Method, device, and program for supporting retinal photocoagulation

The support device and method utilize image analysis and machine learning to optimize retinal photocoagulation by accurately determining the need for additional laser irradiation and setting appropriate treatment parameters, addressing variability in photoreceptor cell destruction.

WO2026234877A1PCT designated stage Publication Date: 2026-11-12OSAKA UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
PCT/JP2026/005573
Authority / Receiving Office
WO · WO
Patent Type
Applications
Current Assignee / Owner
Priority Date
2025-05-07
Filing Date
2026-02-17
Publication Date
2026-11-12

AI Technical Summary

Technical Problem

The variability in photoreceptor cell destruction during retinal photocoagulation due to factors like cataracts and retinal surface conditions makes it difficult for doctors, especially less experienced ones, to accurately determine the need for additional laser irradiation, affecting the effectiveness of neovascular suppression.

Method used

A support device and method using image acquisition units to capture retina images before and after laser irradiation, with determination units to assess photoreceptor cell destruction and recommend additional laser settings, utilizing autofluorescence imaging and machine learning models to predict cell damage and optimize treatment.

Benefits of technology

The system provides accurate determination of additional laser irradiation needs and recommended settings, enhancing the effectiveness of retinal photocoagulation by ensuring adequate photoreceptor cell destruction while minimizing visual acuity impact.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure JP2026005573_12112026_PF_FP_ABST
    Figure JP2026005573_12112026_PF_FP_ABST
Patent Text Reader

Abstract

This support device (1) for supporting retinal photocoagulation for a patient comprises: a first image acquisition unit (3) that acquires a first image of the retina of the patient after a first period of time has elapsed since the retina was irradiated with laser light; and a determination unit (4) that uses at least the first image to determine whether additional laser light irradiation is necessary.
Need to check novelty before this filing date? Find Prior Art

Description

Method, apparatus and program for supporting retinal photocoagulation

[0001] The present invention relates to a method, apparatus and program for supporting retinal photocoagulation.

[0002] Diabetic retinopathy is a retinal disease caused by damage to the capillaries of the retina due to diabetes. When vascular disorders progress and capillaries become ischemic, immature blood vessels called neovessels develop to supply nutrients to photoreceptor cells, leading to proliferative diabetic retinopathy. Neovessels are extremely fragile and prone to bleeding; when bleeding occurs, a scab-like membrane is formed, making the retina prone to tearing or detachment. As the condition further progresses, neovascular glaucoma develops, in which the anterior chamber angle is occluded and aqueous humor circulation is impeded, leading to elevated intraocular pressure, which carries a risk of extreme visual acuity loss or blindness.

[0003] Therefore, a treatment called (pan)retinal photocoagulation (PRP) is performed for patients with proliferative diabetic retinopathy and its preceding stage, severe non-proliferative diabetic retinopathy. In retinal photocoagulation, the ischemic retina is irradiated with laser light to destroy photoreceptor cells, thereby suppressing the development and proliferation of neovessels.

[0004] If the degree and range of photoreceptor cell destruction are small, the effect of suppressing the development and proliferation of neovessels decreases, but if the degree and range of photoreceptor cell destruction are excessively increased, the impact on the patient's visual acuity also becomes significant. For this reason, for example, Table 48.7 of Non-Patent Document 1 defines reference values such as the intensity, irradiation time, and number of irradiations of the laser light applied to the retina.

[0005] Stephen J. Ryan, et al., "RETINA", Seventh Edition, Volume II, Mosby, 2023, p.1127

[0006] However, in actual laser irradiation during retinal photocoagulation, the degree of photoreceptor cell destruction varies due to factors such as opacity of the intermediate light-transmitting body caused by cataracts and the condition of the retinal surface. Therefore, even if laser irradiation is performed in accordance with the conditions described in Non-Patent Document 1, it is not always possible to destroy photoreceptor cells to a sufficient extent to obtain the neovascular suppression effect. For this reason, additional laser irradiation may be performed in some cases.

[0007] However, the decision of whether or not to perform additional laser irradiation varies among doctors, and it is particularly difficult for less experienced doctors to make an accurate judgment.

[0008] This invention has been made in view of the above problems, and aims to appropriately support retinal photocoagulation.

[0009] To solve the above problems, the present invention includes the following embodiments. 1. A support device for assisting retinal photocoagulation for a patient, comprising: a first image acquisition unit that acquires a first image of the retina after a first period has elapsed since laser light irradiation of the patient's retina; and a determination unit that determines whether additional laser light irradiation is necessary, using at least the first image. 2. The support device according to Claim 1, wherein the first image is an autofluorescence image. 3. The support device according to Claim 1, further comprising a second image acquisition unit that acquires a second image of the retina after a second period shorter than the first period has elapsed since laser light irradiation, or before laser light irradiation, wherein the determination unit acquires a third image of the retina after a third period longer than the first period has elapsed, using the first and second images, and determines whether additional laser light irradiation is necessary, using the third image. 4. The support device according to Claim 3, wherein the second image is an autofluorescence image. 5. The support device according to any one of claims 1 to 4, further comprising a recommendation setting determination unit that determines a recommended setting for the additional laser light irradiation when the determination unit determines that additional laser light irradiation is necessary. Claim 6. The support device according to claim 5, wherein the determination unit determines that additional laser light irradiation is necessary when the index value of the degree of destruction of photoreceptor cells in the retina is less than a predetermined target value. Claim 7. The support device according to claim 6, wherein the recommendation setting determination unit determines the recommended setting based on the difference between the index value and the target value. Claim 8. A support device for assisting retinal photocoagulation for a patient, comprising: an image acquisition unit that acquires an image of the retina before laser light irradiation of the patient's retina; and a recommendation setting determination unit that determines a recommended setting for the laser light irradiation using the image. Claim 9. A support method for assisting retinal photocoagulation for a patient, comprising: a first image acquisition step of acquiring a first image of the retina after a first period has elapsed since laser light irradiation of the patient's retina; and a determination step of determining whether additional laser light irradiation is necessary using at least the first image.Item 10. A support method for assisting retinal photocoagulation for a patient, comprising: an image acquisition step of acquiring an image of the patient's retina before laser light irradiation to the retina; and a recommendation setting determination step of determining a recommendation setting for laser light irradiation using the image. Item 11. A support program for assisting retinal photocoagulation for a patient, comprising: a first image acquisition step of acquiring a first image of the retina after a first period has elapsed since laser light irradiation to the patient's retina; and a determination step of determining whether additional laser light irradiation is necessary, using at least the first image, by causing a computer to execute these steps. Item 12. A support program for assisting retinal photocoagulation for a patient, comprising: an image acquisition step of acquiring an image of the patient's retina before laser light irradiation to the retina; and a recommendation setting determination step of determining a recommendation setting for laser light irradiation using the image.

[0010] According to the present invention, retinal photocoagulation can be appropriately supported.

[0011] This is a block diagram showing the configuration of the support device according to the first embodiment. This is a flowchart showing the processing procedure of the support method according to the first embodiment. (a) to (c) are examples of autofluorescence images of the retina immediately before surgery, 1.5 months after surgery, and 2 years after surgery, respectively. (a) and (b) are cross-sectional images of the retina immediately after irradiation with a 400 μm diameter laser beam for 100 ms and 12 months after irradiation, respectively. (a) and (b) are cross-sectional images of the retina immediately after irradiation with a 400 μm diameter laser beam for 20 ms and 12 months after irradiation, respectively. (a) is an example of an autofluorescence image, and (b) is an image in Figure 6(a) in which the areas where photoreceptor cells were destroyed are filled in red. (a) is an example of an autofluorescence image with an automatically set evaluation range, and (b) is an example of an autofluorescence image with a manually set evaluation range. (a) is an example of an autofluorescence image with a manually set evaluation area, and (b) is an example of an autofluorescence image with an automatically set evaluation area. This is an example of recommended settings output to the display. This is a block diagram showing the configuration of the support device according to the second embodiment. This is a flowchart showing the processing procedure of the support method according to the second embodiment. This is a block diagram showing the configuration of the support device according to the third embodiment. This is a flowchart showing the processing procedure of the support method according to the third embodiment. (a) is an example of a fundus image of the retina before laser light irradiation, and (b) is an example of an autofluorescence image two years after laser light irradiation of the retina.

[0012] Embodiments of the present invention will be described below with reference to the accompanying drawings. However, the present invention is not limited to the embodiments described below, and various modifications are possible without departing from its spirit.

[0013] [First Embodiment] (Configuration of the support device) Figure 1 is a block diagram showing the configuration of the support device 1 according to the first embodiment of the present invention. The support device 1 is a device that assists in retinal photocoagulation for a patient and can be configured with a general-purpose computer. In this embodiment, the patient is a patient with diabetic retinopathy, but the patient is not particularly limited as long as it is a patient with a disease that is a target for retinal photocoagulation, for example, it may be a patient with central retinal vein occlusion.

[0014] The support device 1 has a hardware configuration that includes a processor such as a CPU or GPU (not shown), a main memory such as DRAM or SRAM (not shown), and an auxiliary storage device 2 such as an HDD or SSD. The auxiliary storage device 2 stores a fundus image database 7 and a support program 8, etc.

[0015] The auxiliary storage device 2 may be externally attached to the support device 1, or it may be located on the cloud.

[0016] The support device 1 comprises a first image acquisition unit 3, a determination unit 4, a recommended setting determination unit 5, and an output unit 6 as functional blocks. These functional blocks may be implemented in hardware by logic circuits formed on an integrated circuit, but in this embodiment, they are implemented in software by the processor of the support device 1 reading the support program 8 into the main memory and executing it. The support program 8 may be downloaded to the support device 1 via a communication network such as the Internet, or the support program 8 may be recorded on a computer-readable non-temporary recording medium such as a CD-ROM and transferred to the support device 1 via the recording medium.

[0017] (Processing Procedure of the Support Method) Figure 2 is a flowchart showing the processing procedure of the support method according to the first embodiment. The support method has steps S1 to S6, of which the support device 1 executes steps S2 to S6. That is, the support program 8 causes the support device 1 to execute steps S2 to S6.

[0018] In step S1, the patient undergoes retinal photocoagulation surgery. In this surgery, photoreceptor cells (mainly rod cells) are destroyed by irradiating the ischemic area of ​​the patient's retina with laser light. The settings for laser light irradiation are not particularly limited, but in this embodiment, they are set to satisfy the following conditions specified in Non-Patent Literature 1. Some of the conditions specified in Non-Patent Literature 1 are shown in Table 1.

[0019]

[0020] To observe the condition of the retina after surgery, fundus images are taken using a fundus camera immediately after surgery and during subsequent examinations. The captured images are stored in the fundus image database 7. In this embodiment, the images stored in the fundus image database 7 are autofluorescence images, but this is not particularly limited in the present invention.

[0021] Figure 3(a) shows an example of autofluorescence images of the retina, with (a) to (c) being images taken immediately before surgery, one and a half months after surgery, and two years after surgery, respectively. As shown in Figure 3(b), the laser irradiation marks are white one and a half months after surgery, but as shown in Figure 3(c), most of the irradiation marks have changed to gray to black two years after surgery.

[0022] In autofluorescence images, brightness changes due to autofluorescence from cells that support photoreceptor cells (retinal pigment epithelial cells). In Figure 3(b), autofluorescence is not lost in the period immediately after laser irradiation. However, as time passes and retinal pigment epithelial cells also die, autofluorescence is lost and the area turns black, which is circumstantial evidence that photoreceptor cells have died. On the other hand, if photoreceptor cells are not destroyed by laser irradiation, as shown by the dashed lines in Figures 3(b) and (c), the autofluorescence image of the irradiation site is white for a short period after laser irradiation, but changes to gray over time. In other words, it is difficult to determine whether photoreceptor cells have been destroyed based on the brightness of the irradiation site immediately after treatment until a certain period of time has passed and autofluorescence from retinal pigment epithelial cells has disappeared.

[0023] Furthermore, comparing Figures 3(b) and 3(c), it can be seen that the size of many of the irradiation marks has changed. In other words, the size of the area where photoreceptor cells are destroyed by laser light irradiation may change over time.

[0024] Figures 4(a) and 4(b) show cross-sectional images of the retina immediately after irradiation with a 400 μm diameter laser beam for 100 ms and 12 months after irradiation, respectively. Figures 5(a) and 5(b) show cross-sectional images of the retina immediately after irradiation with a 400 μm diameter laser beam for 20 ms and 12 months after irradiation, respectively. As shown, the area of ​​photoreceptor cell damage may shrink over time. Conversely, if the laser beam intensity is high, the area of ​​photoreceptor cell damage may expand over time.

[0025] In step S2 (first image acquisition step) shown in Figure 2, the first image acquisition unit 3 of the support device 1 accesses the fundus image database 7 to acquire a first image of the retina after a first period has elapsed since laser light irradiation of the patient's retina. In this embodiment, the first period is two years, but it is not particularly limited as long as it is a period during which the brightness from the irradiation marks decreases when photoreceptor cells are destroyed, and the size of the destroyed photoreceptor cell area does not change significantly.

[0026] In step S3 (determination step), the determination unit 4 of the support device 1 determines whether additional laser light irradiation is necessary using the first image. In this embodiment, the determination unit 4 calculates an index value of the degree of destruction of photoreceptor cells in the retina (Photocoagulation Index) from the first image, which is an autofluorescence image, and determines that additional laser light irradiation is necessary if the calculated index value is less than a predetermined target value.

[0027] First, with reference to Figures 6 and 7, we will explain how to calculate an index value for the degree of photoreceptor cell destruction in the retina.

[0028] Figure 6(a) shows an example of an autofluorescence image, in which the detection unit 4 identifies irradiation marks with a brightness below a certain value as areas where photoreceptor cells have been destroyed. In Figure 6(b), the areas where photoreceptor cells have been destroyed are shown in red.

[0029] Furthermore, the determination unit 4 calculates an index value representing the proportion of the evaluation range in the retina where photoreceptor cells have been destroyed. The method for setting the evaluation range in the retina is not particularly limited, but a certain range corresponding to the imaging magnification in the autofluorescence image may be automatically set as the evaluation range, as shown in the elliptical area in red in Figure 7(a), or the user may manually set the evaluation range to an area that they have determined should be irradiated with a laser, as shown in the area in red in Figure 7(b).

[0030] Furthermore, as shown below, the difference in index values ​​between automatically and manually set evaluation ranges is not significant enough to hinder the implementation of the invention.

[0031] Figure 8(a) is an example of an autofluorescence image with a manually set evaluation region, and Figure 8(b) is an example of an autofluorescence image with an automatically set evaluation region. In Figures 8(a) and (b), the five images are of the same retina with different viewing directions, and the index value is shown below each image. The index values ​​when the evaluation region is manually set range from 0.3810626 to 0.395707166 (average: 0.387617162), and the index values ​​when the evaluation region is automatically set range from 0.375936957 to 0.402384173 (average: 0.384875956), showing that the difference between the two is small. Therefore, the evaluation region can be set either manually or automatically.

[0032] As described above, the determination unit 4 determines that additional laser light irradiation is necessary if the calculated index value is less than a predetermined target value. The specific numerical value of the target value is not particularly limited, but can be determined appropriately by comprehensively considering the effect of suppressing neovascularization by photoreceptor cell destruction and the impact on visual acuity. For example, the target value can be determined based on the degree of photoreceptor cell destruction expected when laser light is irradiated under the conditions specified in Non-Patent Document 1.

[0033] If the indicator value is less than the predetermined target value, it is determined that additional laser light irradiation is necessary (YES in step S4), and the process proceeds to step S5. If the indicator value is equal to or greater than the predetermined target value, it is determined that additional laser light irradiation is not necessary (NO in step S4), and the process proceeds to step S6.

[0034] In step S5 (recommended setting determination step), the recommended setting determination unit 5 of the support device 1 determines the recommended settings for additional laser light irradiation. In this embodiment, the recommended setting determination unit 5 determines the recommended settings based on the difference between the index value and the target value calculated in step S3. More specifically, the recommended setting determination unit 5 calculates irradiation conditions such that the index of the degree of photoreceptor cell destruction reaches the target value when laser light irradiation is performed again, based on the settings of laser light irradiation performed in the past, the index value of the degree of photoreceptor cell destruction caused by that laser light irradiation, and the difference between the index value and the target value, and determines these irradiation conditions as the recommended settings.

[0035] In step S6 (output step), the output unit 6 of the support device 1 outputs the determination result from the determination unit 4 and / or the recommended setting from the recommended setting determination unit 5. If step S5 is performed, the output unit 6 outputs the recommended setting. If step S5 is not performed, the output unit 6 outputs only the determination result.

[0036] Figure 9 shows an example of the recommended settings displayed on the screen. While Figure 9 also displays the current indicator and target values, these are optional. If step S5 is not performed, i.e., if it is determined that additional laser irradiation is unnecessary, the display will show a message indicating that additional laser irradiation is not required.

[0037] The output method of the output unit 6 is not particularly limited. For example, the output unit 6 may output the judgment result and / or recommended settings as sound via a speaker or print them on paper via a printer.

[0038] (Summary) As described above, the support device 1 according to this embodiment automatically determines whether additional laser irradiation is necessary using a first image of the retina after a first period has elapsed since laser irradiation. Doctors and others can decide whether additional laser irradiation is necessary by referring to the determination result by the support device 1, so that the support device 1 can appropriately support retinal photocoagulation.

[0039] Further, when it is determined that additional laser light irradiation is necessary, the support device 1 determines recommended settings for the additional laser light irradiation. A doctor or the like can refer to the recommended settings from the support device 1 to determine the conditions for the additional laser light irradiation when performing the additional laser light irradiation.

[0040] [Second Embodiment] In the first embodiment described above, the necessity or the like of additional laser light irradiation is determined using an image after a long period of time (for example, 2 years) has elapsed since the laser light irradiation. However, in the second embodiment described below, an improved mode that can output the necessity of additional laser light irradiation and / or recommended settings even earlier will be described.

[0041] (Configuration of Support Device) FIG. 10 is a block diagram showing the configuration of a support device 1' according to the second embodiment of the present invention. The hardware configuration of the support device 1' is the same as that of the support device 1 shown in FIG. 1.

[0042] The support device 1' includes, as functional blocks, a first image acquisition unit 3, a second image acquisition unit 3', a determination unit 4, a recommended setting determination unit 5, and an output unit 6. An fundus image database 7, a support program 8, and a trained model 9 are stored in the auxiliary storage device 2 of the support device 1'. That is, the support device 1' has a configuration in which the support device 1 shown in FIG. 1 is further provided with the second image acquisition unit 3' as a functional block, and the trained model 9 is further stored in the auxiliary storage device 2.

[0043] (Processing Procedure of Support Method) FIG. 11 is a flowchart showing the processing procedure of the support method according to the second embodiment. This flowchart is obtained by adding step S2' between step S2 and step S3 in the flowchart shown in FIG. 2.

[0044] The processing in step S1 is the same as that in the first embodiment.

[0045] In step S2 (first image acquisition step), the first image acquisition unit 3 of the support device 1' accesses the fundus image database 7, and acquires a first image of the retina after a first period has elapsed since the laser light irradiation to the patient's retina. The first period in the present embodiment is shorter than the first period in the first embodiment, and is, for example, one and a half months. That is, as in the image shown in FIG. 3(b), the first image in the present embodiment has the luminance of the irradiation mark not significantly reduced, and it is difficult for human vision to determine whether or not the photoreceptor cells in the irradiation mark are destroyed, and whether or not the size of the destroyed portion of the photoreceptor cells changes with further passage of time.

[0046] In step S2' (second image acquisition step), the second image acquisition unit 3' of the support device 1' acquires a second image of the retina after a second period has elapsed since the laser light irradiation. The second period is not particularly limited as long as it is a period shorter than the first period, but is preferably as short a time as possible (immediately after irradiation).

[0047] In step S3 (determination step), the determination unit 4 of the support device 1' uses the first image and the second image to acquire a third image of the retina after a third period has elapsed, and uses the third image to determine the necessity of additional laser light irradiation. The third period is longer than the first period (one and a half months), for example, two years (the same as the first period in the first embodiment), but is not particularly limited as long as it is a period in which the luminance from the irradiation mark decreases when the photoreceptor cells are destroyed, and the size of the destroyed portion of the photoreceptor cells generally stops changing.

[0048] The method for acquiring the third image is not particularly limited, but in this embodiment, the determination unit 4 acquires the third image using a trained model 9. The trained model 9 is a predictive model generated by machine learning using autofluorescence images obtained from past examinations. Specifically, training data is created by labeling images taken immediately after laser irradiation and images taken one and a half months after laser irradiation with images taken two years after laser irradiation as ground truth data, and the trained model 9 is generated by training an artificial intelligence model such as a neural network using the training data. When the determination unit 4 receives the first and second images as input, the trained model 9 outputs a third image, which is an autofluorescence image taken two years after laser irradiation, as a predicted image.

[0049] The determination unit 4 uses the third image acquired from the trained model 9 to determine whether additional laser light irradiation is necessary. Specifically, it calculates an index value of the degree of photoreceptor cell destruction in the retina from the third image, and if the calculated index value is less than a predetermined target value, it determines that additional laser light irradiation is necessary. The method for calculating the index value is the same as in the first embodiment, so a detailed explanation is omitted.

[0050] Furthermore, the processing in steps S4 to S6 shown in Figure 11 is the same as in the first embodiment. That is, if the index value is less than a predetermined target value, it is determined that additional laser light irradiation is necessary (YES in step S4), and the process proceeds to step S5. If the index value is equal to or greater than the predetermined target value, it is determined that additional laser light irradiation is not necessary (NO in step S4), and the process proceeds to step S6. In step S5, the recommended setting determination unit 5 of the support device 1' determines the recommended setting for additional laser light irradiation based on the difference between the index value and the target value calculated in step S3. In step S6, the output unit 6 of the support device 1' outputs the determination result from the determination unit 4 and / or the recommended setting from the recommended setting determination unit 5.

[0051] (Summary) As described above, the support device 1' according to this embodiment generates a third image as a predicted image, which is obtained after a third period of time has elapsed since laser irradiation, using the first image obtained after a first period of time has elapsed since laser irradiation and the second image obtained after a second period of time has elapsed since laser irradiation. The third image is then used to automatically determine whether additional laser irradiation is necessary. The only images actually required for processing by the support device 1' are the first and second images, both of which are obtained within a short period of time (for example, within one and a half months) after laser irradiation. Therefore, in this embodiment, it is possible to determine whether additional laser irradiation is necessary and / or to determine recommended settings even earlier.

[0052] [Third Embodiment] In the first and second embodiments described above, the necessity of additional laser irradiation is determined using the image after laser irradiation. In the third embodiment described below, a method is described in which the recommended settings for laser irradiation are made using the image before laser irradiation.

[0053] (Configuration of the support device) Figure 12 is a block diagram showing the configuration of the support device 1" according to the third embodiment of the present invention. The hardware configuration of the support device 1" is the same as that of the support device 1 shown in Figure 1 and the support device 1' shown in Figure 10.

[0054] Support device 1" comprises an image acquisition unit 13, a recommended setting determination unit 15, and an output unit 6 as functional blocks. The auxiliary storage device 2 of support device 1" stores a fundus image database 7, a support program 8, and a trained model 19.

[0055] (Processing procedure of the support method) Figure 13 is a flowchart showing the processing procedure of the support method according to the third embodiment.

[0056] In step S11 (image acquisition step), the image acquisition unit 13 of the support device 1 acquires an image of the patient's retina before laser light irradiation.

[0057] In step S12 (recommended setting determination step), the recommended setting determination unit 15 of the support device 1" determines the recommended settings for laser irradiation using the images acquired in step S11. In this embodiment, the recommended setting determination unit 15 determines the recommended settings using a trained model 19. The trained model 19 is a predictive model generated by machine learning using fundus images, autofluorescence images, and laser irradiation setting records obtained from past examinations.

[0058] The method for generating the trained model 19 is described below. Figure 14(a) is a fundus image of the retina before laser irradiation, and Figure 14(b) is an autofluorescence image of the same retina two years after laser irradiation. By performing the same processing on this autofluorescence image as in the determination unit 4 in the first embodiment, it is possible to determine whether or not the photoreceptor cells have been sufficiently destroyed (whether or not additional laser irradiation is necessary). Furthermore, if the photoreceptor cells have not been sufficiently destroyed, by performing the same processing as in the recommended setting determination unit 5 in the first embodiment, the recommended settings for additional laser irradiation can be determined. As a result, if the photoreceptor cells have been sufficiently destroyed, it is determined that the settings at the time of the previous laser irradiation were appropriate. On the other hand, if the photoreceptor cells have not been sufficiently destroyed, it is determined that the irradiation conditions assuming that the previous laser irradiation and the additional laser irradiation are performed at the same time are appropriate.

[0059] In this way, appropriate settings for a single laser irradiation can be derived from the fundus image before laser irradiation and the autofluorescence image two years after laser irradiation, as shown in Figures 14(a) and (b). Then, by labeling the fundus image before laser irradiation and the autofluorescence image two years after laser irradiation with the above-mentioned appropriate settings as ground truth data, training data is created, and by training an artificial intelligence model such as a neural network using the training data, a trained model 19 is generated. When the trained model 19 receives an image before laser irradiation as input to the recommended setting determination unit 5, it outputs appropriate settings for laser irradiation as predicted data. The recommended setting determination unit 15 determines the output from the trained model 19 as the recommended settings.

[0060] In step S13 (output step), the output unit 6 of the support device 1" outputs the recommended settings determined by the recommended setting determination unit 15.

[0061] In step S14, laser light is irradiated onto the ischemic area of ​​the patient's retina based on the recommended settings output in step S13.

[0062] (Summary) As described above, the support device 1" according to this embodiment automatically determines the recommended settings for laser irradiation using an image of the retina before laser irradiation. Doctors and others can perform laser irradiation by referring to the recommended settings of the support device 1", thereby increasing the possibility of sufficiently destroying photoreceptor cells with a single laser irradiation. This eliminates the need for additional laser irradiation and contributes to shortening the treatment period.

[0063] [Additional Notes] The present invention is not limited to the embodiments described above, and various modifications are possible within the scope of the claims. Forms obtained by appropriately combining the technical means disclosed in the embodiments are also included in the technical scope of the present invention.

[0064] For example, in the second embodiment, the second image is an image taken after a second period has elapsed since laser irradiation (for example, immediately after irradiation), but it may also be an image taken before laser irradiation (for example, immediately before irradiation). In this case, the first image may also be an image taken immediately after laser irradiation. That is, the determination unit 4 may use the first image taken after a first period has elapsed since laser irradiation (for example, immediately after irradiation or one and a half months later) and the second image taken before laser irradiation to acquire a third image taken after a third period has elapsed (for example, two years later), and use the three images to determine whether additional laser irradiation is necessary.

[0065] 1. Support device 1' Support device 1” Support device 2. Auxiliary storage device 3. First image acquisition unit 3' Second image acquisition unit 4. Judgment unit 5. Recommended setting determination unit 6. Output unit 7. Fundus image database 8. Support program 9. Trained model 13. Image acquisition unit 15. Recommended setting determination unit 19. Trained model

Claims

1. A support device for assisting retinal photocoagulation in a patient, comprising: a first image acquisition unit that acquires a first image of the retina after a first period has elapsed since laser light irradiation of the patient's retina; and a determination unit that determines whether additional laser light irradiation is necessary using at least the first image.

2. The support device according to claim 1, wherein the first image is a spontaneous fluorescence image.

3. The support device according to claim 1, further comprising a second image acquisition unit that acquires a second image of the retina after a second period shorter than the first period has elapsed since the laser light irradiation, or before the laser light irradiation, wherein the determination unit acquires a third image of the retina after a third period longer than the first period has elapsed using the first and second images, and determines whether or not the additional laser light irradiation is necessary using the third image.

4. The support device according to claim 3, wherein the second image is a spontaneous fluorescence image.

5. The support device according to any one of claims 1 to 4, further comprising a recommendation setting determination unit that determines a recommended setting for the additional laser light irradiation when the determination unit determines that additional laser light irradiation is necessary.

6. The support device according to claim 5, wherein the determination unit determines that additional laser light irradiation is necessary if the index value of the degree of destruction of photoreceptor cells in the retina is less than a predetermined target value.

7. The support device according to claim 6, wherein the recommended setting determination unit determines the recommended setting based on the difference between the index value and the target value.

8. A support device for assisting retinal photocoagulation in a patient, comprising: an image acquisition unit that acquires an image of the patient's retina before laser light irradiation to the retina; and a recommendation setting determination unit that determines recommended settings for laser light irradiation using the image.

9. A method for assisting retinal photocoagulation in a patient, comprising: a first image acquisition step of acquiring a first image of the retina after a first period has elapsed since laser light irradiation of the patient's retina; and a determination step of determining whether additional laser light irradiation is necessary using at least the first image.

10. A support method for assisting retinal photocoagulation in a patient, comprising: an image acquisition step of acquiring an image of the patient's retina before laser light irradiation of the retina; and a recommendation setting determination step of determining a recommendation setting for laser light irradiation using the image.

11. A support program for assisting retinal photocoagulation for a patient, the support program causing a computer to perform a first image acquisition step of acquiring a first image of the retina after a first period has elapsed since laser light irradiation of the patient's retina, and a determination step of determining whether additional laser light irradiation is necessary using at least the first image.

12. A support program for assisting retinal photocoagulation in a patient, the support program causing a computer to perform an image acquisition step of acquiring an image of the patient's retina before laser light irradiation of the retina, and a recommendation setting determination step of determining recommended settings for laser light irradiation using the image.