Apparatus and method for intraocular lens selection using postoperative measurements
Through a combination device of a biometer and an automatic optometry device, combined with multiple intraocular lens selection formulas, a super formula is formed, which solves the clinical dilemma caused by formula differences in the prior art, improves the accuracy and reliability of lens selection, and is suitable for different Type of eyes.
Patent Information
- Application Number
- CN201980091214.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2018-12-06
- Filing Date
- 2019-12-06
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2039-12-06
AI Technical Summary
The existing intraocular lens calculation formulas vary greatly within certain input parameters, resulting in clinical dilemma when selecting appropriate formulas, and the prior art is difficult to effectively utilize the advantages and disadvantages of each formula to obtain more accurate results. .
Using a combination device of a biometer and an automatic optometry device, the lens selection algorithm is adjusted by obtaining preoperative and postoperative measurement values, and the difference between target refractive and postoperative refractive values are determined automatically or autonomously calibrated preoperative measurement values. , combine the advantages of multiple formulas to form a super formula to improve accuracy.
By combining multiple intraocular lens selection formulas, a more accurate super formula is formed, which reduces inconsistencies between formulas, improves the accuracy and reliability of lens selection, is suitable for different types of eyes, and enhances the accuracy of clinical applications .
Smart Images

Figure CN113382673B_ABST
Abstract
Description
[0001] Claiming priority under 35 USC § 119
[0002] This application claims priority to U.S. Provisional Application No. 62 / 776,320, filed on December 6, 2018, entitled “APPARATUS AND METHOD FOR INTRAOCULAR LENS SELECTION USING POST-OPERATIVE MEASUREMENTS,” which is assigned to the assignee of the present application and is incorporated herein by reference in its entirety. Technical Field
[0003] Aspects of the invention relate to systems, devices, and methods for selecting an intraocular lens. Background Art
[0004] There are currently multiple intraocular lens (IOL) calculation formulas. For example, the SRK I formula13 can be considered a first-generation formula. More recent third-generation formulas (such as the SRK / T formula) have had a theoretical component added to improve their accuracy, which has increased the complexity of the formula compared to the SRK I formula. Other IOL calculation formulas include the Hoffer Q, Holladay I, Haigis, and SRK / T formulas. The Koch adjustment can also be used to adjust any of these formulas. Although the existing formulas give similar results within a certain range of input parameters, there are also large differences within the specified input parameter range.
[0005] Each formula has been shown to work best for certain input parameters. The input parameters can include ocular measurement parameters such as axial length, corneal power, white to white distance, gender, anterior chamber depth, preoperative refraction and / or lens thickness. In one aspect, at least two ocular measurement parameters can be used. In one aspect, the at least two parameters can include axial length and corneal power. For example, one particular formula may be more suitable for "shorter" eyes, while another particular formula may be more suitable for "longer" eyes. In addition, "adjustments" to these formulas can be used to obtain better results. "Adjustments" can include any additional factors applied to the IOL calculation formula.
[0006] Current techniques involve selecting a formula to determine lens power and potentially comparing the results to those obtained using another formula. Few ophthalmologists are aware of the data and literature supporting the use of one formula over another. Although the use of a specific formula may be debatable, in certain circumstances (e.g., a specific measurement of axial length or corneal power), one formula is generally considered superior to another. Summary of the invention
[0007] Aspects of the present disclosure may include an apparatus and method for selecting an intraocular lens using a combination of a biometer and an autorefractor. In one aspect, the biometer and the autorefractor are combined into a single apparatus for obtaining preoperative and postoperative measurements of a patient receiving an intraocular lens. The apparatus may recommend an intraocular lens based on the preoperative measurements and the target refraction. The apparatus may correlate the patient's preoperative measurements with the postoperative measurements. For example, the apparatus may determine the difference between the target refraction and the postoperative refraction. The apparatus may automatically or autonomously calibrate future preoperative measurements, or adjust a lens selection algorithm based on the postoperative measurements.
[0008] In one aspect, the present disclosure provides a device for selecting an intraocular lens. The device may include a biometer and an auto-optometer. The biometer may be configured to obtain at least two ocular measurement parameters of an eye. The auto-optometer may be configured to obtain postoperative refraction of the eye. The device may also include a user interface configured to obtain lens selection parameters of the eye, a memory, and a processor communicatively coupled to the biometer, the user interface, the auto-optometer, and the memory. The processor may be configured to determine an intraocular lens power using a formula based on at least two ocular measurement parameters. The processor may be configured to associate at least two ocular measurement parameters, the intraocular lens power, and the postoperative refraction as a training set.
[0009] In one aspect, the present disclosure provides a method for selecting an intraocular lens and / or an apparatus and system for the method. The method may include obtaining at least two ocular measurement parameters of an eye through a biometer. The method may include obtaining a lens selection parameter of the eye. The method may include determining an intraocular lens power using a formula based on at least two ocular measurement parameters. The method may include obtaining a postoperative refraction of the eye from an auto-optometer coupled to the biometer in communication. The method may include associating at least two ocular measurement parameters, the intraocular lens power, and the postoperative refraction as a training set.
[0010] In another aspect, the present disclosure provides a non-transitory computer-readable medium storing computer-executable instructions. The instructions may cause a computer to obtain at least two ocular measurement parameters of an eye through a biometer. The instructions may cause a computer to obtain a lens selection parameter of the eye. The instructions may cause a computer to determine an intraocular lens power using a formula based on at least two ocular measurement parameters. The instructions may cause a computer to obtain a postoperative refraction of the eye from an auto-optometer communicatively coupled to the biometer. The instructions may cause a computer to associate at least two ocular measurement parameters, intraocular lens power, and postoperative refraction as a training set.
[0011] Other advantages and novel features associated with various aspects of the present invention will be set forth in part in the following description and in part will become more apparent to those skilled in the art after examination of the following or through practice. BRIEF DESCRIPTION OF THE DRAWINGS
[0012] In the attached picture:
[0013] Figures 1A-1C Illustrated is a graphical representation of ocular measurement areas where existing formulas differ and represent clinical dilemmas.
[0014] Figure 2A-2C A three-dimensional surface including ideal or near-ideal portions of a plurality of intraocular lens selection formulas for use in accordance with aspects of the present disclosure is illustrated.
[0015] Figure 3 Various features of an example computer system for use with aspects of the present disclosure are illustrated.
[0016] Figure 4 An example system diagram illustrating various hardware components and other features used in accordance with aspects of the present disclosure.
[0017] Figure 5 An example input user interface for use in accordance with aspects of the present disclosure is illustrated.
[0018] Figure 6 An example output user interface used in accordance with aspects of the present disclosure is illustrated.
[0019] Figure 7 A graph representing a dataset of example eyes is illustrated.
[0020] Figure 8 A three-dimensional surface representing learned variations of a formula according to aspects of the present disclosure is illustrated.
[0021] Fig. 9 is a schematic diagram of an example computer system according to one aspect of the present disclosure.
[0022] Fig.10 is a flow chart of an example method for recommending intraocular lens power according to one aspect of the present disclosure.
[0023] Fig.11 is a schematic diagram of an example apparatus according to one aspect of the present disclosure.
[0024] Fig.12 is a flow chart of a second example method for recommending an intraocular lens power according to one aspect of the present disclosure.
[0025] Fig.13 It is a graphic Fig.11 A conceptual diagram of an example usage scenario of an example device. DETAILED DESCRIPTION
[0026] Aspects of the present disclosure may include systems, devices, and methods for selecting intraocular lenses using a three-dimensional metasurface.The metasurface may represent portions of multiple lens selection formulas based on a range of measurements that best fit the respective intraocular lens selection formula.
[0027] The present disclosure provides a new intraocular lens calculation formula, which is based on a combination of existing intraocular lens (IOL) calculation formulas, using graphical analysis to determine the range of input parameters, such as the portion of the range over which the existing intraocular lens calculation formula is most accurate. These individual formulas have never been proposed before or considered as three-dimensional objects that can be combined and overlapped according to a set of determined "standards." The calculation formula for the intraocular lens will graphically "pick" the parts of each formula that have been proven or considered to be the most accurate, and combine them into an overall formula. When other formulas are proposed or additional data is obtained, the combination or "super" formula can be further (e.g., iteratively) adjusted and / or optimized.
[0028] Figures 1A to 1C A Ladas-Siddiqui diagram is shown that highlights areas of clinical agreement and difference between formulas over specific ranges of corneal power and axial length. Figure 1A ) indicates that, within the entire range of input parameters, all formulas differ from at least one of the other four formulas by more than 0.5 diopters. As shown in Figure 104 ( Figure 1B ), when the tolerance for divergence between formulas is increased to 1.0 diopters in predicted IOL power (a clinically undesirable level), a region of correspondence exists between all formulas tested. As shown in Figure 106 ( Figure 1C ), the corresponding area increases further when the tolerance is increased to 1.5 diopters. Therefore, resolving these areas of discordance is of high clinical relevance and requires not only the use of each formula but also a detailed understanding of its specific strengths and weaknesses. Describing the comparison between IOL formulas in this way illustrates the specific input parameter ranges where there are differences between the formulas, thereby enabling a more precise understanding of how the formulas differ.
[0029] FIG. 2A to FIG. 2C A super formula based on a combination of the advantages and disadvantages of each IOL formula is illustrated. The super formula combines the ideal, most accurate output parts of each IOL formula into one combined super formula. FIG. 2A to FIG. 2C The graphical results of this super-formula are illustrated, showing the progressive development and evolution of a single polyhedral surface comprising the most accurate portions of multiple individual formulas based on a specified range of input variables. Figure 2Ais a chart showing the regions where different IOL selection formulas were determined to be optimal. Figure 2B is a chart showing the interrelationships between the regions of different IOL selection formulas. Figure 2C A surface of a combined superformula is shown. In one aspect, although the surface is illustrated as using two eye measurements to represent the output of the formula, an IOL selection formula or system can use more than two eye measurements. An IOL selection system or formula with multiple inputs can be viewed as a multi-dimensional hypersurface.
[0030] Another technique for determining intraocular lens power includes the use of machine learning. For example, a neural network can be trained using the results of a previous operation. The neural network can then be provided with the patient's preoperative measurements. The neural network will attempt to classify the patient's preoperative measurements based on the training set to determine the lens power. This approach may be limited based on the training data, and some groups of preoperative measurements may be considered out of bounds and the neural network may not produce results. In addition, the neural network may produce anomalous results due to random correlations in the training data. Furthermore, changes in the training data may require retraining of the neural network.
[0031] The present disclosure describes methods, devices, and systems for combining and using any number of multiple formulas into a single formula using the ideal parts of each constituent formula based on, for example, theoretical and / or empirical information. In addition, the present disclosure describes a technique for improving a formula-based method using a combination of formulas and machine learning. For example, in one aspect, a neural network can be trained to determine the estimated error that may result from using a calculated intraocular lens power. The estimated error can be used to adjust lens selection parameters, such as a target refractive power or an A-constant, and the formula can be used to recalculate the intraocular lens power using the adjusted lens selection parameters.
[0032] Aspects of the present disclosure may be implemented using hardware, software, or a combination thereof, and may be implemented in one or more computer systems or other processing systems. In one aspect of the present disclosure, features are directed to one or more computer systems capable of performing the functions described herein. Figure 3 One example of such a computer system 300 is shown.
[0033] Computer system 300 includes one or more processors, such as processor 304. Processor 304 is connected to a communication infrastructure 306 (e.g., a communication bus, a crossbar switch, or a network). Various software aspects are described based on this example computer system. After reading this description, it will become apparent to those skilled in the relevant art how to implement various aspects of the present disclosure using other computer systems and / or architectures.
[0034] The computer system 300 may include a display interface 302 that forwards graphics, text, and other data from the communication infrastructure 306 (or from a graphics processing unit (GPU) 332) for display on a display unit 330. For example, the display interface 302 may forward a graphical rendering of a metasurface from the processor 304 to the display unit 330. The computer system 300 also includes a main memory 308, preferably a random access memory (RAM), and may also include an auxiliary memory 310. The auxiliary memory 310 may include, for example, a hard disk drive 312 and / or a removable storage drive 314, which represents a floppy disk drive, a tape drive, an optical disk drive, a universal serial bus (USB) flash drive, etc. The removable storage drive 314 reads and / or writes to a removable storage unit 318 in a known manner. The removable storage unit 318 represents a floppy disk, a tape, an optical disk, a USB flash drive, etc., which is read and written by the removable storage drive 314. It is understood that the removable storage unit 318 includes a computer-usable storage medium having computer software and / or data stored therein.
[0035] Alternative aspects of the present disclosure may include secondary memory 310, and may include other similar means of allowing computer programs or other instructions to be loaded into computer system 300. Such means may include, for example, a removable storage unit 322 and an interface 320. Such examples may include a program cartridge and cartridge interface (such as used in video game devices), a removable memory chip (such as an erasable programmable read-only memory (EPROM) or a programmable read-only memory (PROM)) and an associated socket, and other removable storage units 322 and interfaces 320 that allow software and data to be transferred from a removable storage unit 322 to computer system 300.
[0036] The computer system 300 may also include a communication interface 324. The communication interface 324 allows software and data to be transmitted between the computer system 300 and an external device. Examples of the communication interface 324 may include a modem, a network interface (e.g., an Ethernet card), a communication port, a personal computer memory card international association (PCMCIA) slot and card, etc. The software and data transmitted via the communication interface 324 are in the form of a signal 328, which may be an electronic, electromagnetic, optical or other signal that can be received by the communication interface 324. These signals 328 are provided to the communication interface 324 via a communication path (e.g., a channel) 326. The path 326 carries the signal 328 and may be implemented using wires or cables, optical fibers, telephone lines, cellular links, radio frequency (RF) links and / or other communication channels. In this document, the terms "computer program medium" and "computer-usable medium" generally refer to media such as a removable storage drive 380, a hard disk installed in a hard disk drive 370, and a signal 328. These computer program products provide software to the computer system 300. Aspects of the present disclosure relate to such computer program products.
[0037] In one aspect, the computer system 300 may include an eye measurement device 350. The eye measurement device 350 may determine one or more eye measurement parameters. The eye measurement device may include any device for measuring an eye. For example, the eye measurement device 350 may measure the axial length and corneal power of the eye. In one aspect, the eye measurement device 350 may also measure the white-to-white distance, anterior chamber depth, preoperative refraction, and / or lens thickness. The eye measurement device 350 may also receive input of eye measurement parameters (e.g., gender). The axial length may be the distance from the corneal surface to the retina. The corneal power may be the refractive power of the cornea. As another example, the eye measurement device 350 may measure the anterior chamber depth of the eye. In one aspect, the eye measurement device 350 may be an ultrasonic device. In another aspect, the eye measurement device 350 may be an optical biometer. A device called a biometer may be used. and various optical biometers of IOL MASTER. On the other hand, the eye measurement device 350 may include an intraoperative abberrometry device. The intraoperative abberrometry device can measure the refractive properties of the eye during surgery. For example, the intraoperative abberrometry device can provide information about the sphere, cylinder, and axis of the eye. In addition, the eye measurement device may include a postoperative measurement device, such as a wavefront analyzer. The eye measurement device 350 can be communicatively coupled to the processor 304 via the communication infrastructure 306, the communication interface 324, and / or the communication path 326.
[0038] Computer programs (also referred to as computer control logic) are stored in the main memory 308 and / or the secondary memory 310. The computer programs may also be received via the communication interface 324. Such computer programs, when executed, enable the computer system 300 to perform features in accordance with various aspects of the present disclosure, as discussed herein. Specifically, the computer programs, when executed, enable the processor 304 to perform features in accordance with various aspects of the present disclosure. Thus, such computer programs represent a controller of the computer system 300.
[0039] In one aspect of the present disclosure, when the present disclosure is implemented using software, the software can be stored in a computer program product and loaded into the computer system 300 using a removable storage drive 314, a hard disk drive 312, or a communication interface 320. When executed by the processor 304, the control logic (software) causes the processor 304 to perform the functions described herein. In another aspect of the present disclosure, the system is implemented primarily in hardware using, for example, hardware components such as an application specific integrated circuit (ASIC). It is apparent to those skilled in the relevant art that a hardware state machine is implemented to perform the functions described herein.
[0040] In still another aspect of the present disclosure, the present disclosure may be implemented using a combination of hardware and software.
[0041] Figure 4 4 shows a communication system 400 that can be used according to aspects of the present disclosure. The communication system 400 includes one or more access devices 460 (also interchangeably referred to herein as one or more "users") and one or more terminals 442 and / or other input devices (e.g., eye measurement devices 466). In one aspect, the eye measurement device 466 can be similar to the eye measurement device 350 ( Figure 3 ). The ocular measurement device 466 may also be configured to communicate with the network 444. In one aspect of the present disclosure, for example, the data for use is input and / or accessed after being received from an input device, for example, by the accessor 460 via a terminal 442, which may be, for example, a personal computer (PC), a minicomputer, a mainframe computer, a microcomputer, a telephone device or a wireless device, a personal digital assistant ("PDA"), or a handheld wireless device (e.g., a wireless phone) coupled to a server 443, which may be, for example, a PC, a minicomputer, a mainframe computer, a microcomputer, or other devices having a processor and a data repository and / or connected to the data repository via, for example, a network 444 (e.g., the Internet or an intranet) and / or a wireless network and couplers 445, 446, 464. Couplers 445, 446, 464 include, for example, wired, wireless, or fiber optic links. In another aspect of the present disclosure, the methods and systems of the present disclosure may include one or more features that operate in a standalone environment, such as on a single terminal.
[0042] In one aspect, server 443 may be computer system 300 ( Figure 3 ). In one aspect, for example, server 443 can be configured to perform the methods described herein. For example, server 443 can obtain measurements, such as axial length and corneal power measurements, from terminal 442 and / or other input devices. The measurements can be input by accessor 460 or by visual measurement device 350 ( Figure 3 ). The server 443 may determine a relevant portion of the metasurface that includes an ideal portion of a plurality of intraocular lens selection formulas based on a range of axial length and corneal power measurements that best fit the respective intraocular lens selection formulas. In addition, the server 443 may determine an intraocular lens power based on a formula for the relevant portion of the metasurface.
[0043] Figure 5 An example input user interface (UI) 500 used according to aspects of the present invention is illustrated. For example, the user interface 500 can be implemented by the server 443 and displayed on the terminal 442. The user interface 500 can allow the user to input preoperative measurements of one or more eyes and target parameters. The user interface 500 allows simultaneous calculation and mapping of both eyes. The input data may include corneal curvature (K1, K2 and K index), axial length, corneal index and anterior chamber depth (ACD). The K1 and K2 values can be average values of the magnification of the front surface of the eye including different angles. The K index can be a standard refractive index of the cornea. The K index used by different surgeons or different countries may be slightly different. The anterior chamber depth is a measurement of the original lens position before removal. Then select the lens A-constant and the desired postoperative refraction, and then perform an optimization calculation. The lens A-constant is a characteristic of the lens, which can be specified by the manufacturer and can affect the total refraction. The target refraction can be a target specified by the surgeon. For example, a target refraction of 0.0 can be used.
[0044] The user interface 500 may include a right eye input area 510 and a left eye input area 520. Each input area 510, 520 may include an input area for a specific measurement value or parameter. For example, the input area 510 may include an axial length area 511, a K1 area 512, a K2 area 513, a K index area 514, and an optical ACD area 515. The input area 510 may also include lens selection parameters, including an A-constant area 516 and a target refraction 517. The input areas 510, 520 may also include a help icon (e.g., "?") that provides a description of the measurement value or parameter including an allowed range. Some areas (e.g., the K index area 514) may use a drop-down menu to select a value. Additional areas that may be included in the user interface 500 include intraoperative aberration measurements, such as sphere, cylinder, and axial length.
[0045] Additionally, the user interface 500 may include a surgeon area 530, a patient area 531, and a patient ID area 532. The server 443 may generate a record for the surgeon and the patient based on the areas 530, 531, 532. The user interface 500 may also include an import option 540 that may allow a user to upload a file (e.g., a spreadsheet) including measurements and parameters for one or more patients. A dedicated toric calculator and a post-LASIK calculator may also be included.
[0046] Figure 6 An example output user interface 600 for use according to aspects of the present invention is illustrated. The user interface 600 can display a recommended intraocular lens (IOL) power 610. The user interface 600 can display multiple IOL power values corresponding to available lenses. The user interface 600 can also display an estimated final refraction (REF) 620 associated with each intraocular lens power value. In addition, the user interface 600 can include a graphical representation 630 of a hypersurface for selecting intraocular lens power. Relevant points on the hypersurface based on the measured values can be indicated. Including a three-dimensional graphic helps the surgeon immediately know whether his patient has a typical or unusual eye. The user interface 600 can include an output area for each eye.
[0047] Figure 7 A graph 700 representing a data set of an example eye is illustrated. The data set may be collected from one or more surgeons. As shown, while most eyes have average corneal curvature and axial length, there are outliers.
[0048] Figure 8 A three-dimensional surface representing learned changes to the formula is illustrated. The original Ladas hypersurface has been developed to improve accuracy for eyes of various sizes, including eye types that were previously thought to present difficulties, such as small eyes with short axial lengths. The circles represent data points of actual postoperative results that allowed the creation of an optimized hypersurface 810 compared to the original hypersurface 820, with a large portion of the original hypersurface 820 overlapping with the new surface.
[0049] In one aspect, a metasurface or another formula can be customized for one or more surgeons. For example, validated results from a surgeon can be analyzed against a metasurface or formula to determine if any patterns in the surgeon's surgery can be detected.
[0050] Now turn to Fig. 9, the example system 900 can recommend an intraocular lens power based on a formula and a trained deep learning machine 912. For example, the system 900 can be implemented on a server 443. The system 900 can communicate with one or more terminals 442 via the above-mentioned user interfaces 500, 600. The system 900 may include a formula component 910 for determining an intraocular lens power using a formula based on at least two eye measurement parameters, a deep learning machine 912 (e.g., a neural network 920) for determining an estimated error of the formula, and a user interface 500 for obtaining at least two eye measurement parameters and a target refractive power or A-constant of the eye. The formula component 910 can also adjust the target refractive power or A-constant based on the estimated error, and re-determine the intraocular lens power based on the formula and the adjusted target refractive power or A-constant. The system 900 may also include one or more training sets 930. The training set 930 may include a postoperative data set including two or more preoperative measurements and parameters (e.g., axial length zone 511, K1 zone 512, K2 zone 513, K index zone 514, optical ACD zone 515, and intraoperative aberration measurements, such as sphere, cylinder, and axial), a selected intraocular lens power, and one or more lens selection parameters (e.g., postoperative refraction or A-constant). The training set 930 may be used to train one or more deep learning machines 912 for estimating the error of the intraocular lens power determined by formula 910, as explained in further detail below.
[0051] The system 900 may include an administrative portal 940 for controlling access to the system 900. For example, the administrative portal 940 may allow an administrative user to generate a training set 930 based on validated results 962. Validated results 962 may be uploaded in the form of a database or spreadsheet. An administrative user may select a combination of measurements and parameters for a training set 930. An administrative user may combine uploaded validated results with any existing training set 930. The system 900 may generate a new neural network 920 based on a new or updated training set 930. The system 900 may provide the administrative user with statistics about the neural network. For example, the neural network may be associated with input boundaries and associated values. The administrative user may also configure access control 950 to manage user accounts for different end users. User accounts may be associated with saved patient data 960. Additionally, user accounts may be associated with customized neural networks 920. For example, a customized neural network 920 may be trained with validated results 962 specifically from a particular surgeon, practice group, or lens manufacturer. Customized neural network 920 may help control for unknown or unmeasurable factors that affect a particular surgeon, practice group, or lens manufacturer.
[0052] The formula component 910 can implement an intraocular lens power determination formula. The intraocular lens power determination formula may include any deterministic technique for generating an intraocular lens power based on two or more eye measurements. For example, the formula component 910 may include software executed by a processor to determine the intraocular lens power according to a formula using input values from the user interface 500. For example, the formula component 910 may implement one or more of the following formulas: the Hoffer Q formula, the Holladay I formula, the Haigis formula, the SRK / T formula, the Barrett Universal II formula, or an adjustment to any of these formulas. In one aspect, the formula component 910 may implement the Ladas super formula to select calculations from one or more of the above formulas. For example, the formula component 910 may determine a relevant portion of a hypersurface including an ideal or near-ideal portion of a plurality of intraocular lens selection formulas based on a range of axial length and corneal power measurements that best fit the respective intraocular lens selection formulas. The formula component 910 may provide the determined intraocular lens power value to the neural network 920 together with all input measurements and parameters. In one aspect, formula component 910 can be implemented as a machine learning formula, such as Hill-RBF.
[0053] The learning machine 912 can use deep learning techniques to predict the error of the formula component 910 based on a training set 930 including postoperative results. The learning machine 912 can be implemented by, for example, a neural network 920, which can utilize Python-based tensor flow. In one aspect, the learning machine 912 may include a computer processor, such as a processor 304, which is programmed to execute instructions for developing the neural network 920 based on the network structure (e.g., the number and type of layers). Once the learning machine 912 has trained the neural network 920 (or other learning machine), the processor configured with the trained learning machine 912 can determine the prediction error of the formula component 910 based on the ocular measurement parameters. The neural network 920 can receive multiple digital inputs to predict a single digital output. In one implementation, the neural network can receive three digital inputs (axial length, K, and ACD) and output an error value. The neural network 920 can be trained by one or more training sets 930. The training set 930 can be considered as labeled data because the training set 930 can include postoperative refraction, which can be used to determine the accuracy or error of the formula. Thus, when the neural network 920 receives the set of digital inputs, the neural network 920 can predict the estimated error of the formula component 910. The learning machine 912 can be implemented using different types of learning machines. For example, the learning machine 912 can use any combination of supervised and unsupervised learning techniques. The learning machine 912 can be constructed as, for example, an artificial neural network, a convolutional neural network, a Bayesian network, or other deep learning models.
[0054] The formula component 910 can then adjust the formula inputs based on the prediction error. Specifically, the formula component 910 can adjust the target refraction or A-constant based on the prediction error. For example, the new target refraction can be set to the difference between the target refraction input by the user and the neural network prediction error. As a numerical example, for an eye with an AL of 25, a K of 45, and an ACD of 3, the neural network 920 can predict an error of 0.25. That is, when the formula component 910 calculates an eye with an AL of 25, a K of 45, an ACD of 3, and an A-constant of 119 (regardless of what the user selects for this purpose) and a target refraction of -0.5, the eye is recalculated using the current formula component 910 but using a new target refraction value of -0.5-0.25=-0.75 instead of the target refraction input by the user. In other words, the formula component 910 can adjust a component of the formula input (e.g., the target refraction) by subtracting the neural network prediction error from the input value.
[0055] In one aspect, formula component 910 can limit adjustments to the formula by neural network prediction error. For example, in out-of-bounds situations where neural network 920 does not have good training data, neural network 920 can produce extreme values. Formula component 910 can limit the value of the prediction error. For example, formula component 910 can limit the neural network prediction error to no more than + / - 0.5.
[0056] In one aspect, neural network 920 can be adjusted based on new eye measurements. For example, in one embodiment, neural network 920 is provided with pre- and post-operative refractive measurements of a patient who previously underwent laser-assisted in situ keratomileusis (LASIK). The post-refractive measurements can be considered as formula errors due to the previous refractive surgery. Based on the difference between the pre-operative measurements and the post-refractive measurements, neural network 920 can correct the results provided by IOL formula component 910.
[0057] Fig.10 is a flow chart illustrating an example method 1000 for providing a recommended intraocular lens power. The method 1000 may be performed by the system 900 .
[0058] In block 1010, method 1000 includes obtaining at least two ocular measurement parameters and a lens selection parameter for an eye. In one aspect, for example, Ul 500 may obtain at least two ocular measurement parameters and a lens selection parameter for an eye. In one embodiment, Ul 500 may obtain parameters for both eyes of a patient. In another embodiment, the measurement parameters may be obtained from an ocular measurement device 466. The ocular measurement parameters may include, for example, axial length, corneal power, corneal power index, and anterior chamber depth. In one aspect, the ocular measurement parameters may include intraoperative aberration measurements, such as spherical, cylindrical, and axial.
[0059] In block 1020, method 1000 includes determining an intraocular lens power based on a formula using at least two ocular measurement parameters. For example, formula component 910 can determine the intraocular lens power using a formula based on at least two ocular measurement parameters. For example, in block 1022, determining the intraocular lens power based on the formula can optionally include determining a relevant portion of a metasurface that includes an ideal or near-ideal portion of a plurality of intraocular lens selection formulas based on a range of at least two ocular measurement parameters that best fits each individual intraocular lens selection formula.
[0060] In block 1030, method 1000 may include determining an estimated error of a formula using a deep learning machine that has been trained with validated postoperative results including postoperative refraction corresponding to intraocular lens power. In one aspect, for example, neural network 920 may determine an estimated error of a formula. Neural network 920 may have been trained with a training set 930 including validated postoperative results including postoperative refraction corresponding to intraocular lens power. Validated postoperative results may be obtained from a measurement device such as an autorefractor or a wavefront analyzer.
[0061] In block 1040, method 1000 includes adjusting the lens selection parameter based on the estimated error. In one aspect, for example, formula component 910 can adjust the lens selection parameter based on the estimated error. For example, formula component 910 can subtract the estimated error from the lens selection parameter input by the user.
[0062] In block 1050, method 1000 includes redetermining the intraocular lens power based on the formula and the adjusted lens selection parameters. In one aspect, for example, formula component 910 can redetermine the intraocular lens power based on the formula and the adjusted lens selection parameters. For example, block 1050 can also include optional block 1022.
[0063] In block 1060, method 1000 may optionally include presenting a metasurface including a relevant portion on a display device. For example, U1 600 may present a metasurface including a relevant portion including at least two measurement parameters and an intraocular lens power.
[0064] Now go to Fig.11, the example device 1100 can combine various parts of the system 900 with a medical diagnostic device to provide a single device that implements all or part of the system 900. For example, the device 1100 can recommend an intraocular lens power based on a formula and a trained deep learning machine 1112. It should be understood that although in the example implementation, the device 1100 can physically include multiple components within a single housing 1190, the device 1100 can also be implemented as interconnected components that can be physically located in the same location or not. For example, in one aspect, in addition to or as an alternative to using postoperative measurements locally, the verified results from multiple devices 1100 can provide postoperative measurements to a database located in a network. Some components of the device 1100 can be implemented as computer executable instructions stored on a computer readable medium such as a memory 1106. These instructions can be executed by a processor 1104. In one aspect, the processor 1104 and the memory 1106 can be located within the housing 1190. Additionally, device 1100 may include a display 1108, which may display user interface 1102. Display 1108 may be a touch-sensitive screen that receives input from a user. Alternative or additional input / output devices may be included.
[0065] In one aspect, the example device 1100 may include a biometer 1170 and an auto-optometer 1180 for obtaining eye measurements. The biometer 1170 may obtain physical characteristics for the eye, such as, but not limited to, corneal power, axial length, anterior chamber depth, corneal power index, white-to-white distance, and / or lens thickness. The auto-optometer 1180 may obtain optical measurements for the eye, such as, but not limited to, refraction, sphere, cylinder, and axial of the eye. In one aspect, the auto-optometer 1180 may perform other visual assessment functions, including aberration measurement, topography measurement, keratometry, and pupil measurement. For example, the auto-optometer 1180 may include a wavefront analyzer, or wavefront analyzer. Traditionally, biometers and auto-optometers are separate devices used for different purposes. For example, a biometer may be used to obtain measurements of an intraocular lens, while an auto-optometer may be used to estimate a patient's eyeglass or contact lens prescription. In one aspect, the biometer 1170 and the auto-refractor 1180 can be located in separate sensor heads of the device 1100. The device 1100 can include a single chin rest for positioning the patient relative to one of the sensor heads. Each sensor head can be moved relative to the chin rest to position the head to obtain corresponding measurements. It should be understood that various alternative physical arrangements of the biometer 1170 and the auto-refractor 1180 can be constructed.
[0066] In one aspect, the biometer 1170 and the autorefractor 1180 can store the measurements in the patient data storage device 1160. The patient data storage device 1160 can be a computer memory, preferably a non-volatile computer memory, such as a hard drive, a solid state drive, an EEPROM, etc. The biometer 1170 and the autorefractor 1180 can access the patient file in the patient data storage device 1160 and record the measurements directly. Such automatic recording can reduce manual transcription errors. In an alternative embodiment, the patient data storage device 1160 can be stored externally, such as on a physician's patient management system or a network storage system, in which case the device 1100 can be in electronic communication with the external memory.
[0067] The device 1100 may include a user interface 1102. The user interface 1102 may guide a user (eg, a technician) to obtain measurements of a patient by operating the biometer 1170 and the autorefractor 1180. The user interface 1102 may also include a user interface similar to the user interface 500 ( Figure 5 ) and 600( Figure 6 ). User interface 1102 can automatically import measurements from patient data storage device 1160 into input area 510. User interface 1102 can receive user input for A-constant area 516 and target refraction area 517. User interface 1102 can provide the same information as in user interface 600.
[0068] Formula component 1110 can be similar to formula component 910. Formula component 1110 can receive ocular measurements directly from biometer 1170 or from patient data storage device 1160. Formula component 1110 can implement any intraocular lens power determination formula described herein or known in the art.
[0069] The formula component 1110 can provide the determined intraocular lens power value along with all input measurements and parameters to the neural network 1120. The learning machine 1112 can be similar to the learning machine 912 and include, for example, the neural network 1120. The learning machine 1112 can use deep learning techniques to predict the error of the formula component 1110 based on a training set 1130 including postoperative results. In one aspect, the training set 1130 can include postoperative results received from the autorefractor 1180. In one aspect, the training set 1130 can include postoperative results only from the autorefractor 1180, so that the trained learning machine 1112 is specific to the device 1100. That is, by training the learning machine 1112 based on input measurements and postoperative results from a single device, the device 1100 can be calibrated to correct previous errors. In another aspect, the training set 1130 can be combined with other verified results 1162 (e.g., results from other devices 1100 that can be remotely located). The management portal 1140 can receive and authenticate the validated results 1162, for example, from a trusted network service. In one aspect, a user can access the access control 1150 via the user interface 1102 to specify which training set 1130 to use.
[0070] On the other hand, some functions of the device 1100 can be performed via a network service 1142. For example, the formula component 1110 can be regularly updated based on the results of machine learning performed remotely. For example, the system 900 can provide a network service 1142. The system 900 can, for example, periodically generate updated formulas based on the learning machine 912, and provide the updated formulas to the formula component 1110 via the network service 1142. In this case, the learning machine 1112 and the training set 1130 can be located remotely (e.g., as the learning machine 912 and the training set 930). The management portal 1140 can be used to receive the updated formula component 1110.
[0071] In another aspect, the device 1100 may retain a local learning machine 1112, which may be trained by the network service 1142 or the system 900. The device 1100 may send the relevant pre-operative and post-operative measurements to the network service 1142 via the management portal 1140. The system 900 may then train the learning machine 912 based on the training set 930, which includes the relevant pre-operative and post-operative measurements of the device 1100. The system 900 may then provide the trained learning machine 1112 to the device 1100 for installation. Thus, the device 1100 may utilize the trained learning machine 1112 without training and without access to the verified results of other devices 1100, which may include confidential data or protected health information (PHI).
[0072] Fig.12 1 is a flow chart illustrating an example method 1200 for providing a recommended intraocular lens power. The method 1200 may be performed by the apparatus 1100. The method 1200 may include some blocks similar to the method 1000. It should be understood that the methods 1000 and 1200 may be combined. For the sake of brevity, the description of some repeated blocks is omitted. In addition, as described above, the system 900 may perform some optional blocks of the method 1200.
[0073] In block 1210, method 1200 includes acquiring, by a biometer, at least two ocular measurement parameters for an eye. In one aspect, for example, the biometer 1170 can acquire at least two ocular measurement parameters and a lens selection parameter for the eye. In one embodiment, the biometer 1170 can acquire parameters for both eyes of the patient. The ocular measurement parameters can include, for example, axial length, corneal power, corneal power index, and anterior chamber depth. In one aspect, the ocular measurement parameters can include intraoperative aberration measurements, such as spherical, cylindrical, and axial.
[0074] In block 1220, method 1200 includes obtaining lens selection parameters for the eye. In one embodiment, user interface 1102 can obtain lens selection parameters for the eye. For example, the lens selection parameters can be a target refraction for the eye after insertion of an intraocular lens. The lens selection parameters can be input by a technician or an ophthalmologist.
[0075] In block 1230, the method 1200 includes determining the intraocular lens power using a formula based on at least two ocular measurement parameters.For example, the formula component 1110 can determine the intraocular lens power using a formula based on at least two ocular measurement parameters.
[0076] In block 1240, the method 1200 may include obtaining a post-operative refraction of the eye from an auto-optometer communicatively coupled to the biometer. In one aspect, for example, the auto-optometer 1180 may obtain the post-operative refraction of the eye. The auto-optometer 1180 may be coupled to the biometer 1170 such that the same device is used to obtain at least two ocular measurement parameters and the post-operative refraction. Additionally, the measurements may be stored in the common patient data storage device 1160.
[0077] In block 1250, the method 1200 may include associating at least two ocular measurement parameters, intraocular lens power and postoperative refraction, as a training set. Since the biometer 1170 and the autorefractor 1180 are communicatively coupled, the training set may be associated directly from the device without the need for manual data entry, which may result in transcription errors. In addition, consistency may be improved by generating multiple data sets using a known pair of the biometer 1170 and the autorefractor 1180.
[0078] In box 1260, the method 1200 optionally includes training a deep learning machine using the postoperative refraction of the eye and the intraocular lens power to determine the estimated error of the formula. In one aspect, the device 1100 can train the learning machine 1112 using the postoperative refraction of the eye and the intraocular lens power. The learning machine 1112 can be trained to estimate the error of the formula component 1110 for a specific set of input parameters including at least two eye measurement parameters and a lens selection parameter. The learning machine 1112 can be trained by providing a training set labeled with postoperative refraction as a result. It should be understood that the learning machine 1112 can be trained with training data from a previous process. Since the postoperative refraction is not available, at least two eye measurement parameters of the current process may not be included in the training data. Once the postoperative refraction is available, the learning machine 1112 can be further trained or retrained using the complete training set. In one aspect, box 1260 can be performed by an external system such as system 900, which can communicate with the device 1100 via a network service 1142. The device 1100 may receive the trained learning machine 1112 via the network service 1142 .
[0079] In block 1270, method 1200 may optionally include determining an estimated error of the formula using a deep learning machine trained with validated postoperative results including postoperative refraction corresponding to intraocular lens power. In one aspect, for example, learning machine 1112 may determine an estimated error of the formula. As described above, learning machine 1112 may have been trained with training set 1130 or 930 including validated postoperative results including postoperative refraction corresponding to intraocular lens power.
[0080] In block 1280, method 1200 can optionally include adjusting the lens selection parameters based on the estimated error. In one aspect, for example, formula component 1110 can adjust the lens selection parameters based on the estimated error. For example, formula component 1110 can subtract the estimated error from the lens selection parameters input by the user.
[0081] In block 1290, the method 1200 can optionally include redetermining the intraocular lens power based on the formula and the adjusted lens selection parameters. In one aspect, for example, the formula component 1110 can redetermine the intraocular lens power based on the formula and the adjusted lens selection parameters.
[0082] Fig.131 is a conceptual diagram illustrating an example use scenario of an example device 1100. The device 1100 may be referred to as a self-calibrating biometer. The device 1100 may automatically collect objective data that may be used to calibrate the device 1100. For example, deep learning may be used to analyze postoperative results acquired via the autorefractor 1180 to improve the IOL calculation of the device 1100. As additional data is collected, the device 1100 may be continuously improved.
[0083] Although various aspects of the present disclosure have been described in conjunction with its examples, it will be appreciated by those skilled in the art that changes and modifications may be made to various aspects of the present disclosure described above without departing from the scope of the present disclosure. Other aspects will be apparent to those skilled in the art from consideration of the specification or from practice of various aspects of the disclosure disclosed herein.
Claims
1. A device for selecting an intraocular lens, comprising: a biometer configured to obtain at least two ocular measurement parameters for the eye; a user interface configured to obtain lens selection parameters for the eye; an autophore configured to obtain a postoperative refraction of the eye; Memory; as well as a processor communicatively coupled to the biometer, the autorefractor, the user interface, and the memory, and configured to: determining an intraocular lens power for an intraocular lens for the eye based on a formula using the at least two ocular measurement parameters; associating the at least two ocular measurement parameters, the intraocular lens power, and the postoperative refraction into a training set; determining an estimated error of the formula applied to the eye using a deep learning machine trained with validated postoperative results including postoperative refraction corresponding to intraocular lens power; adjusting the lens selection parameter based on the estimated error; as well as The final intraocular lens power is re-determined based on the formula and the adjusted lens selection parameters.
2. The apparatus of claim 1, wherein the processor is configured to: The deep learning machine is retrained using the training set to determine an estimated error of the formula for another program.
3. The apparatus of claim 1, wherein the processor is configured to include the final intraocular lens power in the training set.
4. The apparatus of claim 1, wherein the lens selection parameter is one of a target refraction or an A-constant.
5. The apparatus of claim 1, wherein the at least two ocular measurement parameters are selected from the group consisting of: axial length, corneal power, corneal power index, and anterior chamber depth.
6. The device according to claim 1, wherein the selection formula of the lens includes one or more of the following: Hoffer Q formula, Holladay I formula, Haigis formula and SRK / T formula, Barrett Universal II formula.
7. The apparatus of claim 1, wherein the ocular measurement parameters include intraoperative aberration measurements.
8. The apparatus according to claim 1, further comprising: A display device, wherein the processor is configured to present the intraocular lens power on a relevant portion of a metasurface that includes an ideal or nearly ideal portion of a plurality of intraocular lens selection formulas, wherein the plurality of intraocular lens selection formulas are based on a range of the at least two ocular measurement parameters that best suits each individual intraocular lens selection formula.
9. A method for selecting an intraocular lens, comprising: obtaining at least two ocular measurement parameters of the eye by a biometer; obtaining lens selection parameters for the eye; determining an intraocular lens power for an intraocular lens for the eye based on a formula using the at least two ocular measurement parameters; obtaining a post-operative refraction of the eye from an autorefractor communicatively coupled to the biometer; associating the at least two ocular measurement parameters, the intraocular lens power, and the postoperative refraction into a training set; determining an estimated error of the formula applied to the eye using a deep learning machine trained with validated postoperative results including postoperative refraction corresponding to intraocular lens power; adjusting the lens selection parameter based on the estimated error; as well as The final intraocular lens power is re-determined based on the formula and the adjusted lens selection parameters.
10. The method according to claim 9, further comprising The deep learning machine is retrained using the training set to determine an estimated error of the formula for another program.
11. The method of claim 9, wherein associating the at least two ocular measurement parameters, the intraocular lens power, and the postoperative refraction as a training set comprises: The final intraocular lens power is included in the training set.
12. The method of claim 9, wherein the lens selection parameter is one of a target refraction or an A-constant.
13. The method of claim 9, wherein the at least two ocular measurement parameters are selected from the group consisting of: axial length, corneal power, corneal power index, and anterior chamber depth.
14. The method according to claim 9, wherein the lens selection formula includes one or more of the following formulas: Hoffer Q formula, Holladay I formula, Haigis formula, SRK / T formula, Barrett Universal II formula.
15. The method of claim 9, wherein the ocular measurement parameters include intraoperative aberration measurements.
16. A non-transitory computer-readable medium storing computer-executable instructions, the non-transitory computer-readable medium comprising instructions for causing a computer to execute: obtaining at least two ocular measurement parameters of the eye by a biometer; obtaining lens selection parameters for the eye; determining an intraocular lens power for an intraocular lens for the eye based on a formula using the at least two ocular measurement parameters; obtaining a post-operative refraction of the eye from an autorefractor communicatively coupled to the biometer; associating the at least two ocular measurement parameters, the intraocular lens power, and the postoperative refraction into a training set; determining an estimated error of the formula applied to the eye using a deep learning machine trained with validated postoperative results including postoperative refraction corresponding to intraocular lens power; adjusting the lens selection parameter based on the estimated error; as well as The final intraocular lens power is re-determined based on the formula and the adjusted lens selection parameters.
17. The non-transitory computer readable medium of claim 16, further comprising instructions that cause the computer to: The deep learning machine is retrained using the training set to determine an estimated error of the formula for another program.
