Selecting intraocular lens power based on integrating finite element modeling with machine learning

Abstract

A system for selecting an intraocular lens (IOL) for implantation in an eye includes a controller adapted to selectively execute a finite element model (FEM) and a machine learning module. The controller is adapted to receive input data including one or more biometric parameters of the eye. Based on the input data, a plurality of capsule parameters are extracted via the machine learning module. The controller is adapted to determine an axial displacement factor, in part, based on the plurality of capsule parameters via the FEM. The axial displacement factor takes into account the predicted axial offset of the IOL after implantation in the eye. When calculating lens constant parameters using the FEM and machine learning module, the axial displacement factor can be incorporated into the final IOL power.

Description

Background Technology

[0001] This disclosure generally relates to the selection of intraocular lenses (IOLs) for implantation in the eye. More specifically, it relates to the selection of IOLs based on post-implantation axial offset using integrated finite element modeling and machine learning modules. The human lens is normally transparent, allowing light to pass through easily. However, many factors can cause areas within the lens to become cloudy and dense, negatively impacting visual quality. This can be corrected through cataract surgery—that is, the selection of an IOL for implantation in a patient's eye. Indeed, cataract surgery is a common procedure performed worldwide. A key driver of clinical outcomes in cataract surgery is the selection of an appropriate IOL power to achieve optimal refractive results. Currently, several calculators exist that use various types of preoperative information related to the patient's eye to select or determine the power of the lens to be implanted. While significant progress has been made in power calculation, challenges remain regarding the eye's location outside the range of average biometric parameters. Summary of the Invention

[0002] This document discloses a system for selecting the power of an intraocular lens (IOL) for implantation in an eye. The system includes a controller having one or more processors and a tangible, non-transitory memory on which instructions are recorded. The controller is adapted to selectively execute a finite element model and a machine learning module. The controller is adapted to receive input data, including one or more biometric parameters of the eye, which may be stored as tabular data and / or images. Multiple capsule parameters corresponding to the lens capsule of the eye are extracted from the input data via the machine learning module. The controller is adapted to determine an axial displacement factor, in part, based on these multiple capsule parameters via the finite element model. The axial displacement factor takes into account the predicted axial offset of the IOL after implantation in the eye. The controller is adapted to recommend the IOL power based on the axial displacement factor using one or more lens constant formulas.

[0003] In some embodiments, the finite element model is tensor-based. The controller may be adapted to execute the finite element model using multiphysics software. The controller may be adapted to select the intraocular lens based on a recommended intraocular lens power.

[0004] In some embodiments, multiple capsule parameters include capsule diameter and capsule thickness. These multiple capsule parameters may further include a capsule skew factor based on the capsule thickness. For example, the capsule skew factor may be determined as the ratio of the distance in the Y direction from the centroid of the capsule profile to the equatorial plane of the capsule profile in the Y direction to the capsule profile's thickness, the capsule profile being a cross-sectional profile of the lens capsule cut through the anterior and posterior poles of the lens capsule. The capsule skew factor can be a positive number between 0 and 0.4. The capsule skew factor is zero when the Y position in the equatorial plane (e.g., the Y coordinate) is exactly midway between the anterior and posterior poles. When the Y coordinate in the equatorial plane is not midway between the anterior and posterior poles, the capsule skew factor is greater than zero, indicating that the corresponding mass of the lens (as enclosed by the lens capsule) is relatively large on the posterior side of the equatorial plane.

[0005] This paper discloses a method for selecting an intraocular lens (IOL) for implantation in an eye using a system having a controller with at least one processor and at least one non-transitory tangible memory. The method includes selectively executing a finite element model and a machine learning module via the controller. The method includes receiving input data via the controller, the input data including one or more biometric parameters of the eye. The method includes extracting multiple capsule parameters corresponding to the lens capsule of the eye based on the input data through the execution of the machine learning module. The method includes determining an axial displacement factor, in part based on the multiple capsule parameters, through the execution of the finite element model. The axial displacement factor takes into account the predicted axial offset of the IOL after implantation in the eye. The method includes recommending the IOL power based on the axial displacement factor using one or more lens constant formulas.

[0006] This document discloses a system for selecting an intraocular lens (IOL) for implantation in an eye. The system includes a controller having one or more processors and a tangible, non-transitory memory on which instructions are recorded. The controller is configured to selectively execute a finite element model and a machine learning module via the execution of these instructions by the one or more processors. For example, the execution of these instructions by the one or more processors causes the controller to receive input data, including one or more biometric parameters of the eye, and to extract, based on the input data, multiple capsule parameters corresponding to the lens capsule of the eye via the machine learning module. The controller is adapted to determine an axial displacement factor, in part based on the multiple capsule parameters, via the finite element model. The controller is adapted to adjust the lens power of the IOL, in part based on the axial displacement factor. The axial displacement factor is a power correction feature that takes into account the predicted axial displacement of the IOL after implantation in the eye.

[0007] The above-described features and advantages, as well as other features and advantages, of this disclosure will become apparent from the following detailed description of the best mode for implementing this disclosure, taken in conjunction with the accompanying drawings. Attached Figure Description

[0008] Figure 1 This is a schematic diagram of a system for selecting an artificial lens to be implanted in the eye, the system having a controller;

[0009] Figure 2 This is a schematic cross-sectional view of an example eye;

[0010] Figure 3A and Figure 3B According to the first and second embodiments, it can be derived from Figure 1 A schematic flowchart of the corresponding methods executed by the controller;

[0011] Figure 4 This is a schematic diagram showing an artificial lens implanted in the lens capsule;

[0012] Figure 5 It is a schematic diagram showing a set of cyst profiles with the same thickness and diameter, which are aligned at their equatorial positions.

[0013] Figure 6 This is a schematic diagram showing another set of cyst outlines;

[0014] Figure 7 It is a schematic example graph showing the axial displacement of the intraocular lens on the vertical axis and the lens diameter on the horizontal axis;

[0015] Figure 8 This is a schematic example illustrating how the machine learning module can be used to replace the finite element analysis module; and

[0016] Figure 9 It is a schematic example curve showing the postoperative equivalent spherical power on the vertical axis and the lens power on the horizontal axis.

[0017] Representative embodiments of this disclosure are shown by way of non-limiting example in the accompanying drawings and are described in more detail below. However, it should be understood that the novel aspects of this disclosure are not limited to the specific forms shown in the above-enumerated drawings. Rather, this disclosure will cover modifications, equivalents, combinations, sub-combinations, substitutions, groupings, and alternatives falling within the scope of this disclosure, such as those covered by the appended claims. Detailed Implementation

[0018] Prior to cataract surgery, ophthalmologists utilize a variety of algorithms to plan intraocular lens (IOL) replacement for optimal visual acuity correction. Power calculation formulas typically assume that the effective lens position of the IOL 12 is approximately aligned with the preoperative capsular bag equator. However, the shape and size of the capsular bag vary within patient populations, making it challenging to achieve optimal refractive outcomes for non-average sizes (e.g., long eyes, short eyes, and / or myopic eyes). Furthermore, after implantation, the implanted IOL can move anteriorly or posteriorly along the axial direction based on the lens's haptic and mechanical characteristics, thus affecting refractive outcomes. One or more embodiments of this disclosure may utilize one or a combination of finite element modeling and machine learning to facilitate the selection of the IOL for implantation.

[0019] Referring now to the accompanying drawings to aid in further discussion of embodiments of this disclosure. The figures in the drawings are intended to aid explanation and are not necessarily intended to be completely accurate. For example, the figures are not necessarily drawn to scale. Furthermore, although the figures depict anatomical components, such components are intended to be representative only and are not necessarily anatomically accurate.

[0020] Referring to the accompanying drawings, similar reference numerals refer to similar parts. Figure 1 A system 10 for selecting an intraocular lens to be implanted in eye E is shown schematically. Figure 2 An example of eye E is shown in the diagram. It should be understood that the artificial lens 12 can take many different forms and includes multiple components and / or alternative components. Figure 1 In the illustrated embodiment, the intraocular lens 12 includes an optical region 14 connected to support structures 16 configured to support the positioning and retention of the intraocular lens 12.

[0021] refer to Figure 1 The controller C is adapted to selectively execute the machine learning module 20 and the finite element model 22. The controller C can be configured to communicate with various entities, such as one or more imaging devices 24, and a user interface 26. The imaging device 24 may include an optical coherence tomography device, a digital or analog microscope, a camera system (e.g., capturing one-dimensional or three-dimensional images or videos), an ultrasound machine, a magnetic resonance imaging machine, or other imaging devices available to those skilled in the art. Additionally, the controller C can communicate with a database 28 storing input data related to the eye. The controller C is adapted to receive input data, which includes one or more biometric parameters of the eye E. In some examples, the biometric parameters may be in tabular or photographic form. Based on the input data, the machine learning module 20 extracts information related to the lens capsule 32 of the eye E (also...). Figure 2 The text provides more detailed information about the multiple capsule parameters 30 related to the lens capsule (50).

[0022] The controller C is adapted to determine the axial displacement factor in part based on multiple capsule parameters 30 via a finite element model 22. The finite element model 22 can be configured to study the interaction between the intraocular lens 12 and lens capsules 50 (and / or lens capsules 32) of various sizes and shapes. The finite element model 22 can also be replaced by a machine learning module trained to replicate the output of the finite element model 22 given the same inputs. This may be useful, for example, when the finite element model 22 is too time-consuming or computationally expensive. The axial displacement factor takes into account the expected axial displacement of the intraocular lens 12 after implantation in the eye E.

[0023] System 10 provides a power calculator for the intraocular lens 12 and an adjustment for a recommended power of the intraocular lens 12, taking into account various capsular size and shape and / or other biometric parameters. As described below, the intraocular lens power, taking into account axial displacement, can be recommended in various ways. Reference Figure 1 System 10 includes a controller C having at least one processor P and at least one memory M (or a non-transitory tangible computer-readable storage medium) on which instructions for executing the first method 100 and / or the second method 150 are recorded. Methods 100 and 150 are respectively... Figure 3A and Figure 3B The figures shown are described below with reference to these figures.

[0024] The first method 100 applies to IOL power formulas that require a lens constant. In this method, axial displacement is taken into account by adjusting the lens constant formula used to recommend implant power. In the second method 150, an axial displacement factor is combined with a baseline power calculation. Here, the power of the intraocular lens 12 is selected in part based on the baseline IOL (intraocular lens) power calculation and the axial displacement factor. Additional machine learning modules can be used to combine the axial displacement factor and the baseline power calculation to recommend the final IOL power. Therefore, system 10 optimizes, improves, or enhances the selection of the intraocular lens 12 for a large patient population.

[0025] Now for reference Figure 3A This shows that it can be generated by Figure 1 The flowchart illustrates method 100 executed by controller C. Method 100 does not need to be applied in the specific order listed herein, and some boxes may be omitted. Memory M may store the controller-executable instruction set, and processor P may execute the controller-executable instruction set stored in memory M.

[0026] according to Figure 3AIn box 102, controller C is configured to receive input data, including one or more biometric parameters of eye E. Biometric parameters may include anterior chamber depth, ciliary process diameter, sulcus-to-sulcus diameter, corneal power expressed as a corneal curvature value, etc. Biometric parameters can be obtained from preoperative images of eye E. Alternatively, surgeons, technicians, or other healthcare professionals may manually input one or more biometric parameters of eye E.

[0027] according to Figure 3A In box 104, controller C is configured to extract multiple cyst parameters based on input data via machine learning module 20. Figure 2 An example cross-sectional image of an eye E is shown to help illustrate the capsule parameters. The eye E includes a lens capsule 50, a cornea 52, an iris 54, and a pupil 56, wherein the eye E is oriented such that the cornea 52 is located at the top of the illustration. The lens capsule 50 may be a thin elastic membrane surrounding the lens of the eye E, such that the lens is enclosed within the lens capsule 50. In this disclosure, references to dimensions corresponding to the lens capsule 50 may refer to the size of the space occupied by the lens capsule 50, and not necessarily to the size of the membrane of the lens capsule 50 itself. Therefore, references to dimensions corresponding to the lens capsule 50 may generally correspond to the size of the lens encapsulated by the lens capsule 50.

[0028] The lens capsule 50 has parameters such as capsule thickness 60 (e.g., the depth of the entire lens capsule 50 from its front end to its rear end), capsule diameter 62 (e.g., which may refer to the maximum width of the lens capsule 50 in a direction orthogonal to the capsule thickness 60), and / or capsule wall thickness. Capsule parameters may include capsule thickness 60, capsule diameter 62, capsule wall thickness, or other parameters. In some embodiments, multiple capsule parameters may include a capsule skew factor that captures the asymmetry of the lens capsule 50, as referenced using the equatorial plane 64 of the lens capsule 50. The equatorial plane 64 may be a plane that intersects the lens capsule 50 at a position corresponding to the capsule diameter 62, such that the equatorial plane 64 can be positioned at the widest portion of the lens capsule 50. In other words, the equatorial plane 64 may be a plane defined by a line connecting the intersection point of a curve defining the front and a curve defining the rear of the lens capsule 50 and perpendicular to the page extension. Figure 2 In the example shown, the cross-sectional image can be referred to using a horizontal X-axis, a vertical Y-axis, and a corresponding Z-axis orthogonal to the X and Y axes (e.g., passing through the page). Using this reference frame, the equatorial plane 64 can be correspondingly set as the XZ plane at a specific location along the Y-axis.

[0029] In these and other embodiments, and as follows regarding Figure 5 and / or Figure 6 As described in further detail, in some embodiments, the skewness can be indicated as a skew factor. As indicated above, the skew factor can be based on the position of the centroid of the lens capsule 50 (e.g., the geometric center of the lens capsule 50, which, assuming uniform mass density, could also be the centroid within the lens capsule 50) relative to the equatorial plane 64. For example, the skew factor can be determined using the position of the centroid (also referred to as the “centroid position”) along a vertical axis (e.g., the Y-axis shown) relative to the vertical axis (or “Y-axis”) of the equatorial plane 64. Using this determination technique, the skew factor can be a positive number between 0 and 0.4 and typically falls between 0 and 0.2.

[0030] Machine learning module 20 may include any suitable artificial intelligence model. For example, machine learning module 20 may include one or more neural networks or other machine learning algorithms trained to extract capsule parameters from biometric data included in the input data. For example, machine learning module 20 may be configured to process raw biometric imaging data, such as optical coherence tomography, ultrasound images, and / or other ocular imaging modalities.

[0031] To extract multiple cyst parameters, the machine learning module 20 can first preprocess the input biometric data, for example, by applying image enhancement techniques, segmentation algorithms, or feature detection methods. The preprocessed data can then be fed into one or more trained neural networks. These neural networks can be convolutional neural networks or other architectures suitable for image analysis and feature extraction.

[0032] A neural network can be trained on a large dataset of labeled biometric images to identify and measure key anatomical features of the lens capsule 50. For example, the neural network can be trained to identify the anterior and posterior poles of the lens capsule 50, measure the capsule diameter 62 and / or capsule thickness 60, and / or determine the location of the centroid.

[0033] Based on the extracted features, the machine learning module 20 can calculate cyst parameters, such as the cyst skew factor. In these and other embodiments, the machine learning module 20 can output a set of quantitative cyst parameters, including cyst diameter, cyst thickness, cyst wall thickness, cyst skew factor, and other relevant parameters.

[0034] according to Figure 3AIn frame 106, controller C is adapted to determine an axial displacement factor, in part, based on multiple capsule parameters via finite element model 22. The axial displacement factor can take into account the predicted axial offset of the intraocular lens 12 after implantation in the eye E. In these and other embodiments, the axial displacement factor can represent how much the intraocular lens 12 is expected to move along the optical axis of the eye E after implantation. This predicted movement may be influenced by the capsule parameters. By incorporating this axial displacement prediction into the lens power calculation, the system can be able to recommend an intraocular lens power that takes into account the expected postoperative position of the lens. This can achieve more accurate refractive results by adjusting the lens power to compensate for the predicted axial offset after implantation.

[0035] In some embodiments, due to the computational complexity of the finite element model, a machine learning module (e.g., machine learning module 20) can be trained to simulate the finite element model 22, allowing the machine learning module to act as a proxy for the finite element model 22. Therefore, in this disclosure, references to "finite element model" may also refer to a machine learning module that has been configured (e.g., trained) to simulate the finite element model.

[0036] In some embodiments, the controller C is adapted to predict the effective lens position of the intraocular lens 12 by incorporating capsule parameters, which may or may not be extracted from biometric parameters (e.g., via machine learning module 20). Alternatively or concurrently, the finite element model 22 may be configured to predict the effective lens position of the intraocular lens 12 relative to the anterior or posterior surface of the cornea, taking into account the settlement of the IOL loop at the equatorial position of the capsule and subsequent axial displacement, both of which may be influenced by the capsule parameters. Multiple capsule parameters are fed into the finite element model 22 to estimate the axial displacement of the intraocular lens 12.

[0037] For example, typically, finite element model 22 uses a numerical technique called the finite element method to simulate a physical system. In some embodiments, the finite element model is tensor-based. As understood by those skilled in the art, tensor-based models employ tensor mathematics to represent the properties and behavior of a physical system. Tensors are mathematical objects that summarize scalars, vectors, and matrices. For example, tensor variables such as stress and strain can be represented at integration points within each element in the finite element analysis, thereby capturing direction dependence. Controller C may employ multiphysics software to execute finite element model 22. As understood by those skilled in the art, multiphysics simulation software refers to software capable of simulating models from different physical domains (e.g., interactive physics models), such as COMSOL multiphysics software. In some embodiments, due to the computational complexity of the finite element model, machine learning module 20 may be trained to simulate finite element model 22.

[0038] As a more specific example of this disclosure, the finite element model 22 may include a computational model that divides the lens capsule 50 and the intraocular lens 12 into small discrete elements. These elements may be interconnected to form a mesh representation of the capsule-IOL system. The finite element model 22 may incorporate the material properties of the lens capsule 50 and the intraocular lens 12, as well as boundary conditions and loading scenarios, to simulate the post-implantation behavior of the intraocular lens 12.

[0039] To determine the axial displacement factor, the finite element model 22 can use extracted capsule parameters as input. These parameters may include capsule diameter, thickness, wall thickness, and skew factor. The finite element model 22 can then simulate the interaction between the intraocular lens 12 and the lens capsule 50 under physiological conditions indicated by the capsule parameters. This simulation can take into account factors such as capsule elasticity, IOL loop compression, and gravitational effects.

[0040] The finite element analysis performed by finite element model 22 can solve for the deformation and stress distribution in the capsule-IOL system. Based on this solution, the axial displacement of the intraocular lens 12 relative to its initial position can be calculated. This displacement value can be used as an axial displacement factor for adjusting IOL diopter calculations. As mentioned above, in some embodiments, the displacement value can be relative to the position of the equatorial plane 64 of the lens capsule 50.

[0041] Figure 7 This is a schematic example graph showing a set of 400 lens capsule traces obtained via finite element model 22. The vertical axis 404 can correspond to... Figure 2 The Y-axis indicates axial displacement, and the horizontal axis 402 can correspond to... Figure 2 The X-axis indicates the cyst diameter. This group 400 includes traces 410, 412, 414, 416, 418, and 420. Traces 410, 412, 414, 416, 418, and 420 correspond to the following cyst thicknesses (in mm) and cyst skew factors: [3.8 mm, 0.11], [3.8 mm, 0.14], [4.8 mm, 0.11], [4.3 mm, 0.07], [4.3 mm, 0.11], and [4.8 mm, 0.14]. Figure 7 The axial posterior displacement of the intraocular lens (relative to the equatorial plane) is shown to vary with the size and shape of the capsule.

[0042] In some embodiments, finite element model 22 is adapted to create a capsule finite element seed database using experimental data, which covers the human physiological range of lens diameter, lens thickness, and effective lens position of the target intraocular lens. The experimental data may be stored in database 28. The interaction and movement of the intraocular lens 12 after implantation vary with the shape and size of the capsule, thus affecting the effective lens position of the implanted intraocular lens 12. Finite element model 22 can be extended to cover every combination of capsule sizes observed clinically. For example, database 28 may be updated over time with clinical observations or other data, such that finite element model 22 may include additional traces or observations. In these and other embodiments, the experimental data can then be used to train machine learning algorithms to rapidly predict displacements of new patient-specific capsule parameters without rerunning the full finite element simulation each time.

[0043] according to Figure 3A In box 108, method 100 includes recommending implant power using one or more lens constant formulas based on axial displacement factors (according to box 106) learned from finite element model 22 and / or machine learning module 20 for various eye types or capsular structures. In other words, the axial displacement factor can be used to adjust the lens constants input to the power calculator. Any suitable lens constant formula available to those skilled in the art can be used.

[0044] Now for reference Figure 3B This shows that it can be generated by Figure 1 The flowchart shows the execution of method 150 by controller C. Method 150 does not need to be applied in the specific order listed herein, and some boxes may be omitted. Figure 3B In box 152, controller C is configured to acquire input data, which includes one or more biometric parameters of eye E. Biometric parameters may include anterior chamber depth, ciliary process diameter, sulcus-to-sulcus diameter, corneal power expressed as corneal curvature values, etc. These biometric parameters can be obtained from preoperative images of eye E.

[0045] according to Figure 3BIn box 154, controller C is adapted to perform a basic IOL (intraocular lens) power calculation to obtain the basic optical power of intraocular lens 12. The IOL power calculator includes, but is not limited to, the Barrett general formula, the Kane formula, etc. Controller C can input biometric parameters (from box 102) into optical design software using an optical eye model available to those skilled in the art. Controller C can employ ray tracing techniques that track the propagation of light through the optical eye model to predict the IOL power based on one or more lens constants. The basic IOL power is selected as the power that minimizes or reduces spherical aberration and other optical aberrations by focusing light directly onto the retina of the eye E.

[0046] according to Figure 3B In box 156, controller C is configured to extract multiple cyst parameters (e.g., as described above) based on input data via machine learning module 20. The multiple cyst parameters include... Figure 2 The capsule thickness 60 and capsule diameter 62 are shown. Several capsule parameters include a capsule skew factor, which captures the asymmetry within the lens capsule relative to the equatorial plane 64.

[0047] according to Figure 3B In block 158, method 150 includes determining an axial displacement factor of the intraocular lens 12. In some embodiments, determining the axial displacement factor may include determining the expected axial displacement of the intraocular lens 12 by performing a finite element model 22, based in part on a plurality of capsule parameters, as described above.

[0048] In these and other embodiments, method 150 may include an adjustment factor for determining (e.g., calculating) the focal power of the intraocular lens 12 based on an axial displacement factor. Alternatively or additionally, the adjustment factor may be determined based on other factors that may be contained within or based on biometrics and / or size of the capsule, which may be determined based on knowledge acquired in capsule parameters. The adjustment factor may indicate the amount of adjustment that can be made to the focal power of the lens.

[0049] Figure 4 This is a schematic diagram showing an artificial lens 212 implanted in the lens capsule 210 based on a finite element model 22. (See diagram 22 for details.) Figure 4 As shown, the intraocular lens 212 is not aligned with or is asymmetrical relative to the equatorial plane 220 of the lens capsule 210. Note that because the intraocular lens 212 is somewhat oblique to the equatorial plane, the line referencing the equatorial plane 220 is not necessarily a true representation of the equatorial portion of the lens capsule 210. Figure 4In the illustrated embodiment, the capsule deflection factor is biased towards the posterior pole (above zero). Or in other words, because most of the intraocular lens 212 is positioned towards the anterior region of the capsule 210, the capsule deflection factor is biased towards the posterior pole. In the example shown, in Figure 4 In the middle, the rear direction is downward and the front direction is upward. Finite element model 22 can be replaced by a machine learning module to speed up computation time and reduce computational requirements, as described above.

[0050] Return to Figure 3B ,according to Figure 3B In block 160, method 150 includes ultimately determining the lens power of the intraocular lens 12 based on an axial displacement factor and / or a lens power adjustment factor. In these and other embodiments, ultimately determining the lens power may include adjusting the base IOL power based on the axial displacement factor. Alternatively or additionally, the intraocular lens 12 may be selected in part based on the adjusted base IOL power. In some embodiments, additional machine learning models may be used to combine the axial displacement factor and the base IOL power.

[0051] Figure 5 A set of capsule contours 300 aligned at an equatorial plane 310 is shown. In other words, the equatorial planes of the different capsule contours 300 are aligned such that the equatorial plane 310 represents the equatorial plane of all capsule contours. Capsule contours 312, 314, 316, and 318 of lenses with the same capsule thickness (along the vertical axis (or also called the Y-axis) 304) and the same capsule diameter (along the horizontal axis (or also called the X-axis) 302) but with four different capsule skew factor values ​​are shown. Capsule contours 312, 314, 316, and 318 define corresponding centroids 322, 324, 326, and 328, respectively. As indicated elsewhere in this disclosure, a centroid can generally be understood as the center of mass of a geometric object with uniform density. Each of the corresponding centroids 322, 324, 326, and 328 defines a corresponding X-coordinate 330 and a corresponding Y-coordinate 332 (e.g., the Y-coordinate 334 of centroid 328). The lens capsule defines corresponding capsule contours (e.g., capsule contours 312, 314, 316, 318), which are cross-sectional contours of different lens capsules cut through corresponding anterior and posterior poles. For example, Figure 5The figures show an example relative position of the anterior pole 342 and an example relative position of the posterior pole 344 of the capsule contour 312. Alternatively or alternatively, the equatorial plane 310 may be located midway between the anterior pole 342 and the posterior pole 344 of the capsule contour 312, such that the equatorial plane 310 of the capsule contour 312 may be located at the center of the thickness of the capsule contour 312. Note that other equatorial planes 310 of other capsule contours (e.g., capsule contours 314, 316, and / or 318) may not be located midway between their respective anterior and posterior poles, such that their respective equatorial planes 310 are not located at the center of the capsule thickness of such a lens capsule. It should be understood that the figures are not drawn to scale.

[0052] In some embodiments, the cyst skew factor can be defined as the ratio of the distance in the Y direction from the equatorial plane to the centroid (e.g., centroid 322) of the corresponding cyst profile (e.g., cyst profile 312) to the cyst thickness 60 (e.g., the distance between the corresponding anterior and posterior poles of the corresponding cyst profile). Further, when the equatorial plane is positioned midway between the corresponding anterior and posterior poles, the centroid position can be aligned with the corresponding equatorial plane of their respective cyst profiles. However, when the equatorial plane is not positioned midway between the corresponding anterior and posterior poles, the centroid position can be misaligned with the corresponding equatorial plane of their respective cyst profiles.

[0053] When the Y-axis position of the corresponding centroid of a given cyst contour is exactly midway between the corresponding anterior and posterior poles, the cyst skew factor can be zero. For example, the cyst skew factor of cyst contour 312 can be zero because the equatorial plane 310 and Y-coordinate 332 of the corresponding centroid 322 of cyst contour 312 are exactly midway between the anterior pole 342 and the posterior pole 344. However, the cyst skew factors of other cyst contours 314, 316, and 318 can be non-zero because their corresponding centroids 324, 326, and 328, as well as the equatorial plane 310 (as indicated above, representing the position of the corresponding equatorial plane), may not be midway between their corresponding anterior and posterior poles.

[0054] Figure 6 This is a schematic diagram showing another set of capsule contours 350, including capsule contours 352, 354, 356, and 358. A set of capsule contours 350 is shown, exhibiting the same capsule thickness (along the vertical axis (or also called the Y-axis) 360) and the same capsule diameter (along the horizontal axis (or also called the X-axis) 362) but with four different capsule skew factor values. Figure 6In this configuration, cyst contours 352, 354, 356, and 358 are aligned at anterior pole 364 and posterior pole 366, corresponding to all cyst contours in the same set of cyst contours 350. Cyst contours 352, 354, 356, and 358 define a variable-displacement structure; for example, cyst contours 352 and 358 define equatorial planes 370 and 372, respectively, axially offset by a distance 374. In one example, the cyst has a height / thickness of approximately 5.6 mm along the vertical axis 360, a width (to their respective vertices) of approximately 5 mm along the horizontal axis 362, and an axial offset distance 374 of approximately 1.4 mm.

[0055] When the capsule deflection factor is zero, the Y-coordinate of the corresponding equatorial plane (and corresponding centroid) is exactly midway between the anterior pole 364 and the posterior pole 366. For example, the capsule deflection factor of capsule contour 358 can be zero because the Y-coordinate of its corresponding equatorial plane 372 can be zero. As the capsule deflection factor increases, the corresponding equatorial plane moves toward the anterior pole 364, the centroid is set to be away from the corresponding equatorial plane in the posterior direction, and the lens capsule (see...) Figure 4 The mass on the anterior side of the sac plane decreases compared to the posterior side. For example, the skew factor of sac contour 352 can be greater than that of sac contours 354, 356, and 358 because the equatorial plane 370 corresponding to sac contour 352 can be furthest from and closest to the centers of the anterior pole 364 and posterior pole 366 compared to the corresponding equatorial poles of sac contours 354, 356, and 358. As indicated elsewhere, the sac skew factor can be between 0 and 0.4, and typically between 0 and 0.2.

[0056] Machine learning module 20 may include a neural network trained using a training dataset. The training process occurs in a closed-loop or iterative manner, where the neural network is trained until a specific criterion is met, for example, until the difference between the network output and the real data falls below a certain threshold. The neural network achieves convergence as a predefined loss function associated with the training dataset is minimized. Convergence marks the completion of training. System 10 can be configured to be "adaptive" and periodically updated after collecting additional training data for machine learning module 20. It should be understood that system 10 is not limited to a specific neural network method.

[0057] Figure 8 An example is provided on how to use the machine learning module 20 to simulate the finite element analysis model 22. (Reference) Figure 8Example graphs illustrating various patient data distribution surfaces, including a first data region R1, a second data region R2, and a third data region R3, are shown in three spatial dimensions. The data distribution surfaces are patient-specific finite element models shown along a first axis 450 (capsule thickness in mm), a second axis 452 (capsule diameter in mm), and a third axis 454 (refractive power or optical diopter). Points in each region represent the input (first axis 450 and second axis 452) and output (third axis 454) of the finite element model for a range of capsule deviation values. These surfaces represent machine learning modules 20 trained on finite element model 22 within the same data range.

[0058] Figure 9 This diagram shows a comparison between baseline IOL power and corrected or adjusted baseline IOL power. Line 510 depicts the uncorrected power. Line 520 depicts the adjusted baseline IOL power. The horizontal axis 502 indicates IOL power (refractive power), while the vertical axis 504 indicates the predicted postoperative IOL equivalent spherical power (refractive power). The adjusted baseline IOL power can be output to the lens selection module (see [link to module]). Figure 1 ), to be used for selecting the artificial lens 12 for implantation in eye E.

[0059] In summary, System 10 utilizes Machine Learning Module 20 to selectively execute Finite Element Model 22 to determine the axial displacement factor used to enhance the calculation of base IOL (intraocular lens) power, thereby optimizing, adjusting, or improving refractive outcomes. By first generating seed data (for the effective lens position) using Finite Element Model 22, System 10 can be extended to various types of IOLs and their target populations. Subsequently, Finite Element Model 22 can suggest power correction for each specific patient's combination of capsule parameters in a clinical setting without having to regenerate patient-specific Finite Element Model data.

[0060] Figure 1 The various components of system 10 can communicate via wireless network 34. Network 34 can be a bus implemented in various ways, such as a serial communication bus in the form of a local area network (LAN). The LAN can include, but is not limited to, a control area network (CAN), a control area network with flexible data rates (CAN-FD), Ethernet, Bluetooth, Wi-Fi, and other forms of data connection. Network 34 can be a wireless local area network (LAN) that links multiple devices using a wireless distribution method, a wireless metropolitan area network (MAN) that connects several wireless LANs, or a wireless wide area network (WAN). Other types of connections can also be used.

[0061] refer to Figure 1The controller C can be configured to receive and transmit data via a mobile application 36, which can be installed on a smartphone, laptop, tablet, desktop computer, or other electronic device and may include a touchscreen interface or I / O device (such as a keyboard or mouse). The circuitry and components of the mobile application (“app”) are available to those skilled in the art. The controller C can interact with cloud unit 40 and / or remote server 38 and is configured to share data across all clinical locations employing system 10. In some embodiments, the mobile application 36 is used to interface with the remote server 38 in cloud unit 40, perform method 100 on the remote server, and provide feedback of recommended lens correction factors / features to the mobile application 36 running on processor P. Cloud unit 40 may include one or more servers hosted on the Internet to store, manage, and process data. Remote server 38 may be a private or public information source maintained by an organization such as a research institute, company, university, and / or hospital.

[0062] Figure 1 The controller C includes a computer-readable medium (also called a processor-readable medium) that includes a non-transitory (e.g., tangible) medium involved in providing data (e.g., instructions) that can be read by a computer (e.g., by the computer's processor). Such a medium can take many forms, including, but not limited to, non-volatile and volatile media. Non-volatile media can include, for example, optical discs or magnetic disks, and other persistent storage. Volatile media can include, for example, dynamic random access memory (DRAM), which can constitute main memory. Such instructions can be transmitted via one or more transmission media, including coaxial cables, copper wires, and optical fibers, including wires containing a system bus connected to the computer's processor. Some forms of computer-readable media include, for example, floppy disks, floppy disks, hard disks, magnetic tapes, other magnetic media, CD-ROMs, DVDs, other optical media, physical media, RAM, PROMs, EPROMs, FLASH-EEPROMs, other memory chips or cartridges, or other media that can be read by a computer.

[0063] The lookup tables, databases, data repositories, or other data stores described in this article can include various mechanisms for storing, accessing, and retrieving a variety of data, including hierarchical databases, a set of files in a file storage system, application databases in proprietary formats, relational database management systems (RDBMS), etc. Each such data store can be included in a computing device employing a computer operating system (such as one of the aforementioned operating systems) and can be accessed via a network in one or more of various ways. File systems can be accessed from the computer operating system and can include files stored in various formats. RDBMS can employ Structured Query Language (SQL), as well as languages ​​used to create, store, edit, and execute stored programs, such as the PL / SQL language mentioned above.

[0064] The flowcharts shown illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each box in the flowchart or block diagram may represent a module, segment, or portion of code, including one or more executable instructions for implementing a specific logical function(s). It should also be noted that each box in the block diagram and / or flowchart illustrations, and combinations of boxes in the block diagram and / or flowchart illustrations, may be implemented by a system based on dedicated hardware or a combination of dedicated hardware and computer instructions that performs the specified function or action. These computer program instructions may also be stored in a computer-readable medium that can direct a controller or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable medium produce an article of art including instructions for implementing the functions / actions specified in the flowchart and / or block diagram boxes.

[0065] The numerical values ​​of parameter items (e.g., quantities or conditions) in this specification (including the appended claims) should be understood to be modified by the term “about” in each corresponding instance, regardless of whether “about” actually appears before the numerical value. “About” indicates that the numerical value is slightly imprecise (the value is somewhat close to precise; about or fairly close to the value; almost close). If the imprecision provided by “about” is not otherwise understood in the art to have this common meaning, then “about” as used herein at least indicates the variation that may arise from common methods of measuring and using such parameter items. Furthermore, the disclosure of ranges includes the disclosure of each value throughout the range or of further subdivided ranges. Each value within a range and the endpoints of the range are thus disclosed as separate embodiments.

[0066] The detailed description and accompanying drawings are supportive and descriptive of this disclosure, but the scope of this disclosure is defined only by the claims. While some best modes and other embodiments for implementing the claimed disclosure have been described in detail, various alternative designs and embodiments exist to practice the disclosure as defined in the appended claims. Furthermore, the features of the embodiments shown in the drawings or the various different embodiments mentioned in this specification are not necessarily to be construed as embodiments independent of each other. Rather, each feature described in one of these examples of embodiments may be combined with one or more other desired features from other embodiments to produce other embodiments not described in words or with reference to the drawings. Accordingly, such other embodiments fall within the scope of the appended claims.

[0067] For example, the subject matter of this disclosure is illustrated by various aspects described below. For convenience, various examples of aspects of this disclosure are described as numbered examples (1, 2, 3, etc.). These examples are provided as examples and do not limit the disclosure. Unless the context otherwise requires, aspects of the various implementations described herein may be omitted, substituted for aspects of other implementations, or combined with aspects of other implementations. For example, one or more aspects of Example 1 below may be omitted, substituted for another example (e.g., Example 2), or one or more aspects of each example may be combined with aspects of another example. The following is a non-limiting overview of some exemplary implementations presented herein.

[0068] Example 1. A system for selecting an artificial lens for implantation in an eye, the system comprising: A controller having a processor and a tangible, non-transitory memory on which instructions are recorded; The controller is configured to selectively execute finite element modeling and machine learning modules, and the instructions executed by the one or more processors cause the controller to: Obtain input data, which includes one or more biometric parameters of the eye; Based on the input data, the machine learning module extracts multiple capsule parameters corresponding to the lens capsule of the eye; The axial displacement factor is determined in part based on the plurality of capsule parameters via the finite element model, the axial displacement factor taking into account the predicted axial displacement of the intraocular lens after implantation in the eye; and The intraocular lens power is recommended based on the axial displacement factor using one or more lens constant formulas.

[0069] Example 2. A system as described in Example 1, wherein the finite element model is tensor-based, and the controller is adapted to execute the finite element model using multiphysics software.

[0070] Example 3. A system as described in Example 1 or Example 2, wherein the controller is adapted to select the intraocular lens based on a recommended intraocular lens power.

[0071] Example 4. The system as described in any one of Examples 1 to 3, wherein the plurality of cyst parameters include cyst diameter and cyst thickness.

[0072] Example 5. The system as described in Example 4, wherein the plurality of cyst parameters include a cyst skew factor based on the cyst thickness.

[0073] Example 6. The system as described in Example 5, wherein the capsule skew factor is the ratio of the Y-coordinate of the centroid of the capsule profile to the capsule thickness, the capsule profile being the cross-sectional profile of the lens capsule cut through the anterior and posterior poles of the lens capsule.

[0074] Example 7. The system as described in Example 6, wherein the sac skew factor is between 0 and 0.2.

[0075] Example 8. The system as described in Example 6, wherein the sac skew factor is zero when the Y-coordinate in the equatorial plane is exactly midway between the anterior pole and the posterior pole.

[0076] Example 9. The system as described in Example 8, wherein the capsule skew factor is greater than zero when the Y-coordinate in the equatorial plane is not between the anterior and posterior poles and the corresponding mass of the lens of the eye is relatively large on the anterior and posterior sides of the equatorial plane.

[0077] Example 10. A method for selecting an artificial lens for implantation in an eye using a system having a controller with at least one processor and at least one non-transitory tangible memory, the method comprising: The finite element model and machine learning modules are selectively executed via the controller; The controller receives input data, which includes one or more biometric parameters of the eye. Based on the input data, the machine learning module extracts multiple capsule parameters corresponding to the lens capsule of the eye. An axial displacement factor is determined in part based on the plurality of capsule parameters via the execution of the finite element model, the axial displacement factor taking into account the predicted axial displacement of the intraocular lens after implantation in the eye; and The intraocular lens power is recommended based on the axial displacement factor using one or more lens constant formulas.

[0078] Example 11. The method as described in Example 10, further comprising: If the finite element model is selected to be tensor-based, the controller is adapted to execute the finite element model using multiphysics software.

[0079] Example 12. The method as described in Example 10 or 11 further includes: The machine learning module is selected as a neural network.

[0080] Example 13. The method of any one of Examples 10 to 12, further comprising: The plurality of cyst parameters are selected to include cyst diameter and cyst thickness.

[0081] Example 14. The method as described in Example 13, further comprising: The plurality of cyst parameters are selected to include the cyst skew factor.

[0082] Example 15. The method as described in Example 14 further includes: The capsule skew factor is selected as the ratio of the Y-coordinate of the centroid of the capsule contour to the capsule thickness, wherein the capsule contour is the outer contour of the lens cut through the anterior and posterior poles of the lens capsule.

[0083] Example 16. The method as described in Example 15, wherein the sac skew factor is between 0 and 0.2.

[0084] Example 17. The method as described in Example 15 further includes: When the Y-coordinate of the equatorial region is exactly midway between the anterior pole and the posterior pole, the skewing factor is selected to be zero.

[0085] Example 18. The method as described in Example 17 further includes: The capsule deviation factor is selected to be greater than zero when the Y-coordinate of the equatorial plane is not between the anterior and posterior poles and the corresponding mass of the lens of the eye is relatively large on the posterior side of the equatorial plane.

[0086] Example 19. A system for selecting an artificial lens for implantation in an eye, the system comprising: A controller having one or more processors and tangible non-transitory memory, on which instructions are recorded; The controller is configured to selectively execute a first machine learning module and a second machine learning model trained to simulate a finite element model, wherein the instructions are executed by the one or more processors to cause the controller to: Receive input data, the input data including one or more biometric parameters of the eye; Based on the input data, multiple cyst parameters are extracted via the first machine learning module; The axial displacement factor is determined in part based on the plurality of bladder parameters via a second machine learning model trained to simulate the finite element model; and The lens power of the intraocular lens is adjusted in part based on the axial displacement factor, which is a power correction feature that takes into account the predicted axial displacement of the intraocular lens after implantation in the eye.

Claims

1. A system for selecting an artificial lens for implantation in an eye, the system comprising: A controller having one or more processors and tangible non-transitory memory, on which instructions are recorded; The controller is configured to selectively execute finite element modeling and machine learning modules, and the instructions executed by the one or more processors cause the controller to: Obtain input data, which includes one or more biometric parameters of the eye; Based on the input data, the machine learning module extracts multiple capsule parameters corresponding to the lens capsule of the eye; The axial displacement factor is determined in part based on the plurality of capsule parameters via the finite element model, the axial displacement factor taking into account the predicted axial displacement of the intraocular lens after implantation in the eye; and The intraocular lens power is recommended based on the axial displacement factor using one or more lens constant formulas.

2. The system of claim 1, wherein, The finite element model is based on tensors, and the controller is adapted to execute the finite element model using multiphysics software.

3. The system of claim 1, wherein, The controller is adapted to select the intraocular lens based on a recommended intraocular lens power.

4. The system of claim 1, wherein, The multiple cyst parameters include cyst diameter and cyst thickness.

5. The system as described in claim 4, wherein, The plurality of cyst parameters include a cyst skew factor based on the cyst thickness.

6. The system of claim 5, wherein, The capsule skew factor is the ratio of the Y-coordinate of the centroid of the capsule profile to the capsule thickness, and the capsule profile is the cross-sectional profile of the lens capsule cut through the anterior and posterior poles of the lens capsule.

7. The system of claim 6, wherein, The cyst skew factor is between 0 and 0.

2.

8. The system of claim 6, wherein, The sac skew factor is zero when the Y-coordinate in the equatorial plane is exactly midway between the anterior and posterior poles.

9. The system of claim 8, wherein, The capsule skew factor is greater than zero when the Y-coordinate in the equatorial plane is not between the anterior and posterior poles and the corresponding mass of the lens of the eye is relatively large on the posterior side of the equatorial plane.

10. A method for selecting an artificial lens for implantation in an eye using a system having a controller with at least one processor and at least one non-transitory tangible memory, the method comprising: The finite element model and machine learning modules are selectively executed via the controller; The controller receives input data, which includes one or more biometric parameters of the eye. Based on the input data, the machine learning module extracts multiple capsule parameters corresponding to the lens capsule of the eye. The axial displacement factor is determined in part based on the plurality of capsule parameters via the execution of the finite element model, the axial displacement factor taking into account the predicted axial displacement of the intraocular lens after implantation in the eye; as well as The intraocular lens power is recommended based on the axial displacement factor using one or more lens constant formulas.

11. The method of claim 10, further comprising: If the finite element model is selected to be tensor-based, the controller is adapted to execute the finite element model using multiphysics software.

12. The method of claim 10, further comprising: The machine learning module is selected as a neural network.

13. The method of claim 10, further comprising: The plurality of cyst parameters are selected to include cyst diameter and cyst thickness.

14. The method of claim 13, further comprising: The plurality of cyst parameters are selected to include the cyst skew factor.

15. The method of claim 14, further comprising: The capsule skew factor is selected as the ratio of the Y-coordinate of the centroid of the capsule profile to the capsule thickness, wherein the capsule profile is the cross-sectional profile of the lens capsule cut through the anterior and posterior poles of the lens capsule.

16. The method of claim 15, wherein, The cyst skew factor is between 0 and 0.

2.

17. The method of claim 15, further comprising: When the Y-coordinate in the equatorial plane is exactly midway between the anterior pole and the posterior pole, the skewing factor is selected to be zero.

18. The method of claim 17, further comprising: The capsule deviation factor is selected to be greater than zero when the Y-coordinate of the equatorial plane is not between the anterior and posterior poles and the corresponding mass of the lens of the eye is relatively large on the posterior side of the equatorial plane.

19. A system for selecting an artificial lens for implantation in an eye, the system comprising: A controller having one or more processors and tangible non-transitory memory, on which instructions are recorded; The controller is configured to selectively execute a first machine learning module and a second machine learning model trained to simulate a finite element model, wherein the instructions are executed by the one or more processors to cause the controller to: Receive input data, the input data including one or more biometric parameters of the eye; Based on the input data, multiple cyst parameters are extracted via the first machine learning module; The axial displacement factor is determined in part based on the plurality of bladder parameters via a second machine learning model trained to simulate the finite element model; and The lens power of the intraocular lens is adjusted in part based on the axial displacement factor, which is a power correction feature that takes into account the predicted axial displacement of the intraocular lens after implantation in the eye.

Images