Predicting treatment outcomes for neovascular age-related macular degeneration using baseline characteristics

Abstract

A method and system for predicting a treatment outcome is provided. Three-dimensional imaging data of a subject's retina is received. A first output is generated using a deep learning system and the three-dimensional imaging data. The first output and baseline data are received as inputs for a symbolic model. The symbolic model uses the inputs to predict a treatment outcome for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD).

Description

[Technical field]

[0001] Inventor: Neha Anegondi, Jian Dai, Michael Kawczynski, Yusuke Kikuchi CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority to U.S. Provisional Patent Application No. 63 / 172,063, entitled “Treatment Outcome Prediction for Neovascular Age-Related Macular Degeneration using Baseline Characteristics,” filed April 7, 2021, the entirety of which is incorporated herein by reference.

[0002] The present specification is generally directed to predicting the treatment prognosis of a subject diagnosed with age-related macular degeneration.More specifically, the present specification provides a method and system for predicting the treatment prognosis of a subject diagnosed with neovascular age-related macular degeneration (nAMD) using baseline data determined for the subject. [Background technology]

[0003] Age-related macular degeneration (AMD) is a disease that affects the central area of ​​the retina of the eye, called the macula. AMD is the leading cause of vision loss in subjects over the age of 50. Neovascular AMD (nAMD) is one of two advanced stages of AMD. In nAMD, new and abnormal blood vessels grow uncontrollably under the macula. This type of growth can cause swelling, bleeding, fibrosis, other problems, or a combination thereof. Treatment of nAMD typically involves anti-vascular endothelial growth factor (anti-VEGF) therapy (e.g., anti-VEGF drugs such as ranibizumab). The response of the retina to such treatments is at least in part subject-specific, such that different subjects may respond differently to the same type of anti-VEGF drug. Furthermore, anti-VEGF therapy is typically administered by intravitreal injection, which is expensive and can itself cause complications (e.g., blindness). Summary of the Invention

[0004] In one or more embodiments, a method for predicting a treatment outcome is provided. Three-dimensional imaging data of a subject's retina is received. A first output is generated using a deep learning system and the three-dimensional imaging data. The first output and baseline data are received as inputs for a symbolic model. The symbolic model uses the inputs to predict a treatment outcome for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD).

[0005] In one or more embodiments, a method is provided for predicting a treatment prognosis for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD). A first predicted prognosis is generated using a deep learning system and three-dimensional imaging data of the subject's retina. A second predicted prognosis is generated using a symbolic model and baseline data of the subject. The first predicted prognosis and the second predicted prognosis are used to predict a treatment prognosis for the subject undergoing treatment for nAMD.

[0006] In one or more embodiments, a system for administering anti-vascular endothelial growth factor (anti-VEGF) treatment for a subject diagnosed with neovascular age-related macular degeneration (nAMD) comprises a memory containing a machine-readable medium including machine-executable code and a processor coupled to the memory, the processor configured to execute the machine-executable code that causes the processor to receive three-dimensional imaging data of the subject's retina, generate a first output using a deep learning system and the three-dimensional imaging data, receive the first output and baseline data as inputs for a symbolic model, and predict, via the symbolic model, a treatment prognosis for the subject undergoing treatment for neovascular age-related macular degeneration (nAMD) using the inputs.

[0007] In some embodiments, a system is provided that includes one or more data processors and a non-transitory computer-readable storage medium that includes instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of one or more methods disclosed herein.

[0008] In some embodiments, a computer program product is provided that is tangibly embodied in a non-transitory machine-readable storage medium and includes instructions configured to cause one or more data processors to perform some or all of one or more of the methods disclosed herein.

[0009] Some embodiments of the present disclosure include a system including one or more data processors. In some embodiments, the system includes a non-transitory computer-readable storage medium including instructions that, when executed on the one or more data processors, cause the one or more data processors to perform some or all of the methods and / or some or all of the processes disclosed herein. Some embodiments of the present disclosure include a computer program product tangibly embodied in a non-transitory machine-readable storage medium including instructions configured to cause the one or more data processors to perform some or all of the methods and / or some or all of the processes disclosed herein.

[0010] The terms and expressions which have been employed are used as terms of description rather than of limitation, and in the use of such terms and expressions there is no intention of excluding any equivalents or portions of the features shown and described, but it is recognized that various modifications are possible within the scope of the invention claimed.

[0011] Thus, while the claimed invention has been specifically disclosed by embodiments and optional features, it is to be understood that modifications and variations of the concepts disclosed herein may be employed by those skilled in the art, and such modifications and variations are deemed to be within the scope of the present invention as defined by the appended claims.

[0012] For a more complete understanding of the principles disclosed herein and their advantages, reference should now be made to the following descriptions taken in conjunction with the accompanying drawings, in which: [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 is a block diagram of a prediction system in accordance with various embodiments. [Diagram 2] 1 is a flow chart of a process for predicting treatment outcome according to various embodiments. [Diagram 3] 1 is a flow chart of a process for predicting treatment outcome according to various embodiments. [Figure 4] 1 is a flow chart of a process for predicting treatment outcome according to various embodiments. [Diagram 5] 1 is a table showing performance data for model stacking and model averaging approaches in predicting treatment outcomes according to one or more embodiments. [Figure 6] 1 is a table showing performance data for model stacking and model averaging approaches in predicting treatment outcomes according to one or more embodiments. [Figure 7] FIG. 1 is a block diagram of a computer system according to one or more embodiments. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] It should be understood that the figures are not necessarily drawn to scale, and that objects within the figures are not necessarily drawn to scale relative to each other. The figures are representations intended to bring clarity and understanding to the various embodiments of the devices, systems, and methods disclosed herein. Wherever possible, the same reference numbers are used throughout the drawings to refer to the same or like parts. Further, it should be understood that the drawings are not intended to limit the scope of the present teachings in any way.

[0015] I. Overview Determining a subject's response to age-related macular degeneration (AMD) treatment, and often to neovascular AMD (nAMD) treatment, among others, can include determining the subject's visual acuity response, the subject's reduction in foveal thickness, or both. A subject's visual acuity is the sharpness of the subject's vision, which can be measured by the subject's ability to distinguish letters or numbers at a given distance. Visual acuity is often confirmed by visual acuity testing and measured according to a standard Snellen chart. Retinal images can provide information that can be used to estimate a subject's visual acuity. For example, optical coherence tomography (OCT) images can be used to estimate a subject's visual acuity at the time the OCT image is captured. Foveal thickness, also referred to as central retinal thickness (CST), can be defined as the average macular thickness in a central 1 mm diameter area. CST can also be measured using OCT images.

[0016] However, in certain cases, such as clinical trials, it may be desirable to be able to predict a subject's future response to AMD treatment (e.g., nAMD treatment). For example, it may be desirable to predict whether a subject's vision will improve at a selected time period after treatment (e.g., 6 months after treatment, 9 months after treatment, 12 months after treatment, 24 months after treatment, etc.). Furthermore, it may be desirable to classify any such improvement in vision. In some cases, it may be desirable to predict whether a subject will experience a decrease in CST (e.g., any decrease in CST, or a decrease above a selected threshold). Such prediction and classification may allow for a treatment regimen to be individualized for a given subject. For example, predictions regarding a subject's vision response to a particular AMD treatment may be used to customize the injection dosage, the interval at which injections are administered, or both. Furthermore, such predictions may improve clinical trial screening, pre-screening, or both, by allowing for the exclusion of subjects predicted to be poorly responsive to treatment.

[0017] Thus, various embodiments described herein provide methods and systems for predicting a treatment outcome for a subject in response to an AMD treatment (e.g., nAMD treatment). In particular, baseline data is input into a symbolic model and used to predict the outcome of a subject who has undergone such treatment. The outcome may include, for example, but not limited to, a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central retinal thickness, a predicted decrease in central retinal thickness, or a combination thereof. In some embodiments, the input sent to the symbolic model includes both the baseline data and an output (e.g., a previously generated predicted outcome) generated based on three-dimensional imaging data (e.g., OCT imaging data). For example, the OCT imaging data may be processed by a deep learning system to generate a predicted outcome that is combined with the baseline data. In this manner, the baseline data and this predicted outcome are fused to form an input sent to the symbolic model.

[0018] In other embodiments, the symbolic model may be used to generate a first output using baseline data, and the deep learning system may be used to generate a second output using three-dimensional imaging data. These two outputs are combined, fused, or otherwise integrated to form a prognostic output that includes or indicates a predicted treatment outcome. For example, the first output and the second output may be a first predicted outcome and a second predicted outcome, respectively. A weighted average (e.g., an equally weighted average) of these two predicted outcomes may be used as the final treatment outcome for the subject.

[0019] Recognizing and taking into account the importance and usefulness of the method and system that can provide the above-mentioned improvements, the embodiments described herein provide a method and system for predicting visual acuity response to AMD treatment (e.g., nAMD treatment). More specifically, the embodiments described herein provide a method and system for processing baseline data using a symbolic model to predict the treatment prognosis of the subject undergoing nAMD treatment at a selected time point (e.g., 6 months, 9 months, 12 months, 24 months, etc.) after the baseline time point. The baseline time point can be, for example, but not limited to, the first day of treatment. The use of the method and system described herein can have the technical effect of reducing the overall computational resources and / or time required to predict the treatment prognosis of the subject undergoing nAMD treatment. Furthermore, the use of the method and system can make it possible to predict the treatment prognosis of the subject more efficiently and accurately compared to other methods and systems.

[0020] Additionally, the embodiments described herein may facilitate the creation of personalized treatment regimens for individual subjects to ensure appropriate dosages and / or intervals between treatment administrations (e.g., injections). Among other things, the embodiments described herein may aid in the generation of accurate, efficient, and convenient personalized treatment or administration schedules, and may improve clinical cohort selection or clinical trial design.

[0021] II. Prediction of prognosis for neovascular age-related macular degeneration (nAMD) treatment II.A. Exemplary Prediction Systems for Predicting AMD Treatment Prognosis FIG. 1 is a block diagram of a prediction system 100 according to various embodiments. The prediction system 100 is used to predict the treatment outcome of one or more subjects for AMD treatment. The AMD treatment, which may be an nAMD treatment, may include, for example, but is not limited to, an anti-VEGF treatment, an antibody treatment, another type of treatment, or a combination thereof. The anti-VEGF treatment may include, for example, ranibizumab, which may be administered by intravitreal injection. The antibody treatment may be, for example, a monoclonal antibody treatment directed against vascular endothelial growth factor (VEGF) and an angiopoietin 2 inhibitor. In one or more embodiments, the antibody treatment includes faricimab.

[0022] The prediction system 100 includes a computing platform 102, a data storage 104, and a display system 106. The computing platform 102 may take a variety of forms. In one or more embodiments, the computing platform 102 includes a single computer (or computer system) or multiple computers in communication with each other. In other examples, the computing platform 102 takes the form of a cloud computing platform. In some examples, the computing platform 102 takes the form of a mobile computing platform (e.g., a smartphone, a tablet, a smartwatch, etc.).

[0023] The data storage 104 and the display system 106 are each in communication with the computing platform 102. In some examples, the data storage 104, the display system 106, or both may be considered part of or integrated with the computing platform 102. Thus, in some examples, the computing platform 102, the data storage 104, and the display system 106 may be separate components that communicate with each other, while in other examples some combination of these components may be integrated.

[0024] The predictive system 100 includes a data analyzer 108, which may be implemented using hardware, software, firmware, or a combination thereof. In one or more embodiments, the data analyzer 108 is implemented on the computing platform 102. The data analyzer 108 processes a set of inputs 110 using a model system 112 to predict (or generate) a prognostic output 114.

[0025] The model system 112 may include any number or combination of artificial intelligence or machine learning models. In one or more embodiments, the model system 112 includes a first prognosis prediction model 116 and a second prognosis prediction model 118. In one or more embodiments, the first prognosis prediction model 116 includes a deep learning system that may include, for example, one or more neural networks, at least one of which is a deep learning neural network (or deep neural network) (DNN). In one or more embodiments, the second prognosis prediction model 118 includes a symbolic model, which includes one or more models that use symbolic learning or symbolic inference. For example, the second prognosis prediction model 118 may include, but is not limited to, at least one of a linear model, a random forest model, an extreme gradient boosting (XGBoost) algorithm, or another type of model or algorithm.

[0026] In one or more embodiments, the set of inputs 110 sent to the model system 112 may be received at least in part from a source external to the prediction system 100 over one or more communications links (e.g., a wired communications link, a wireless communications link, an optical communications link, etc.). In one or more embodiments, the set of inputs 110 is retrieved at least in part from the data storage 104.

[0027] The set of inputs 110 for the model system 112 may include baseline data 120. In one or more embodiments, the set of inputs 110 may further include three-dimensional imaging data 122. The baseline data 120 includes data obtained at a baseline time point. The baseline time point may be, for example, a time point prior to treatment or a time point contemporaneous with the first treatment administration (e.g., day 1 of treatment).

[0028] The baseline data may include, for example, but is not limited to, at least one of demographic data, a baseline visual acuity measurement, a baseline CST measurement, a baseline low luminance defect (LLD), a treatment group, or any other type of baseline measurement. The demographic data may include, for example, but is not limited to, age, sex, or another type of demographic metric. The baseline visual acuity measurement may be, for example, a best corrected visual acuity (BCVA) measurement. The baseline CST measurement may be, for example, in micrometers. The LLD may be the difference between the baseline BCVA measurement and the baseline low luminance visual acuity (LLVA) measurement.

[0029] The three-dimensional imaging data 122 may include OCT imaging data, data extracted from an OCT image (e.g., an OCT en-face image), tabular data extracted from an OCT image, some other form of imaging data, or a combination thereof. The OCT imaging data may include, for example, a spectral domain OCT (SD-OCT) B-scan. The three-dimensional imaging data 122 may be imaging data at a baseline time point for the subject prior to treatment or concurrent with the first treatment administration.

[0030] The model system 112 processes the set of inputs 110 to predict at least one treatment outcome 124 for a subject who has undergone or will undergo an nAMD treatment. The treatment outcome 124 may include, for example, but is not limited to, a predicted visual acuity measurement (e.g., predicted BCVA), a predicted change in visual acuity (e.g., predicted change in BCVA), a predicted CST, a predicted reduction in CST, or any other type of treatment outcome for a subject undergoing treatment. The treatment outcome 124 may be generated at a selected time point after the baseline time point. For example, the treatment outcome 124 may be predicted at n months after the baseline time point, where n months is selected as a month between 3 months and 30 months after the baseline time point. In one or more embodiments, the treatment outcome 124 may be predicted for a time such as, but not limited to, 6 months, 9 months, 12 months, 24 months, or some other time after treatment. Examples of how the model system 112 can be used to predict the treatment outcome 124 are described in more detail below in Figures 2-4.

[0031] The data analyzer 108 may use the treatment prognosis 124 to form the prognosis output 114. The prognosis output 114 may include, for example, the treatment prognosis 124. In one or more embodiments, the prognosis output 114 includes multiple treatment prognoses at multiple time points after treatment (e.g., a 6-month treatment prognosis, a 9-month treatment prognosis, and a 12-month treatment prognosis).

[0032] In one or more embodiments, the prognostic output 114 includes other information generated based on the treatment outcome 124. For example, the prognostic output 114 may include an individualized treatment regimen for a given subject based on the predicted treatment outcome 124. In some examples, the prognostic output 114 may include a customized injection dosage, one or more intervals at which injections are administered, or both. The prognostic output 114 may include instructions to modify or supplement the type of treatment administered to the subject, in some cases, based on the predicted treatment outcome 124 indicating that the subject will not have a desired response to the treatment. In this manner, the prognostic output 114 may be used to improve overall treatment management.

[0033] In one or more embodiments, at least a portion of the prognostic output 114, or a graphical representation of at least a portion of the prognostic output 114, is displayed on the display system 106. In some embodiments, at least a portion of the prognostic output 114, or a graphical representation of at least a portion of the prognostic output 114, is sent to a remote device 126 (e.g., a mobile device, a laptop, a server, the cloud, etc.).

[0034] II.B. Exemplary Methods for Predicting AMD Treatment Prognosis 2 is a flow chart of a process 200 for predicting a treatment outcome according to various embodiments. In one or more embodiments, the process 200 is performed using the prediction system 100 illustrated in FIG.

[0035] Step 202 includes receiving three-dimensional imaging data of a subject's retina. The three-dimensional imaging data 122 in FIG. 1 may be one example of an implementation of the three-dimensional imaging data in step 202. The three-dimensional imaging data may include OCT imaging data, data extracted from an OCT image (e.g., an OCT en-face image), tabular data extracted from an OCT image, some other form of imaging data, or a combination thereof. The OCT imaging data may include, for example, a spectral domain OCT (SD-OCT) B-scan. The three-dimensional imaging data may be imaging data at a baseline time point for the subject prior to treatment or concurrent with a first treatment administration.

[0036] Step 204 includes generating a first output using a deep learning system and the three-dimensional imaging data. The first prognosis prediction model 116 illustrated in FIG. 1 may be one example of an implementation of the deep learning system used in step 204. The deep learning system may be comprised of one or more neural networks. In one or more embodiments, the first output generated in step 204 is a predicted prognosis (e.g., a predicted treatment prognosis). For example, the deep learning system may be trained to predict a treatment prognosis based on one or more OCT images generated at a baseline time point for the subject.

[0037] Step 206 includes receiving the first output and baseline data as input for a symbolic model. The second prognostic model 118 illustrated in FIG. 1 may be an example of an implementation of the symbolic model used in step 206. The symbolic model may be implemented, for example, using at least one of a linear model, a random forest model, an XGBoost algorithm, or another type of symbolic learning model. The baseline data 120 in FIG. 1 may be an example of an implementation of the baseline data in step 206. The baseline data may include, for example, at least one of attribute data (e.g., age, sex, etc.), baseline visual acuity measurements (e.g., baseline BCVA), baseline central retinal thickness (CST) measurements, baseline low light defect (LLD), or treatment group.

[0038] Step 208 includes predicting (or generating) a treatment outcome for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD) using the inputs by the symbolic model. The treatment outcome may include, for example, but is not limited to, at least one of a predicted visual acuity measurement (e.g., predicted BCVA), a predicted change in visual acuity, a predicted CST, a predicted reduction in CST, or another indicator of the subject's response to treatment. The treatment outcome predicted in step 208 may be at a selected time point after treatment, such as, but not limited to, 6 months, 9 months, 12 months, 24 months, or some other time after treatment.

[0039] In various embodiments, the treatment prognosis predicted (or generated) in step 208 includes a visual acuity response (VAR) output, which is a value or score that identifies a predicted change in the subject's visual acuity. For example, the VAR output can be a value or score that classifies the subject's visual acuity response in terms of a predicted level of improvement (e.g., letter improvement) or level of decline (e.g., vision loss). As one specific example, the VAR output can be a predicted numerical change in BCVA that is subsequently processed and identified as belonging to one of a number of different classes of BCVA change, each class of BCVA change corresponding to a different range of letter improvement. As another example, the VAR output can be the predicted change class itself. In yet another example, the VAR output can be a predicted change in some other measure of visual acuity. In other embodiments, the VAR output can be a value or representation output that requires one or more additional processing steps to arrive at a predicted change in visual acuity. For example, the VAR output can be the subject's predicted future BCVA at a certain time period after treatment (e.g., at 9 months, at 12 months). An additional processing step or steps may include calculating the difference between the predicted future BCVA and the baseline BCVA to determine a predicted change in vision.

[0040] The process 200 may optionally include step 210. Step 210 includes generating a prognostic output based on the treatment prognosis. The prognostic output 114 in FIG. 1 may be one example of an implementation of the prognostic output in step 210. The prognostic output may include, for example, a treatment prognosis, or multiple treatment prognoses at multiple time points after treatment (e.g., a 6-month treatment prognosis, a 9-month treatment prognosis, and a 12-month treatment prognosis).

[0041] In one or more embodiments, the prognostic output includes other information generated based on the treatment prognosis. For example, the prognostic output may include a personalized treatment regimen for a given subject based on the predicted treatment prognosis. In some examples, the prognostic output may include a customized injection dosage, one or more intervals at which the injection is administered, or both. The prognostic output may include instructions to change or supplement the type of treatment administered to the subject, in some cases, based on the predicted treatment prognosis indicating that the subject will not have the desired response to the treatment. In this way, the prognostic output can be used to improve overall treatment management.

[0042] 3 is a flow chart of a process 300 for predicting a treatment outcome according to various embodiments. In one or more embodiments, the process 300 is performed using the prediction system 100 illustrated in FIG.

[0043] Step 302 includes generating a first output using a deep learning system and three-dimensional imaging data of a subject's retina. The three-dimensional imaging data 122 in FIG. 1 may be one example of an implementation of the three-dimensional imaging data in step 302. The three-dimensional imaging data may include OCT imaging data, data extracted from an OCT image (e.g., an OCT en-face image), tabular data extracted from an OCT image, some other form of imaging data, or a combination thereof. The OCT imaging data may include, for example, a spectral domain OCT (SD-OCT) B-scan. The three-dimensional imaging data may be image data at a baseline time point for the subject prior to treatment or concurrent with administration of a first treatment.

[0044] The first output in step 302 may include a first predicted outcome (first predicted treatment outcome). For example, a deep learning system may be trained to generate the first predicted outcome based on the three-dimensional imaging data.

[0045] Step 304 includes generating a second output using the symbol model and baseline data. The baseline data 120 in FIG. 1 may be one example of an implementation of the baseline data in step 304. The baseline data may include, for example, at least one of demographic data (e.g., age, sex, etc.), baseline visual acuity measurements (e.g., baseline BCVA), baseline central retinal thickness (CST) measurements, baseline low light defect (LLD), or treatment group.

[0046] The second output in step 304 may include a second predicted prognosis (second predicted treatment prognosis). For example, a symbolic model may be trained to generate a second predicted prognosis based on the baseline data. The second prognosis prediction model 118 illustrated in FIG. 1 may be one example of an implementation of the symbolic model used in step 304. The symbolic model may be implemented, for example, using at least one of a linear model, a random forest model, an XGBoost algorithm, or another type of symbolic learning model.

[0047] Step 306 includes predicting a treatment outcome for a subject undergoing treatment for nAMD using the first output and the second output. In one or more embodiments, step 306 includes predicting a treatment outcome as a weighted average (e.g., an equally weighted average) of the first output (e.g., the first predicted outcome) and the second output (e.g., the second predicted outcome). In some embodiments, the first predicted outcome generated by the deep learning system may be weighted over the second predicted outcome generated by the symbolic model. In other embodiments, the second predicted outcome generated by the symbolic model may be weighted over the first predicted outcome generated by the deep learning system.

[0048] The process 300 may optionally include step 308. Step 308 may include generating a prognostic output based on the treatment prognosis. The prognostic output 114 in FIG. 1 may be one example of an implementation of the prognostic output in step 308. The prognostic output may include, for example, a treatment prognosis, or multiple treatment prognoses at multiple time points after treatment (e.g., a 6-month treatment prognosis, a 9-month treatment prognosis, and a 12-month treatment prognosis).

[0049] In one or more embodiments, the prognostic output includes other information generated based on the treatment prognosis. For example, the prognostic output may include a personalized treatment regimen for a given subject based on the predicted treatment prognosis. In some examples, the prognostic output may include a customized injection dosage, one or more intervals at which the injection is administered, or both. The prognostic output may include instructions to change or supplement the type of treatment administered to the subject, in some cases, based on the predicted treatment prognosis indicating that the subject will not have the desired response to the treatment. In this way, the prognostic output can be used to improve overall treatment management.

[0050] 4 is a flow chart of a process 400 for predicting a treatment outcome according to various embodiments. In one or more embodiments, the process 400 is performed using the prediction system 100 illustrated in FIG.

[0051] Step 402 includes receiving baseline data as an input for a symbolic model. The baseline data 120 in FIG. 1 may be an example of an implementation of the baseline data in step 402. Furthermore, the second prognostic model 118 illustrated in FIG. 1 may be an example of an implementation of the symbolic model used in step 206. The symbolic model may be implemented, for example, using at least one of a linear model, a random forest model, an XGBoost algorithm, or another type of symbolic learning model. In one or more embodiments, the baseline data includes a baseline visual acuity measurement (e.g., a baseline BCVA). The baseline visual acuity measurement may be generated using three-dimensional imaging data (e.g., OCT imaging data) and a deep learning system.

[0052] Step 404 includes processing the baseline data using a symbolic model. The symbolic model may use any number of symbolic artificial intelligence learning methods to process the baseline data. In some embodiments, step 404 includes processing the baseline data and a previously generated treatment prognosis received from another system (e.g., a deep learning system).

[0053] Step 406 includes predicting, with the symbolic model, a treatment outcome for the subject undergoing treatment for nAMD based on processing of the baseline data. The prognosis output 124 in FIG. 1 may be one example of an implementation of the treatment outcome.

[0054] III. Exemplary Experimental Data The first study was conducted using data from 185 eyes treated with a nAMD treatment (e.g., faricimab). The data was obtained from the AVENUE clinical trial (NCT02484690) for subjects randomized into four faricimab treatment groups. Data for a particular eye included baseline data and post-treatment data. The baseline data included demographic data (age, sex), baseline BCVA, baseline CST, hypointense defects, and treatment group. The data further included SD-OCT imaging data (e.g., B-scan) of the eye. The post-treatment data included complete BCVA data with CST at 9 months post-treatment. The data was split into 80% training data and 20% test data.

[0055] The treatment outcomes were predicted using a deep learning system (e.g., an example of an implementation of the first prognostic prediction model 116 in FIG. 1) and various symbolic models (e.g., an example of an implementation of the second prognostic prediction model 118 in FIG. 1). The treatment outcomes were defined in two ways: functional and anatomical. The functional part of the treatment outcomes included VAR output (e.g., BCVA letter score at 9 months). The anatomical part of the treatment outcomes included CST reduction rate from baseline to 9 months, which was converted to a true / false binary variable (e.g., true indicates CST reduction rate is greater than 35%). The threshold value for the binary variable (e.g., 35%) was selected based on the mean or median CST reduction rate of the subjects.

[0056] The primary metric for the functional portion of treatment outcome was the coefficient of determination (R 2 The primary metric for the anatomical component of treatment outcomes was the area under the receiver operating characteristic (AUROC) curve. Secondary metrics included precision, accuracy, and recall. Evaluation of model performance involved 5-fold cross-validation.

[0057] In a model stacking approach involving two stages with a given symbolic model, a deep learning system was first used to generate a predicted prognosis in the first stage. This predicted prognosis was also used as one of the input features for the symbolic model in the second stage together with the baseline data. Five-fold cross-validation was used to tune the hyperparameters of the deep learning system and the symbolic model. In the first stage, five-fold CV was used to tune the hyperparameters of the deep learning system. In the i (i=1, 2, 3, 4, 5) iterations of the five-fold cross-validation in the second stage, the predictions of the deep learning system from the i iterations of the five-fold cross-validation in the first stage were used as one of the input features in combination with the baseline data. A total of six models were developed using the model stacking approach.

[0058] In the model averaging approach, for a given symbolic model, the predicted prognosis generated by the deep learning system and the predicted prognosis generated by the symbolic model were averaged together (e.g., with equal weighting) to generate a predicted treatment prognosis. A total of six models were developed using the model averaging approach.

[0059] To calculate the test data performance metrics, the symbolic model was retrained on the entire training data with the optimal hyperparameters found in the 5-fold cross-validation. The deep learning systems were used in an ensemble fashion, i.e., the average of five deep learning systems (i.e., from each 5-fold cross-validation) was used.

[0060] FIG. 5 is a table showing performance data for model stacking and model averaging approaches in predicting treatment outcomes according to one or more embodiments. Treatment outcomes include predicted BCVA at 9 months. The benchmark model identifies each individual model used. For model stacking, the identified model is a symbolic model stacked with a deep learning system. For model averaging, the identified model is a symbolic model whose output is averaged with the output of a deep learning system.

[0061] 6 is a table showing performance data for model stacking and model averaging approaches in predicting treatment outcomes according to one or more embodiments. Treatment outcomes include CST reduction rate classification, where a true or positive classification indicates a CST reduction rate of more than 35%. The benchmark model identifies each individual model used. For model stacking, the identified model is a symbolic model stacked with a deep learning system. For model averaging, the identified model is a symbolic model whose output is averaged with the output of a deep learning system.

[0062] IV. Computer Implemented Systems FIG. 7 is a block diagram illustrating an example of a computer system according to various embodiments. The computer system 700 may be an example of one implementation of the computing platform 102 described above in FIG. 1. In one or more examples, the computer system 700 may include a bus 702 or other communication mechanism for communicating information and a processor 704 coupled with the bus 702 for processing information. In various embodiments, the computer system 700 may also include a memory, which may be a random access memory (RAM) 706 or other dynamic storage device, coupled to the bus 702 for determining instructions to be executed by the processor 704. The memory may also be used to store temporary variables or other intermediate information during execution of instructions to be executed by the processor 704. In various embodiments, the computer system 700 may further include a read only memory (ROM) 708 or other static storage device coupled to the bus 702 for storing static information and instructions for the processor 704. A storage device 710, such as a magnetic disk or optical disk, may be provided and coupled to the bus 702 for storing information and instructions.

[0063] In various embodiments, the computer system 700 can be coupled via the bus 702 to a display 712, such as a cathode ray tube (CRT) or liquid crystal display (LCD), for displaying information to a computer user. An input device 714, including alphanumeric and other keys, can be coupled to the bus 702 for communicating information and command selections to the processor 704. Another type of user input device is a cursor control 716, such as a mouse, joystick, trackball, gesture input device, eye gaze-based input device, or cursor direction keys, for communicating directional information and command selections to the processor 704 and for controlling cursor movement on the display 712. This input device 714 typically has two degrees of freedom in two axes, a first axis (e.g., x) and a second axis (e.g., y), that allow the device to specify a position in a plane. However, it should be understood that input devices 714 that allow three-dimensional (e.g., x, y, and z) cursor movement are also contemplated herein.

[0064] Consistent with a particular implementation of the present teachings, results can be provided by computer system 700 in response to processor 704 executing one or more sequences of one or more instructions contained in RAM 706. Such instructions can be read into RAM 706 from another computer-readable medium or computer-readable storage medium, such as storage device 710. Execution of the sequences of instructions contained in RAM 706 can cause processor 704 to perform the processes described herein. Alternatively, hardwired circuitry can be used in place of or in combination with software instructions to implement the present teachings. Thus, implementation of the present teachings is not limited to any specific combination of hardware circuitry and software.

[0065] The terms "computer-readable medium" (e.g., data store, data storage, data storage device, etc.) or "computer-readable storage medium" as used herein refer to any medium that participates in providing instructions to processor 704 for execution. Such media may take many forms, including but not limited to, non-volatile media, volatile media, and transmission media. Examples of non-volatile media may include, but are not limited to, optical, solid-state, and magnetic disks, such as storage device 710. Examples of volatile media may include, but are not limited to, dynamic memory, such as RAM 706. Examples of transmission media may include, but are not limited to, coaxial cables, copper wire, and fiber optics, including the wires that comprise bus 702.

[0066] Common forms of computer readable media include, for example, a floppy disk, a flexible disk, a hard disk, a magnetic tape or any other magnetic medium, a CD-ROM, any other optical medium, punch cards, paper tape, any other physical medium with a pattern of holes, RAM, PROM, and EPROM, flash EPROM, any other memory chip or cartridge, or any other tangible medium from which a computer can read.

[0067] In addition to computer-readable media, instructions or data can be provided as signals on a transmission medium included in a communication device or system to provide a sequence of one or more instructions to the processor 704 of the computer system 700 for execution. For example, a communication device may include a transceiver having signals indicative of instructions and data. The instructions and data are configured to cause one or more processors to implement the functions outlined in the disclosure herein. Representative examples of data communication transmission connections include, but are not limited to, a telephone modem connection, a wide area network (WAN), a local area network (LAN), an infrared data connection, an NFC connection, an optical communication connection, and the like.

[0068] It should be understood that the methods, flowcharts, diagrams, and accompanying disclosure described herein can be implemented using computer system 700 as a stand-alone device or on a distributed network of shared computer processing resources, such as a cloud computing network.

[0069] The methods described herein may be implemented by various means depending on the application. For example, the methods may be implemented in hardware, firmware, software, or any combination thereof. In a hardware implementation, the processing unit may be implemented in one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processors (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described herein, or combinations thereof.

[0070] In various embodiments, the methods of the present teachings may be implemented as firmware and / or software programs and applications written in conventional programming languages ​​such as C, C++, Python, etc. When implemented as firmware and / or software, the embodiments described herein may be implemented on a non-transitory computer-readable medium having a program stored thereon that causes a computer to perform the methods described above. It should be understood that the various engines described herein may be provided on a computer system, such as computer system 700, whereby processor 704 will receive instructions provided by any one or combination of memory components RAM 706, ROM 708, or storage device 710, and user input provided via input device 714 to perform the analysis and determination provided by these engines.

[0071] V. Illustrative Definitions and Context The present disclosure is not limited to these exemplary embodiments and applications or to the manner in which the exemplary embodiments and applications operate or are described herein. Further, the figures may show simplified or partial views, and dimensions of elements in the figures may be exaggerated or out of proportion.

[0072] Additionally, when the terms "on," "attached to," "connected to," "coupled to," or similar terms are used herein, an element (e.g., a component, material, layer, substrate, etc.) may be "on," "attached to," "connected to," or "coupled to" another element, regardless of whether the element is directly on, directly attached to, connected to, or coupled to another element, or whether there are one or more intervening elements between the one element and the other element. Additionally, when a reference is made to a list of elements (e.g., elements a, b, c), such reference is intended to include any one of the listed elements by itself, any combination of less than all of the listed elements, and / or all combinations of the listed elements. The division of sections herein is merely for ease of review and is not intended to limit any combination of the elements discussed.

[0073] The term "subject" may refer to a subject of a clinical trial, a person receiving treatment, a person undergoing anti-cancer drug therapy, a person being monitored for remission or recovery, a person undergoing preventive health analysis (e.g., by medical history), or a patient of any other person or subject. In various instances, "subject" and "patient" may be used interchangeably herein.

[0074] Unless otherwise specified, scientific and technical terms used in connection with the present teachings described herein shall have the meanings commonly understood by those skilled in the art.Furthermore, unless otherwise indicated by context, singular terms shall include the plural and plural terms shall include the singular.In general, the nomenclature and techniques utilized in connection with chemistry, biochemistry, molecular biology, pharmacology and toxicology are described herein and are well known and commonly used in the art.

[0075] As used herein, "substantially" means sufficient to function for the intended purpose. Thus, the term "substantially" allows for minor and slight variations from an absolute or perfect state, dimension, measurement, or result, etc., that would be expected by one of ordinary skill in the art, but that do not significantly affect overall performance. When used in reference to a number or parameter or a characteristic that can be expressed as a number, "substantially" can mean within 10 percent.

[0076] The term "plurality" means two or more.

[0077] As used herein, the term "plurality" can be 2, 3, 4, 5, 6, 7, 8, 9, 10, or more.

[0078] As used herein, the term "set of" means one or more. For example, a set of items includes one or more items.

[0079] As used herein, the phrase "at least one of" when used in conjunction with a list of items means that one or more different combinations of the listed items may be used, and in some cases, only one of the items in the list is used. The items may be specific objects, things, steps, operations, processes, or categories. In other words, "at least one of" means that any combination or number of items from the list may be used, but not all of the items in the list may be used. For example, but not limited to, "at least one of item A, item B, or item C" means item A, item A and item B, item B, item A, item B, and item C, item B and item C, or item A and C. In some cases, "at least one of item A, item B, or item C" means, but is not limited to, two items A, one item B, and ten items C, four items B, and seven items C, or any other suitable combination.

[0080] As used herein, a "model" may include one or more algorithms, one or more mathematical techniques, one or more machine learning algorithms, or a combination thereof.

[0081] As used herein, "machine learning" may be the practice of using algorithms to analyze data, learn from it, and then make decisions or predictions about something in the world. Machine learning uses algorithms that can learn from data without relying on rule-based programming.

[0082] As used herein, an "artificial neural network" or "neural network" (NN) may refer to a mathematical algorithm or computational model that mimics an interconnected group of artificial neurons that processes information based on a connectionist approach to computation. A neural network, sometimes called a neural net, may use one or more layers of linear units, nonlinear units, or both to predict an output for a received input. Some neural networks include one or more hidden layers in addition to an output layer. The output of each hidden layer may be used as an input to the next layer in the network, i.e., the next hidden layer or the output layer. Each layer of the network generates an output from the received input according to the current value of a set of each parameter. In various embodiments, a reference to a "neural network" may be a reference to one or more neural networks.

[0083] Neural networks may process information in two ways. For example, neural networks may process information in a training mode when they are being trained, and in an inference (or prediction) mode when they put what they have learned into practice. Neural networks may learn through a feedback process (e.g., backpropagation) that allows the network to adjust the weight factors (which modify the network's behavior) of individual nodes in intermediate hidden layers so that the output matches the output of the training data. In other words, neural networks may learn by being fed training data (training examples), and eventually learn how to arrive at the correct output even when presented with a new range or set of inputs. The neural network may include, for example, but is not limited to, at least one of a feedforward neural network (FNN), a recurrent neural network (RNN), a modular neural network (MNN), a convolutional neural network (CNN), a residual neural network (ResNet), an ordinary differential equation neural network (Neural-ODE), or another type of neural network.

[0084] VI. Description of the Preferred Embodiments Embodiment 1. A method for predicting a treatment outcome, comprising receiving three-dimensional imaging data of a subject's retina, generating a first output using a deep learning system and the three-dimensional imaging data, receiving the first output and baseline data as input for a symbolic model, and predicting, by the symbolic model, a treatment outcome for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD) using the input.

[0085] Embodiment 2. The method of embodiment 1, wherein the three-dimensional imaging data comprises optical coherence tomography (OCT) imaging data.

[0086] Embodiment 3. The method of embodiment 1 or embodiment 2, wherein the baseline data includes at least one of demographic data, baseline visual acuity measurements, baseline central area retinal thickness measurements, baseline hypoluminal defect, or treatment group.

[0087] Embodiment 4. The method of embodiment 3, wherein the attribute data includes at least one of age or gender.

[0088] Embodiment 5. The method of any one of embodiments 1-4, wherein the treatment prognosis includes at least one of a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central zone retinal thickness, or a predicted decrease in central zone retinal thickness.

[0089] Embodiment 6. The baseline data includes a baseline visual acuity measurement, 6. The method of any one of embodiments 1-5, further comprising identifying a baseline visual acuity measurement using the first output.

[0090] Embodiment 7. The method of any one of embodiments 1-6, wherein the treatment outcome is predicted at n months after the baseline time point, where n months is selected as a month between 3 and 30 months after the baseline time point.

[0091] Embodiment 8. The method of any one of embodiments 1-7, wherein the treatment comprises a monoclonal antibody directed against vascular endothelial growth factor and an angiopoietin 2 inhibitor.

[0092] Embodiment 9. The method of any one of embodiments 1-8, wherein the treatment comprises faricimab.

[0093] Embodiment 10. A method for predicting a treatment prognosis for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD), comprising: generating a first predicted prognosis using a deep learning system and three-dimensional imaging data for the subject's retina; generating a second predicted prognosis using a symbolic model and baseline data for the subject; and predicting a treatment prognosis for the subject undergoing treatment for nAMD using the first predicted prognosis and the second predicted prognosis.

[0094] Embodiment 11. The method of embodiment 10, wherein predicting comprises predicting the treatment outcome as a weighted average of the first predicted treatment outcome and the second predicted treatment outcome.

[0095] Embodiment 12. The method of embodiment 10 or embodiment 11, wherein the three-dimensional imaging data includes optical coherence tomography (OCT) imaging data.

[0096] Embodiment 13. The method of any one of embodiments 10-12, wherein the baseline data includes at least one of demographic data, baseline visual acuity measurements, baseline central area retinal thickness measurements, baseline hypoluminal defect, or treatment group.

[0097] Embodiment 14. The method of embodiment 13, wherein the attribute data includes at least one of age or gender.

[0098] Embodiment 15. The method of any one of embodiments 10-14, wherein each of the first predicted treatment outcome, the second predicted treatment outcome, and the treatment outcome includes at least one of a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central retinal thickness, or a predicted decrease in central retinal thickness.

[0099]

[0033] Embodiment 16. A system for administering anti-vascular endothelial growth factor (anti-VEGF) treatment to a subject diagnosed with neovascular age-related macular degeneration (nAMD), comprising: a memory containing a machine-readable medium comprising machine-executable code; and a processor coupled to the memory, the processor comprising: receiving three-dimensional imaging data relating to a subject's retina; generating a first output using the deep learning system and the three-dimensional imaging data; receiving the first output and baseline data as input for a symbolic model; The system is configured to execute machine-executable code that uses an input from the symbolic model to predict a treatment prognosis for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD).

[0100] Embodiment 17. The system of embodiment 16, wherein the three-dimensional imaging data includes optical coherence tomography (OCT) imaging data.

[0101] Embodiment 18. The system of embodiment 16 or embodiment 17, wherein the baseline data includes at least one of demographic data, baseline visual acuity measurements, baseline central area retinal thickness measurements, baseline hypoluminal defect, or treatment group.

[0102] Embodiment 19. The system of any one of embodiments 16-18, wherein the treatment prognosis includes at least one of a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central zone retinal thickness, or a predicted decrease in central zone retinal thickness.

[0103] Embodiment 20. The system of any one of embodiments 16 to 18, wherein the treatment comprises faricimab.

[0104] Embodiment 21. A method for predicting a treatment outcome, comprising receiving baseline data as input for a symbolic model, processing the baseline data using the symbolic model, and predicting, by the symbolic model, a prognosis for a subject undergoing the treatment based on the processing of the baseline data.

[0105] Embodiment 22. The method of embodiment 21, wherein the baseline data includes baseline visual acuity measurements, and further comprising generating the baseline visual acuity measurements using the three-dimensional imaging data and a deep learning system.

[0106] VII. Further Considerations Headings and subheadings between sections and subsections in this document are included merely to improve readability and do not imply that features may be combined across the sections and subsections, and therefore, the sections and subsections do not describe separate embodiments.

[0107] Some embodiments of the present disclosure include a system including one or more data processors. In some embodiments, the system includes a non-transitory computer-readable storage medium containing instructions that, when executed on the one or more data processors, cause the one or more data processors to perform a part or all of one or more methods and / or a part or all of one or more processes disclosed herein. Some embodiments of the present disclosure include a computer program product tangibly embodied in a non-transitory machine-readable storage medium containing instructions configured to cause one or more data processors to perform a part or all of one or more methods and / or a part or all of one or more processes disclosed herein.

[0108] The terms and expressions employed are used as terms of description and not of limitation, and in the use of such terms and expressions there is no intention to exclude any equivalents or portions of the features shown and described, recognizing that various modifications are possible within the scope of the claimed invention. Thus, although the claimed invention has been specifically disclosed by embodiments and optional features, it should be understood that modifications and variations of the concepts disclosed herein may be employed by those skilled in the art, and such modifications and variations are deemed to be within the scope of the invention as defined by the appended claims.

[0109] Although the present teachings are described in connection with various embodiments, it is not intended that the present teachings be limited to such embodiments. On the contrary, the present teachings encompass various alternatives, modifications, and equivalents, as will be understood by those skilled in the art. In describing various embodiments, the specification may present a method and / or process as a particular sequence of steps. However, to the extent that the method or process does not depend on the particular order of steps described herein, the method or process should not be limited to the particular order of steps described, and as can be readily understood by those skilled in the art, the order may be varied and still remain within the spirit and scope of the various embodiments. It should be understood that various changes may be made in the function and arrangement of elements (e.g., elements in a block or schematic diagram, elements in a flow diagram, etc.) without departing from the spirit and scope as set forth in the appended claims.

[0110] In the following description, specific details are set forth to provide a thorough understanding of the embodiments. However, it will be understood that the embodiments may be practiced without these specific details. For example, circuits, systems, networks, processes, and other components may be shown as components in block diagram form in order not to obscure the embodiments with unnecessary detail. In other instances, well-known circuits, processes, algorithms, structures, and techniques may be shown without unnecessary detail in order to avoid obscuring the embodiments.

Claims

1. 1. A method for predicting treatment outcome, comprising: Receiving three-dimensional imaging data relating to a subject's retina; generating a first predicted treatment outcome using a deep learning system and the three-dimensional imaging data; receiving the first predicted treatment outcome and baseline data as input for a symbolic model; and predicting, with the symbolic model, a second treatment outcome for the subject undergoing treatment for neovascular age-related macular degeneration (nAMD) using the inputs; The method, wherein the first and second treatment outcomes include at least one of a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central retinal thickness, or a predicted decrease in central retinal thickness.

2. The method of claim 1 , wherein the three-dimensional imaging data comprises optical coherence tomography (OCT) imaging data.

3. 3. The method of claim 1 or 2, wherein the baseline data includes at least one of demographic data, a baseline visual acuity measurement, a baseline central retinal thickness measurement, a baseline hypoluminal defect, or a treatment group.

4. The method of claim 3 , wherein the attribute data includes at least one of age or gender.

5. 3. The method of claim 1 or 2, wherein the treatment outcome comprises a predicted central zone retinal thickness or a predicted decrease in central zone retinal thickness.

6. the baseline data includes a baseline visual acuity measurement; The method of claim 1 or 2, further comprising generating the baseline visual acuity measurement using the three-dimensional imaging data and a deep learning system.

7. 3. The method of claim 1 or 2, wherein the treatment outcome is predicted at n months after the baseline time point, the n months being selected as a month between 3 and 30 months after the baseline time point.

8. 3. The method of claim 1 or 2, wherein the treatment comprises a monoclonal antibody directed against vascular endothelial growth factor and an angiopoietin 2 inhibitor.

9. 3. The method of claim 1 or 2, wherein the treatment comprises faricimab.

10. 1. A method for predicting treatment outcome for a subject undergoing treatment for neovascular age-related macular degeneration (nAMD), comprising: generating a first predicted treatment outcome using a deep learning system and three-dimensional imaging data relating to the subject's retina; generating a second predicted treatment outcome using the symbolic model and baseline data for the subject; and and predicting a treatment outcome for the subject receiving treatment for the nAMD using the first predicted treatment outcome and the second predicted treatment outcome; The method, wherein each of the first predicted treatment outcome, the second predicted treatment outcome and the treatment outcome includes at least one of a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central retinal thickness, or a predicted decrease in central retinal thickness.

11. The predicting step comprises:

11. The method of claim 10, comprising predicting the treatment outcome as a weighted average of the first predicted treatment outcome and the second predicted treatment outcome.

12. The method of claim 10 or 11, wherein the three-dimensional imaging data comprises optical coherence tomography (OCT) imaging data.

13. 12. The method of claim 10 or 11, wherein the baseline data includes at least one of demographic data, a baseline visual acuity measurement, a baseline central retinal thickness measurement, a baseline hypoluminal defect, or a treatment group.

14. The method of claim 13 , wherein the attribute data includes at least one of age or gender.

15. 12. The method of claim 10 or 11, wherein each of the first predicted treatment outcome, the second predicted treatment outcome, and the treatment outcome comprises a predicted central regional retinal thickness or a predicted decrease in central regional retinal thickness.

16. 1. A system for predicting a treatment outcome of an anti-vascular endothelial growth factor (anti-VEGF) treatment for a subject diagnosed with neovascular age-related macular degeneration (nAMD), comprising: a memory containing a machine-readable medium including machine-executable code; a processor coupled to the memory, the processor comprising: Receiving three-dimensional imaging data relating to a subject's retina; generating a first predicted treatment outcome using a deep learning system and the three-dimensional imaging data; receiving the first predicted treatment outcome and baseline data as input for a symbolic model; and using the input with the symbolic model to predict a second treatment outcome for the subject undergoing treatment for neovascular age-related macular degeneration (nAMD); The system, wherein the first and second treatment outcomes include at least one of a predicted visual acuity measurement, a predicted change in visual acuity, a predicted central retinal thickness, or a predicted decrease in central retinal thickness.

Images