JP7393798B2 - Diagnosis support device, diagnosis support method, and diagnosis support program - Google Patents

Description

The present invention relates to a technique for supporting the diagnosis of vascular disease, and particularly to a technique for supporting the diagnosis of arterial dissection.

Arterial dissection is defined as blood flow within a blood vessel that enters the blood vessel wall and spreads between the walls. The blood vessel wall of an artery is composed of an internal elastic lamina, tunica media, and tunica adventitia from the inside. Arteries are divided into elastic arteries, whose tunica media is rich in elastic fibers, and muscular arteries, whose tunica media contains almost no elastic fibers. It is classified into arteries. In muscular arteries, the tunica media (smooth muscle media) is soft, so if the internal elastic lamina is torn for some reason, blood flows into the smooth muscle media layer, forming a false lumen and causing arterial dissection.

Muscular arteries include, for example, cerebral arteries, brachial arteries, femoral arteries, popliteal arteries, etc., and arterial dissection is particularly common in cerebral arteries. According to national survey data described in Non-Patent Document 1, it has been shown that the vertebral artery and basilar artery are the most common sites of cerebral artery dissection in Japanese people.

In the clinical diagnosis of cerebral artery dissection, imaging diagnosis is particularly important in addition to the presence of antecedent headache and changes in the shape of blood vessels over time. Specifically, by proving the presence of an intraluminal flap (hereinafter referred to as a flap) and a false intraluminal hematoma, which are specific to arterial dissection, using angiography, MRA (CE-MRA) original images, CTA original images, etc. , a diagnosis of cerebral artery dissection is made (Non-Patent Document 2). In particular, the presence of a flap is considered a reliable indicator of cerebral artery dissection.

In addition, in a survey conducted by the present inventors on 10 patients in whom changes in the shape of blood vessels over time and intraluminal hematomas were confirmed, structures such as flaps of vertebrobasilar artery dissection and intraluminal hematomas were depicted. 3T fat-suppressed T1-weighted images using local excitation (HR vfl-TSE method) were shown to be more useful than original MRA and CTA images.

Akira Yamaura, et al. 3 others, "Nationwide survey of non-traumatic intracranial dissecting arterial lesions (first report)", Stroke Surgery, 1998, 26 (2), p. 79-86 Jun Goto, “Cerebral artery dissection”, Journal of the Japanese Society of Internal Medicine, June 10, 2009, Volume 98, No. 6, p. 91-98

However, currently, image diagnosis of arterial dissection is performed only by MRI, without cerebral angiography or histological proof. Furthermore, the ability to detect flaps and false intraluminal hematomas depends on the qualitative evaluations of doctors and technicians who interpret cerebrovascular images, so there is room for bias. Furthermore, even when using the resolution of the HR vfl-TSE method, there were cases in which structures within the dissection cavity were insufficiently visualized, and the diagnostic accuracy of image interpretation could not be said to be high.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a diagnostic support device that improves the accuracy of diagnosing vascular diseases.

The inventors of the present invention have made extensive studies to solve the above problems, and have discovered that it is possible to predict a disease based on the brightness distribution in a line segment set in a cross-sectional image of a blood vessel.

That is, the present invention includes the inventions described in the following sections.
Item 1.
A diagnostic support device that supports diagnosis of vascular disease,
an image acquisition unit that acquires an image including a cross section of the blood vessel;
a line segment setting unit that sets one or more line segments that cross the cross section;
a brightness distribution calculation unit that calculates a brightness distribution in the line segment;
A diagnostic support device equipped with
Item 2.
a curve creation unit that creates a brightness distribution curve representing the brightness distribution;
a display unit that displays the brightness distribution curve;
Item 2. The diagnostic support device according to Item 1, further comprising:
Item 3.
3. The diagnosis support device according to item 1 or 2, further comprising a prediction unit that predicts the presence or absence of the disease based on the brightness distribution.
Item 4.
4. The diagnosis support device according to item 3, wherein the prediction unit predicts the presence or absence of the disease based on a brightness change pattern in the line segment.
Item 5.
5. The diagnosis support device according to item 3 or 4, wherein the prediction unit predicts the presence or absence of the disease based on a difference between a maximum value and a minimum value of brightness in the line segment.
Item 6.
4. The diagnosis support device according to item 3, wherein the prediction unit makes predictions using artificial intelligence.
Section 7.
7. The diagnostic support device according to any one of Items 1 to 6, wherein the blood vessel is a cerebral artery.
Section 8.
8. The diagnostic support device according to item 7, wherein the disease is arterial dissection.
Item 9.
9. The diagnosis support device according to any one of Items 1 to 8, wherein the line segment setting unit sets four or more line segments that intersect with each other at at least one location.
Item 10.
10. The diagnostic support device according to item 9, wherein the line segment passes through a central region of the cross section.
Item 11.
11. The diagnosis support device according to any one of Items 1 to 10, wherein the image is a 3T fat-suppressed T1-weighted image using local excitation.
Item 12.
11. The diagnosis support device according to any one of Items 1 to 10, wherein the image is a CTA original image.
Item 13.
11. The diagnosis support device according to any one of Items 1 to 10, wherein the image is an MRA original image.
Section 14.
A diagnostic support method for supporting the diagnosis of vascular disease, the method comprising:
an image acquisition step of acquiring an image including a cross section of the blood vessel;
a line segment setting step of setting one or more line segments that cross the cross section;
a brightness distribution calculation step of calculating a brightness distribution in the line segment of the cross section;
A diagnostic support method with
Item 15.
A diagnostic support program that operates a computer as a diagnostic support device that supports the diagnosis of vascular disease,
an image acquisition unit that acquires an image including a cross section of the blood vessel;
a line segment setting unit that sets one or more line segments that cross the cross section, and
a brightness distribution calculation unit that calculates a brightness distribution in the line segment of the cross section;
A diagnostic support program that runs a computer as a diagnostic support program.

According to the present invention, the accuracy of diagnosing vascular diseases can be improved.

1 is a block diagram showing the configuration of a diagnostic support system according to an embodiment of the present invention. (a) is a cross-sectional view showing an example of a patent type arterial dissection, and (b) to (d) are HR vfl-TSE images, MRA original images, and CTA original images corresponding to the cross-sections, respectively. . (a) is a cross-sectional view showing an example of arterial dissection with intraluminal hematoma formation, and (b) to (d) are HR vfl-TSE images, MRA original images, and CTA original images corresponding to the cross-sections, respectively. It is an image. (a) shows the state in which four line segments are set in the HR vfl-TSE image shown in FIG. 2(b), and (b) to (e) each of the four line segments. It is a graph showing a brightness distribution curve. (a) shows the state in which four line segments are set in the HR vfl-TSE image shown in FIG. 3(b), and (b) to (e) each of the four line segments. It is a graph showing a brightness distribution curve. (a) shows a state in which four line segments are set in a cross-sectional image of a normal artery in which an artifact appears, and (b) to (e) show each brightness distribution curve for each of the four line segments. This is a graph showing. (a) shows the state in which four line segments are set in the MRA original image shown in Fig. 2 (c) or the CTA original image shown in Fig. 2 (d), and (b) to (e) respectively , is a graph showing each brightness distribution curve in each of the four line segments. (a) shows the state where four line segments are set in the MRA original image shown in FIG. 3(c), and (b) to (e) each show the brightness distribution in each of the four line segments. It is a graph showing a curve. (a) shows the state in which four line segments are set in the CTA original image shown in FIG. 3(d), and (b) to (e) each show the brightness distribution of each of the four line segments It is a graph showing a curve. 1 is a flowchart showing a procedure of a diagnosis support method according to an embodiment of the present invention. It is a block diagram showing the composition of the diagnostic support system concerning the modification of one embodiment of the present invention. FIG. 6 is a box plot showing the distribution of brightness differences in each brightness distribution curve of patent type arterial dissection, false intraluminal hematoma formation type arterial dissection, and atherosclerosis.

Embodiments of the present invention will be described below with reference to the accompanying drawings. Note that the present invention is not limited to the embodiments described below.

(overall structure)
FIG. 1 is a block diagram showing the configuration of a diagnostic support system 1 according to an embodiment of the present invention. The diagnosis support system 1 includes an imaging device 2 and a diagnosis support device 3.

The imaging device 2 is a device that captures medical images. As the imaging device 2, an MRI device or a CT device can be used, but a 3 Tesla MRI device is suitable. The 3 Tesla MRI system uses local excitation technology and variable flip angle Turbo Spin Echo method (hereinafter referred to as ``vfl-TSE method'') to produce 3T fat-suppressed T1-weighted images (hereinafter referred to as ``HR'') using High resolution 3D vfl-TSE method. vfl-TSE image).

In this embodiment, the imaging device 2 captures an HR vfl-TSE image (3T fat-suppressed T1-weighted image using local excitation) including a cross section of a cerebral artery of a subject. Further, the imaging device 2 is communicably connected to the diagnosis support device 3 , and images captured by the imaging device 2 are input to the diagnosis support device 3 .

(Diagnostic support device)
The diagnosis support device 3 can be configured with, for example, a general-purpose personal computer, and has a hardware configuration including a CPU (not shown), a main storage device (not shown), an auxiliary storage device 31, an input section 32, and a display section. It is equipped with 33. In the diagnostic support device 3, the CPU reads various programs stored in the auxiliary storage device 31 into the main storage device and executes them, thereby performing various calculation processes.

The auxiliary storage device 31 can be configured with, for example, a hard disk drive (HDD) or a solid state drive (SSD). The auxiliary storage device 31 stores medical images captured by the imaging device 2 as well as a diagnostic support program P. The diagnosis support program P may be recorded on a non-temporary computer-readable recording medium such as a CD-ROM, and by causing the diagnosis support device 3 to read the recording medium, the diagnosis support program P can be read by the diagnosis support device 3. You can install it on 3. Alternatively, the code of the diagnostic support program P may be downloaded to the diagnostic support device 3 via a communication network such as the Internet. Note that the auxiliary storage device 31 may be built into the diagnosis support device 3 or may be provided as an external storage device separate from the diagnosis support device 3.

The input unit 32 can be configured with, for example, a keyboard, a mouse, a touch panel, or the like. Further, the display section 33 can be configured by, for example, a liquid crystal display.

The diagnosis support device 3 has a function of supporting diagnosis of vascular disease, and in this embodiment, particularly has a function of supporting diagnosis of arterial dissection of a cerebral artery. In order to realize this function, the diagnosis support device 3 includes a control section 34. The control unit 34 is a functional block realized by the CPU reading out the diagnostic support program P stored in the auxiliary storage device 31 into the main storage device and executing it.

The control unit 34 includes an image acquisition unit 341 that acquires an image including a cross section of a blood vessel, a line segment setting unit 342 that sets one or more line segments that cross the cross section, and a brightness controller that calculates a brightness distribution in the line segment. It includes a distribution calculation section 343 and a curve creation section 344 that creates a brightness distribution curve representing the brightness distribution. The functions of these parts will be explained in more detail below.

(Image acquisition)
The image acquisition unit 341 acquires an image including a cross section of the subject's cerebral artery from the imaging device 2. In this embodiment, a cross section means a plane substantially perpendicular to the short axis (Axial) of a blood vessel.

Hereinafter, a description will be given of how a cross section of an artery is displayed depending on the medical image and the type of arterial dissection. As described above, in this embodiment, an HR vfl-TSE image (3T fat-suppressed T1-weighted image using local excitation), an MRA original image, and a CTA original image are assumed as medical images. Furthermore, arterial dissection is classified into a patent type and an intra-false lumen hematoma formation type, depending on whether or not a thrombus is generated within the false lumen.

FIG. 2(a) shows an example of the cross-sectional structure of an artery in a patent type artery dissection. In patent type arterial dissection, there is a flap dissected from the vessel wall in the lumen, and blood flow exists in both the true lumen and the false lumen, and the artery is The image and CTA original image are displayed as shown in FIGS. 2(b) to 2(d), respectively. That is, in the HR vfl-TSE image, the brightness of the blood flow part is low, whereas in the MRA original image and the CTA original image, the brightness of the blood flow part is high. Furthermore, in all of the HR vfl-TSE image, the MRA original image, and the CTA original image, the luminance of the blood vessel wall and flap is approximately between black and white.

Further, FIG. 3(a) shows an example of the cross-sectional structure of an artery in a false intraluminal hematoma formation type arterial dissection. In intra-false lumen hematoma formation type arterial dissection, there is a flap dissociated from the blood vessel wall in the lumen, blood flow exists in the true lumen, a hematoma exists in the false lumen, and the artery is The HR vfl-TSE image, MRA original image, and CTA original image are displayed as shown in FIGS. 3(b) to 3(d), respectively. That is, in the HR vfl-TSE image and the MRA original image, the brightness of the hematoma part is high, whereas in the CTA original image, the brightness of the hematoma part is low.

(line segment settings)
In this embodiment, the line segment setting unit 342 shown in FIG. 1 sets four line segments that cross the cross section of the cerebral artery and intersect with each other in the central region of the cross section. For example, if the image acquired by the image acquisition unit 341 is an HR vfl-TSE image and the image includes a cross section of an artery in a patent type arterial dissection as shown in FIG. 2(b), the line segment setting unit 342 sets four line segments L1 to L4 as shown in FIG. 4(a).

Line segments L1 to L4 are all straight lines, and intersect with each other in the central region of the cross section. The angle between line segment L1 and line segment L2, the angle between line segment L2 and line segment L3, the angle between line segment L3 and line segment L4, and the angle between line segment L4 and line segment L1 are all 45°. It is. Note that these angles do not necessarily have to be the same.

Furthermore, the line segments L1 to L4 do not necessarily have to be straight lines and do not have to protrude from the cross section. Furthermore, although the line segments L1 to L4 do not necessarily have to intersect at one point, it is preferable that they pass through the central region. Furthermore, if the cross section is not circular, the central region may be specified by, for example, determining the center of gravity of the cross section.

(Brightness distribution calculation/brightness distribution curve creation)
The brightness distribution calculation unit 343 shown in FIG. 1 calculates the brightness distribution in the line segment set by the line segment setting unit 342. For example, as shown in FIG. 4A, when line segments L1 to L4 are set in the arterial cross-sectional image of patent type arterial dissection, the brightness distribution calculation unit 343 The brightness of each pixel up to 1 is calculated, and similar calculations are performed for the other line segments L2 to L4.

The curve creation unit 344 creates a brightness distribution curve (profile curve) representing the brightness distribution calculated by the brightness distribution calculation unit 343. The brightness distribution curve created by the curve creation section 344 is displayed on the display section 33.

(Prediction when using HR vfl-TSE image)
FIGS. 4(b) to 4(e) are graphs showing respective brightness distribution curves in the line segments L1 to L4 set in the HR vfl-TSE image shown in FIG. 4(a), respectively. In each graph, the vertical axis corresponds to brightness, and the horizontal axis corresponds to coordinates along each line segment L1 to L4. In the HR vfl-TSE image, the brightness is low in areas where blood flow is present, and the brightness is medium in the flap. Therefore, in the line segment crossing the flap, the brightness changes from medium (blood vessel wall) to low (blood flow) to medium (flap) to low (blood flow) to medium (blood vessel wall).

Here, since the figures shown in FIG. 4(a) etc. are schematic cross-sectional views, the change in brightness at the boundary part is steep, but in the actual image, the change in brightness at the boundary part is gradual. . Therefore, the actual brightness distribution curve is not an angular curve as shown in FIGS. 4(b) to 4(e), but a rounded curve. Therefore, as shown in FIGS. 4C and 4D, the brightness distribution curve in which the brightness changes from medium to low to medium to low to medium has an ω shape as a whole.

In addition, as shown in FIG. 5(a), when the line segments L1 to L4 are set in the HR vfl-TSE image of the arterial cross section of the arterial dissection with intraluminal hematoma formation type, the brightness of each line segment L1 to L4 is The distribution curves are shown in FIGS. 5(b) to (e), respectively. In the HR vfl-TSE image, the area where the hematoma is present has higher brightness than the flap. Therefore, in the line segment that crosses the flap, the brightness is medium (vessel wall) → low (hematoma) → medium (flap) → high (blood flow) → medium (vessel wall), or medium (vessel wall) → high. Changes from (blood flow) → medium (flap) → low (hematoma) → medium (vascular wall).

Here, as mentioned above, the actual brightness distribution curve is rounded and the width of the flap is small, so the brightness of the part from the blood flow (or hematoma) through the flap to the hematoma (or blood flow) The distribution curve exhibits a ∫ (integral sign) shape. For example, the brightness distribution curves shown in FIGS. 5(c) and 5(d) include a portion exhibiting a horizontally inverted ∫ shape.

As described above, in the HR vfl-TSE image, when a flap exists in the artery of patent type arterial dissection, the brightness distribution curve of at least one of the line segments L1 to L4 exhibits an ω shape. Further, when a flap exists in an artery with intraluminal hematoma formation type arterial dissection, the brightness distribution curve of at least one of the line segments L1 to L4 includes a portion exhibiting a ∫ shape. Therefore, users (doctors, technicians, etc.) can determine whether a flap exists or not and whether arterial dissection is occurring based on whether the brightness distribution curve has an ω shape and whether it includes a ∫ shape portion. type can be predicted.

Note that even in a normal artery without a flap, a high-intensity part called an artifact may appear in the center of the blood flow region in the HR vfl-TSE image, as shown in Figure 6(a). . If the line segments L1 to L4 pass through the artifact, each of the brightness distribution curves of the line segments L1 to L4 will have an ω shape, as shown in FIGS. 6(b) to 6(e).

Here, when fewer than four line segments are set in the cross section where the flap exists, the smaller the number of line segments, the higher the probability that all the line segments will cross the flap. Therefore, when the set number of line segments is less than four, even if the brightness distribution curve has an ω shape, it becomes difficult to determine whether it is due to a flap or an artifact.

On the other hand, when four or more line segments are set, in the cross section where the flap exists, only a portion of the four line segments L1 to L4 are connected to the flap from the false lumen, as shown in FIG. 4(a), for example. The brightness distribution curves of all the line segments L1 to L4 do not have an ω shape. Therefore, by setting four or more line segments, even if an artifact occurs in an image of a normal artery, it is possible to prevent erroneous prediction that a flap exists in the artery. Thereby, the prediction accuracy of arterial dissection can be further improved compared to the case where the set number of line segments is less than four.

(Prediction when using MRA original image or CTA original image)
FIG. 7(a) shows a state in which line segments L1 to L4 are set in the MRA original image or CTA original image of an artery cross section of a patent type arterial dissection, and FIGS. 7(b) to (e) show 7(a) is a graph showing each brightness distribution curve in line segments L1 to L4 shown in FIG. 7(a). In both the MRA original image and the CTA original image, the brightness of the blood flow portion is higher than that of the flap. Therefore, in the line segment from the false lumen to the true lumen via the flap, the brightness changes from medium (blood vessel wall) → high (blood flow) → medium (flap) → high (blood flow) → medium (blood vessel wall). Therefore, as shown in FIGS. 7(c) and (d), the brightness distribution curve in which the brightness changes from medium to high to medium to high to medium has an ω shape that is vertically inverted as a whole (inverted ω shape). present.

FIG. 8(a) shows the state in which line segments L1 to L4 are set in the MRA original image of the arterial cross section of a false intraluminal hematoma formation type arterial dissection, and FIG. 8(b) to (e) respectively 8A is a graph showing each brightness distribution curve in line segments L1 to L4 shown in FIG. 8(a). In the MRA original image, the hematoma area also has higher brightness than the flap, similar to the blood flow area. Therefore, in the line segment from the false lumen to the true lumen via the flap, the brightness changes from medium (vascular wall) → high (hematoma) → medium (flap) → high (blood flow) → medium (vascular wall). Therefore, as shown in FIGS. 8(c) and 8(d), the brightness distribution curve in which the brightness changes from medium to high to medium to high to medium has an inverted ω shape as a whole.

In this way, in the MRA original image, if there is a flap in the artery, at least one of the brightness distribution curves will have an inverted ω shape, regardless of whether it is a patent type arterial dissection or a false intraluminal hematoma formation type arterial dissection. exhibits. Therefore, when using the original MRA image, the presence of a flap can be predicted, but the type of arterial dissection cannot be determined.

Figure 9(a) shows the state in which line segments L1 to L4 are set in the CTA original image of the arterial cross section of a false intraluminal hematoma formation type arterial dissection, and Figures 9(b) to (e) respectively. 9A is a graph showing each brightness distribution curve in line segments L1 to L4 shown in FIG. 9(a). In the CTA original image, the brightness of the hematoma part is lower than that of the flap. Therefore, in the line segment from the false lumen to the true lumen via the flap, the brightness changes from medium (vascular wall) → low (hematoma) → medium (flap) → high (blood flow) → medium (vascular wall). As described above, since the actual brightness distribution curve is a rounded curve, the portion of the brightness distribution curve where the curve changes from medium to high to medium has a Δ (delta) shape. Therefore, as shown in FIGS. 9(c) and 9(d), the brightness distribution curve in which the brightness changes from medium to low to medium to high to medium includes a portion exhibiting a Δ shape.

In this way, in the CTA original image, in the case of patent type arterial dissection, at least one of the brightness distribution curves exhibits an inverted ω shape, as shown in Figures 7(c) and (d), indicating the formation of a false intraluminal hematoma. In the case of type arterial dissection, at least one of the brightness distribution curves includes a portion exhibiting a Δ shape, as shown in FIGS. 9(c) and 9(d). Therefore, the user can predict the presence or absence of a flap and the type of arterial dissection based on whether the brightness distribution curve has an inverted ω shape and whether it includes a Δ-shaped portion. can.

(Diagnosis support method)
FIG. 10 is a flowchart showing the procedure of the diagnosis support method according to this embodiment. The diagnosis support method according to this embodiment is implemented by the diagnosis support system 1.

In step S1, the imaging device 2 captures an image including a cross section of a cerebral artery of the subject.

In step S2 (image acquisition step), the image acquisition unit 341 acquires an image including a cross section of the cerebral artery from the imaging device 2.

In step S3 (line segment setting step), the line segment setting unit 342 sets one or more line segments that cross the cross section included in the image. In this embodiment, four line segments L1 to L4 are set, as shown in FIG. 4(a) and the like.

In step S4 (brightness distribution calculation step), the brightness distribution calculation unit 343 calculates the brightness distribution in the line segments L1 to L4.

In step S5, the curve creation unit 344 creates a brightness distribution curve representing the calculated brightness distribution.

In step S6, the display unit 33 displays the created brightness distribution curve.

(Summary)
As described above, in this embodiment, the line segment setting unit 342 sets one or more line segments that cross the cross section of an image including a cross section of an artery acquired by the image acquisition unit 341, and the brightness distribution calculation unit 343 calculates the brightness distribution in the line segment, a curve creation unit 344 creates a brightness distribution curve representing the brightness distribution, and the display unit 33 displays the brightness distribution curve. Thereby, the user can predict the presence or absence of cerebral artery dissection based on the shape of the brightness distribution curve.

When the image is an HR vfl-TSE image, the user determines whether a flap is present or not based on whether the brightness distribution curve has an ω shape and whether it includes a ∫ shape portion, and The type of cerebral artery dissection can be predicted. When the image is a CTA original image, the user can determine whether a flap exists or not and whether the brain The type of arterial dissection can be predicted. If the image is an MRA original image, the user cannot predict the type of cerebral artery dissection, but based on whether the brightness distribution curve has an inverted omega shape, the user can determine whether a flap is present or not, that is, whether a cerebral artery dissection exists or not. It is possible to predict whether dissociation has occurred.

Therefore, by using the diagnostic support device 3, it is possible to improve the accuracy of diagnosing cerebral artery dissection compared to diagnosis based only on image interpretation, which tends to vary depending on the skill of the doctor or technician.

[Modified example]
Modifications of this embodiment will be described below. In addition, in this modification, members having the same functions as the members already described are given the same reference numerals, and the description thereof will be omitted.

FIG. 11 is a block diagram showing the configuration of a diagnostic support system 1' according to a modification of this embodiment. The diagnosis support system 1' includes an imaging device 2 and a diagnosis support device 3'. That is, the diagnosis support system 1' has a configuration in which the diagnosis support device 3 in the diagnosis support system 1 shown in FIG. 1 is replaced with a diagnosis support device 3'.

The diagnosis support device 3' includes a CPU (not shown), a main storage device (not shown), an auxiliary storage device 31, an input section 32, a display section 33, and a control section 34'. . That is, the diagnosis support apparatus 3' has a configuration in which the control section 34 in the diagnosis support apparatus 3 shown in FIG. 1 is replaced with a control section 34'.

The control section 34' includes an image acquisition section 341, a line segment setting section 342, a brightness distribution calculation section 343, and a prediction section 345. That is, the control section 34' has a configuration in which the curve creation section 344 in the control section 34 shown in FIG. 1 is replaced with a prediction section 345.

The prediction unit 345 predicts the presence or absence of a disease, particularly the presence or absence of cerebral artery dissection, based on the brightness distribution calculated by the brightness distribution calculation unit 343. That is, in the above embodiment, the user predicted the presence or absence of cerebral artery dissection based on the shape of the brightness distribution curve displayed on the display unit 33, but in this modification, the prediction unit 345 Similar criteria are used to predict the presence or absence of cerebral artery dissection.

For example, if the medical image is an HR vfl-TSE image and the brightness changes from medium to low to medium to low to medium in at least one of the set line segments (i.e., if a brightness distribution curve is created , a curve exhibiting an ω-shape as a whole), the prediction unit 345 predicts that it is a patent type arterial dissection. Similarly, the medical image is an HR vfl-TSE image, and in at least one of the set line segments, the brightness partially changes from low → medium → high or high → medium → low (i.e., When a brightness distribution curve is created, if the curve includes a portion exhibiting a ∫ shape, the prediction unit 345 predicts that it is a false intraluminal hematoma formation type arterial dissection.

In addition, the medical image is an MRA original image, and in at least one of the set line segments, the brightness changes from medium → high → medium → high → medium (that is, when a brightness distribution curve is created, the brightness of the entire (a curve exhibiting an inverted ω shape), the prediction unit 345 predicts that it is an arterial dissection. Note that when the medical image is an MRA original image, the prediction unit 345 does not predict whether the type of arterial dissection is a patent type or a false intraluminal hematoma formation type.

In addition, the medical image is a CTA original image, and in at least one of the set line segments, the brightness changes from medium → high → medium → high → medium (that is, when creating a brightness distribution curve, the brightness of the entire (a curve exhibiting an inverted ω shape), the prediction unit 345 predicts that it is a patent type arterial dissection. Similarly, the medical image is a CTA source image, and in at least one of the set line segments, the brightness partially changes from medium to high to medium (i.e., when a brightness distribution curve is created, △ (a curve including a portion exhibiting a shape), the prediction unit 345 predicts that it is a false intraluminal hematoma formation type arterial dissection.

In this way, the prediction unit 345 can predict the presence or absence of arterial dissection based on the brightness change pattern in the line segment. The prediction result by the prediction unit 345 is displayed on the display unit 33. Thereby, the diagnosis support device 3' can support the user's diagnosis of cerebral artery dissection.

As a further modification, it is preferable that the prediction unit 345 performs prediction using artificial intelligence (AI). In this case, for medical images of a large number of subjects, the image acquisition section 341, line segment setting section 342, and brightness distribution calculation section 343 are used to calculate the brightness distribution of each line segment set in the cross section of the artery. Training data is created by associating the distribution data with the definitive diagnosis results for each subject. Based on this training data, machine learning is performed using, for example, a neural network, and the prediction unit 345 is realized by a learned algorithm. Note that the artificial intelligence is not limited to neural networks, and may be based on any known artificial intelligence technology.

In general, the prediction accuracy improves as the number of line segments set in the cross section of the artery increases, but in the user's prediction described in the above embodiment, the number of brightness distribution curves increases, making the prediction even more difficult. On the other hand, in prediction by the prediction unit 345 (particularly prediction using artificial intelligence), highly accurate prediction is possible by setting a large number of line segments.

Note that atherosclerosis is a disease in which cross-sectional images of blood vessels resemble arterial dissection. In blood vessels that have developed atherosclerosis, plaque (athema) is formed in which cholesterol and other substances accumulate. Therefore, the cross-sectional image of the blood vessel approximates the image in FIG. 5(a) in which the hematoma in the false lumen is replaced with cholesterol or the like. Therefore, in an HR vfl-TSE image, if the brightness partially changes from low to medium to high or high to medium to low in at least one of the set line segments, and a brightness distribution curve is created, false Similar to intraluminal hematoma formation type arterial dissection, the curve includes a ∫-shaped portion.

Here, the brightness distribution curve of atherosclerosis is different from the brightness distribution curve of false intraluminal hematoma formation type arterial dissection due to the difference between the maximum and minimum brightness values (hereinafter referred to as the brightness difference (signal difference)). tends to be small (see Figure 12). Therefore, when the luminance partially changes from low to medium to high or from high to medium to low in at least one line segment set in the blood vessel portion of the HR vfl-TSE image, the prediction unit 345 further calculates Based on the brightness difference in the brightness distribution curve, it is possible to predict whether it is a false intraluminal hematoma formation type arterial dissection or atherosclerosis.

Moreover, as shown in FIG. 12, the brightness difference of the brightness distribution curve is
There is a tendency for pseudoluminal hematoma formation type arterial dissection > atherosclerosis > patent type arterial dissection. Therefore, the prediction unit 345 can predict the presence or absence of the above three diseases based only on the brightness difference of the brightness distribution curve. However, in order to improve prediction accuracy, the prediction unit 345 preferably predicts the presence or absence of a disease based on both the brightness change pattern and the brightness difference in the line segment.

[Additional notes]
The present invention is not limited to the above-mentioned embodiments, and various changes can be made within the scope shown in the claims, and forms obtained by appropriately combining the technical means disclosed in the embodiments are also within the scope of the present invention. included in the target range.

Examples of the present invention will be described below, but the present invention is not limited thereto.
In this example, HR vfl-TSE images were obtained for 26 patients with patent type arterial dissection, 28 patients with intraluminal hematoma formation type arterial dissection, and 15 patients with atherosclerosis. Then, four line segments were set in the blood vessel portion in each image, and the brightness distribution in the line segments was calculated. Furthermore, for each brightness distribution, the difference between the maximum value and the minimum value of brightness (brightness difference) was calculated.

FIG. 12 is a boxplot showing the distribution of brightness differences in the brightness distribution curves of patent type arterial dissection (ω), false intraluminal hematoma formation type arterial dissection (∫), and atherosclerosis (p∫). . As shown in the same figure, the brightness differences in the brightness distribution curves of patent type arterial dissection, false intraluminal hematoma formation type arterial dissection, and atherosclerosis have interquartile ranges that do not overlap with each other;
There is a tendency for pseudoluminal hematoma formation type arterial dissection > atherosclerosis > patent type arterial dissection. Therefore, it was found that the brightness difference in the brightness distribution curve can be used as an index for predicting the presence or absence of the above three diseases.

In the above embodiment, the diseases to be predicted are cerebral artery dissection and atherosclerosis, but there is no particular limitation as long as it is a vascular disease.

1 Diagnosis support system 1' Diagnosis support system 2 Imaging device 3 Diagnosis support device 3' Diagnosis support device 31 Auxiliary storage device 32 Input section 33 Display section 34 Control section 34' Control section 341 Image acquisition section 342 Line segment setting section 343 Brightness distribution Calculation unit 344 Curve creation unit 345 Prediction unit L1 to L4 Line segment P Diagnosis support program

Claims (9)

A diagnostic support device that supports diagnosis of vascular disease,
an image acquisition unit that acquires an image including a cross section of the blood vessel;
a line segment setting unit that sets one or more line segments that cross the cross section;
a brightness distribution calculation unit that calculates a brightness distribution in the line segment;
a curve creation unit that creates a brightness distribution curve representing the brightness distribution;
a display unit that displays the brightness distribution curve;
Equipped with
the vascular disease is cerebral artery dissection ;
The image is a diagnosis support device, wherein the image is a 3T fat-suppressed T1-weighted image using local excitation, a CTA original image, or an MRA original image . The diagnosis support device according to claim 1 , further comprising a prediction unit that predicts the presence or absence of the disease based on the brightness distribution. The diagnosis support device according to claim 2 , wherein the prediction unit predicts the presence or absence of the disease based on a pattern of changes in brightness in the line segment. The diagnosis support device according to claim 2 or 3, wherein the prediction unit predicts the presence or absence of the disease based on a difference between a maximum value and a minimum value of brightness in the line segment. The diagnosis support device according to claim 3, wherein the prediction unit makes predictions using artificial intelligence. The diagnosis support device according to any one of claims 1 to 5 , wherein the line segment setting section sets four or more line segments that intersect with each other at at least one location. The diagnostic support device according to claim 6 , wherein the line segment passes through a central region of the cross section. A diagnostic support method for supporting the diagnosis of vascular disease, the method comprising:
an image acquisition step of acquiring an image including a cross section of the blood vessel;
a line segment setting step of setting one or more line segments that cross the cross section;
a brightness distribution calculation step of calculating a brightness distribution in the line segment of the cross section;
a curve creation step of creating a brightness distribution curve representing the brightness distribution;
a display step of displaying the brightness distribution curve;
Equipped with
the vascular disease is cerebral artery dissection ;
The diagnosis support method , wherein the image is a 3T fat-suppressed T1-weighted image using local excitation, a CTA original image, or an MRA original image . A diagnostic support program that operates a computer as a diagnostic support device that supports the diagnosis of vascular disease,
an image acquisition unit that acquires an image including a cross section of the blood vessel;
a line segment setting unit that sets one or more line segments that cross the cross section ;
a brightness distribution calculation unit that calculates a brightness distribution in the line segment of the cross section;
a curve creation unit that creates a brightness distribution curve representing the brightness distribution, and
a display unit that displays the brightness distribution curve;
run the computer as
the vascular disease is cerebral artery dissection ;
The image is a diagnosis support program in which the image is a 3T fat-suppressed T1-weighted image using local excitation, a CTA original image, or an MRA original image .

Images