Method and system for determining intraocular lens parameters for ophthalmic surgery using emulated finite element analysis models

Abstract

Certain aspects of the present disclosure provide techniques for performing ophthalmic surgical procedures, such as cataract surgery. One example method for determining one or more IOL parameters for intraocular lenses (IOLs) to be used in the cataract surgical procedure includes using a fusion machine learning model to generate a recommendation including one or more IOL parameters for an IOL to be used in the cataract surgery based at least in part on a first predicted lens behavior of each of the one or more IOLs by an emulated finite element analysis (EFEA) model and a second predicted lens behavior of each of the one or more IOLs by an IOL power calculator (IPC) machine learning model.

Description

[Technical Field]

[0001] Introduction Aspects of the present disclosure relate to ophthalmic surgery, and more particularly to determining IOL parameters for an intraocular lens (IOL) to be used during cataract surgery for a patient using an emulated finite element (FEA) model. As defined herein, IOL parameters include at least one of the type, size, and power of the IOL to be inserted into the patient's eye during cataract surgery. [Background technology]

[0002] Ophthalmic surgery generally includes a variety of procedures performed on a person's eye. These surgical procedures may include, among other procedures, cataract surgery, which is a procedure in which the crystalline lens, or natural lens, of a human eye is removed and replaced with a synthetic lens, also known as an IOL, to correct vision problems resulting from opacification of the natural lens.

[0003] Selecting the correct IOL for a patient is crucial in achieving optimal postoperative visual outcomes. IOLs are available in a variety of types, powers, and sizes and can be selected based on measurements of the anatomical parameters of a patient's eye. Anatomical parameters of a person's eye, such as axial length (i.e., the distance between the anterior corneal surface and the retina), corneal thickness, anterior chamber depth (i.e., the distance between the anterior corneal surface and the anterior lens surface), and corneal diameter (i.e., the distance between the borders of the cornea and the sclera on both sides of the eye), generally influence the selection of IOL parameters made in planning and performing cataract surgery for a patient. Planning and performing cataract surgery, as defined herein, includes determining the correct IOL parameters to improve the patient's vision. For example, based on measurements of the patient's anatomical parameters, the surgeon may attempt to determine IOL parameters that are likely to restore the patient's vision. The surgeon then selects an IOL from a set of IOLs that matches the determined IOL parameters. The surgeon then places or implants the selected IOL into the patient's capsular bag.

[0004] Measurements of anatomical parameters for a particular patient may often fall within a known distribution (e.g., a normal distribution two standard deviations from the overall mean, between lower and upper limits that a set percentage of patients fall into, and within which approximately 95% of patient measurements fall), and therefore planning and performing cataract surgery for such a patient may be a relatively straightforward task. However, if one or more anatomical parameters for a particular patient deviate from the known distribution or are otherwise outliers (hereinafter "abnormal"), planning and performing cataract surgery for such a patient may become a more complicated task. In addition, the combination of anatomical parameters may make treating the eye a complex task.

[0005] Thus, an IOL power calculator (IPC) may be used to determine IOL parameters for each particular patient. The IPC can provide a prediction of the postoperative refractive outcome based on a model or formula that takes as input measurements of the patient's anatomical parameters and provides the patient's predicted postoperative refractive outcome for one or more different IOLs. Using the prediction from the IPC, the surgeon can attempt to select the IOL most likely to restore the patient's vision, such as an IOL with the lowest predicted postoperative refractive error.

[0006] While such IPC models or formulas have been developed for non-accommodative IOLs, they may not function adequately for accommodative IOLs. Accommodative IOLs differ from standard "static" IOLs because they can change focal length. Accommodative IOLs may be fluid-filled, allowing the lens to dynamically change shape. Accommodative IOLs have flexible "arms" called haptics, which use the eye's muscle movements to change focus from far to near. This adjustment allows incoming light rays to be properly focused on the retina. The eye accommodates when gazing at a close object. During accommodation, the eye's ciliary muscles contract, causing the zonules to relax, thickening the shape of the natural lens. A thicker lens has a steeper curvature that allows incoming light rays from near objects to be better focused on the retina. With an accommodative IOL implanted in a patient's eye, contraction of the ciliary muscles causes the flexible haptics of the accommodative IOL to bend, thereby moving the focal region of the accommodative IOL forward. The focal region in an anterior position provides additional near focusing or "accommodation" capabilities. These complex interactions of the accommodative IOL with components of the patient's eye may not be captured by current cataract planning and IPC, which may lead surgeons to select IOL parameters that may result in surgical failure. Summary of the Invention [Problem to be solved by the invention]

[0007] Therefore, there is a need for techniques to accurately determine IOL parameters for a patient based, at least in part, on measurements of anatomical parameters for the patient's eye. [Means for solving the problem]

[0008] Certain embodiments provide a method for determining one or more IOL parameters for an intraocular lens (IOL) to be used in a cataract surgery procedure. The method generally includes: generating, using one or more ophthalmic imaging devices, a plurality of data points associated with measurements of a plurality of anatomical parameters of the eye to be treated; generating, using a machine learning model trained to emulate a finite element analysis (FEA) model, a first predicted lens behavior based at least in part on the plurality of data points associated with the measurements of the plurality of anatomical parameters of the eye to be treated and one or more IOL parameters for each of the one or more IOLs; generating, using an IOL power calculator machine learning model, a second predicted lens behavior based at least in part on at least a subset of the plurality of data points associated with the measurements of the plurality of anatomical parameters of the eye to be treated and one or more IOL parameters for each of the one or more IOLs; and generating, using a fusion machine learning model, a recommendation including one or more IOL parameters for an IOL to be used in the cataract surgery based at least in part on the first and second predicted lens behaviors.

[0009] Aspects of the present disclosure provide means, apparatus, processors, and computer-readable media for performing the methods described herein.

[0010] To the accomplishment of the foregoing and related ends, the one or more aspects comprise the features hereinafter fully described and particularly pointed out in the claims. The following description and the annexed drawings set forth in detail certain illustrative features of the one or more aspects. These features are indicative, however, of but a few of the various ways in which the principles of the various aspects may be employed.

[0011] The accompanying drawings depict certain aspects of one or more embodiments and therefore should not be considered as limiting the scope of the disclosure. [Brief explanation of the drawings]

[0012] [Figure 1]1 illustrates an example environment in which an emulated finite element analysis (EFEA) model, an IPC machine learning (ML) model, and a fused ML model, according to certain aspects described herein, are trained and deployed for use in generating recommendations, including recommended IOL parameters, for cataract surgery for a patient based at least on measurements of the patient's anatomical parameters. [Figure 2] FIG. 1 is a diagram of an eye model and eye properties. [Figure 3] 1 is a flow diagram illustrating example operations for training an EFEA model according to certain embodiments described herein. [Figure 4] 1 illustrates the use of an EFEA model and an IPC ML model to train a fusion ML model, according to certain embodiments described herein. [Figure 5] 1 illustrates example operations that may be performed by a computing device to generate and output recommendations, including IOL parameters, for a patient's cataract surgery based on a patient's anatomical parameter data points, in accordance with certain aspects described herein. [Figure 6] 1 illustrates an example computing device in which embodiments of the present disclosure may be implemented, according to certain aspects described herein. DETAILED DESCRIPTION OF THE INVENTION

[0013] For clarity, where possible, the same reference numerals have been used to denote identical elements common to the figures, and it is contemplated that elements and features of one embodiment may be beneficially incorporated in other embodiments without specific indication.

[0014] As mentioned above, cataract surgery is a surgical procedure in which a diseased lens is replaced with an IOL. Typically, the diseased lens is a lens that has developed a cataract, which is a clouding of the natural lens that adversely affects the patient's vision (e.g., the patient experiences faded vision, blurred or double vision, halos around point light sources, or other adverse effects). The IOL may be selected to replace the patient's natural lens and restore or at least improve the patient's vision. The determination of a set of IOL parameters for an IOL to be implanted in a patient's eye is influenced by measurements of the patient's anatomical parameters. More specifically, as mentioned above, based on the characteristics and / or measurements of the patient's eye, a surgeon may attempt to determine a set of IOL parameters likely to restore the patient's vision. For example, some IOLs may enable optimization of near or distance vision, or may be used to compensate for a patient's natural corneal astigmatism.

[0015] Accommodative IOLs differ from non-accommodative IOLs in that they can dynamically change shape to change focal length and accommodate the patient's eye movements. In the human eye, the lens is surrounded by the circular ciliary muscle. The ciliary band connects the ciliary muscle to the lens capsule, which surrounds the lens. When the eye gazes at a distant object, the ciliary muscle relaxes and the ciliary band tightens, allowing the lens to have a flatter curve and better focus incoming light rays from distant objects onto the retina. In young eyes, accommodation is essentially instantaneous and easy. As the eye ages, the lens becomes less flexible, leading to a loss of near vision, a characteristic sign of presbyopia in people over the age of 40. Such patients may be well candidates for the use of accommodative lenses. When the accommodative IOL is implanted in a patient's eye, contraction of the patient's ciliary muscles causes the flexible haptics of the accommodative IOL to bend, thereby moving the focal region of the accommodative IOL forward and increasing its focusing ability.

[0016] As discussed above, the interaction of an accommodative IOL with a patient's eye can be complex and may not be accounted for by existing lens fitting procedures and IPC. Accordingly, aspects presented herein provide systems and methods that use FEA to model the interaction of a patient's eye with an IOL to better predict IOL behavior in order to predict postoperative outcomes and recommend IOL parameters to the patient. In particular, FEA can be used to train an FEA to inform the selection of an accommodative IOL for a patient based on the patient's anatomical characteristics and / or measurements.

[0017] Finite element analysis can be used to generate predictions for both static and accommodative IOLs. FEA uses a numerical technique called the finite element method (FEM) to simulate physical phenomena (e.g., the interaction of an IOL in a human eye). FEA can involve creating a mesh of up to millions of smaller elements (e.g., components of the eye) that together form the shape of a structure (e.g., a patient's eye). Calculations are performed for each element, and the individual results are combined to provide the final result for the structure. FEM uses partial differential equations (PDEs) and / or matrices to describe complex physical behaviors, such as structural, fluid, and thermal behavior (e.g., the behavior of an accommodative IOL in a patient's eye). These PDEs or matrices can be solved to calculate relevant quantities to estimate the behavior of the IOL under various conditions.

[0018] FEA can be used to model how an accommodative IOL behaves after implantation in a patient's capsular bag to understand how different capsule geometries affect the dynamics of the accommodative IOL and to predict a patient's post-operative outcome for a given IOL. According to embodiments of the present disclosure, an FEA model is designed and solved to predict the behavior of one or more lenses after implantation in a patient's eye using pre-operative measurements of one or more specific patients and one or more lens designs. The FEA model can be fine-tuned by comparing the predictions output by the FEA model with clinical data across a large design space of patient pre-operative measurements and IOL designs to predict how the IOL power of different sized fluid accommodative IOLs will change when implanted in different capsules in different patient eyes with different geometries.

[0019] However, FEA models can be computationally expensive, e.g., they can take up to 24 hours or even longer to run. Furthermore, FEA models may require expertise to generate, configure, and run the FEA model.

[0020] Accordingly, aspects of the present disclosure provide for the training and use of an emulated FEA model. The use of the emulated FEA model enables a faster, more efficient approach to utilizing FEA modeling to predict postoperative outcomes and recommend IOL parameters for patients. In some embodiments, the emulated FEA model is trained to match or closely match the output of the FEA model. The trained emulated FEA model may be computationally less expensive than the FEA model (e.g., the model may run in seconds) and does not require expert operation. Thus, the emulated FEA model is portable and can be easily deployed in a clinical environment for a clinician or surgeon to fit an IOL to a patient for implantation.

[0021] In some embodiments, the IPC is used in addition to the emulating FEA model to provide predicted postoperative outcomes for a patient for a given set of IOLs with different IOL powers, types, and / or sizes. The IPC can be a standard IPC (e.g., Bartlett IPC, Haigis IPC, Hill-RBF IPC, SRK (Sanders-Retzlaff, Kraff) IPC, etc.). In some other embodiments, the IOL power calculator is a machine learning (ML)-based IOL power calculator. As discussed in more detail herein, the ML-based IPC can be trained across a large design space of patient preoperative measurements and IOL designs to predict postoperative refractive error for various accommodative IOLs for each patient based on the patient's anatomical measurements and / or characteristics.

[0022] In some embodiments, the emulated FEA model and the ML-based IPC model are used to generate and train a fusion model. In some embodiments, the fusion model is also an ML-based model. The fusion model receives outputs from both the emulated FEA model and the ML-based IPC model as inputs to the fusion model, and then outputs recommended IOL parameters. The fusion model can be trained using inputs from the emulated FEA model and the IOL power calculator to make IOL parameter recommendations. These recommendations can be compared with clinical data to refine the fusion model.

[0023] An example of a computing environment for ophthalmic surgical treatment planning using the EFEA model 1 illustrates an example computing environment in which models are trained and used in generating recommendations, including IOL parameters, for a patient's cataract surgery. Typically, these models are trained using a corpus of training data, including records corresponding to past patient data, and can be deployed for use in planning a current patient's cataract surgery, including generating IOL parameters. As defined herein, a new or current patient (hereinafter, "current patient") is generally a patient undergoing cataract surgery to replace a defective natural crystalline lens. As discussed in more detail below, the recommended IOL parameters for the current patient may be generated by a fusion model trained to generate the recommended IOL parameters.

[0024] The past patient data for each past patient includes patient demographic information, recorded data points related to measurements of anatomical parameters, desired outcomes, actual treatment data such as actual IOL parameters of the IOL implanted in the patient's eye or other information about the past patient's treatment, and treatment outcome data (e.g., postoperative refractive error and parameters indicative of the past patient's satisfaction or dissatisfaction with treatment (e.g., patient satisfaction score)). It is noted herein that the actual treatment administered to a patient may differ from the recommended treatment.

[0025] Using these models, a vast amount of past patient data can be utilized to generate recommended IOL parameters for the current patient. This vast amount of past patient data, in part, represents the knowledge and past experience of other surgeons who have performed similar surgeries on similar patients. Using the systems and methods described herein, surgeons can utilize this vast amount of past patient data to identify, for the current patient, IOL parameters that will result in an optimal surgical outcome for that patient. Thus, the present technology advances the field of medicine by enabling better IOL parameter selection, such as during cataract surgery, which ultimately results in improved visual acuity after IOL placement.

[0026] Various approaches can be used to train and deploy a model that generates IOL parameters for a current patient. One example deployment is shown in FIG. 1 . However, other deployments are also contemplated. For example, FIG. 1 illustrates a deployment scheme in which a model is trained on a remote server 120 and deployed to a user console 130 used by a surgeon during cataract surgery planning. In another example deployment, a model can be trained and deployed on a remote server 120 accessible via, for example, a computing system, an imaging device, and / or a surgical console. In another example deployment, a model can be trained on a remote server 120 and deployed to an imaging device 110 used preoperatively and / or intraoperatively. However, it should be recognized that various other approaches can be envisioned for training and deploying a model that generates IOL parameters for a current patient, and the deployment scheme shown in FIG. 1 is a non-limiting, illustrative example.

[0027] FIG. 1 illustrates an example computing environment 100 in which one or more imaging devices 110, a server 120, a user console 130, and a past patient data repository 140 are connected via a network to train one or more models used in generating recommended IOL parameters for a current patient based at least in part on data points related to measurements of the current patient's anatomical parameters provided by the one or more imaging devices 110.

[0028] The imaging device 110 generally represents a variety of devices capable of generating data points associated with one or more measurements of anatomical parameters of a patient's eye. As used herein, anatomical parameters of the eye may refer to established optical parameters. FIG. 2 is a diagram of an eye model 200 illustrating various optical parameters, according to certain embodiments described herein. As shown in FIG. 2 , the optical parameters include an axial length (e.g., the distance from the anterior corneal surface 202A of the cornea 202 to the retina 204), a central corneal thickness (CT) measurement of the cornea 202, an anterior chamber depth (AD) of the anterior chamber 208 (e.g., the distance from the apex of the posterior cornea 202P to the apex of the anterior lens surface 206A of the lens 206), a lens thickness (LT) of the lens 206, a lens diameter (LD) (also referred to as the lens equatorial diameter) of the lens 206, a lens volume of the lens 206, a lens surface area of ​​the lens 206, the curvature of the anterior corneal surface 202A, the curvature of the posterior lens surface 206P, a corneal diameter (WD) (e.g., The parameters may include one or more lens characteristic dimensions, such as, but not limited to, the curvature and astigmatism of the anterior corneal surface 202A, the shape of the anterior corneal surface 202A, the depth of the vitreous humor 210, and other optical parameters associated with the patient's eye, including the patient's lens characteristic dimensions, the geometry of the lens capsule, the ciliary muscle 212, and the ciliary zone (e.g., but not limited to, the ciliary body area, the ciliary body process moment, the ciliary body length, etc.).

[0029] These optical parameters are examples of anatomical parameters measured by imaging device 110, however, imaging device 110 may measure additional anatomical parameters that may be used in the model.

[0030] In some embodiments, the data generated by imaging device 110 for each patient may be stored in repository 140 as patient data, which may include raw measurements of the patient's anatomical parameters, biometric information derived from the measurements, and / or imaging data including two-dimensional cross-sectional images showing the cornea, iris, lens, and retina, three-dimensional images of the eye, two-dimensional topographical maps of the eye, or other types of imaging data. In general, computing environment 100 may include any number of imaging devices 110, which may be used to generate different types of data that may be used as input to one or more models to generate recommended IOL parameters.

[0031] The imaging devices 110 may include (1) preoperative measurement and / or imaging devices used in a clinic and / or (2) intraoperative measurement and / or imaging devices. In one example, one of the imaging devices 110 may be an optical coherence tomography (OCT) device capable of generating optical images. Various types of OCT devices may be used. As an example, an OCT device may be used to generate a two-dimensional cross-sectional image of a current patient's eye from which measurements of various anatomical parameters may be derived. The two-dimensional cross-sectional image may show the positions of the cornea, lens, and retina on a two-dimensional plane (e.g., with the cornea on one side of the two-dimensional cross-section and the posterior surface of the retina on the other side of the two-dimensional cross-section). From this two-dimensional cross-sectional image of the current patient's eye, the OCT device may derive various measurements.

[0032] For example, an OCT device can generate axial length measurements, central corneal thickness measurements, anterior chamber depth measurements, lens thickness measurements, and other related measurements from cross-sectional images. In some embodiments, an OCT device can generate one-dimensional data measurements (e.g., from a central point) or generate three-dimensional measurements from which additional information, such as a map of tissue thickness, can be generated. Examples of OCT devices are described in more detail in U.S. Pat. No. 9,618,322, which discloses a "Process for Optical Coherence Tomography and Apparatus for Optical Coherence Tomography," and U.S. Patent Application Publication No. 2018 / 0104100, which discloses an "Optical Coherence Tomography Cross View Image," both of which are incorporated herein by reference in their entireties.

[0033] Another one of the imaging devices 110 may be a keratometer. Generally, a corneal curvature may be measured by reflecting a light pattern, such as a ring-shaped illuminated dot, off the current patient's eye and capturing the reflected light pattern. The keratometer may perform image analysis on the reflected pattern (compared to the pattern output by the imaging device 110 of the reflection from the current patient's eye) to measure or otherwise determine the values ​​of various anatomical parameters. These anatomical parameters may include, for example, curvature information and aberration information of the anterior corneal surface of the current patient's eye. The curvature information may be, for example, a general curvature measurement, maximum curvature and axis information identifying the axis along which the maximum curvature occurs, and minimum curvature and axis information identifying the axis along which the minimum curvature occurs.

[0034] Another one of the imaging devices 110 can be a topography device that measures the topography of the anterior corneal surface. The topography device can generate a detailed topography map of the base shape using analysis of reflected light patterns distributed across the corneal area. For example, because the cornea is typically spherical or nearly spherical, the topography map can show deviations from the base shape, in which different colors represent the amount of deviation from the base shape at discrete points along the cornea.

[0035] Another exemplary imaging device 110 can be an intraoperative aberrometer. One example of an intraoperative aberrometer is the Ora™ with Verifeye™ (Alcon Inc., Switzerland), which is described in more detail in part in commonly owned U.S. Pat. No. 7,883,505, disclosing an “Integrated Surgical Microscope and Wavefront Sensor,” and U.S. Pat. No. 8,784,443, disclosing a “Real-Time Surgical Reference Indicium Apparatus and Methods for Astigmatism Correction,” both of which are incorporated herein by reference. The ORA™ can capture, among other measurements, total ocular refraction measurements that account for total ocular astigmatism, including surgically induced astigmatism and posterior corneal astigmatism, as well as post-myopic PRK / LASIK and long-short ocular measurements, to provide guidance for adjusting lens selection and placement for all eye types.

[0036] The imaging device 110 may also include a rotating camera (e.g., a Scheimpflug camera), a magnetic resonance imaging (MRI) device, an ophthalmoscope, an optical biometer, a three-dimensional stereo digital microscope (such as the NGENUITY® 3D visualization system (Alcon Inc., Switzerland)).

[0037] Server 120 generally represents a computing device or a collection of computing devices on which training data sets can be generated and used to train one or more models for generating recommended IOL parameters. Server 120 is communicatively coupled to a past patient data repository 140 (hereinafter "repository 140"), in which past patient records are stored. In certain embodiments, repository 140 may be or include a database server for receiving information from server 120, user console 130, and / or imaging device 110 and storing the information in corresponding patient records in a structured and organized manner.

[0038] In certain embodiments, each patient record in repository 140 includes information such as the patient's demographic information, data points associated with anatomical parameter measurements, actual treatment data associated with the patient's cataract surgery, and treatment outcome data. For example, each patient's demographic information may include the patient's age, gender, ethnicity, and so forth. Data points associated with pre-operative and / or intra-operative measurements of anatomical parameters may include raw data generated by, or measurements derived from, an OCT device, a keratometer, a profilometer, or the like, as described above. As described further below, the actual treatment data may include the actual IOL parameters (e.g., IOL type, IOL power, IOL size) of the IOL used for the patient, as well as any additional relevant information related to the patient's treatment. For example, the actual treatment data may indicate how cataract surgery was performed for the patient, the tools used in the treatment, and other information related to specific procedures performed during surgery. Each patient record also includes treatment outcome data, which may include outcome parameters, such as various data points indicating patient satisfaction with treatment, such as a binary indication of satisfaction or dissatisfaction with the surgical outcome, visual acuity level measured after treatment, or otherwise.

[0039] Server 120 uses these past patient records to generate datasets for use in training ML models that can recommend IOL parameters to surgeons for treating current patients. More specifically, as shown in FIG. 1 , server 120 includes one or more training data generators 122 (hereinafter “TDG 122”) and one or more model trainers 124. TDG 122 retrieves data from repository 140 to generate datasets that model trainer 124 uses to train FEA model 125, EFEA model 126, IPC ML model 127, and / or fusion model 128. It should be understood that FEA model 125, EFEA model 126, IPC ML model 127, and fusion model 128 may be trained using the same, overlapping, or different datasets. Furthermore, FEA model 125, EFEA model 126, IPC ML model 127, and fusion model 128 may be trained by different model trainers 124. The generation and training of the FEA model 125 and the EFEA model 126 are discussed in more detail below with respect to Figure 3. The generation and training of the IPC ML model 127 and the fusion model 128 are discussed in more detail below with respect to Figure 4.

[0040] Model trainer 124 includes or refers to one or more algorithms configured to use the training data set to train FEA model 125, EFEA model 126, IPC ML model 127, and fusion model 128. In certain embodiments, a trained model refers to a function, e.g., with weights and parameters, used to make IOL-related predictions. The IOL-related predictions may include predicted postoperative refractive error, predicted optimal IOL parameters, and IOL power per frame.

[0041] The model trainer 124 can train one or more of the EFEA model 126, the IPC ML model 127, and the fusion model 128 using one or more ML algorithms. ML algorithms may generally include supervised learning algorithms, unsupervised learning algorithms, and / or semi-supervised learning algorithms. Unsupervised learning is a type of machine learning algorithm used to draw inferences from a database of input data without labeled answers. Supervised learning is the ML task of learning a function that maps inputs to outputs based on example input-output pairs, for example. Supervised learning algorithms may generally include regression algorithms, classification algorithms, decision trees, various types of neural networks, etc.

[0042] In some embodiments, the model trainer 124 can train deep learning models. These deep learning models may include, for example, convolutional neural networks (CNNs), adversarial learning algorithms, generative networks, or other deep learning algorithms that can learn relationships between data sets that may not be explicitly defined in the data used to train such models. The deep learning model may, for example, map inputs to different neurons in one or more layers of the deep learning model (e.g., if the model is generated using a neural network), where each neuron in the model represents a new feature learned over time in an internal representation of the input. These neurons may then be mapped to outputs representing recommended IOL parameters, as described above.

[0043] After the models are trained, the model trainer 124 may deploy one or more of the trained models to the user console 130 for use in predicting and recommending IOL parameters for the current patient. In some embodiments, the EFEA model 126, the IPC ML model 127, and the fusion model 128 are deployed on the user console 130, but the FEA model 125 is not. For example, as discussed herein, the FEA model 125 may be computationally expensive and may require an expert to generate and run. Instead, as discussed in more detail below with respect to FIG. 3, the FEA model 125 may be used to generate the less computationally expensive and simpler to operate EFEA model 126, which is then deployed on the user console 130. As noted above, FIG. 1 illustrates only one example of when the EFEA model 126, the IPC ML model 127, and the fusion model 128 may be deployed. In other examples, the EFEA model 126, the IPC ML model 127, and the fusion model 128 may be deployed to or executed on the server 120. In yet other examples, the model trainer 124 may deploy the EFEA model 126, the IPC ML model 127, and the fusion model 128 to one or more of the imaging devices 110.

[0044] User console 130 generally represents a computing device or system communicatively coupled to server 120, repository 140, and / or imaging device 110. In particular embodiments, user console 130 may be a desktop computer, laptop computer, tablet computer, smartphone, or other computing device. For example, user console 130 may include or be associated with a computing system used in a surgeon's office or clinic. In another example, user console 130 may be a surgical console used by a surgeon in an operating room to perform cataract surgery on a current patient.

[0045] 1 , the trained model is deployed by server 120 to user console 130 to predict, for a current patient, IOL parameters that will optimize the patient's surgical outcome. As shown, user console 130 includes a patient data recorder (“PDR”) 132 and an IOL recommendation generator (“IRG”) 134. Note that while in FIG. 1 , TDR 132 and IRG 134 execute on user console 130, in certain other embodiments, TDR 132 and IRG 134 may execute on one or more other computing systems, such as server 120, imaging device 110, and / or any computing system capable of communicating with one or more of server 120, imaging device 110, and / or historical patient data repository 140.

[0046] IRG 134 generally refers to a software module or set of software instructions or algorithms, including EFEA model 126, IPC ML model 127, and fusion model 128, that takes an input set about the current patient and generates recommended IOL parameters as output. In certain embodiments, IRG 134 is configured to receive an input set from at least one of repository 140, imaging device 110, a user interface of user console 130, and other computing devices that a medical team may use to record information about the current patient. In certain embodiments, IRG 134 outputs the recommended IOL parameters to a display device communicatively coupled to user console 130, prints the recommended IOL parameters, sends one or more electronic messages including the recommended IOL parameters to a destination device (e.g., a connected device such as a tablet, smartphone, wearable device, etc.).

[0047] The input set described above includes data points associated with measurements of anatomical parameters of a new patient's eye provided by imaging device 110. These data points are provided to user console 130 for use as inputs to the model, and may also be provided to server 120 and / or repository 140.

[0048] Pre-operatively or intra-operatively, the user console 130 retrieves data points (e.g., from the repository 140 or from the temporary memory of the user console 130) associated with pre-operative and / or intra-operative measurements of anatomical parameters and other patient information associated with the current patient (e.g., demographic information for the current patient, etc.) to use as input to models stored in the user console 130. The data points and other patient information may be provided to the IRG 134, which runs the inputs through the EFEA model 126, the IPC ML model 127, and the fusion model 128 to generate recommended IOL parameters. The user console 130 then outputs the recommendations generated by one or more of the models.

[0049] The TDR 132 receives or generates treatment data related to the treatment provided to the current patient. As described above, the treatment data may include the actual IOL parameters used for the current patient and any additional relevant information, such as the method used to perform cataract surgery. As previously defined, actual IOL parameters refer to the type, power, and size of the IOL actually implanted in the current patient's eye by the surgeon. If the surgeon does not follow the recommended IOL parameters, the actual IOL parameters will differ from the recommended IOL parameters. In certain such cases, the TDR 132 may receive treatment data as user input to the user interface of the user console 130. If the surgeon follows the recommended IOL parameters, the actual IOL parameters will be the same as the recommended IOL parameters. In certain such cases, the TDR 132 treats the recommended IOL parameters as actual IOL parameters recorded as part of the treatment data. In the embodiment of FIG. 1, the TDR 132 transmits the actual IOL parameters to the repository 140 and / or the server 120. The repository 140 then augments the current patient's record with the actual IOL parameters used in the treatment.

[0050] The TDR 132 generally allows a user of the user console 130 to provide post-operative information identifying the surgical outcome of a treatment. Although the TDR 132 is illustrated as executing on the user console 130, one skilled in the art will appreciate that the TDR 132 can execute on a computing device separate from the user console 130.

[0051] The TDG 122 continues to enrich the training dataset with information about patients for whom the deployed model provided recommended IOL parameters. In certain embodiments, the TDG 122 enriches the dataset whenever information about a new (i.e., current) patient becomes available. In other certain embodiments, the TDG 122 enriches the dataset with batches of new patient records, which may improve resource efficiency. The TDG 122 may convert records in the repository 140 that include data points associated with the current patient's measurements of one or more anatomical parameters, actual treatment data, and treatment outcome data into new samples in the training dataset. The model trainer 124 uses the new training dataset to retrain one or more of the models. More generally, whenever a new (i.e., current) patient is treated, information about the new patient is stored in the TDG 122's repository 140 to supplement the training dataset, and the model trainer 124 uses the supplemented training dataset to retrain one or more of the models.

[0052] Example of how to generate an emulated FEA model 3 is a flow diagram illustrating example operations 300 for training a simulated finite element analysis model (e.g., EFEA model 126) according to certain embodiments described herein. The operations 300 may be performed by one or more model trainers 124 shown in FIG.

[0053] As shown, the operations 300 begin at block 302 by generating a FEA model. In some embodiments, the FEA model 125 is generated by the model trainer 124. Generating the FEA model may include generating a mathematical model of the interaction between components of a human eye (e.g., eye model 200) and an IOL, such as an accommodative lens. In some embodiments, the FEA model includes a set of PDEs and / or matrices that can be solved to calculate one or more quantities related to the interaction of the accommodative IOL with components of the human eye. The variables of the mathematical model include pre-operative anatomical parameters of the eye. For example, the variables may include any of the anatomical parameters measured by the imaging device 110. In particular, the anatomical parameters include, but are not limited to, the parameters described above in connection with the eye, lens feature dimensions, and lens capsule geometry. The lens feature dimensions may be used to define a complete pre-operative profile of the lens capsule. The use of additional anatomical parameters may improve the accuracy of the FEA model.

[0054] In some embodiments, the FEA model can be solved for one or more sets of anatomical parameters to generate data indicative of how the IOL may behave within the capsular bag given different parameters and characteristics of the IOL. The FEA model also receives one or more parameters of the IOL as input. For example, the type, size, and / or label power of the IOL can be input into the FEA model along with each set of anatomical parameters. As used herein, the IOL label power can be the preoperative power of the IOL.

[0055] For an input set of a patient's preoperative anatomical parameters and IOL parameters, the FEA model predicts the postoperative behavior of each input IOL after implantation into the patient's capsular bag. In some embodiments, the FEA model predicts the IOL's postoperative resting state and effective lens position (e.g., the IOL's postoperative position relative to the cornea and retina), and thus the IOL's refractive outcome. For example, based on the postoperative effective lens position, the power needed to ensure that light is properly focused on the retina can be determined. The IOL's postoperative resting state can determine the pressure applied to the IOL (e.g., by the haptics as they flex in response to pressure), which changes the IOL's shape and size and thus affects the IOL's power. Therefore, by predicting the postoperative resting state and effective lens position, the FEA model can predict the IOL's sizing factor and postoperative refractive outcome. As discussed in more detail with respect to FIG. 3, the FEA model can output a predicted IOL power per frame. As used herein, frame refers to a measurement instant. The FEA model therefore outputs predicted IOL power at different times, thereby enabling modeling of how an accommodative IOL will interact with a patient's capsular bag after implantation in the patient's eye.

[0056] As shown, operations 300 may continue at block 304 by fine-tuning the FEA model with clinical data. As shown, fine-tuning the FEA model with clinical data at block 304 includes inputting anatomical parameters of past patients and one or more IOL parameters into the FEA model at block 306 to predict IOL behavior based on the input. In some embodiments, the anatomical parameters of past patients input into the FEA model include values ​​for each anatomical parameter used to build the FEA model. In some embodiments, the anatomical parameters of past patients input into the FEA model include values ​​for a subset of the anatomical parameters used to build the FEA model. In one exemplary embodiment, the anatomical parameters of past patients include preoperative lens feature dimensions of past patients, and the IOL parameters are one or more IOL label powers. Based on the input, the FEA model outputs predicted IOL behavior. In one exemplary embodiment, the predicted IOL behavior includes IOL power per frame.

[0057] As shown, fine-tuning the FEA model using clinical data in block 304 includes comparing the clinical data for the same anatomical parameters with the FEA model's predicted IOL behavior in block 308. For example, if the FEA model 125 outputs a predicted IOL power per frame for an input set of anatomical parameters and IOL parameters, the predicted IOL power per frame is compared to clinical results (e.g., treatment results in repository 140), which may include actual IOL power per frame for patients with the same set of anatomical parameters and IOL parameters. In block 310, the FEA model is adjusted based on the comparison. In some embodiments, the FEA model is adjusted by varying one or more parameters of the FEA model. The parameters of the FEA model may include, but are not limited to, Zonal tension, Zonal dynamics, capsule elasticity, friction, and / or other parameters of the FEA model.

[0058] The FEA model 125 can be fine-tuned across many sets of anatomical and IOL parameters to find a FEA model that accurately predicts IOL behavior. As used herein, a FEA model that accurately predicts IOL behavior is capable of predicting IOL behavior at a specified success threshold level. In some embodiments, the FEA model 125 is fine-tuned across the full range or a wider range of anatomical values ​​expected to be seen in the clinic.

[0059] 3, after fine-tuning the FEA model, operations 300 continue at block 312 by generating an emulated FEA model (e.g., EFEA model 126). The EFEA model is generated to match (e.g., closely match within a specified threshold) the FEA model. In some embodiments, the EFEA model is a machine learning model.

[0060] The operations 300 continue at block 314 by training the EFEA model to replicate the output of the FEA model. In some embodiments, the EFEA model 126 is trained by the model trainer 124 shown in FIG. 1 . In some embodiments, the model trainer 124 refers to an AI / ML learning algorithm or a combination of AI / ML learning algorithms for training an AI / ML model. Examples of AI / ML learning algorithms include various types of optimization algorithms, such as gradient descent, stochastic gradient descent, nonlinear conjugate gradient, etc., and the EFEA model 126 may be a Gaussian process regression (GPR) model, a linear regression model, an autoregressive integrated moving average (ARIMA) model, a neural network, etc. For input values ​​for which the FEA model 125 was run, the EFEA model 126 is expected to return the same output value with zero uncertainty. For input values ​​for which the FEA model 125 was not run, the EFEA model 126 predicts, along with estimated uncertainties, what the FEA model 125 would output.

[0061] 3 , training the EFEA model to replicate the output of the FEA model in block 314 includes inputting anatomical parameters of past patients and one or more IOL parameters into the EFEA model and predicting IOL behavior based on the input in block 316. To train EFEA model 126 to emulate FEA model 125, the input parameters are those input parameters implemented by FEA model 125.

[0062] Training the EFEA model to match the EFEA model at block 314 further includes comparing the predicted IOL behavior of the EFEA model to the predicted IOL behavior of a FEA model using the same inputs at block 316. As described above, in some embodiments, the predicted IOL behavior by FEA model 125 and EFEA model 126 is in IOL power per frame. Thus, in some embodiments, EFEA model 126 may be trained to reproduce the FEA-predicted IOL power curve per frame. Training the EFEA model to match the EFEA model at block 314 further includes adjusting the EFEA model (e.g., adjusting the model weights) based on the comparison at block 318.

[0063] The TDG 122 can generate one training data set (or multiple training data sets) used to train the EFEA model 126. The training data set for the EFEA model 126 can include mapping demographic information and / or data points associated with measurements of anatomical parameters for each of several past patients to corresponding predictions of IOL behavior by the FEA model 125. The model trainer 124 trains the EFEA model 126 based on the training data set. For example, the model trainer 124 trains the EFEA model 126 to generate predicted IOL behavior (e.g., IOL power per frame) for given inputs including anatomical parameters and IOL parameters. The EFEA model 126 then provides the predicted IOL behavior based at least in part on the inputs. The repository 140 can include records of predictions by the FEA model 125 of IOL behavior for the same set of inputs. The predicted IOL behavior provided by EFEA model 126 can be compared to a record of the output provided by FEA model 125 for the same set of inputs to minimize the discrepancy between the predictions provided by EFEA model 126 and the output provided by FEA model 125. Once EFEA model 126 has been trained to the point where it returns predictions that are the same or nearly the same (e.g., within an acceptable margin of error) as the outputs provided by FEA model 125 for the same set of inputs, EFEA model 126 may be able to extrapolate and provide predicted IOL behavior even for inputs that were not run through FEA model 125. As a result, training EFEA model 126 according to embodiments described herein addresses technical deficiencies in using FEA models to predict IOL behavior within the lens capsule by significantly reducing the amount of resources (e.g., computing resources) that would have to be utilized to predict IOL behavior over a wide range of inputs using a FEA model.

[0064] As described above, FEA model 125 can be computationally expensive and require an expert to create, configure, and run it. However, EFEA model 126 can return predictions in a fraction of a second with minimal software and hardware requirements. Thus, according to an embodiment of the present disclosure, EFEA model 126, rather than FEA model 125, is used as part of IOL recommendation generator 134 to generate recommended IOL parameters. Thus, after training EFEA model 126, EFEA model 126 may be deployed to one or more server computers, user consoles, imaging devices integrated with computing devices, etc. For example, as shown in FIG. 1 , EFEA model 126 may be deployed to user console 130.

[0065] An example of how to train a fusion model According to embodiments of the present disclosure, an EFEA (e.g., EFEA model 126) can be used in addition to an IPC ML base model (e.g., IPC ML model 127) to train a fusion model to generate predicted IOL behavior and / or recommended IOL parameters.

[0066] FIG. 4 illustrates the use of the EFEA model 126 and the IPC ML model 127 to train the fusion model 128, according to certain embodiments described herein. The training of the EFEA model 126 was discussed above with respect to FIG. 3. To train the fusion model 128, the EFEA model 126 can be executed, and the output of the EFEA model 126 can be used as input to the fusion model 128. The EFEA model 126 uses input 402 to generate predicted IOL behavior. The input 402 to the EFEA model 126 includes patient anatomical parameters 404 and IOL parameters 406. As described above, the patient anatomical parameters 404 that may be input to the EFEA model 126 may include all or a subset of the anatomical parameters used to generate the FEA model 125. The IOL parameters 406 may include IOL type, IOL size, and / or IOL label power. Based on the input, the EFEA model 126 outputs predicted IOL behavior. As described above, the predicted IOL behavior may include predicted postoperative refractive error, predicted postoperative effective lens position, and / or predicted IOL power frames (which may include predicted preoperative, intraoperative, and postoperative IOL powers). In an exemplary embodiment, the patient's anatomical parameters 604 include lens characteristic dimensions, the IOL parameters 406 include at least IOL label powers for one or more IOLs, and the predicted IOL behavior includes at least a predicted IOL power per frame for each of the one or more IOLs.

[0067] The output from the IPC, in addition to the output from the EFEA model 126, is also used as input to the fusion model 128. As mentioned above, the IPC can be a standard IPC (e.g., Bartlett IPC, Haigis IPC, Hill-RBF IPC, SRK IPC, etc.). In some embodiments, an IPC ML model 127 is used.

[0068] TDG 122 may generate one training dataset (or multiple training datasets) used to train IPC ML model 127. The training dataset for IPC ML model 127 may include mapping demographic information and / or data points associated with measurements of anatomical parameters for each of several past patients to corresponding clinical outcomes for post-operative outcomes. In some embodiments, the training dataset for IPC ML model 127 involves the same pre-operative anatomical parameters used for EFEA model 126. In some embodiments, the training dataset for IPC ML model 127 uses a subset of the anatomical parameters as EFEA model 126. That is, in some embodiments, FEA model 125 and EFEA model 126 may use all the anatomical parameters of IPC ML model 127 and additional modeled anatomical parameters.

[0069] The model trainer 124 trains the IPC ML model 127 based on the training dataset. For example, the model trainer 124 trains the IPC ML model 127 to generate predicted post-operative outcomes for given inputs 408 including patient anatomical parameters 410 and IOL parameters 412. In some embodiments, the patient anatomical parameters 410 of the inputs 408 to the IPC ML model 127 are the same patient anatomical parameters 404. In some embodiments, the patient anatomical parameters 410 of the inputs 408 to the IPC ML model 127 are a subset of the patient anatomical parameters 404. The IPC ML model 127 outputs predicted post-operative outcomes based at least in part on the inputs. In some embodiments, the predicted post-operative outcomes are predicted post-operative refractive error, predicted post-operative effective lens position, and / or predicted IOL power. In some embodiments, the IPC ML model 127 outputs predicted IOL / frame / IOL based at least in part on the inputs.

[0070] To train and refine the IPC ML model 127, the predicted postoperative outcomes by the IPC ML model 127 can be compared with historical clinical data records in the treatment outcome repository 140 for the same input set to determine whether the predictions by the IPC ML model 127 provided satisfactory results. Based on the comparison, the model trainer 124 can tune the IPC ML model 127 to refine the model. The IPC ML model 127 can be trained over a large set of anatomical values ​​and IOL parameters.

[0071] Once trained, the IPC ML model 127 may be deployed to one or more server computers, user consoles, imaging devices integrated with computing devices, etc. For example, as shown in FIG. 1, the IPC ML model 127 may be deployed to a user console 130.

[0072] Fusion model 128 may be another ML-based model trained with inputs from both EFEA model 126 and IPC ML model 127 to generate recommended IOL parameters. In some embodiments, fusion model 128 may be trained to weight inputs from EFEA model 126 differently than inputs from IPC ML model 127. In some embodiments, fusion model 128 may weight inputs from EFEA model 126 and / or IPC ML model 127 differently for different data points.

[0073] The output 414 from the fusion model 128 includes at least one or more recommended IOL parameters 416. As discussed herein, the one or more recommended IOL parameters 416 may include a recommended IOL type, IOL power, and / or IOL size.

[0074] In some other embodiments, the output 414 from the fusion model 128 may include one or more predicted post-operative outcomes, such as predicted post-operative refractive error for each of the one or more IOLs.

[0075] Prior to deployment, the fusion model 128 may undergo a training period, which involves the model trainer 124 adjusting the weights of the fusion model 128 to reduce the calculated loss between outputs 414, which may represent predicted postoperative refractive error. The fusion model 128 may be trained to recommend IOL parameters that result in the lowest predicted postoperative refractive error. The treatment outcome data may indicate the actual postoperative refractive error measured on the patient postoperatively. The treatment outcome data, which may be provided as part of each historical patient record, may generally indicate the actual IOL used on the patient, postoperative measurements (e.g., the actual postoperative refractive error measured on the patient postoperatively), and / or a postoperative client satisfaction score.

[0076] Once trained, the fusion model 128 may be deployed to one or more server computers, user consoles, imaging devices integrated with computing devices, etc. For example, as shown in FIG. 1 , the fusion model 128 may be deployed to a user console 130. After deployment, the fusion model 128 may continue to be trained in the same manner using treatment outcome data recorded for current / new patients by the TDR 132 and added to the repository 140.

[0077] An example of how to perform cataract eye surgery based on recommendations generated using the fusion model 5 illustrates example operations 500 that may be performed by a computing device to generate and output recommendations including IOL parameters for cataract surgery for a patient based on the patient's anatomical parameters, according to certain aspects described herein. Operations 500 may be performed by a clinical device, such as user console 130.

[0078] The operations 500 may begin at block 502 by using an EFEA model (e.g., EFEA model 126) to generate a first predicted lens behavior (e.g., IOL power per frame) for each of a set of IOLs given first preoperative anatomical parameters of a patient's eye (e.g., patient anatomical parameters 404) and one or more parameters of each of the one or more IOLs (e.g., IOL parameters 406). In an exemplary embodiment, for a current patient in the clinic, a clinician enters the patient's preoperative anatomical parameters, including the patient's lens characteristic dimensions, into the user console 130. In some embodiments, the user console 130 itself measures the patient's anatomical parameters. In some embodiments, the imaging device 110 measures the patient's anatomical parameters, and the imaging device 110 provides the measured anatomical parameters to the user console 130 or displays the measured anatomical parameters to the clinician, who manually enters the anatomical parameters into the user console 130. In some embodiments, the patient's anatomical parameters have been previously measured and stored, and the user console 130 retrieves the stored anatomical parameters, or the clinician retrieves the stored anatomical parameters and manually enters the anatomical parameters into the user console 130. In some embodiments, the one or more IOLs comprise a set of accommodative IOLs, such as a set of IOLs in stock or available to the clinician. The set of IOLs may be IOLs covering a range of IOL powers with specific IOL power step sizes.

[0079] The EFEA model outputs a first predicted lens behavior for each of the one or more IOLs. In this exemplary embodiment, based on the patient's input pre-operative anatomical parameters, the EFEA model 126 outputs a plurality of frames of predicted IOL power for each of the input IOLs and pre-operative anatomical parameters. The plurality of frames may include at least frames from the start of insertion of the IOL into the patient's eye until the IOL reaches a rest state.

[0080] The operations 500 may continue at block 504 by using an IPC ML model (e.g., IPC ML model 127) to generate a second predicted lens behavior (e.g., post-operative refractive outcome) for each of the set of IOLs given second one or more pre-operative anatomical parameters of the patient's eye (e.g., patient anatomical parameters 410) and one or more parameters of each of the one or more IOLs (e.g., IOL parameters 412). In some embodiments, the anatomical parameters input to the IPC ML model 127 are the same as or a subset of the anatomical parameters input to the EFEA model 126 at block 502. The IPC ML model outputs a second predicted post-operative outcome for each of the one or more IOLs. In some embodiments, based on the input preoperative anatomical parameters and IOL parameters, the IPC ML model 127 outputs a predicted postoperative refractive error at each of the one or more IOLs, a predicted IOL power at each of the one or more IOLs, a predicted IOL power per frame at each of the one or more IOLs, or a combination thereof.

[0081] The operations 500 may continue at block 506 by using a fusion model (e.g., fusion model 128) to generate one or more recommended IOL parameters (e.g., recommended IOL parameters 416) given the first and second predicted lens behaviors for each of the one or more IOLs. In some embodiments, the fusion model 128 outputs a recommended IOL type, power, and / or size based on inputs from the EFEA model 126 and the IPC ML model 127.

[0082] An example system that performs cataract surgery based on recommendations generated using a fusion model. 6 illustrates an example computing device 600 that uses an EFEA model (e.g., EFEA model 126, etc.) to assist in the performance of an ophthalmic procedure, such as cataract surgery, according to certain aspects described herein. For example, computing device 600 may be user console 130 shown in FIG. 1.

[0083] As shown, computing device 600 includes a user interface 602 , one or more input / output (I / O) interfaces 604 , a network interface 606 , and a control module 608 .

[0084] The user interface 602 may be a graphical user interface (GUI) through which a user interacts with the computing device 600. The I / O interface 604 may allow various I / O devices (e.g., a keyboard, a display, a mouse device, pen input, etc.) to be connected to the computing device 600. The network interface 606 may connect the computing device 600 to a network (which may be a local network, an intranet, the Internet, or any other group of computing devices communicatively connected to one another).

[0085] Control module 608 includes memory 610, a central processing unit (CPU) 622, and storage 624. CPU 622 may retrieve and execute programming instructions stored in memory 610. CPU 622 may represent a single CPU, multiple CPUs, a single CPU with multiple processing cores, etc.

[0086] Memory 610 represents volatile memory such as random access memory, and / or non-volatile memory such as non-volatile random access memory, phase change random access memory, etc. Memory 610 may include input parameters 612, an EFEA model 614, an ML IPC model 616, a fusion model 618, and treatment data 620.

[0087] The computing device 600 may receive input from the user interface 602 and / or the I / O device interface 604 and store input parameters 612 in the memory 610. The CPU 622 may retrieve the input parameters 612 and run the EFEA model 614 and the ML IPC model 616 using the input parameters 612 to generate first and second predicted post-operative outcomes. The CPU 622 may run the fusion model 618 using the first and second predicted post-operative outcomes to generate recommended IOL parameters for use in, or during, the patient's cataract surgery. The computing device 600 may receive treatment data from the user interface 602 and / or the I / O device interface 604 and store treatment data 620 in the memory 610.

[0088] Additional considerations The above description is provided to enable those skilled in the art to practice various embodiments described herein. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments. For example, changes may be made to the function and arrangement of elements discussed without departing from the scope of the disclosure. In various examples, various actions or elements may be omitted, substituted, or added as appropriate. Features described with respect to some examples may be combined in several other examples. For example, an apparatus may be implemented or a method may be practiced using any number of aspects described herein. Furthermore, the scope of the present disclosure is intended to cover similar apparatuses or methods that are implemented using structure, functionality, or structure and functionality in addition to or other than various aspects of the disclosure described herein. It should be understood that any aspect of the disclosure disclosed herein may be implemented by one or more elements recited in the claims.

[0089] As used herein, a phrase referring to "at least one of" a list of items refers to any combination of those items, including single members. As an example, "at least one of a, b, or c" is intended to cover a, b, c, ab, ac, bc, and abc, as well as any combination with multiples of the same element (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or a, b, and c in any other order).

[0090] As used herein, the term "determining" encompasses a variety of acts. For example, "determining" may include calculating, computing, processing, deriving, investigating, querying (e.g., querying a table, database, or other data structure), ascertaining, etc. "Determining" may also include receiving (e.g., receiving information), accessing (e.g., accessing data in a memory), etc. "Determining" may also include resolving, selecting, choosing, establishing, etc.

[0091] The methods disclosed herein include one or more steps or actions for achieving the method. Method steps and / or actions may be interchangeable with one another without departing from the claims. In other words, unless a specific order of steps or actions is specified, the order and / or use of specific steps and / or actions may be modified without departing from the claims. Furthermore, various actions of the methods described above may be performed by any suitable means capable of performing the corresponding functions. These means may include various hardware and / or software elements and / or modules, including, but not limited to, circuits, application specific integrated circuits (ASICs), or processors. In general, where there are actions illustrated in figures, those actions may include corresponding means-plus-function elements that are similarly numbered.

[0092] The various illustrative logic blocks, modules, and circuits described in connection with this disclosure may be implemented or performed by a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device (PLD), discrete gate or transistor logic, discrete hardware elements, or any combination thereof designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but alternatively, the processor may be any commercially available processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, such as a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration.

[0093] The processing system may be implemented with a bus architecture. The bus may include any number of interconnected buses and bridges, depending on the particular application and overall design constraints of the processing system. The bus may interconnect various circuits, including, among other things, a processor, machine-readable media, and input / output devices. A user interface (e.g., keypad, display, mouse, joystick, etc.) may also be connected to the bus. The bus may also connect various other circuits, such as timing sources, peripherals, voltage regulators, power management circuits, etc., which are known in the art and will not be described further. The processor may be implemented with one or more general-purpose and / or special-purpose processors. Examples include microprocessors, microcontrollers, DSP processors, and other circuits capable of executing software. Those skilled in the art will recognize how to best implement the described functionality of a processing system, depending on the particular application and the overall design constraints imposed on the overall system.

[0094] If implemented in software, the functions described above may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Software should be construed broadly to mean instructions, data, or any combination thereof, whether referred to as software, firmware, middleware, microcode, hardware description language, or otherwise. Computer-readable media includes both computer storage media and communication media, such as any medium that facilitates transfer of a computer program from one place to another. A processor may be responsible for managing the bus and general processing, including the execution of software modules stored on the computer-readable storage medium. The computer-readable storage medium may be coupled to the processor such that the processor can read information from and write information to the storage medium. Alternatively, the storage medium may be integral to the processor. By way of example, computer-readable media may include a transmission line, a carrier wave modulated with data, and / or a computer-readable storage medium on which instructions are stored separately from a wireless node, all of which may be accessed by the processor via a bus interface. Alternatively or additionally, the computer-readable medium, or any portion thereof, may be integrated into the processor, such as in the case of a cache and / or general-purpose register file. Examples of machine-readable storage media include, for example, RAM (random access memory), flash memory, ROM (read-only memory), PROM (programmable read-only memory), EPROM (erasable programmable read-only memory), EEPROM (electrically erasable programmable read-only memory), registers, magnetic disks, optical disks, hard drives, or any other suitable storage medium or any combination thereof. The machine-readable medium may be embodied in a computer program product.

[0095] A software module may include a single instruction or many instructions and may be distributed across several different code sections, across different programs, and across multiple storage media. A computer-readable medium may include many software modules. A software module contains instructions that, when executed by a device such as a processor, cause a processing system to perform various functions. A software module may include a transmitting module and a receiving module. Each software module may reside on a single storage device or may be distributed across multiple storage devices. For example, a software module may be loaded from a hard drive into RAM when a trigger event occurs. During execution of a software module, a processor may load some of the instructions into a cache to speed access. One or more cache lines may then be loaded into a general-purpose register file for execution by the processor. When referring to the functionality of a software module, it is understood that such functionality is implemented by the processor when executing instructions from that software module.

[0096] The following claims are not limited to the embodiments set forth herein but are to be accorded the full scope consistent with the language of the claims. In the claims, when an element is referred to in the singular, it does not mean "one and only one" unless specifically stated otherwise, but rather "one or more." The term "some" refers to one or more unless specifically stated otherwise. No element of a claim is to be construed under the provisions of 35 U.S.C. § 112(f) unless the element is expressly recited using the phrase "means for," or, in the case of a method claim, unless the element is recited using the phrase "step of." All structural and functional equivalents of the elements of various aspects described throughout this disclosure that are known or later become known to those skilled in the art are expressly incorporated herein by reference and are intended to be encompassed by the claims. Furthermore, nothing disclosed herein is intended to be made available to the public, regardless of whether such disclosure is expressly recited in the claims.

Claims

1. 1. A method for determining one or more intraocular lens (IOL) parameters for an IOL to be used in a cataract surgical procedure, comprising: generating a plurality of data points relating to measurements of a plurality of anatomical parameters of the eye being treated using one or more ophthalmic imaging devices; generating a first predicted lens behavior based at least in part on the plurality of data points associated with the measurements of the plurality of anatomical parameters and one or more IOL parameters for each of the one or more IOLs using a machine learning model trained to emulate a finite element analysis (FEA) model; generating a second predicted lens behavior based at least in part on at least a subset of the plurality of data points associated with the measurements of the plurality of anatomical parameters of the eye being treated and the one or more IOL parameters of each of one or more IOLs using an IOL power calculator machine learning model; generating, using a fused machine learning model, a recommendation including one or more IOL parameters for the IOL to be used in the cataract surgery based at least in part on the first and second predicted lens behaviors; and A method comprising:

2. The method of claim 1 , wherein the plurality of anatomical parameters comprises one or more lens characteristic dimensions of the eye.

3. The method of claim 1 , wherein the one or more IOL parameters include at least one of an IOL type, an IOL size, or an IOL power.

4. generating an FEA model using a finite element method (FEM) based on a set of data points related to anatomical parameters of at least one previous patient and at least one set of one or more IOL parameters; using the FEA model to generate a predicted lens behavior based at least in part on the set of data points related to the anatomical parameters of the at least one past patient and the at least one set of one or more IOL parameters; adjusting the FEA model based on a comparison of the predicted lens behavior and observed lens behavior for an IOL having the at least one set of IOL parameters implanted in the eye of the previous patient at the set of data points associated with the anatomical parameters; The method of claim 1 further comprising:

5. using the machine learning model to generate a predicted lens behavior based at least in part on a set of data points related to anatomical parameters of at least one past patient and at least one set of one or more IOL parameters; adjusting weights associated with the machine learning model based on a comparison of the predicted lens behavior output by the machine learning model with the predicted lens behavior output by the FEA model for the set of data points related to the anatomical parameter and the at least one set of the one or more IOL parameters; The method of claim 1 , further comprising training the machine learning model to emulate the FEA model by:

6. The method of claim 5 , wherein the predicted lens behavior includes IOL power per frame.

7. generating a predicted lens behavior based at least in part on a set of data points related to anatomical parameters of at least one past patient and at least one set of one or more IOL parameters using the IOL power calculator machine learning model; adjusting the IOL power calculator machine learning model based on a comparison of the predicted lens behavior and the observed lens behavior for an IOL using the at least one set of one or more IOL parameters implanted in the eye of the previous patient using the set of data points; 10. The method of claim 1, further comprising training the IOL power calculator machine learning model by:

8. The method of claim 7 , wherein the predicted lens behavior comprises a predicted post-operative refractive outcome.

9. using the fusion model to generate one or more recommended IOL parameters for past patients based at least in part on a third predicted lens behavior by the machine learning model trained to emulate the FEA model and a fourth predicted lens behavior by the IOL power calculator machine learning model; adjusting the fusion model based on a comparison of the one or more recommended IOL parameters with treatment outcome data of IOLs having the one or more recommended IOL parameters and implanted in the patient's eye in the past; The method of claim 1 , further comprising training the fusion model by:

10. 1. A system for determining one or more intraocular lens (IOL) parameters for an IOL to be used in a cataract surgical procedure, comprising: one or more ophthalmic imaging devices configured to generate a plurality of data points relating to measurements of a plurality of anatomical parameters of the eye being treated; a memory that stores a machine learning model trained to emulate a finite element analysis (FEA) model, an IOL power calculator machine learning model, and a fusion machine learning model; at least one processor coupled to the memory; wherein the at least one processor using the machine learning model trained to emulate the FEA model to generate a first predicted lens behavior based at least in part on the plurality of data points associated with the measurements of the plurality of anatomical parameters and one or more IOL parameters for each of one or more IOLs; generating a second predicted lens behavior based at least in part on at least a subset of the plurality of data points associated with the measurements of the plurality of anatomical parameters of the eye being treated and the one or more IOL parameters of each of one or more IOLs using the IOL power calculator machine learning model; using the fused machine learning model to generate a recommendation, including one or more IOL parameters, for the IOL to be used in the cataract surgery based at least in part on the first and second predicted lens behaviors; and A system configured to:

11. The system of claim 10 , wherein the plurality of anatomical parameters includes one or more lens characteristic dimensions of the eye.

12. The system of claim 11 , wherein the one or more IOL parameters include at least one of an IOL type, an IOL size, or an IOL power.

13. The at least one processor obtaining a FEA model generated using a finite element method (FEM) based on a set of data points associated with anatomical parameters of at least one previous patient and at least one set of one or more IOL parameters; using the FEA model to generate a predicted lens behavior based at least in part on the set of data points related to the anatomical parameters of the at least one past patient and the at least one set of one or more IOL parameters; adjusting the FEA model based on a comparison of the predicted lens behavior and observed lens behavior for an IOL having the at least one set of IOL parameters implanted in the eye of the previous patient at the set of data points associated with the anatomical parameters; The system of claim 11 , further configured to:

14. The at least one processor using the machine learning model to generate a predicted lens behavior based at least in part on a set of data points related to anatomical parameters of at least one past patient and at least one set of one or more IOL parameters; adjusting weights associated with the machine learning model based on a comparison of the predicted lens behavior output by the machine learning model with the predicted lens behavior output by the FEA model for the set of data points related to the anatomical parameter and the at least one set of the one or more IOL parameters; The system of claim 11 , further configured to train the machine learning model to emulate the FEA model by:

15. The system of claim 14 , wherein the predicted lens behavior includes IOL power per frame.

16. The at least one processor generating a predicted lens behavior based at least in part on a set of data points related to anatomical parameters of at least one past patient and at least one set of one or more IOL parameters using the IOL power calculator machine learning model; adjusting the IOL power calculator machine learning model based on a comparison of the predicted lens behavior and the observed lens behavior for an IOL using the at least one set of one or more IOL parameters implanted in the eye of the previous patient using the set of data points; The system of claim 11 , further configured to train the IOL power calculator machine learning model by:

17. The system of claim 16 , wherein the predicted lens behavior comprises a predicted post-operative refractive outcome.

18. The at least one processor using the fusion model to generate one or more recommended IOL parameters for past patients based at least in part on a third predicted lens behavior by the machine learning model trained to emulate the FEA model and a fourth predicted lens behavior by the IOL power calculator machine learning model; adjusting the fusion model based on a comparison of the one or more recommended IOL parameters with treatment outcome data of IOLs having the one or more recommended IOL parameters and implanted in the patient's eye in the past; The system of claim 11 , further configured to train the fusion model by:

Images