosed in Japanese Patent No. 4077430 described above is suitable. Then, a correction means for correcting the shading of the entire X-ray image so that the shading of the sample image in the X-ray image matches a predetermined reference value is further provided. Then, the bone density is evaluated based on the shading of the specific region in the corrected image after the correction by the correction means. That is, according to this configuration, the sample is used as an index so that the shading of the image of the sample matches a predetermined reference value, that is, the shading of the image of the sample should be the original shading. The shading of the entire X-ray image including the image of the specimen is corrected so as to be the degree. Then, the bone density is evaluated based on the shading of the specific region in the corrected image. As a result, the bone density can be accurately evaluated regardless of the difference in the shading of the entire original (before correction) X-ray image.

When such a correction means is provided, the above-mentioned bone quality calculation means is not a variation in the shading of the specific region in the X-ray image, but a specific region in the corrected image after the correction by the correction means. The variation in the degree of shading may be obtained. More specifically, the bone quality calculation means obtains the relative ratio between the standard deviation of the shading of the specific region in the corrected image and the average value of the shading of the specific region in the corrected image, and further, the bone quality evaluation means , The bone quality may be evaluated based on the relative ratio.

Another aspect of the present invention is the invention of the method corresponding to the device invention, and another aspect is the invention of the program corresponding to the device invention.

It is a figure which shows the overall schematic structure of the fracture risk evaluation system which concerns on one Embodiment of this invention. It is a figure which shows one aspect of the X-ray film after X-ray photography in the same embodiment. It is an external perspective view of the aluminum block as a specimen in the same embodiment. It is a figure which shows one aspect of the X-ray film before X-ray photography in the same embodiment. It is a schematic diagram which shows the state when X-ray photography is performed in the same embodiment. It is a list which shows the examination data of the subject in the same embodiment. It is a graph based on the inspection data of FIG. It is a graph different from FIG. 7 based on the inspection data of FIG. It is a graph different from FIG. 8 based on the inspection data of FIG. It is a graph different from FIG. 9 based on the inspection data of FIG. It is a graph different from FIG. 10 based on the inspection data of FIG. It is a list which shows the correlation between each parameter in the inspection data of FIG. It is a graph which shows the result of ROC (Receiver Operating Characteristic) analysis based on the inspection data of FIG. It is a list showing the evaluation criteria of the fracture risk in the same embodiment. It is a schematic diagram which shows one aspect of the main screen displayed on the display in the same embodiment. It is a schematic diagram which shows a mode different from FIG. 15 of the main screen. It is a schematic diagram showing still another aspect from FIG. 16 of the main screen. It is a schematic diagram which shows one aspect of the sub-screen displayed on the display in the same embodiment. It is a schematic diagram which shows a mode different from FIG. 18 of the sub-screen. It is a schematic diagram conceptually showing the mode of the table in the same embodiment. It is a schematic diagram which shows the aspect different from FIG. 19 of the sub-screen in the same embodiment. It is a schematic diagram which shows the aspect which is further different from FIG. 17 of the main screen in the same embodiment. It is a schematic diagram showing still another aspect from FIG. 22 of the main screen. It is a flowchart which outlines a part of the fracture risk evaluation program in the same embodiment.

An embodiment of the present invention will be described by taking the fracture risk evaluation system 10 for dentistry shown in FIG. 1 as an example. The fracture risk evaluation system 10 is based on the invention disclosed in Japanese Patent No. 4077430 described above, and in particular, is based on the bone density evaluation system according to the second embodiment. Hereinafter, the description of the same parts as those disclosed in the same gazette will be omitted as appropriate.

As shown in FIG. 1, the fracture risk evaluation system 10 according to the present embodiment includes a personal computer (hereinafter referred to as “PC”) 12. The PC 12 functions as a fracture risk assessment device by executing a fracture risk assessment program provided by an appropriate storage medium 14 such as a DVD-ROM (Digital Versatile Disk Read Only Memory). The PC 12 includes an input device 12a as a command input means and a display 12b as a display means. Although not shown, the input device 12a includes a keyboard and a pointing device such as a mouse. Further, a film scanner 16 as an image input means and a color printer 18 as a printing means are connected to the PC 12.

According to the fracture risk evaluation system 10, the X-ray image of the lower jaw bone of the subject (patient) captured on the X-ray film 20 is read by the film scanner 16 and converted into digital image data. The converted image data is input to the PC 12 and stored in a hard disk as a storage means (not shown) in the PC 12 in, for example, an 8-bit or 24-bit bitmap format. In the same manner as this, a plurality of image data obtained from a plurality of subjects are stored in the hard disk. Some of these image data were obtained from the same subject on different days.

An X-ray image of the mandible of the subject is taken on the X-ray film 20 as described above. Specifically, as shown in FIG. 2, the first premolar 30, the canines 32 adjacent thereto, and the second premolar Images of the two premolars 34 and the alveolar bone 36 that supports these teeth 30-34 have been taken. In addition to the images of these teeth 30 to 34 and the alveolar bone 36, for example, a horizontally long rectangular reference bar 38 is also photographed above these images.

Here, the reference bar 38 is also referred to as a block made of aluminum as a sample as shown in FIG. 3 (or also referred to as a “phantom”. Hereinafter, this block will also be described with the same reference numerals as the reference bar 38. It is an image of.). The block 38 has a rectangular flat bottom surface 38a, and has a stepped shape in which the thickness dimension (dimension in the direction perpendicular to the bottom surface 38a) changes stepwise in the longitudinal direction of the bottom surface 38a. The height dimension (step) ΔT of each step is constant, and the length dimension (depth) P of each step is also constant. The number of stages is about 7 to 9 stages. Further, an aluminum foil 40 as an X-ray blocking means for more reliably blocking the X-ray 50 (see FIG. 5) described later is attached to the upper surface of the uppermost portion. The length dimension (dimension in the longitudinal direction of the bottom surface 38a) L of the block 38 is about 20 [mm], and the width dimension (dimension in the lateral direction of the bottom surface 38a) W is about 10 [mm]. is there. Further, the thickness dimension of the lowermost portion (height dimension from the bottom surface 38a to the upper surface of the lowermost step portion) Ta is around 1 [mm], and the thickness dimension of the uppermost step portion (from the bottom surface 38a to the uppermost step portion). Height dimension to the upper surface) Tb is about 6 [mm] to 8 [mm].

The block 38 is attached to the imaging surface (exposure surface) 20a of the rectangular X-ray film 20 as shown in FIG. 4 before the X-ray imaging. Specifically, the block 38 has its longitudinal direction (of the bottom surface 38a) along the lateral direction of the X-ray film 20 and its bottom surface 38a facing the imaging surface 20a of the X-ray film 20. , It is attached to the portion near the upper edge, which is one edge of the photographing surface 20a, with an appropriate adhesive. Then, at the time of X-ray photography, the X-ray film 20 to which the block 38 is attached has an upper edge (the edge on the side close to the portion to which the block 38 is attached) as shown in FIG. ) Is directed toward the upper jaw side (the surface side of the paper surface in FIG. 5) of the subject (not shown), and the imaging surface 20a is directed outward, and the teeth 30 to 34 and (shown in FIG. 5) to be imaged. It is installed inside the alveolar bone 36 (not), that is, in the mouth. At this time, the X-ray film 20 is supported by a film holder (not shown). Then, X-rays 50 are emitted from the X-ray photographing apparatus (X-ray irradiator) 52 outside the mouth toward the photographing surface 20a of the X-ray film 20. As a result, as shown in FIG. 2, the images of each tooth 30 to 34, the image of the alveolar bone 36, and the image of the block 38 (reference bar 38) are photographed side by side on one X-ray film 20. Will be done.

In the X-ray film 20, a portion having a low transmittance for X-rays is shown brighter (whiter), and a portion having a higher transmittance for the X-ray 50 is shown darker (blackish). For example, the reference bar 38, which is an image of the block 38, is indicated by the shading (brightness / darkness) according to the thickness dimension of the block, and more specifically, the portion corresponding to the portion having the large thickness dimension. It is shown brighter, and the portion corresponding to the portion having the smaller thickness dimension is shown darker. In particular, since the X-ray 50 is hardly transmitted to the portion to which the aluminum foil 40 is attached, the portion corresponding to this portion is shown to be the brightest (so to speak, pure white). Further, regarding each tooth 30 to 34, especially the enamel and dentin parts have a relatively high transmittance to X-ray 50, so the parts corresponding to these parts are shown relatively brightly. Is done. On the other hand, the portion corresponding to the alveolar bone 36 is shown by the shade according to the bone density (bone mineral density) of the alveolar bone 36, and more specifically, the larger the bone density, the brighter the bone density. The smaller it is, the darker it is.

Based on the X-ray image taken on the X-ray film 20, the PC12 evaluates the bone density and bone quality of the subject based on the image data, and the risk of osteoporotic fracture in the subject. evaluate. Specifically, as shown by the broken line 60 in FIG. 2, a specific region corresponding to the alveolar bone 36 portion around the lower half of the root 30a of the first premolar 30 is set as the evaluation target region, and the evaluation target region 60 Bone density is evaluated from the shading of the image. At the same time, the bone quality is evaluated from the variation in the shading of the image of the evaluation target area 46. Furthermore, the risk of fracture is evaluated based on the evaluation results of bone density and bone quality.

The upper edge of the evaluation target area 60 is set at a position corresponding to half (= D / 2) of the section D from the cervical portion to the root apex of the first premolar 30. Then, the lower edge of the evaluation target area 60 is set at a position corresponding to the root apex of the first premolar 30. Further, the right edge of the evaluation target region 60 is aligned with a position corresponding to the middle between the opposite edges of the first premolar 30 and the canine 32 on the straight line L connecting the cervical portions of the teeth 30 to 34. Be done. Similarly, the left edge of the evaluation target region 60 is aligned with the position corresponding to the middle between the edges of the first premolar 30 and the second premolar 34 on the straight line L. However, the region corresponding to the root 30a of the first premolar 30 is excluded from the evaluation target region 60.

It is also disclosed in Japanese Patent No. 4077430 described above that the specific region corresponding to the alveolar bone 36 portion around the lower half of the root 30a of the first pretumor tooth 30 is set as the evaluation target region 60. This is because the bone density of the alveolar bone 36 in the portion corresponding to the evaluation target region 60 is closely related to the bone density of the whole body. Incidentally, the upper portion of the alveolar bone 36 is easily affected by periodontal disease, and for example, when suffering from the periodontal disease, the bone density of this portion decreases. In order to eliminate the influence of such periodontal disease, the region corresponding to the upper portion of the alveolar bone 36 is excluded from the evaluation target region 60. The component of the root 30a itself of the first premolar 30 is not particularly related to the bone density of the whole body. Therefore, even if the region corresponding to the tooth root 30a is included in the evaluation target region 60, there is no particular effect on the evaluation result of bone density. However, in the present embodiment, it is excluded from the evaluation target area 60 as described above. Further, in the evaluation of bone quality, the same evaluation target area 60 is set following the evaluation of bone density.

By the way, according to the bone density evaluation system according to the second embodiment in the above-mentioned Japanese Patent No. 4077430, the X-ray image similar to FIG. 2 in the present embodiment, strictly speaking, the shading of the reference bar with respect to the image data. The shading of the entire X-ray image including the reference bar is corrected so that the degree is the shading that should be. Then, the bone density of the alveolar bone is evaluated based on the average value of the shading of the evaluation target region in the corrected X-ray image, and the bone density of the whole body is evaluated. In this way, the bone density of the whole body can be evaluated based on the average value of the density of the evaluation target area. The average value of the density of the evaluation target area is the bone density measured by the DXA method described above, that is, the whole body. This is because it shows a strong relationship with the bone density of. Incidentally, in the current osteoporosis diagnosis, the osteoporosis diagnosis is routinely performed based on the T value. The T value is the ratio of the bone density of the subject to this YAM as a percentage, with the average value of the bone density of so-called young people aged 20 to 44 (Young Adult Mean; hereinafter referred to as "YAM") as 100. It is the expressed value (strictly referred to as "young adult comparison percentage"). As a matter of course, the average value of the shading of the evaluation target region shows a strong correlation with this T value.

Similarly, in the present embodiment as well, with respect to the X-ray image shown in FIG. 2, strictly speaking, for the image data, the reference bar 38 is set so that the shading of the reference bar 38 is the original shading. The shading of the entire X-ray image including is corrected. Then, the bone density of the alveolar bone 36 is evaluated based on the average value of the shading of the evaluation target region 60 in the corrected X-ray image, and the bone density of the whole body is evaluated.

Further, in the present embodiment, the standard deviation of the shading of the evaluation target area 60 is obtained, and in addition, the coefficient of variation obtained by dividing this standard deviation by the average value of the shading of the evaluation target area 60 is obtained. Be done. This coefficient of variation represents the variation in the shading of the evaluation target region 60, but since it is a so-called dimensionless number, it is not the absolute variation of the shading itself, but the relative of the shading to the average value of the shading. Represents the variation. Then, the bone quality of the alveolar bone 36 is evaluated based on this coefficient of variation, and eventually the bone quality of the whole body is evaluated. The basis for evaluating the bone quality of the whole body based on the coefficient of variation will be described in detail later.

Then, in the present embodiment, the fracture risk is evaluated based on the evaluation result of bone density and the evaluation result of bone quality. The evaluation result of the fracture risk is displayed on the display 12b shown in FIG. 1 together with the evaluation result of the bone density and the evaluation result of the bone quality, and is printed on an appropriate piece of paper 22 such as a chart by the color printer 18 as needed. Will be done.

Here, a total of 30 postmenopausal women aged 50 to 69 years (mean age 59.37 years) were taken as subjects, and for these subjects, the measured values of the bone mineral density of the lumbar spine by the DXA method (hereinafter, “L-BMD””). ), The average value of the shading of the evaluation target area 60 (hereinafter sometimes referred to as “al-BMD”), the standard deviation D of the shading of the evaluation target area 60, and the evaluation target area. FIG. 6 shows the results of examining the coefficient of variation CV of 60 shades, the above-mentioned T value, the presence or absence of a history of osteoporotic fracture, and the number of minute injuries in the evaluation target area 60. The average value of the shading of the evaluation target region 60, that is, al-BMD, correlates with the bone density of the alveolar bone 36 as described above, and eventually correlates with the bone density of the whole body. The coefficient of variation CV is a value obtained by dividing the standard deviation D by al-BMD as a percentage. Further, the "T value" is a value when the above-mentioned YAM is 1.011, that is, a value obtained by dividing L-BMD by the YAM of 1.011 and expressing it as a percentage. Regarding "previous fracture", if this value is 1, it means that there is a history of osteoporotic fracture, and if this value is 0, it means that there is no history of the osteoporotic fracture. Further, the “number of minute damages” is the number of minute damages actually visually confirmed from the evaluation target area 60 in the X-ray image.

A graph showing the relationship between L-BMD, al-BMD, and the coefficient of variation CV with respect to the T value, for example, based on the T value in FIG. 6, is shown in FIG. Specifically, FIG. 7A shows a subject having a T value of less than 70%, a subject having a T value of 70% or more and less than 80%, and a subject having a T value of 80% or more. The average value of L-BMD is represented by a graph. Similarly, FIG. 7B shows subjects with a T value of less than 70%, subjects with a T value of 70% or more and less than 80%, and subjects with a T value of 80% or more, respectively. The average value of al-BMD is represented by a graph. Then, FIG. 7C shows the coefficient of variation CV for each of the subjects having a T value of less than 70%, the subjects having a T value of 70% or more and less than 80%, and the subjects having a T value of 80% or more. It is a graph showing the average value of. A subject having a T value of less than 70% is a subject who is suspected of having osteoporosis in the current diagnosis of osteoporosis based on the T value. A subject having a T value of 70% or more and less than 80% is a subject who is suspected of having osteopenia in the same diagnosis, and a subject having a T value of 80% or more is normal in the same diagnosis. It is a subject who is supposed to be.

As can be seen from FIG. 7, especially as can be seen from FIG. 7A, the larger the T value, the larger the L-BMD, that is, the L-BMD is in a proportional relationship with the T value. This is natural from the above-mentioned relationship between the L-BMD and the T value. As can be seen from FIG. 7B, the larger the T value, the larger the al-BMD, that is, the al-BMD is substantially proportional to the T value. This means that al-BMD can be an index for evaluating the bone density of the whole body, that is, the bone density of the whole body can be evaluated based on the al-BMD. On the other hand, as can be seen from FIG. 7C, the coefficient of variation CV is smaller as the T value is larger, that is, it is generally inversely proportional to the T value. From this, it can be seen that the coefficient of variation CV has some relation to bone strength apart from bone density.

Next, using the presence or absence of a history of fracture in FIG. 6 as a reference, the relationship between the T value, al-BMD, and the coefficient of variation CV with respect to the presence or absence of a history of fracture is shown in FIG. Specifically, FIG. 8A is a graph showing the average value of the T values for each of the subjects having a history of fracture and the subjects having no history of fracture. In addition, when the relationship of L-BMD with respect to the presence or absence of a history of fracture is represented by a graph, it is the same as in FIG. 8A. Then, FIG. 8B is a graph showing the average value of al-BMD for each of the subjects having a history of fracture and the subjects having no history of fracture. Further, FIG. 8C is a graph showing the average value of the coefficient of variation CV for each of the subjects having a history of fracture and the subjects having no history of fracture.

As can be seen from FIG. 8, in particular, as can be seen from FIG. 8A, the T value is smaller for subjects with a history of fracture, and the T value is larger for subjects without a history of fracture. This means that the larger the T value, that is, the higher the bone density, the lower the risk of fracture, in other words, the higher the bone strength. As can be seen from FIG. 8B, similarly, with respect to al-BMD, the subject with a history of fracture has a smaller al-BMD, and the subject without a history of fracture has a larger al-BMD. On the other hand, the coefficient of variation CV is larger for subjects with a history of fracture and smaller for subjects without a history of fracture, as can be seen from FIG. 8 (c). From this, it seems that the coefficient of variation CV has some relation to the fracture risk, that is, the bone strength.

Further, using the T value in FIG. 6 as a reference, the relationship between the ratio of subjects with a history of fracture to this T value and the number of microinjuries is shown in FIG. 9 as a graph. Specifically, FIG. 9A shows a subject having a T value of less than 70%, a subject having a T value of 70% or more and less than 80%, and a subject having a T value of 80% or more. This is a graph showing the ratio of subjects with a history of fracture (ratio of the number of people). Then, FIG. 9B shows the number of minute injuries for each of the subjects having a T value of less than 70%, the subjects having a T value of 70% or more and less than 80%, and the subjects having a T value of 80% or more. It is a graph showing the average value of.

According to FIG. 9, in particular, according to FIG. 9A, subjects with a T-value of 70% or more and less than 80% had a history of fracture than subjects with a T-value of less than 70%. The ratio is large. This means that the higher the T-value, that is, the higher the bone density, the lower the risk of fracture, in other words, the higher the bone strength. According to FIG. 9B, the number of microdamages is larger in the subjects having a T value of 70% or more and less than 80% than in the subjects having a T value of less than 70%. This microinjury is one of the bony elements as described above. From this, it can be seen that bone quality including microinjuries is an independent factor of bone strength, in other words, fracture risk, apart from bone mineral density.

In addition, using al-BMD in FIG. 6 as a reference, the relationship between the ratio of subjects with a history of fracture to al-BMD and the number of microinjuries is shown in FIG. Specifically, FIG. 10A shows a subject having an al-BMD of less than 71.4, a subject having an al-BMD of 71.4 or more and less than 82.0, and a subject having an al-BMD of 80.2 or more. It is a graph showing the ratio of the subjects and the subjects who have a history of fracture for each of them. Then, FIG. 10B shows a subject having an al-BMD of less than 71.4, a subject having an al-BMD of 71.4 or more and less than 82.0, and a subject having an al-BMD of 80.2 or more. The average value of the number of microscopic damages for each of the above is shown in a graph. A subject having an al-BMD of less than 71.4 corresponds to a subject having a T value of less than 70%, that is, a subject who is suspected of having osteoporosis in the current diagnosis of osteoporosis based on the T value, in other words. For example, it corresponds to a subject who is "needs attention" in the second embodiment of the above-mentioned Japanese Patent No. 4077430. A subject having an al-BMD of 71.4 or more and less than 82.0 corresponds to a subject having a T value of 70% or more and less than 80%, that is, osteopenia in the current osteoporosis diagnosis based on the T value. In other words, it corresponds to a subject who is suspected of having "Caution" in the second embodiment of Japanese Patent No. 4077430. Further, a subject having an al-BMD of 80.2 or more corresponds to a subject having a T value of 80% or more, that is, a subject who is considered to be normal in the current diagnosis of osteoporosis based on the T value, in other words, a patent. Corresponds to a subject who is considered "normal" in the second embodiment of Japanese Patent Application Laid-Open No. 4077430. The 71.4 and 82.0 boundary values (threshold values) for al-BMD here are empirically determined, including comparisons with the 70% and 80% boundary values for the T value.

According to FIG. 10, in particular, according to FIG. 10 (a), a subject having an al-BMD of 71.4 or more and less than 82.0 had a fracture than a subject having an al-BMD of less than 71.4. The proportion of subjects with a history is high. This means that the higher the al-BMD, that is, the higher the bone density, the lower the risk of fracture, in other words, the higher the bone strength. This is similar to what is meant by the relationship between the T value shown in FIG. 9A and the proportion of subjects with a history of fractures. According to FIG. 10B, the number of microdamages was smaller in the subjects having al-BMD of 71.4 or more and less than 82.0 than those of the subjects having al-BMD of less than 71.4. The difference between the two is small. From this, it can be seen that bone quality including microinjuries is an independent factor of bone strength, in other words, fracture risk, which is different from bone mineral density.

Further, using the coefficient of variation CV in FIG. 6 as a reference, the relationship between the ratio of subjects with a history of fracture to the coefficient of variation CV and the number of microinjuries is shown in FIG. Specifically, FIG. 11A shows a subject having a coefficient of variation CV of 15.4% or more, a subject having a coefficient of variation CV of 12.2% or more and less than 15.4%, and a subject having a coefficient of variation CV of 12. It is a graph showing the ratio of subjects with less than 2% and subjects with a history of fracture for each. Then, FIG. 11B shows a subject having a coefficient of variation CV of 15.4% or more, a subject having a coefficient of variation CV of 12.2% or more and less than 15.4%, and a subject having a coefficient of variation CV of less than 12.2%. It is a graph showing the average value of the number of microscopic damages for each of the subjects. The boundary values of 15.4% and 12.2% for the coefficient of variation CV here are values determined based on the relationship between the T value and the coefficient of variation CV shown in FIG. 7 (c), for example. is there. That is, the boundary value of 15.4% is the mean value (17.13%) of the coefficient of variation CV for subjects with a T value of less than 70% (that is, subjects suspected of having osteoporosis), and the T value is It is a value determined based on the mean value (13.66%) of the coefficient of variation CV for subjects of 70% or more and less than 80% (that is, subjects suspected of having bone loss), and is detailed. Is an intermediate value between these two average values. The boundary value of 12.2% is the average value (13.66%) of the coefficient of variation CV for subjects whose T value is 70% or more and less than 80%, and the subject whose T value is 80% or more (that is, normal). It is a value determined based on the average value (10.66%) of the coefficient of variation CV for the subject), and more specifically, it is an intermediate value between these two average values.

According to FIG. 11, in particular, according to FIG. 11A, the larger the coefficient of variation CV, the larger the proportion of subjects with a history of fracture, and the smaller the coefficient of variation CV, the smaller the proportion of subjects with a history of fracture. That is, the larger the coefficient of variation CV, the greater the risk of fracture, and the smaller the coefficient of variation CV, the smaller the risk of fracture. According to FIG. 11B, the larger the coefficient of variation CV is, the larger the number of fine damages is, and the smaller the coefficient of variation CV is, the smaller the number of fine damages is. From this, it can be seen that the coefficient of variation CV correlates with the number of microinjuries, that is, correlates with the bone quality including the number of microinjuries. For example, a subject having a coefficient of variation CV of 15.4% or more can be evaluated as having a fragile bone quality, so to speak, “needs attention”. Then, it can be evaluated that the subject having a coefficient of variation CV of 12.2% or more and less than 15.4% has a slightly fragile bone quality and requires some "caution". Furthermore, subjects with a coefficient of variation of less than 12.2% can be evaluated as having tough bones, so to speak, "normal".

FIG. 12 shows the correlation coefficient between each parameter in FIG. The numerical value in the upper part of each numerical value column in FIG. 12 is the correlation coefficient, and the numerical value in the lower part is the significance probability in the two-sided test. As can be seen from FIG. 12, each parameter including the coefficient of variation CV correlates with each other. That is, it can be seen that the coefficient of variation CV correlates with other parameters L-BMD, al-BMD, the presence or absence of a history of fracture, and the number of microinjuries.

Further, FIG. 13 shows the results of ROC analysis of the ability to discriminate the presence or absence of a history of fracture for each of L-BMD, al-BMD and the coefficient of variation CV. As can be seen from FIG. 13, it is recognized that all of L-BMD, al-BMD and the coefficient of variation CV have the ability to discriminate the presence or absence of a history of fracture. Incidentally, it can be seen that the discriminating ability by al-BMD is the highest (the farthest from the linear reference line).

As described above, the coefficient of variation CV correlates with the bone quality including the number of minute injuries, and the coefficient of variation CV has the ability to discriminate the presence or absence of a history of fracture. Therefore, in the present embodiment, the coefficient of variation CV is based on the coefficient of variation CV as described above. The bone quality of the whole body is evaluated. Furthermore, the risk of fracture is evaluated based on the evaluation result of this bone quality and the evaluation result of bone density by al-BMD. Specifically, as shown in FIG. 14, it is considered to be normal only when the coefficient of variation CV is less than 12.2% and al-BMD is 82.0% or more, and "normal". Is evaluated. Even if the coefficient of variation is less than 12.2%, if al-BMD is 71.4% or more and less than 82.0%, it is considered that osteopenia is suspected and it is called "caution". Evaluation is given. Further, even if the al-BMD is 82.0% or more, even if the coefficient of variation CV is 12.2% or more and less than 15.4%, it is considered that osteopenia is suspected, and "Caution". Is evaluated. Other than this, all of them are suspected of having osteoporosis and are evaluated as "need attention".

Now, it is assumed that a plurality of image data are stored in the hard disk in the PC 12 as described above, and in particular, image data of four or more X-ray images obtained from the same subject on different days are stored. .. When the above-mentioned fracture risk evaluation program is started in this state, the main screen 100 as shown in FIG. 15 is displayed on the display 24 of the PC 12.

That is, a horizontally long title bar 102 is displayed at the top of the main screen 100. Then, in the title bar 102, for example, a horizontally written character string (here, the character string “Medical record No. 00001”) 104 representing the medical record number of the subject is displayed left-justified. At the same time, a button with an “x” mark for closing the main screen 100, a so-called close button 105, is displayed at the right end portion in the title bar 102. Then, a horizontally long menu bar 106 similar to the title bar 102 is displayed below the title bar 102. In the menu bar 106, a plurality of character strings 108, 108, ... Representing the contents of the menu that can be operated on the main screen 100 are displayed in a horizontal row. Further, a horizontally long toolbar 110 is displayed below the menu bar 106. In this toolbar 110, a plurality of tool buttons 112, 112, ... Which are stylized menus in the menu bar 106 are displayed in a horizontal row. Then, a rectangular frame area 114 is displayed below the toolbar 110.

In the upper portion of the frame area 114, four substantially square picture boxes 116, 116, ... Are displayed in a horizontal row. Below each of these picture boxes 116, 116, ..., "Image 1", "Image 2", representing the heading of the X-ray image displayed in each of the picture boxes 116, 116, ... As will be described later. The four character strings 118, 118, ... "Image 3" and "Image 4" are displayed in this order from the left side to the right side. Further, below each character string 118, a character string 120 representing al-BMD for the X-ray image corresponding to the character string 118, a character string 122 representing the fluctuation coefficient CV, and an evaluation result of fracture risk are displayed. The character string 124 to be represented is displayed in this order from the upper side to the lower side.

In addition, on the right side of the arrangement of the picture boxes 116, 116, ..., Four radio buttons 126, 126, ... For selecting which of the picture boxes 116, 116, ... Are enabled. Is displayed in a vertical row. Then, on the right side of each radio button 124, a character string 128 representing "image A" (A; a numerical value of any of 1 to 4) corresponding to the radio button 124 is displayed. Further, above the arrangement of the radio buttons 126, 126, ..., The character string 130 "image selection" indicating the function of the radio buttons 126, 126, ... Is displayed. On the other hand, below the arrangement of the radio buttons 124, 124, ..., The character string 132 representing the reference mean value SMb described later and the character string 134 representing the reference standard deviation SDb described later are arranged from the upper side to the lower side. It is displayed in this order.

Further, below the character strings 132 and 134 representing the reference mean value SMb and the reference standard deviation SDb, and to the right of the sequence of the character strings 120, 120, ... Representing the above-mentioned al-BMD, there is a fracture risk. Two edit boxes 136 and 138 for inputting two boundary values for the al-BMD related to the evaluation of are displayed side by side. Then, above the edit boxes 136 and 138, the character strings 140 and 142 representing the headings of the edit boxes 136 and 138 are displayed as "caution" and "need attention". That is, in the edit box 136 on the left side, a value that is a boundary between "normal" and "caution" for al-BMD is input, and the above-mentioned value "82.0" is input as a default value. Then, in the edit box 138 on the right side, a value serving as a boundary between "caution" and "need attention" for al-BMD is input, and the above-mentioned value "71.4" is input as a default value. Similarly, on the right side of the sequence of the character strings 122, 122, ... Representing the coefficient of variation CV described above, two for inputting two boundary values for the coefficient of variation CV related to the evaluation of fracture risk. The edit boxes 144 and 146 are displayed. That is, in the edit box 144 on the left side, a value that is a boundary between "normal" and "caution" for the coefficient of variation CV is input, and the above-mentioned value "12.2" is input as a default value. Then, in the edit box 146 on the right side, a value serving as a boundary between "caution" and "need attention" for the coefficient of variation CV is input, and the above-mentioned value "15.4" is input as a default value.

Further, a histogram display area 148 is displayed in the lower portion of the frame area 114. In the histogram display area 148, vertical scale lines 150, 150, ... Are drawn at regular intervals in the direction along the horizontal axis. Further, below each scale line 150, a character string 152 representing the indicated value of the scale line 150 is displayed.

The main screen 100 has substantially the same configuration as that of the second embodiment in the above-mentioned Japanese Patent No. 4077430, and has character strings 120, 120, ... Representing al-BMD and characters representing the coefficient of variation CV. The parts of columns 122, 122, ... And the character strings 124, 124, ... Representing the evaluation result of the fracture risk are different, and the four edit boxes 136, 138, 144 and 146 and the two characters accompanying them are different. Only the parts of columns 140 and 142 are different.

On the main screen 100 shown in FIG. 15, for example, the radio button 126 corresponding to "image 1" is turned ON (clicked) by the operation of the pointing device described above, that is, the picture box 116 corresponding to the "image 1" is effective. It is assumed that it will be converted. Then, in this state, it is assumed that arbitrary image data is read from the hard disk by operating the pointing device. Then, as shown in FIG. 16, the X-ray image 154 according to the image data is displayed in the picture box 116 of the “image 1”. The X-ray image 154 is displayed in gray scale. If the original image data is color data, the image data is converted to grayscale data when read from the hard disk. The command for reading this image data is stored in, for example, the "File" menu in the menu bar 106. The image data can also be read by operating the corresponding tool button 112 in the toolbar 110. The procedure for reading the image data is the same as that of the second embodiment in Japanese Patent No. 4077430 described above.

For the other "image 2", "image 3" and "image 4", when the image data is read in the same manner as this, as shown in FIG. 17, the X-ray image 154 according to the read image data 154, ... Are displayed in the corresponding picture boxes 116, 116, ... The X-ray images 154, 154, ... Displayed in these picture boxes 116, 116, ... Are all of the same subject, for example, "image 1", "image 2", "image 3". And the shooting date is the oldest in the order of "image 4" (the shooting date of "image 1" is the oldest).

After the X-ray images 154, 154, ... Are displayed in the picture boxes 116, 116, ... In this way, for example, the radio button 126 corresponding to "image 1" is turned on, and in this state, the "image" is displayed. When the pointing device performs an operation for starting the evaluation of the fracture risk for the X-ray image 140 of 1 ”, a sub screen (dialog box) 200 as shown in FIG. 18 is displayed on the display 12b. The command for starting the evaluation of the fracture risk is stored in, for example, the "edit" menu in the menu bar 106.

As shown in FIG. 18, a horizontally long title bar 202 is displayed at the top of the sub screen 200. Then, in the title bar 202, for example, a horizontally written character string representing the medical record number and the currently activated information "image A" (here, the character "medical record No. 000001-image 1"). Column) 204 is displayed left justified. At the same time, a close button 205 for closing the sub screen 200 is displayed at the right end portion in the title bar 202. Then, a rectangular frame area 206 is displayed below the title bar 202.

In the left side area of the frame area 206, for example, a rectangular area selection area 208 is displayed with a small space below the frame area 206. Then, in the area selection area 208, an enlarged image 210 which is an enlargement of the X-ray image 154 of the currently activated “image A” (here, “image 1”) is displayed.

Further, below the area selection area 208, in the lower left corner of the frame area 206, the individual character strings 212, 212, ... Of "upper", "lower", "left" and "right" are placed. The four radio buttons 214, 214, ... Attached are displayed in two rows and two columns. Then, above these radio buttons 214, 214, ... (Character strings 212, 212, ...), A character string 216 "area extraction" indicating the function of the radio buttons 214, 214, ... Is displayed. Further, a check box 220 with the character string 218 "reference check" is displayed on the right side of the group of radio buttons 214, 214, ... (Character strings 212, 212, ...).

On the other hand, four radio buttons 222, 222, ... Are displayed in a vertical row on the right side of the area selection area 208 and on the upper side of the frame area 206. Then, on the right side of each of these radio buttons 222, 222, ..., Individual character strings representing the headings of the respective radio buttons 222, 222, ..., For example, "reference bar", "highest value (white)", The character strings "minimum value (black)" and "evaluation target" 224, 224, ... Are displayed. Further, above each of these radio buttons 222, 222, ... (Character strings 224, 224, ...), A character string 226 called "selection range" indicating the function of each radio button 222, 222, ... Is displayed. .. Immediately after the sub screen 200 is displayed, all of the radio buttons 222, 222, ... Are in the OFF state.

A note area 230 for displaying the note 228 is provided on the right side of the group of the radio buttons 222, 222, ... On the upper side of the frame area 206. Immediately after the sub screen 200 is displayed, in other words, when all the radio buttons 222, 222, ... Are in the OFF state, the character string "unprocessed" is displayed as the note 228.

Further, below the note area 230, two edit boxes 232 and 234 are displayed side by side. Then, above these edit boxes 232 and 234, two character strings 236 and 238 representing the headings of the edit boxes 232 and 234, which are "reference maximum value" and "reference minimum value", are displayed. One of the edit boxes 232 and 234, for example, the edit box 232 with the heading (character string) 236 of "reference maximum value", is input with the reference maximum value Ysmax described later, and is "180" as the default value. Is entered. On the other hand, in the edit box 234 with the heading 238 of "reference minimum value", the reference minimum value Ysmin described later is input, and the value "30" is input as the default value.

In addition, below each edit box 232 and 234, a horizontally elongated rectangular reference bar display area 240 is displayed. Then, on the upper left side of the reference bar display area 240, a character string 242 representing the heading of the reference bar display area 240 called "reference bar" is displayed.

Further, a substantially square evaluation target display area 244 is displayed on the lower left side of the reference bar display area 240. Then, on the upper left side of the evaluation target display area 244, a character string 246 representing the heading of the evaluation target display area 244 called "evaluation target" is displayed.

A substantially square histogram display area 248 is displayed on the right side of the evaluation target display area 244. In the histogram display area 248, vertical scale lines 250, 250, ... Are drawn at regular intervals in the direction along the horizontal axis. Further, above the histogram display area 248, three radio buttons 252, 252, ... Are displayed in a horizontal row. At the same time, on the right side of each of these radio buttons 252,252 ..., individual character strings representing the respective headings of the radio buttons 252,252 ... "Original image", "bar", and "target". 254, 254, ... Are displayed. Further, above each radio button 252, 252 ... (character string 254, 254, ...), A character string 256 called "histogram" representing the function of each radio button 252, 252, ... Is displayed.

The sub screen 200 also has substantially the same configuration as that of the second embodiment in the above-mentioned Japanese Patent No. 4077430, and the radio button 222 with the above-mentioned character string 224 "evaluation target" is turned on. Only the content of note 228 displayed in the note area 230 when it is done is different.

On this sub-screen 200, for example, as shown in FIG. 19, the heading 224 of the four radio buttons 222, 222, ... Attached to the character 226 "selection range" by the operation of the pointing device described above is 224. It is assumed that the radio button 222 marked with is turned on. Then, on the enlarged image 210 in the area selection area 208, as shown by the broken line 258 in FIG. 19, the area corresponding to the reference bar 38 shown in FIG. 2 by the operation of the pointing device (also “reference” for this area. It is referred to as "bar".) It is assumed that 260 is selected in a rectangular shape. At this time, in the longitudinal direction of the reference bar 260 (horizontal direction in FIG. 19), one side of the rectangular region 258 coincides with one end edge of the reference bar 260 and faces one side of the rectangular region 258. It is important that the rectangular region 258 is selected so that the other side of the rectangular region 258 is aligned with the other edge of the reference bar 260. In the lateral direction of the reference bar 260 (vertical direction in FIG. 19), the remaining sides (upper side and lower side) of the rectangular region 258 are inside the peripheral edge of the reference bar 260, and FIG. It is sufficient that the region corresponding to each of the uppermost portion and the lowermost portion of the block 38 shown in the above (at least one portion thereof) is included in the rectangular region 258.

When the rectangular region 258 is selected in this way, so to speak, when the reference bar region 258 is set, the shading in the set reference bar region 258, for example, the brightness Y [i, j] ([i, j]). ]; Coordinates of each pixel constituting the enlarged image 210), the highest value and the lowest value are detected. Here, as the maximum value, the brightness Y [i, j] of the region basically corresponding to the uppermost portion of the block 38, that is, the region corresponding to the portion to which the aluminum foil 40 described above is attached is detected. To. On the other hand, as the minimum value, the brightness Y [i, j] of the region corresponding to the lowermost portion of the block 38 is basically detected.

The maximum value and the minimum value of the luminance Y [i, j] detected in this manner are recorded in the table 300 shown in FIG. 20 as the luminance maximum value Ybmax and the luminance minimum value Ybmin, respectively. The table 300 is formed in the hard disk when the fracture risk evaluation program is started. In addition, the above-mentioned reference maximum value Ysmax and reference minimum value Ysmin are also recorded in this table 300.

Then, based on the maximum brightness value Ybmax, the minimum brightness value Ybmin, the reference maximum value Ysmax, and the reference minimum value Ysmin recorded in the table 300, the brightness Y [i, of the enlarged image 210 displayed in the area selection area 208 j] is corrected. Specifically, the corrected luminance Y'[i, j] is obtained based on the following equation 1. According to this number 1, the enlarged image is such that the maximum luminance value Ybmax matches the reference maximum value Ysmax as the reference value and the minimum luminance value Ybmin matches the reference minimum value Ysmin as the reference value. The brightness Y [i, j] of the entire 210 is corrected.

Then, the enlarged image 210 in the area selection area 208 is redisplayed based on the corrected luminance Y'[i, j] obtained by the number 1. That is, the enlarged image 210 after the brightness Y [i, j] is corrected is displayed again.

Further, all the pixels of the enlarged image 210 are distributed according to the gradation x of the corrected luminance Y'[i, j], for example, by the gradation x of 256. Then, based on the following equation 2, the frequency Ha'[x] for each gradation x is obtained.

In the number 2, Na is the total number of pixels of the enlarged image 210, and na ′ [x] is the number of pixels of the total number of pixels Na allocated to the gradation x. The frequency Ha'[x] based on this number 2 is also recorded in the above table 300. By dividing the number of pixels na'[x] for each gradation x by the total number of pixels Na in this way, the so-called normalized frequency Ha'[x] can be obtained.

At the same time, the enlarged image 262 of the reference bar area 258 according to the corrected luminance Y'[i, j] is displayed in the reference bar display area 240. As for the reference bar area 258, each pixel constituting the reference bar area 258 is distributed according to the gradation x of the respective luminance Yb'[i, j]. Then, based on the following equation 3, the corrected dioptric power Hb'[x] for each gradation x is obtained.

In the number 3, Nb is the total number of pixels constituting the reference bar region 258, and nb'[x] is the number of pixels of the total number of pixels Nb allocated to the gradation x. The corrected frequency Hb'[x] based on this number 3 is also recorded in the table 300.

Further, based on the corrected power Hb'[x] obtained by the equation 3, the average value Mb'of the corrected brightness Yb'[i, j] in the reference bar region 258 is obtained. Specifically, the corrected average value Mb'is obtained based on the following equation 4.

The corrected mean value Mb'based on this number 4 is also recorded in the table 300. Then, when the currently activated image is "image 1", the corrected average value Mb'based on this equation 4 is set as the reference average value SMb'described later.

Further, the standard deviation Db'of the corrected luminance Yb'[i, j] in the reference bar region 258 is obtained based on the corrected power Hb'[x] and the corrected average value Mb'. Specifically, the corrected standard deviation Db'is obtained based on the following equation 5.

The corrected standard deviation Db'based on this number 5 is also recorded in table 300. Then, when the currently activated image is "image 1", the corrected standard deviation Db'based on this equation 5 is set as the reference standard deviation SDb' described later.

The corrected mean value Mb'and the corrected standard deviation Db'are displayed in the note area 230 as note 228. Further, in the upper part of the note area 230, a character string representing the heading "reference bar" is also displayed as the note 228.

Here, it is assumed that the radio button 252 with the heading 254 of "bar" among the three radio buttons 252, 252, ... Attached to the character string 256 "histogram" described above is turned on. Then, the histogram (curve) 264 according to the corrected frequency Hb'[x] of the reference bar area 258 is displayed in the histogram display area 248. Therefore, from this histogram 264, the distribution of the corrected luminance Yb'[i, j] in the reference bar region 258 can be recognized.

Further, although not shown, when the radio button 252 with the heading 254 of "original image" is turned on, the histogram display area 248 has a histogram according to the corrected dioptric power Ha'[x] of the entire enlarged image 210. Is displayed. Therefore, the distribution of the corrected luminance Y'[i, j] of the entire enlarged image 210 can be recognized from the histogram according to the corrected power Ha'[x]. At this point, even if the radio button 252 with the heading "Target" 254 is turned on, nothing is displayed in the histogram display area 248.

After the correction process is performed in this way, as shown in FIG. 21, the heading 224 of the four radio buttons 222, 222, ... Attached to the character 226 "selection range" ... It is assumed that the radio button 222 marked with is turned on. Then, the boundary line of the reference bar area 258 shown by the broken line in FIG. 19 above disappears. In this state, on the enlarged image 210 of the area selection area 208, as shown by the broken line 266 in FIG. 21, the area corresponding to the evaluation target area 60 shown in FIG. 2 is selected by the operation of the pointing device described above. And. That is, the evaluation target area 266 is selected on the enlarged image 210 of the area selection area 208. As a result, the evaluation target area 266 is set, and the corrected frequency Ho'[x] of the evaluation target area 266 is obtained based on the following equation 6.

In the number 6, No is the total number of pixels constituting the evaluation target region 266, and no'[x] is the number of pixels of the gradation x in the total number of pixels No. The corrected frequency Ho'[x] based on this number 6 is recorded in the above-mentioned table 300.

Then, based on the corrected power Ho'[x] obtained by the equation 6, the average value of the corrected brightness Yo'[i, j] in the evaluation target region 266 is obtained. Specifically, the corrected average value Mo'is obtained based on the following equation 7. The corrected average value Mo'based on the number 7 is treated as the above-mentioned al-BMD. Then, the corrected average value Mo'as this al-BMD is also recorded in the table 300.

Further, the standard deviation Do'of the corrected luminance Yo'[i, j] in the evaluation target region 266 is obtained based on the corrected power Ho'[x] and the corrected average value Mo'. Specifically, the corrected standard deviation Do'is obtained based on the following equation 8. Then, the corrected standard deviation Do'based on the number 8 is also recorded in the table 300.

In addition, the coefficient of variation CV is obtained by dividing the corrected standard deviation Do'based on the number 8 by the corrected mean value Mo'based on the above-mentioned number 7, that is, based on the following number 9. The coefficient of variation CV based on this number 9 is also recorded in table 300.

Furthermore, from the coefficient of variation CV based on the equation 9 and the corrected mean value Mo'that al-BMD based on the above equation 7, the coefficient of variation CV and al-BMD are specifically evaluated as shown in FIG. Fracture risk is assessed by collating with the criteria. This evaluation result RF is also recorded in Table 300.

When the fracture risk evaluation result RF is obtained in this way, as shown in FIG. 21, the enlarged image 268 of the evaluation target area 266 is displayed in the evaluation target display area 244. At the same time, a character string representing al-BMD (corrected mean value Mo') and coefficient of variation CV is displayed as note 228 in the note area 230. In addition, in the upper part of the note area 230, a character string representing the heading "evaluation target" is also displayed as the note 228.

Here, it is assumed that the radio button 252 with the heading 254 of "target" among the three radio buttons 252, 252, ... Attached to the character string 256 of the above-mentioned "histogram" is turned on. Then, in the histogram display area 248, the histogram 264 according to the corrected frequency Ho'[x] of the evaluation target area 266 is displayed. Therefore, from this histogram 264, the distribution of the corrected luminance Yo'[i, j] in the evaluation target region 266 can be recognized.

Since the other operation procedures on the sub screen 200 are not directly related to the main subject of the present invention, their description will be omitted. If necessary, refer to the second embodiment particularly described in Japanese Patent No. 4077430 described above.

When the above-mentioned close button 205 is turned on after the operation on the sub screen 200 is completed, the display on the display 12b returns to the main screen 100 as shown in FIG. 22.

That is, the above-mentioned reference mean value SMb and reference standard deviation SDb are displayed by the character strings 132 and 134 on the right side portion of the frame area 114. Then, the X-ray image 154 in the picture box 116 of the "image 1" is redisplayed based on the corrected luminance Y'[i, j]. In other words, the X-ray image 154 after the brightness Y [i, j] has been corrected is displayed in the picture box 116. Further, the character strings 120, 122 and 124 below the picture box 116 of the "image 1" display the al-BMD, the coefficient of variation CV and the fracture risk evaluation result RF for the "image 1". In FIG. 22, the al-BMD is 88.14, that is, the al-BMD is 82.0 or more, and the coefficient of variation CV is 14.77, that is, the coefficient of variation CV is 14.77. Since it is 12.2% or more and less than 15.4%, the evaluation result RF of "Caution" is given by collation with the evaluation criteria shown in FIG. In addition, the histogram 156 according to the corrected power Ho'[x] is displayed in the histogram display area 148. This histogram 156 is similar to the histogram 264 on the sub-screen 200 shown in FIG.

In the same manner as this, the sub screen 200 is displayed for the other "image A", that is, "image 2", "image 3", and "image 4", and the operation on the sub image 200 is performed. However, on the sub screen 200 other than the "image 1", the correction process is performed in a manner different from that in the "image 1".

That is, when the reference bar area 258 is selected on the sub screen 200 other than the “image 1”, each pixel constituting the reference bar area 258 is divided into different gradations x of the respective luminance Yb [i, j]. It is sorted. Then, the frequency Hb [x] for each gradation x is obtained based on the following equation tens.

In the number 10, Nb is the total number of pixels constituting the reference bar region 258, and nb [x] is the number of pixels allocated to the gradation x in the total number of pixels Nb. The frequency Hb [x] based on this number 10, that is, the normalized frequency Hb [x] is recorded in the above table 300.

Then, based on this normalized frequency Hb [x], the average value Mb and the standard deviation Db of the luminance Yb [i, j] in the reference bar region 258 are obtained. Specifically, the mean value Mb is obtained based on the equation 11, and the standard deviation Db is determined based on the equation 12. These mean values Mb and standard deviation Db are also recorded in Table 300.

Then, the brightness Y [i, j] of the entire enlarged image 210 in the area selection area 208 is corrected based on the following number 13 including the average value Mb and the standard deviation Db, that is, the corrected brightness Y. '[I, j] is required.

In short, for "image 2", "image 3" and "image 4" other than "image 1", the corrected luminance Yb'[i, j] of the reference bar area 258 of the "image 1" is used as a reference, respectively. The brightness Y [i, j] of is corrected, that is, the correction process is performed. The enlarged image 210 in the area selection area 208 is redisplayed based on the corrected luminance Yb'[i, j]. Further, in the reference bar display area 240, an enlarged image 262 of the reference bar area 258 based on the corrected luminance Yb'[i, j] is displayed.

The procedure after that is the same as in the case of "Image 1". That is, when the evaluation target area 266 is selected, the al-BMD and the coefficient of variation CV are finally obtained, and the fracture risk is evaluated.

After the operations on the sub screen 200 are performed for all the "images A" in this way, when the screen is returned to the main screen 100, the main screen 100 becomes as shown in FIG. 23.

That is, the X-ray image 154 in the picture box 116 of each "image A" is redisplayed based on the corrected luminance Y'[i, j]. As a result, the corrected brightness Y'[i, j] of each of the X-ray images 154, 154, ... In each of the picture boxes 116, 116, ... Is almost the same, and particularly corresponds to the block (reference bar) 38. The corrected brightness Y'[i, j] of the region to be corrected becomes uniform. Then, the character strings 120, 122, and 124 below the picture box 116 of each "image A" display the al-BMD, the coefficient of variation CV, and the fracture risk evaluation result RF for the "image A". .. Further, in the histogram display area 148, a histogram 156 according to the corrected frequency Ho'[x] of each "image A", that is, a total of four histograms 156, 156, ... Are displayed. It should be noted that these histograms 156, 156, ... Are displayed in different modes, such as colors and line types, in order to distinguish them from each other.

Here, regarding the operation of the PC12 (strictly speaking, the CPU (Central Processing Unit) in the PC12) according to the above-mentioned fracture risk evaluation program when the sub screen 200 is displayed, the fracture risk evaluation result RF is particularly subjected to correction processing. The operation of the PC 12 until the above is described will be described with reference to FIG. 24.

That is, when the reference bar area 258 on the enlarged image 210 in the area selection area 208 is selected by the operation of the pointing device as described with reference to FIG. 19, the PC 12 proceeds to step S1 and the reference bar area 258 Is set (to itself). Then, the process proceeds to step S3, and the brightness Y [i, j] of the entire enlarged image 210 including the reference bar area 258 is corrected. Specifically, for example, when "image 1" is enabled, the brightness Y [i, j] is corrected based on the above equation 1, that is, the corrected brightness Y'[i, j] is set. Ask. On the other hand, when any one of "image 2", "image 3" and "image 4" other than "image 1" is enabled, the above-mentioned number 3 to the above-mentioned number 3 to the above-mentioned number 3 to the above-mentioned number 3 to the above-mentioned number 3 to more detail The corrected luminance Y'[i, j] is obtained based on the number 13 including the calculation results of the number 5 and the number 10 to 12.

After the corrected luminance Y'[i, j] is obtained in this way, that is, after the correction process is performed, the PC 12 proceeds to step S5. Then, in step S5, when the evaluation target area 266 is selected by the operation of the pointing device as described with reference to FIG. 21, the selected evaluation target area 266 is set (to itself). Then, the PC 12 proceeds to step S7, and based on the above-mentioned number 7, specifically, based on the number 7 including the calculation result of the above-mentioned number 6, the corrected average value Mo'of the evaluation target area 266 is set. Ask. As described above, this corrected average value Mo'is treated as al-BMD.

Further, the PC 12 proceeds to step S9 and obtains the corrected standard deviation Do'of the evaluation target region 266 based on the above equation 8. Then, the process proceeds to step S11, and the coefficient of variation CV of the evaluation target region 266 is obtained based on the above equation 9. Further, the process proceeds to step S13, and the corrected standard deviation Do'obtained in step S9 and the coefficient of variation CV obtained in step S11 are collated with the evaluation criteria shown in FIG. 14, and the evaluation result RF of the fracture risk is given.

As described above, according to the present embodiment, the bone quality and bone density of the subject are evaluated based on the X-ray image of the mandible of the subject taken for dental treatment, and the risk of fracture is evaluated. That is, compared to the above-mentioned conventional technique that presupposes that a relatively large-scale examination called the DXA method is performed, evaluation of bone quality and bone density is realized by a simpler examination for dentistry, which in turn increases the risk of fracture. Evaluation is realized. This is expected to be extremely beneficial in the prevention and treatment of osteoporosis including osteopenia.

Incidentally, the alveolar bone of the mandible, which is the subject of the X-ray image in the present embodiment, starts to decrease in bone density (bone mass) earlier and more rapidly than the lumbar vertebra, which is the subject of the DXA method. Recent studies have shown that it progresses. It has also been found that the bone density of the spine including the lumbar spine may show a false increase due to metamorphic changes due to aging or strenuous exercise. From these facts, according to this embodiment based on the X-ray image of the alveolar bone of the mandible, the bone density is evaluated more accurately than the above-mentioned conventional technique based on the X-ray image of the lumbar spine by the DXA method. As a result, it is expected that the risk of fracture can be evaluated more accurately. This is also expected for the evaluation of bone quality.

Further, since the evaluation of bone quality in the present embodiment is based on the coefficient of variation CV which is a dimensionless number, especially when the shading of the original X-ray image (X-ray film 20) is different for each original X-ray image. Even if there is, the evaluation of the bone quality can be accurately realized regardless of this. That is, although the coefficient of variation CV referred to here represents the variation in the shading of the evaluation target region 60 shown in FIG. 2, it is not the absolute variation in the shading, but the shading of the shading with respect to the average value of the shading. Represents relative variability. On the other hand, the shading of the entire X-ray image including the evaluation target area 60 may change depending on various conditions such as at the time of photographing. When the shading of the entire X-ray image changes in this way, the absolute variation in the shading of the evaluation target region 60 also changes. On the other hand, the coefficient of variation CV does not represent the absolute variation in the shading of the evaluation target region 60 as described above, but represents the relative variation of the shading with respect to the average value of the shading. Even if the shading of the entire X-ray image changes, it is basically constant. Therefore, based on such a coefficient of variation CV, it is possible to accurately evaluate the bone quality. Even if the subjects are different, the shading itself of the evaluation target area 60 and the absolute variation of the shading of the evaluation target area 60 change. In this case as well, the bone quality is evaluated based on the coefficient of variation CV. Can be accurately realized.

Further, in the evaluation of the bone density in the present embodiment, the shading of the entire X-ray image including the reference bar 38 is corrected so that the shading of the reference bar 38 as an index becomes constant, and then this The bone density is evaluated based on the corrected X-ray image. Therefore, even in the evaluation of the bone density, even if the shading of the original X-ray image is different for each of the original X-ray images, the influence thereof can be eliminated and an accurate evaluation can be realized.

It should be noted that this embodiment is a specific example of the present invention, and does not limit the scope of the present invention.

For example, although it was decided that the bone quality evaluation based on the coefficient of variation CV is performed after the above-mentioned correction process, the bone quality evaluation based on the coefficient of variation CV may be performed before the correction process.

Further, the bone quality may be evaluated based on the reciprocal (1 / CV) of the coefficient of variation CV instead of the coefficient of variation CV.

Furthermore, it was decided that the fracture risk would be evaluated based on the coefficient of variation CV and al-BMD, but based on either the coefficient of variation CV or al-BMD, for example, based only on the coefficient of variation CV. , Fracture risk may be assessed. However, it goes without saying that a more accurate fracture risk assessment can be realized based on both the coefficient of variation CV and al-BMD.

In addition, the evaluation criteria for fracture risk shown in FIG. 14 is an example, and is not limited to this. For example, according to FIG. 14, when the coefficient of variation CV is 12.2% or more and less than 15.4%, and the al-BMD is 71.4% or more and less than 82.0%, A rating of "attention required" is given, but instead, a rating of "caution" may be given. Further, the boundary values of 15.4% and 12.2% for the coefficient of variation CV are also examples, and are not limited to these values. Similarly, the boundary values of 71.4 and 82.0 for al-BMD are examples and are not limited to these values. In the present embodiment, the evaluation criteria shown in FIG. 14 are set by the so-called retrospective study, but more appropriate evaluation criteria are set by conducting more clinical experiments including the prospective study. Is expected.

The fracture risk evaluation program installed on the PC 12 may be provided not only by the storage medium 14 such as a DVD-ROM but also through a telecommunication line such as the Internet. Further, the fracture risk assessment program is executed by the PC 12, that is, the fracture risk assessment device is realized by the so-called general-purpose computer called the PC 12, but the fracture risk assessment device is not limited to this, and the fracture risk assessment device is realized by a dedicated device. It may be realized.

10 Fracture risk assessment system 12 Personal computer 14 Storage medium 20 X-ray film

Claims (8)

It is a bone quality evaluation device that evaluates the bone quality of the subject based on the X-ray image of the mandible of the subject.
A bone quality calculation means for calculating the relative ratio between the standard deviation of the density of the specific region corresponding to the specific portion of the mandible in the X-ray image and the average value of the density of the specific region as a bone quality correlation value.
A bone quality evaluation means for evaluating the bone quality based on a comparison result between the bone quality correlation value and the bone quality threshold value by the bone quality calculation means, and
A bone quality evaluation device provided with. The specific portion includes the alveolar bone portion around the first premolar,
The bone quality evaluation device according to claim 1. In the bone quality evaluation device according to claim 1 or 2.
The bone quality evaluation means of the subject is based on the comparison result of the average value of the shading of the specific region as the bone density correlation value and the bone density threshold, and the comparison result of the bone quality correlation value and the bone quality threshold. A bone quality evaluation device that evaluates the risk of fractures. The X-ray image includes an image of a predetermined specimen placed side by side with the mandible.
A correction means for correcting the shading of the X-ray image so that the shading of the image of the sample matches a predetermined reference value is further provided.
The bone quality evaluation means evaluates the bone density correlation value based on the shading of the specific region in the corrected image after the correction by the correction means.
The bone quality evaluation device according to claim 3. It is a bone quality evaluation method for evaluating the bone quality of a subject based on an X-ray image of the mandible of the subject.
A bone quality calculation process for calculating the relative ratio between the standard deviation of the density of the specific region corresponding to the specific part of the mandible in the X-ray image and the average value of the density of the specific region as a bone quality correlation value.
A bone quality evaluation process for evaluating the bone quality based on a comparison result between the bone quality correlation value and the bone quality threshold value obtained by the bone quality calculation process , and
A bone quality evaluation method comprising. In the bone quality evaluation method according to claim 5,
The bone quality evaluation process, based on the comparison result of the comparison between the average value and the bone density threshold gray at the particular region of the bone density correlation value, and the bony correlation value and bony threshold of the subject A bone quality evaluation method for assessing the risk of fracture. It is a bone quality evaluation program for evaluating the bone quality of a subject based on an X-ray image of the mandible of the subject using a computer.
A bone quality calculation procedure for calculating the relative ratio between the standard deviation of the shading of a specific region corresponding to a specific part of the mandible in the X-ray image and the average value of the shading of the specific region as a bone quality correlation value.
The bone quality evaluation procedure for evaluating the bone quality based on the comparison result between the bone quality correlation value and the bone quality threshold value by the bone quality calculation procedure, and
A bone quality evaluation program that causes the above computer to execute. In the bone quality evaluation program according to claim 7.
The bone quality evaluation procedure is based on a comparison result between the average value of the shading of the specific region as a bone density correlation value and the bone density threshold, and a comparison result between the bone quality correlation value and the bone quality threshold. A bone quality assessment program that assesses the risk of fractures.

Images