Systems, apparatus, and methods for treating retinal and macular diseases using artificial intelligence
OCT devices with AI and machine learning models address the subjective nature of AMD treatment by providing personalized plans based on patient data, enhancing treatment efficacy and consistency.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- ラダスジョン グレゴリー
- Filing Date
- 2024-05-07
- Publication Date
- 2026-06-10
Smart Images
Figure 2026518863000001_ABST
Abstract
Description
Technical Field
[0001] Cross - reference to Related Applications This application claims priority to U.S. Patent Application No. 63 / 500,842, filed on May 8, 2023, titled "SYSTEMS, APPARATUS AND METHODS FOR TREATMENT OF RETINAL AND MACULAR DISEASES USING ARTIFICIAL INTELLIGENCE", which is assigned to the assignee of the present application and is hereby incorporated by reference in its entirety.
[0002] Aspects of the present invention relate to systems, apparatuses, and methods for the treatment of retinal and macular diseases using artificial intelligence.
Background Art
[0003] Retinal and macular diseases are the main causes of blindness worldwide, and billions of dollars are spent annually on the treatment of these diseases. Examples of retinal and macular diseases include, for example, age - related macular degeneration (AMD), diabetic macular edema (DME), edema due to retinal vein occlusion (RVO), and myopic choroidal neovascularization (CNV). Also, retinal and macular diseases may include exudation due to retinal vascular diseases.
[0004] The visual prognosis of retinal and macular diseases has improved over the past decade through the use of anti - vascular endothelial growth factor (anti - VEGF) injections that can stabilize or improve the edema associated with these diseases. There are various formulations of anti - VEGF drugs, with various characteristics such as efficacy, duration of effect, and cost in specific patients.
[0005] The measurement and morphology of optical coherence tomography (OCT) are the main factors when determining treatment strategies, although other factors such as changes in visual acuity, age, and other comorbidities may also be considered.
Summary of the Invention
[0006] Aspects of this disclosure may include apparatus and methods for generating treatment plans for retinal and macular diseases.
[0007] In some embodiments, the techniques described herein relate to methods for treating retinal and macular diseases, the methods comprising: performing optical coherence tomography (OCT) on a patient to quantify an initial level of macular edema or retinal exudation; administering a series of anti-vascular endothelial growth factor (anti-VEGF) injections to the patient over a period of time; performing additional periodic OCT on the patient after each anti-VEGF injection to determine the treatment response, assessed by subsequent levels of edema or retinal exudation; collecting a set of patient-based training data, the training data comprising initial levels of macular edema or retinal exudation, subsequent levels of edema or retinal exudation, patient information, and treatment information; and applying the training data for a patient to a machine learning model trained on multiple sets of training data for multiple patients to determine a treatment plan for the patient's AMD after a series of anti-VEGF injections.
[0008] In some embodiments, the techniques described herein relate to methods, further comprising the step of training a machine learning model on multiple sets of training data to select a treatment plan that optimizes a patient's vision.
[0009] In some embodiments, the techniques described herein relate to methods, and the patient information includes at least age, visual acuity, and an indication of the presence or absence of one or more other diseases.
[0010] In some embodiments, the techniques described herein relate to methods in which a machine learning model is trained using a gradient boosting library to construct a random forest model.
[0011] In some embodiments, the techniques described herein relate to methods in which the initial level of macular edema is the thickness measured by OCT or the fluid volume measured by OCT.
[0012] In some embodiments, the techniques described herein relate to methods, and the treatment information includes at least the date of each anti-VEGF injection, the drug of each anti-VEGF injection, and the dosage of each anti-VEGF injection.
[0013] In some embodiments, the techniques described herein relate to methods and further include the step of administering an anti-VEGF injection to a patient as indicated by a treatment plan.
[0014] In some cases, the retinal disease or macular disease is age-related macular degeneration (AMD).
[0015] In some aspects, retinal or macular diseases are one or more of the following: age-related macular degeneration (AMD), diabetic macular edema (DME), edema due to retinal vein occlusion (RVO), and myopic choroidal neovascularization (CNV).
[0016] In some embodiments, the technology described herein relates to an optical coherence tomography (OCT) apparatus including: a sensor configured to detect a level of macular edema in the eye; a memory storing computer executable instructions; and a processor coupled to the memory and configured to execute the computer executable instructions, the processor executing the computer executable instructions to cause the OCT apparatus to: perform optical coherence tomography (OCT) on a patient to quantify an initial level of macular edema; receive treatment information for a series of anti-vascular endothelial growth factor (anti-VEGF) injections on a patient over a period of time; perform additional periodic OCT on the patient after each anti-VEGF injection to determine a subsequent level of edema; collect a set of patient-based training data, the training data including an initial level of macular edema, a subsequent level of edema, patient information, and treatment information; and apply the training data on a patient to a machine learning model trained on multiple sets of training data on multiple patients to determine a treatment plan for the patient's AMD after a series of anti-VEGF injections.
[0017] In some embodiments, the technology described herein relates to an OCT device, wherein at least one processor causes the OCT device to execute instructions to train a machine learning model to select a treatment plan that optimizes the patient's visual acuity based on multiple sets of training data.
[0018] In some embodiments, the technology described herein relates to an OCT device, and the patient information includes at least age, visual acuity, and an indication of the presence or absence of one or more other diseases.
[0019] In some embodiments, the techniques described herein relate to OCT equipment, and the machine learning model is trained using a gradient boosting library to model a random forest.
[0020] In some embodiments, the techniques described herein relate to an OCT device, where the initial level of macular edema is the thickness of the retina measured by OCT, or the amount of fluid measured by OCT.
[0021] In some embodiments, the technology described herein relates to an OCT device, and the treatment information includes at least the date of each anti-VEGF injection, the drug of each anti-VEGF injection, and the dosage of each anti-VEGF injection.
[0022] In some embodiments, the technology described herein relates to a computer network including: a memory storing computer executable instructions; and a processor coupled to the memory and configured to execute computer executable instructions, the processor executing the computer executable instructions to cause the processor to: receive training data from multiple optical coherence tomography (OCT) devices, the training data including, for each of multiple patients, an initial level of macular edema or retinal exudation, a subsequent level of macular edema or retinal exudation after each of a series of anti-vascular endothelial growth factor (anti-VEGF) injections to the patient over a period of time, patient information, and treatment information; train a machine learning model on the training data to select a treatment plan that optimizes the patient's predicted visual acuity based on the patient dataset, the patient dataset including an initial level of macular edema, a subsequent level of edema after each of an initial series of anti-VEGF injections to the patient over a period of time, patient medical information, and patient treatment information; and apply the patient dataset for a patient to the machine learning model to determine a treatment plan for the patient's AMD after an initial series of anti-VEGF injections.
[0023] In some embodiments, the technology described herein relates to a computer network, and the patient medical information includes at least age, visual acuity, and an indication of the presence or absence of one or more other diseases.
[0024] In some embodiments, the techniques described herein relate to computer networks, and machine learning models are trained using gradient boosting libraries to construct random forest models.
[0025] In some embodiments, the technology described herein relates to a computer network, and the initial level of macular edema is the thickness measured by OCT or the fluid volume measured by OCT.
[0026] In some embodiments, the technology described herein relates to a computer network, and the treatment information includes at least the date of each anti-VEGF injection, the agent of each anti-VEGF injection, and the dosage of each anti-VEGF injection.
[0027] In some embodiments, the technology described herein relates to a computer network and further includes a patient computer device, and the processor is further configured to output a treatment plan to the patient computer device.
[0028] Further advantages and novel features of aspects of the present invention will be explained in part in the following description, and in part will become more apparent to those skilled in the art upon examination of the following, or upon learning by practice.
Brief Description of the Drawings
[0029] In the drawings:
[0030] [Figure 1] FIG. 1 is a diagram of an exemplary optical coherence tomography (OCT) system configured to recommend treatment for age-related macular degeneration (AMD) based on a trained model according to an aspect of the present disclosure. [Figure 2] FIG. 2 is a diagram of an exemplary treatment process of a patient using the OCT system of FIG. 1 according to an aspect of the present disclosure. [Figure 3] FIG. 3 is a diagram of an exemplary computer system according to an aspect of the present disclosure. [Figure 4] FIG. 4 is a diagram of an exemplary communication system 400 that can be used according to an aspect of the present disclosure. [Figure 5] FIG. 5 is a flowchart illustrating an exemplary method for providing a recommended treatment plan for AMD according to an aspect of the present disclosure.
Modes for Carrying Out the Invention
[0031] In the initial evaluation of patients suspected of having exudative AMD, a thorough clinical examination is performed to confirm that the cause is related to AMD and not due to another disease process. Once confirmed, the patient undergoes optical coherence tomography (OCT) of the macular region to quantify the amount of edema. At this point, the patient is usually initiated with one of the anti-vascular endothelial growth factor (anti-VEGF) drugs, administered three times at one-month intervals. Rarely, a change in medication may be decided during this initial period. OCT to quantify macular edema is measured at each visit to assess improvement or worsening. After the first three injections, a decision is made to discontinue treatment, change medications, or continue treatment with observation or extension. This decision is almost always determined by the amount of edema confirmed by OCT at that particular visit. This pattern continues throughout the patient's life, and each physician selects the optimal treatment for a particular patient based on their own clinical judgment, utilizing their knowledge and experience.
[0032] Current treatment methods for exudative AMD may face several challenges. Firstly, current patient treatment may rely on the subjective clinical judgment of individual physicians, who are limited by their personal experience and published research on treatment options. As treatment options increase, it may become more difficult for each physician to predict the outcomes of various treatment choices. Secondly, patients may change physicians, and the new physician may not be familiar with the factors that influenced the previous treatment decision.
[0033] Aspects of this disclosure may include systems, apparatus, and methods for the treatment of age-related macular degeneration (AMD). An OCT device or a network device connected thereto may be configured to utilize artificial intelligence to recommend treatment for a patient's AMD based on a history of OCT measurements as well as patient and treatment information. In some embodiments, the artificial intelligence may include a machine learning model trained on multiple sets of training data for multiple patients. For example, an OCT device may be communicably connected to a network of other OCT devices that pool training data for training a machine learning model. Furthermore, the network may provide recommended treatments to connected devices (e.g., the patient's mobile device) to inform the patient of treatment options.
[0034] Next, looking at Figure 1, the exemplary OCT system 100 may be configured to recommend treatment for AMD based on a trained model 172. The OCT system 100 may be implemented in an OCT device 110 and an associated computer. The OCT device 110 may include a sensor 112. For example, the sensor 112 can use reflected light waves to generate an image of the back of the eye (e.g., the retina and optic nerve). The sensor 112 measures the amount of light reflected by various layers of the retina and optic nerve. OCT can reveal the thickness, structure, and health of the retina and optic nerve. In some implementations, the OCT images can be used to determine the amount of fluid in the eye.
[0035] The OCT system 100 may further include a computer system. For example, the OCT system 100 may include a processor 104 and memory 106. The processor 104 can store computer executable instructions for performing the processes described herein. In some implementations, the computer system may be integrated with the OCT device. In some implementations, all or part of the computer system may be distributed. For example, in some implementations, the computer system may be a cloud network 180 containing geographically distributed computing resources.
[0036] The OCT system 100 may include an image processing device 120 configured to process images from the sensor 112 to determine one or more properties of the eye. For example, the image processing device 120 can analyze the images to determine the level of macular edema based on the measured thickness. As another example, the image processing device 120 can determine the amount of fluid. The image processing device can store the characteristics of the ocular patient data in the patient data storage 122.
[0037] The OCT system 100 may include a user interface 130. The user interface 130 can guide the user (e.g., a technician) when operating the OCT device 110 to acquire patient measurements. The user interface 130 may also include a user interface (e.g., a graphical user interface) for the user to manually input patient data or to acquire patient data from an external source such as electronic medical records. In some implementations, the user interface 130 can display patient data to the user by, for example, automatically importing measurements from patient data storage 122 into fields of the user interface.
[0038] The OCT system 100 may include a learning machine 170, which is configured to train a model 172 to determine a treatment plan for a patient's AMD after a series of anti-VEGF treatments, based on patient data and training data including AMD treatment information for multiple patients. The learning machine 170 can retrieve patient data from patient data storage 122 and / or a user interface 130. Patient data may include the patient's history (e.g., age, sex, other medical conditions), initial level of macular edema, initial visual acuity, AMD treatment information (e.g., date, medication, dosage), and pre- or post-treatment levels of macular edema and visual acuity. The learning machine 170 can retrieve training data from a training set 160, which may be aggregated based on patients examined using the OCT device 110 and external validation results 184. The training data may include the same types of data as the patient data.
[0039] The learning machine 170 may be a machine learning library that provides tools for training model 172. For example, the learning machine 170 may be a distributed gradient boosting library such as XGBoost. The gradient boosting library may be configured to generate a recommendation system for recommending treatment plans for individual patients. The treatment plans recommended by the recommendation system may be treatment plans that optimize the predicted visual acuity of individual patients based on patient data. The distributed gradient boosting library can combine various machine learning models such as decision trees, bagging classifiers, and random forests. Boosting refers to building a stronger classifier from many weak classifiers by building subsequent models to correct errors in previous models. Therefore, a gradient boosting library can be used to combine multiple classifiers (e.g., decision trees) to recommend treatment plans for patients and provide a stronger classifier. For example, each component classifier may be trained to classify patient data into components of a treatment plan (e.g., recommended medication, dosage, treatment interval). The weights of each component classifier may be selected to optimize visual acuity based on the training data. The gradient boosting library may select from among the component classifiers to optimize visual acuity. Furthermore, Model 172 can be validated using a reserved subset of the training data. Additional validation and / or retraining may be repeated using a new set of training data. Model 172 can output a recommended treatment plan to display 108.
[0040] In another embodiment, some functions of the OCT system 100 may be performed via the cloud network 180. For example, the OCT system 100 may include a management portal 140 that controls communication between the OCT system 100 and the cloud network 180. The management portal 140 may be locally controlled via a user interface 130 and an access control unit 150. For example, the access control unit 150 may allow a specific user to retrieve and upload data to the cloud network 180.
[0041] The cloud network 180 can host various services that may be utilized by the OCT system 100. For example, the cloud network 180 can host a computing system 182 which includes computing resources such as processors and memory that may be used when performing some of the functions described herein. For example, in some implementations, model training may be performed on system 182. In some implementations, the cloud network 180 may include storage for validation results 184. Validation results 184 may be anonymized patient data including, for example, information about AMD treatment (e.g., OCT measurements, treatments performed, visual acuity results, etc.). In some implementations, the OCT system 100 can import the validation results 184 and use them as a training set 160.
[0042] Figure 2 is a diagram of an example treatment process 200 for a patient using the OCT system 100. In one embodiment, the OCT system 100 can perform OCT on the patient at each visit, which may be scheduled regularly (e.g., monthly). In one embodiment, the treatment process 200 may include an initial visit 210, a treatment visit 220, and a subsequent visit 230.
[0043] An initial visit 210 may be a visit to an ophthalmologist before a diagnosis of AMD is made. During the initial visit 210, the OCT system 100 can measure the patient's eye. Measurements by the OCT system 100 may constitute part of the initial diagnosis. The initial diagnosis may be performed by the ophthalmologist. For example, the ophthalmologist may decide whether to treat the AMD during the initial visit 210. The initial diagnosis may include excluding non-AMD conditions that the OCT device 110 and / or model 172 are not trained to diagnose. In some implementations, a set of patient data may not be available during the initial visit 210. For example, prior to anti-VEGF treatment, the patient data may not contain treatment information. In some implementations, if the ophthalmologist diagnoses AMD, the ophthalmologist may further decide how to treat the AMD (e.g., which anti-VEGF drug to use). For example, anti-VEGF drugs may include AVASTIN (bevacizumab), LUCENTIS (ranibizumab), EYLEA (aflibercept), VABYSMO (faricimab-svoa), and BYOOVIZ (ranibizumab-nuna). Anti-VEGF drugs may also include biosimilas of the example anti-VEGF drugs. Again, if there is no information on the patient's past treatments, OCT device 110 and / or model 172 may not make a recommendation.
[0044] Treatment visits 220 may be performed at regular intervals (e.g., one month) after the initial visit 210, according to the treatment plan. Generally, the treatment plan includes three or more treatment visits 220. During treatment visits 220, the OCT system 100 can measure the patient's eye, for example, by determining the thickness of the retina or the amount of fluid using the OCT device 110. Generally, the ophthalmologist can continue treatment with the same medication according to the treatment plan. In some implementations, the OCT system 100 can provide recommendations on whether to change the medication. For example, model 172 can detect a problem with the initially selected medication based on measurements after the initial treatment. Such problems may be relatively rare.
[0045] A follow-up visit 230 may occur after the initial treatment visit 220. This follow-up visit 230 may include measuring the patient's eye using the OCT device 110. In some embodiments, the patient's treatment plan may branch significantly after the initial treatment visit 220, based on the results of the initial visit. Generally, the ophthalmologist may decide to observe the patient without further treatment, extend the treatment, continue with the same medication, or change the medication. In some embodiments, the OCT system 100 can make recommendations regarding the patient's treatment plan. For example, the OCT system 100 can apply training data about the patient to model 172 to determine a treatment plan for the patient's AMD after a series of anti-VEGF injections, based on treatment information regarding the treatment at the initial visit, the patient's eye measurements, and patient information. Recommended treatment plans may include, for example, observing the patient without further treatment, extending the treatment, continuing with the same medication, or changing the medication. In some implementations, if a change in medication is recommended, the OCT system 100 may provide a second recommendation regarding the medication. For example, in some implementations, the second recommendation may include the medication, the dosage of the medication, and the frequency of treatment. In some implementations, the OCT system 100 may include a first model that recommends a treatment plan and a second model that recommends a specific treatment plan for the new medication.
[0046] Aspects of this disclosure may be implemented using hardware, software, or a combination thereof, and may be implemented in one or more computer systems or other processing systems. In certain aspects of this disclosure, the features are intended for one or more computer systems capable of performing the functions described herein. An example of such a computer system 300 is shown in Figure 3.
[0047] The computer system 300 includes one or more processors (e.g., processor 304). The processor 304 is connected to a communication infrastructure 306 (e.g., a communication bus, crossover bar, network). Various software configurations relating to this exemplary computer system are described. By reading this description, those skilled in the art will be able to see how to implement the configurations of this disclosure using other computer systems and / or architectures.
[0048] The computer system 300 may include a display interface 302, which transfers graphics, text, and other data from the communication infrastructure 306 (or from the graphics processing unit (GPU) 332) for display on the display unit 330. For example, the display interface 302 can transfer the graphical rendering of the super surface from the processor 304 to the display unit 330. The computer system 300 also includes main memory 308 (preferably random access memory (RAM)) and may also include secondary memory 310. The secondary memory 310 may include a hard disk drive 312 and / or a removable storage drive 314, such as a floppy disk drive, magnetic tape drive, optical disk drive, or Universal Serial Bus (USB) flash drive. The removable storage drive 314 reads from and writes to the removable storage device 318 in a well-known manner. The removable storage device 318 represents a floppy disk, magnetic tape, optical disk, USB flash drive, etc., and is read from and written to by the removable storage drive 314. Naturally, the removable storage device 318 includes a computer-usable storage medium on which computer software and / or data is stored.
[0049] Alternative embodiments of the present disclosure may include a secondary memory 310, and may also include other similar devices that enable loading computer programs and other instructions into the computer system 300. Examples of such devices include removable storage devices 322 and interfaces 320. Such examples include program cartridges and cartridge interfaces (such as those found in video game machines), removable memory chips (such as EPROMs (erasable programmable read-only memory) and programmable read-only memory (PROMs)) and associated sockets, and other removable storage devices 322 and interfaces 320 that enable the transfer of software and data from the removable storage device 322 to the computer system 300.
[0050] The computer system 300 may include a communication interface 324. The communication interface 324 enables the transfer of software and data between the computer system 300 and external devices. Examples of the communication interface 324 include a modem, a network interface (e.g., an Ethernet card), a communication port, a PCMCIA (Personal Computer Memory Card International Association) slot and card, etc. The software and data transferred via the communication interface 324 are in the form of signals 328, which may be electronic signals, electromagnetic signals, optical signals, or other signals receivable by the communication interface 324. These signals 328 are provided to the communication interface 324 via a communication path (e.g., a channel) 326. This path 326 transmits the signals 328 and may be implemented using wires or cables, optical fibers, telephone lines, cellular links, radio frequency (RF) links, and / or other communication channels.
[0051] Therefore, in one or more implementations, the described functions may be implemented in hardware, software, or any combination thereof. When implemented in software, the functions may be stored or encoded as one or more instructions or codes in a computer-readable medium. Computer-readable mediums include computer storage media and are sometimes called non-temporary computer-readable media. Non-temporary computer-readable media may exclude temporary signals. The storage medium may be any medium that a computer can access. Examples, but not limited to, such computer-readable media include random access memory (RAM), read-only memory (ROM), EEPROM (electrically erasable programmable ROM), optical disk storage, magnetic disk storage, other magnetic storage devices, combinations of the aforementioned types of computer-readable media, and other media that can be used to store computer executable code in the form of computer-accessible instructions or data structures.
[0052] In some embodiments, the computer system 300 may include an OCT device 110 and / or another ocular measuring device 350. The OCT device 110 can determine one or more ocular measuring parameters, such as retinal thickness or fluid volume. In some implementations, the OCT device 110 may include, or be associated with, another ocular measuring device 350 (which may include any device for measuring the eye). For example, the ocular measuring device may include a biometer configured to measure the axial length and corneal force of the eye. In some embodiments, the ocular measuring device may further measure the limbal diameter (white-to-white distance), anterior chamber depth, preoperative refractive value, and / or lens thickness. The ocular measuring device 350 may further receive input for ocular measuring parameters (e.g., gender or sex). Axial length may be the distance from the surface of the cornea to the retina. Corneal force may be the refractive power of the cornea. As another example, the ocular measuring device 350 may measure the anterior chamber depth of the eye. In one embodiment, the ophthalmic measuring instrument 350 may be an ultrasonic instrument. In another embodiment, the ophthalmic measuring instrument 350 may be an optical biometer. Various optical biometers are sold under the names LENSTAR® and IOL MASTER. In another embodiment, the ophthalmic measuring instrument 350 may include an intraoperative aberration measuring instrument. An intraoperative aberration measuring instrument can measure the refractive characteristics of the eye during surgery. For example, an intraoperative aberration measuring instrument can provide information about the spherical, cylindrical, and axial surfaces of the eye. Furthermore, the ophthalmic measuring instrument may include measuring instruments such as a wavefront analyzer or an automatic refractometer. The ophthalmic measuring instrument 350 may be communicatively coupled to the processor 304 via a communication base 306, a communication interface 324, and / or a communication channel 326.
[0053] The computer program (also called computer control logic) is stored in the main memory 308 and / or secondary memory 310. The computer program may be received via the communication interface 324. When executed, such a computer program enables the computer system 300 to perform features relating to the embodiments of this disclosure as described herein. In particular, when executed, the computer program enables the processor 304 to perform features relating to the embodiments of this disclosure. Thus, such a computer program represents the controller of the computer system 300.
[0054] In embodiments of this disclosure in which the disclosure is implemented using software, the software may be stored in a computer program product and may be loaded into a computer system 300 using a removable storage drive 314, a hard drive 312, or a communication interface 320. When the control logic (software) is executed by the processor 304, it causes the processor 304 to perform the functions described herein. In another embodiment of this disclosure, the system is implemented primarily in hardware using hardware components such as application-specific integrated circuits (ASICs). Implementations of a hardware state machine for performing the functions described herein will be obvious to those skilled in the art.
[0055] In yet another aspect of this disclosure, the disclosure may be implemented using a combination of both hardware and software.
[0056] Figure 4 shows a communication system 400 that can be used according to an aspect of the present disclosure. The communication system 400 includes one or more accessors 460 (also interchangeably referred to herein as one or more “users”) and one or more terminals 442 and / or other input devices (e.g., OCT devices 110). The OCT devices 110 may be further configured to communicate with a network 444. In one aspect of the present disclosure, the data used is, for example, received from an input device and then input and / or accessed by accessors 460 via terminals 442 (e.g., PCs, minicomputers, mainframe computers, microcomputers, telephones or wireless devices, personal digital assistants (“PDA”) or handheld wireless devices (e.g., wireless phones)) which are coupled to a server 443 (e.g., personal computers (PCs), minicomputers, mainframe computers, microcomputers, other devices having a processor and a data repository and / or a connection to a data repository) via couplings 445, 446, 464 to the network 444 (such as the Internet or an intranet and / or a wireless network). The couplings 445, 446, and 464 include, for example, wired links, wireless links, or fiber optic links. In another aspect of the Disclosure, the methods and systems of the Disclosure may include one or more features that operate in a standalone environment, such as a single terminal.
[0057] In one embodiment, server 443 may be an example of computer system 300 (Figure 3). In one embodiment, for example, server 443 may be configured to perform the methods described herein. For example, server 443 may obtain measurements such as retinal thickness and fluid volume from OCT device 110. The measurements may be entered by accessor 460 or provided by OCT device 110 (Figure 3). Server 443 may also obtain other patient information, including treatment information. Server 443 may generate a set of training data containing patient information for multiple patients (possibly from different OCT devices 110). Server 443 may train model 172 (Figure 1) based on the training data. In some implementation examples, server 443 may export the trained model to OCT device 110 or another device (e.g., terminal 442). Server 443, OCT device 110 and / or terminal 442 may be configured to apply patient data to machine learning model 172 to determine a treatment plan for a patient's AMD after a series of anti-VEGF injections.
[0058] Figure 5 is a flowchart illustrating an exemplary method 500 for recommending a treatment plan for patients with retinal or macular diseases. For example, the retinal or macular disease may be one or more of the group consisting of age-related macular degeneration (AMD), diabetic macular edema (DME), edema due to retinal vein occlusion (RVO), and myopic choroidal neovascularization (CNV). Method 500 may be performed by system 100.
[0059] In block 510, method 500 includes the step of performing optical coherence tomography (OCT) on a patient to quantify the initial level of macular edema or retinal exudation. In one embodiment, for example, an OCT device 110 can perform OCT on a patient using a sensor 112. System 100 can quantify the initial level of macular edema or retinal exudation using an image processing device 120. For example, the image processing device 120 can analyze images from the OCT device 110 to determine the thickness or amount of fluid in the retina.
[0060] In block 520, method 500 includes the step of administering a series of anti-VEGF injections to a patient over a period of time. For example, an ophthalmologist may administer a series of anti-VEGF injections to a patient during a treatment visit 220. The ophthalmologist can record treatment information using system 100. For example, treatment information for a patient may include the date of the treatment, the anti-VEGF drug, and the dosage of the treatment.
[0061] In block 530, method 500 includes the step of performing additional periodic OCT on the patient after each anti-VEGF injection to determine the treatment response, which is assessed by the subsequent level of edema or retinal exudation. For example, an OCT device 110 may perform additional periodic OCT on the patient at the next treatment visit 220 (for example, before the next anti-VEGF injection). The OCT device 110 can record the subsequent level of edema or retinal exudation.
[0062] In block 540, method 500 includes the step of collecting a set of patient-based training data, the training data including an initial level of macular edema, a subsequent level of edema, patient information, and treatment information. In one embodiment, for example, an OCT device 110 and / or an ophthalmologist may record training data relating to the patient. For example, the OCT device 110 may record an initial level of macular edema or retinal exudation measured at the first visit 210 and a subsequent level of macular edema or retinal exudation measured at each treatment visit 220. An ophthalmologist may record treatment information using system 100. For example, treatment information relating to a patient may include at least the date of the treatment, the anti-VEGF drug, and the dosage of the treatment. Patient information may be entered by a user of the OCT system 100 (e.g., ophthalmologist, assistant, patient), or patient information may be obtained from an electronic medical record. For example, patient information may include at least age, visual acuity, and an indication of the presence or absence of one or more other diseases.
[0063] In some implementations, block 550 optionally includes a step of training a machine learning model based on multiple sets of training data to select a treatment plan that optimizes a patient's visual acuity. For example, the learning machine 170 can train model 172 by executing a machine learning algorithm such as a gradient boosting library. For example, model 172 may include a random forest model.
[0064] In block 560, method 500 includes the step of applying training data about a patient to a machine learning model trained on multiple sets of training data about multiple patients to determine a treatment plan for a patient's retinal or macular disease after a series of anti-VEGF injections. In one embodiment, for example, system 100 may apply training data about a patient to a machine learning model 172. The machine learning model 172 may provide a recommended treatment plan for the patient. For example, a recommended treatment plan may include one of the following: continuation of observation, extension of an existing treatment plan, continuation of the same medication, or change of medication. If medication is changed, the recommended treatment plan may include the new medication, the dosage, and the frequency of treatment.
[0065] In block 570, method 500 may optionally include the step of administering an anti-VEGF injection as indicated in the treatment plan. In one embodiment, for example, an ophthalmologist may administer the anti-VEGF injection as indicated in the treatment plan to the patient.
[0066] While the aspects of this disclosure have been described in relation to examples, it is naturally possible for those skilled in the art to modify and change the aspects of this disclosure described herein without departing from the scope of this specification. Other aspects will become apparent to those skilled in the art from the examination of this specification or from the practice of following the aspects of this disclosure disclosed herein.
Claims
1. A method for treating retinal diseases or macular diseases, The procedure involves performing optical coherence tomography (OCT) on the patient to quantify the initial level of macular edema or retinal exudation, The steps include administering a series of anti-vascular endothelial growth factor (anti-VEGF) injections to the patient over a certain period of time, The steps include: performing additional periodic OCT on the patient after each anti-VEGF injection to determine the treatment response, which is evaluated by the subsequent level of macular edema or retinal exudation; A step of collecting a set of training data based on the patient, wherein the training data includes the initial level of macular edema or retinal exudation, the subsequent level of macular edema or retinal exudation, patient information, and treatment information. The steps include: applying the training data for the patient to a machine learning model trained on multiple sets of training data for multiple patients to determine a treatment plan for the patient's retinal or macular disease after a series of anti-VEGF injections; A method that includes this.
2. A step of training the machine learning model based on multiple sets of training data to select a treatment plan that optimizes the patient's visual acuity, The method according to claim 1, further comprising:
3. The patient information includes at least age, visual acuity, and indication of the presence or absence of one or more other diseases. The method according to claim 1.
4. The aforementioned machine learning model is trained using a gradient boosting library to construct a random forest model. The method according to claim 1.
5. The initial level of macular edema or retinal exudation is the thickness of the retina measured by the OCT, or the amount of fluid measured by the OCT. The method according to claim 1.
6. The aforementioned treatment information includes, at a minimum, the date of each anti-VEGF injection, the drug of each anti-VEGF injection, and the dosage of each anti-VEGF injection. The method according to claim 1.
7. The step of administering the anti-VEGF injection indicated in the treatment plan to the patient, The method according to claim 1, further comprising:
8. The aforementioned retinal or macular disease is age-related macular degeneration (AMD). The method according to claim 1.
9. The aforementioned retinal or macular disease is one or more of the following: age-related macular degeneration (AMD), diabetic macular edema (DME), edema due to retinal vein occlusion (RVO), and myopic choroidal neovascularization (CNV). The method according to claim 1.
10. A sensor configured to detect the level of macular edema or retinal exudation in the eye, Memory containing executable computer instructions, A processor coupled to the memory and configured to execute the computer executable instructions, Optical coherence tomography (OCT) equipment equipped with, The processor executes the computer executable instructions, To quantify the initial level of macular edema or retinal exudation in patients with retinal or macular disease by performing optical coherence tomography (OCT) on the patient, To receive treatment information regarding a series of anti-vascular endothelial growth factor (anti-VEGF) injections administered to the patient over a certain period of time, After each anti-VEGF injection, additional periodic OCT is performed on the patient to determine the treatment response, which is evaluated by the subsequent level of macular edema or retinal exudation. The collection of training data based on the aforementioned patient, wherein the training data includes the initial level of macular edema or retinal exudation, the subsequent level of macular edema or retinal exudation, patient information, and treatment information. Applying the training data for the patient to a machine learning model trained on multiple sets of training data for multiple patients to determine a treatment plan for the patient's retinal or macular disease after a series of anti-VEGF injections, The OCT device is made to perform the following: OCT equipment.
11. The aforementioned processor executes the instruction, Training the machine learning model based on multiple sets of training data to select a treatment plan that optimizes the patient's visual acuity, The OCT device is configured to perform the above-mentioned action. The OCT device according to claim 10.
12. The patient information includes at least age, visual acuity, and indication of the presence or absence of one or more other diseases. The OCT device according to claim 10.
13. The aforementioned machine learning model is trained using a gradient boosting library to model a random forest. The OCT device according to claim 10.
14. The initial level of macular edema or retinal exudation is the thickness of the retina measured by the OCT, or the amount of fluid measured by the OCT. The OCT device according to claim 10.
15. The aforementioned treatment information includes, at a minimum, the date of each anti-VEGF injection, the drug of each anti-VEGF injection, and the dosage of each anti-VEGF injection. The OCT device according to claim 10.
16. The aforementioned retinal or macular disease is age-related macular degeneration (AMD). The OCT device according to claim 10.
17. The aforementioned retinal or macular disease is one or more of the following: age-related macular degeneration (AMD), diabetic macular edema (DME), edema due to retinal vein occlusion (RVO), and myopic choroidal neovascularization (CNV). The OCT device according to claim 10.
18. Memory containing executable computer instructions, A processor coupled to the memory and configured to execute the computer executable instructions, A computer network equipped with, The processor executes the computer executable instructions, Receiving training data from multiple optical coherence tomography (OCT) devices, wherein the training data includes, for each of several patients with retinal disease or macular disease, the initial level of macular edema or retinal exudation, the subsequent level of macular edema or retinal exudation after each of a series of anti-vascular endothelial growth factor (anti-VEGF) injections administered to the patient over a certain period, patient information, and treatment information. Training a machine learning model based on training data to select a treatment plan that optimizes a patient's predicted visual acuity based on a patient dataset, wherein the patient dataset includes the initial level of macular edema of the patient, the subsequent levels of edema after each of an initial series of anti-VEGF injections administered to the patient over a period of time, patient medical information, and patient treatment information. Applying the patient dataset for the patient to the machine learning model to determine a treatment plan for the patient's retinal or macular disease after the initial series of anti-VEGF injections, To cause the processor to execute the above, Computer network.
19. The aforementioned patient medical information includes, at a minimum, age, visual acuity, and indication of the presence or absence of one or more other diseases. The computer network according to claim 18.
20. The aforementioned machine learning model is trained using a gradient boosting library to construct a random forest model. The computer network according to claim 18.
21. The initial level of macular edema is the thickness measured by the OCT, or the fluid volume measured by the OCT. The computer network according to claim 18.
22. The aforementioned treatment information includes, at a minimum, the date of each anti-VEGF injection, the drug of each anti-VEGF injection, and the dosage of each anti-VEGF injection. The computer network according to claim 18.
23. Further equipped with patient computer equipment, The processor is further configured to output the treatment plan to the patient computer device. The computer network according to claim 18.
24. The aforementioned retinal or macular disease is age-related macular degeneration (AMD). The computer network according to claim 18.
25. The aforementioned retinal or macular disease is one or more of the following: age-related macular degeneration (AMD), diabetic macular edema (DME), edema due to retinal vein occlusion (RVO), and myopic choroidal neovascularization (CNV). The computer network according to claim 18.