Artificial intelligence auxiliary presbyopia treatment classification system based on crystalline lens function and structural feature and application thereof

Abstract

The invention relates to an artificial intelligence-based presbyopia classification system, which is used for assisting presbyopia operation decision-making by comprehensively analyzing the function and structural characteristics of crystalline lenses. According to the system, an i-Trace visual function analyzer is used for evaluating a lens function parameter, namely a lens imbalance index (DLS), of a presbyopia crowd, and a lens structure parameter, namely an average lens density (LD-P), is measured through Pentacam and / or sweep-frequency optical coherence tomography (SS-OCT); (Lens density by SS-OCT, LD-S) is adopted in the method. A clustering method assisted by artificial intelligence is adopted, presbyopia groups are divided into a non-crystalline lens surgery group and a crystalline lens surgery group, and an optimal diagnosis boundary value is calculated through a subject working characteristic curve. A scoring system is constructed to find that the classification diagnosis efficiency of the DLI combined with the LD-P is the highest, and the LD-S can be used as an auxiliary diagnosis index. In addition, the invention further comprises an intelligent outpatient service system for presbyopia operation decision making. The invention provides an accurate diagnostic tool and aims to optimize the decision-making process of presbyopia surgery.

Description

Technical Field

[0001] The present invention relates to the technical field of medical auxiliary diagnosis, and in particular to an artificial intelligence (AI)-assisted presbyopia treatment classification system based on lens function and structural characteristics and its application. Background Art

[0002] Presbyopia affects approximately 1.8 billion people worldwide, and the potential productivity loss caused by uncorrected or undercorrected presbyopia is estimated to be US$25.367 billion, accounting for 0.037% of global GDP. 1,2 As the severity of presbyopia increases, the need for surgery to remove lenses and improve quality of life is increasing. 6 There is no consensus on the choice of surgical options for presbyopic people. The timing and criteria for surgery are unclear, leading to different choices among refractive and cataract specialists.

[0003] The key factor in choosing refractive or lens surgery is a comprehensive evaluation of lens function and structure to obtain an objective classification of presbyopia. 4,16 The i-Trace DLI provides an objective assessment of lens function by taking into account higher-order aberrations (HOA), pupil size, and contrast sensitivity data. 17,18 In addition, Pentacam utilizes Scheimpflug images to generate high-resolution 3D images, providing precise evaluation of the lens structure. 19,20 . Swept-Source Optical Coherence Tomography (SS-OCT) can also help assess lens morphology and density. Comprehensive analysis using artificial intelligence may enable accurate classification of presbyopia 21 This invention introduces an innovative AI-based presbyopia classification system that integrates lens functional and structural features to aid surgical decision making. Summary of the invention

[0004] The purpose of the present invention is to address the deficiencies in the prior art and to provide an artificial intelligence-assisted presbyopia treatment classification system based on lens function and structural characteristics and its application.

[0005] In a first aspect, the present invention provides: an artificial intelligence-based presbyopia classification system, characterized in that: it includes constructing a comprehensive measurement method for the lens function, structure and imaging of the presbyopic population; applying an AI-assisted clustering method, integrating lens structure and function indicators, accurately classifying the presbyopic population, and deriving a diagnostic cutoff value; and constructing a scoring system.

[0006] As a preferred example, the comprehensive measurement method of lens function, structure and imaging of healthy and presbyopic people is constructed: including

[0007] (1) Best-corrected distance visual acuity (BCDVA), distance-corrected near visual acuity (DCNVA), addition (ADD), and intraocular pressure (IOP);

[0008] (2) anterior and posterior segment photography and Lens Opacities Classification System III (LOCS III) grading;

[0009] (3) Lens biometry: axial length (AL), anterior chamber depth (ACD), and lens thickness (LT);

[0010] (4) Lens function assessment: measurement of dysfunction lens index (DLI), higher order aberrations (HOA) and modulation transfer function (MTF);

[0011] (5) Lens structure measurement: Pentacam analysis of average lens density (Lens density by Pentacam, LD-P), Pentacam Nucleus Staging (Pentacam Nucleus Staging, PNS), swept-source optical coherence tomography (SS-OCT) to measure the radius of curvature of the anterior and posterior surfaces of the lens, and quantitative analysis of lens density in SS-OCT images (Lens density by SS-OCT, LD-S) by Image J to outline the boundaries of the cortex, adult nucleus, and inner nucleus, and automatically measure the average and maximum density values ​​of the regional area.

[0012] As a preferred example, the artificial intelligence-based presbyopia classification system uses an AI-assisted clustering method to comprehensively consider the lens structure and function indicators to accurately classify the presbyopic population, divide the patients into a non-lens surgery group (NLS) and a lens surgery group (LS), and calculate the optimal diagnostic threshold through the receiver operating characteristic (ROC) curve.

[0013] As a preferred example, the diagnostic specificity and sensitivity of different scoring systems were analyzed through ROC, and the optimal scoring combination with the highest diagnostic efficiency was selected according to the AUC and its corresponding 95% CI. The results showed that the classification diagnostic efficiency of DLI combined with LD-P was the highest, and LD-S could be used as an auxiliary diagnostic indicator. If the patient was classified into the LS group, lens surgery was recommended; otherwise, the patient was classified into the NLS group and refractive surgery was recommended.

[0014] Secondly, a method for automatic and accurate classification of presbyopia was constructed: its characteristics are that the lens structure and function index parameters are input, the AI ​​algorithm is used for accurate calculation, the grouping results are automatically output, and surgical suggestions are provided. This method integrates the contribution of lens function and structure parameters, calculates the contribution of each parameter to the whole and the proportion of individual examination data in the diagnostic threshold, and through ROC curve analysis, those with scores equal to or greater than the critical score are divided into the LS group, otherwise they are divided into the NLS group.

[0015] On the third aspect, the present invention provides an intelligent outpatient system for presbyopia. After the lens structure and function index parameters with the greatest contribution are automatically extracted in the outpatient system, the background performs precise calculations based on the AI ​​algorithm for accurate presbyopia classification, obtains grouping results, and outputs them to the outpatient system. Doctors are guided in the selection of surgical plans based on the grouping results.

[0016] The advantages of the present invention are: The purpose of the present invention is to develop an AI-based presbyopia classification system to comprehensively analyze the functional and structural characteristics of the lens to assist in surgical decision-making.

[0017] Comprehensive analysis: There is currently no consensus on the choice of surgical options for presbyopic people, and the timing and standards of surgery are still unclear. The present invention uses AI to conduct a comprehensive analysis of lens structure and function indicators.

[0018] Accurate classification: Through DLI and lens density values, patients can be accurately classified into NLS and LS groups.

[0019] Efficient diagnosis: Through AI-assisted cluster analysis, diagnostic thresholds, scoring systems, ROC and other screening methods, the best diagnostic threshold solution is found.

[0020] Automation: The developed automatic and accurate classification method provides a convenient automatic diagnosis tool.

[0021] Intelligence: Develop an intelligent presbyopia outpatient system and use the above grouping results for clinical surgical plan selection. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 Examination of lens function, structure and imaging.

[0023] Figure 2 Representative images and schematics of lens function and structure examinations.

[0024] Figure 3 Example of automatic and accurate classification of presbyopia. DETAILED DESCRIPTION

[0025] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content of the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the appended claims of the application equally.

[0026] 1. Methods

[0027] Comprehensive measurement of lens function, structure and imaging in healthy and presbyopic people:

[0028] (1) Best-corrected distance visual acuity (BCDVA), distance-corrected near visual acuity (DCNVA), addition (ADD), and intraocular pressure (IOP);

[0029] (2) anterior and posterior segment photography and Lens Opacities Classification System III (LOCS III) grading;

[0030] (3) Lens biometry: axial length (AL), anterior chamber depth (ACD), and lens thickness (LT);

[0031] (4) Lens function assessment: measurement of dysfunction lens index (DLI), higher order aberrations (HOA) and modulation transfer function (MTF);

[0032] (5) Lens structure measurement: Pentacam analysis of average lens density (Lens density by Pentacam, LD-P), Pentacam Nucleus Staging (Pentacam Nucleus Staging, PNS), swept-source optical coherence tomography (SS-OCT) to measure the radius of curvature of the anterior and posterior surfaces of the lens, and quantitative analysis of lens density in SS-OCT images (Lens density by SS-OCT, LD-S) by Image J to outline the boundaries of the cortex, adult nucleus, and inner nucleus, and automatically measure the average and maximum density values ​​of the regional area.

[0033] Input the lens structure and function index parameters, apply AI algorithm for precise calculation, automatically output grouping results and provide surgical recommendations.

[0034] Application Example 1

[0035] Female, 50 years old, clinically diagnosed with binocular presbyopia, auxiliary examination (right eye / left eye): BCDVA 0.8 / 0.8, DCNVA 0.5 / 0.5, ADD+1.50D / +1.75D, IOP 18 / 19mmHg, axial length 23.2 / 23.3mm, lens function measurement: DLI 8.6 / 9.2, lens density measurement: LD-P 8.7 / 8.75, LD-S 62.00 / 61.75 pixels. The above lens function and structure parameters are input, and the AI ​​algorithm is used for precise calculation. The output result is NLS grouping, and refractive surgery is recommended for presbyopia correction ( Figure 3 B).

[0036] Application Example 2

[0037] Male, 58 years old, clinically diagnosed with binocular presbyopia, auxiliary examination (right eye / left eye): BCDVA 0.6 / 0.5, DCNVA 0.4 / 0.3, ADD+2.25D / +2.75D, IOP 17 / 19mmHg, axial length 22.8 / 22.9mm, lens function measurement: DLI7.2 / 7.15, lens density measurement: LD-P 9.8 / 9.9, LD-S 63.25 / 64.25 pixels. Input the above lens function and structure parameters, apply AI algorithm for precise calculation, output the LS grouping, and recommend lens surgery to correct presbyopia ( Figure 3 C).

[0038] The present invention provides an artificial intelligence-based presbyopia classification system, which utilizes lens function and structural parameters to assist in the decision-making process of refractive or lens surgery. By combining the lens density cutoff value measured by Pentacam and / or SS-OCT with the cutoff value of DLI, the present invention can achieve rapid diagnostic classification. In addition, the present invention also includes an automatic classification method based on an AI algorithm, providing an accurate and automated classification tool to improve the diagnostic efficiency and accuracy of presbyopia. Finally, the present invention establishes an intelligent presbyopia clinic based on the above-mentioned AI algorithm to guide clinical surgical decisions.

[0039] References:

[0040] 1.Fricke TR,Tahhan N,Resnikoff S,et al.Global prevalence ofpresbyopia and vision impairment from uncorrected presbyopia.Ophthalmology2018;125(10):1492-9.doi:10.1016 / j.ophtha.2018.04.013

[0041] 2.Frick KD,Joy SM,Wilson DA,Naidoo KS,Holden BA.The global burden ofpotential productivity loss from uncorrected presbyopia.Ophthalmology 2015;122(8):1706-10.doi:10.1016 / j.ophtha.2015.04.014

[0042] 3.Mercer RN,Milliken CM,Waring GO,Rocha KM.Future trends inpresbyopia correction.J Refract Surg 2021;37(S1).doi:10.3928 / 1081597X-20210408-06

[0043] 4.Wolffsohn JS,Davies LN.Presbyopia:effectiveness of correctionstrategies.Prog Retin Eye Res 2019;68:124-43.doi:10.1016 / j.preteyeres.2018.09.004

[0044] 5.Orman B,Benozzi G.Pharmacological strategies for treatingpresbyopia.Curr Opin Ophthalmol 2021;32(4):319-23.doi:10.1097 / ICU.0000000000000770

[0045] 6.Khurana DA,Swathi N,Rajalakshmi AR.Factors influencing the need andwillingness for presbyopic correction:a cross sectional study from southindia.Sci Rep2023;13(1).doi:10.1038 / s41598-023-50288-w

[0046] 7.Waring GO,Rocha KM.Characterization of the dysfunctional lenssyndrome and areview of the literature.Curr Ophthalmol Rep 2018;6(4):249-55.doi:10.1007 / s40135-018-0190-3

[0048] 8.Fernández J,Rodríguez-Vallejo M,Martínez J,Tauste A, DP.Frompresbyopia to cataracts:a critical review on dysfunctional lens syndrome.JOphthalmol2018;2018:1-10.doi:10.1155 / 2018 / 4318405

[0049] 9.S M.Dysfunctional lens syndrome,a new way to educatepatients.Accessed 2024-2-7.https: / / www.aao.org / eyenet / academy-live / detail / dysfunctional-lens-syndrome-ducat e-patients

[0050] 10.AlióJL,Schimchak P,Negri HP,Montés-MicóR.Crystalline lens opticaldysfunction through aging.Ophthalmology 2005;112(11):2022-9.doi:10.1016 / j.ophtha.2005.04.034

[0052] 11.Ye Y,Zhao J,Zhang Z,et al.Long-term follow-up for monovisionsurgery byimplantable collamer lens v4c implantation for myopia correction inearly presbyopia.Graefe's Archive for Clinical and Experimental Ophthalmology2022.doi:10.1007 / s00417-021-05545-x

[0054] 12.Reinstein DZ,Macgregor C,Archer TJ,Gupta R,Potter JG.A review ofposteriorchamber phakic intraocular lenses.Curr Opin Ophthalmol 2024;35(2):138-46.doi:10.1097 / ICU.0000000000001018

[0055] 13.Ye Y,Zhang Z,Niu L,et al.Binocular imbalance in patients afterimplantablecollamer lens v4c implantation or femtosecond laser-assisted insitu keratomileusisfor myopia with presbyopia.Front Neurosci 2023;17.doi:10.3389 / fnins.2023.120479214.Ferguson TJ,Randleman JB.Cataract surgeryfollowing refractive surgery:principles to achieve optical success andpatient satisfaction.Surv Ophthalmol2024;69(1):140-59.doi:10.1016 / j.survophthal.2023.08.002

[0056] 15.Kohnen T,Berdahl JP,Hong X,Bala C.The novel optical design andclinicalclassification of a wavefront-shaping presbyopia-correctingintraocular lens.Clinical Ophthalmology(Auckland,N.Z.)2023;17:2449-57.doi:10.2147 / OPTH.S40008316.Donaldson PJ,Chen Y,Petrova RS,Grey AC,LimJC.Regulation of lens water content:effects on the physiological optics ofthe lens.Prog Retin Eye Res 2023;95:101152.doi:10.1016 / j.preteyeres.2022.101152

[0057] 17.Li Z,Yu L,Chen D,et al.Dysfunctional lens index serves as a novelsurgerydecision-maker for age-related nuclear cataracts.Curr Eye Res 2019;44(7):733-8.doi:10.1080 / 02713683.2019.1584676

[0058] 18.Martínez-Plaza E,Ruiz-Fortes P,Soto-Negro R,et al.Characterizationofdysfunctional lens index and opacity grade in a healthypopulation.Diagnostics(Basel)2022;12(5):1167.doi:10.3390 / diagnostics12051167

[0059] 19.Xu Y,Ye Y,Xian Y,Niu L,Zhou X,Zhao J.Comparison of corneal andlens densitymeasurements obtained by pentacam and casia2 in myopes.BmcOphthalmol 2023;23(1).doi:10.1186 / s12886-023-03199-3

[0060] 20.Ye Y, Zhao J, Niu L, Shi W, Wang X, Zhou X. Long-term evaluation of anterior lens density after implantable collamer lens v4c implantation inpatients with myopia over40years old.Br J Ophthalmol 2022;106(11):1508-13.doi:10.1136 / bjophthalmol-2021-319205

[0062] 21.Yousefi S, Elze T, Pasquale LR, et al.Monitoring glaucomatous functional loss using an artificial intelligence–enabledashboard.Ophthalmology2020;127(9):1170-8.doi:10.1016 / j.ophtha.2020.03.008

[0064] The above is only a preferred embodiment of the present invention. It should be pointed out that ordinary technicians in this technical field can make several improvements and supplements without departing from the method of the present invention. These improvements and supplements should also be regarded as the scope of protection of the present invention.

Claims

1. A presbyopia classification system based on artificial intelligence (AI), characterized by: This includes building a comprehensive measurement method for the lens function, structure, and imaging of presbyopic people; applying an AI-assisted clustering method to comprehensively classify the lens structure and function indicators to accurately classify the presbyopic population and derive diagnostic cutoffs; and building a scoring system.

2. The artificial intelligence-based presbyopia classification system according to claim 1, characterized in that: The above-mentioned comprehensive measurement method of lens function, structure and imaging of presbyopic people is constructed: including (1) Best-corrected distance visual acuity (BCDVA), distance-corrected near visual acuity (DCNVA), addition (ADD), and intraocular pressure (IOP); (2) anterior and posterior segment photography and Lens Opacities Classification System III (LOCS III) grading; (3) Lens biometry: axial length (AL), anterior chamber depth (ACD), and lens thickness (LT); (4) Lens function assessment: measurement of dysfunction lens index (DLI), higher order aberrations (HOA) and modulation transfer function (MTF); (5) Lens structure measurement: Pentacam analysis of average lens density (Lens density by Pentacam, LD-P), Pentacam Nucleus Staging (Pentacam Nucleus Staging, PNS), swept-source optical coherence tomography (SS-OCT) to measure the radius of curvature of the anterior and posterior surfaces of the lens, and quantitative analysis of lens density of SS-OCT images (Lens density by SS-OCT, LD-S) by Image J to outline the boundaries of the cortex, adult nucleus, and inner nucleus, and automatically measure the average and maximum density values ​​of the regional area.

3. The artificial intelligence-based presbyopia classification system according to claim 1, characterized in that: The AI-assisted clustering method is used to comprehensively consider the lens structure and function indicators to accurately classify the presbyopic population, and the patients are divided into the non-Lens Surgery Group (NLS) and the lens surgery group (LS), and the optimal diagnostic threshold is calculated by the Receiver Operating Characteristic (ROC) curve.

4. The artificial intelligence based presbyopia classification system according to claims 1 and 3, characterized in that: The diagnostic specificity and sensitivity of different scoring systems were analyzed by ROC, and the optimal scoring combination with the highest diagnostic efficiency was selected according to the AUC and its corresponding 95% CI. The results showed that the classification diagnostic efficiency of DLI combined with LD-P was the highest, and LD-S could be used as an auxiliary diagnostic indicator. If the patient was classified into the LS group, lens surgery was recommended; otherwise, the patient was classified into the NLS group, and refractive surgery was recommended.

5. A method for constructing automatic and accurate classification of presbyopia: characterized in that, Input lens structure and function index parameters, apply AI algorithm for accurate calculation, automatically output grouping results and provide surgical recommendations. This method integrates the contribution of lens function and structure parameters, calculates the contribution of each parameter to the overall and the proportion of individual examination data in the diagnostic threshold, and through ROC analysis, those with scores equal to or greater than the critical score are divided into the LS group, otherwise they are divided into the NLS group.

6. A smart presbyopia clinic system, characterized by: The lens structure and function index parameters with the greatest contribution obtained according to claims 1-5 are automatically extracted in the outpatient system, and the background performs precise calculations based on the AI ​​algorithm for accurate presbyopia classification to obtain grouping results, which are output to the outpatient system. Doctors use the grouping results to guide the selection of surgical plans.

Images