Arch height prediction method and system based on parameter model multi-coordination

Abstract

The invention relates to an arch height prediction method and system based on parameter model multi-coordination. The arch height prediction method comprises the following steps: selecting and collecting related parameters of a patient conforming to a sample standard and an implantable phakic intraocular lens of the patient; performing correlation matrix and multiple linear regression analysis on the parameters in the step 1 by utilizing statistical software, and establishing a new model Ma; establishing a comparison model library, wherein the comparison model library comprises a common place and a formula or a model Mb for predicting the arch height; the model Ma and the model Mb are compared to predict an arch height value and an actual arch height value, variance analysis is adopted, an LSD method is used for pairwise comparison, a Bland-Altman method is used for consistency check, and therefore the effectiveness and the accuracy of the model Ma are verified; setting the verified model as Mn, and adding the Mn into the model library in the step 3; repeatedly executing the model for N times until a newest model Mnow is obtained; pre-operation related parameter values are applied to the model M, and a post-operation arch height value is obtained through prediction and analysis. According to the scheme, the comparison model library can be increased along with the increase of the number of the sample groups, so that cyclic comparison verification can be realized, the obtained arch height prediction model has a good prediction effect, and the result is relatively accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of vault height prediction methods, and in particular to a vault height prediction method and system based on multi-coordination of parametric models. Background Art

[0002] The implantation of a posterior chamber refractive phakic intraocular lens with a central hole (Collamer Implantable Contact Lens, ICL V4c) is a safe and effective myopia correction solution. Different from corneal laser and other surgical methods that cut corneal tissue, the ICL implantation does not cut the cornea. Instead, the intraocular lens is implanted into the eye without damaging the cornea, and it has good safety, effectiveness, and stability. The visual quality after correction is better, which is a new trend in myopia correction technology.

[0003] The vault height is an important indicator for the success of ICL implantation. The vault height is the height from the center of the posterior surface of the ICL optical zone to the anterior surface of the natural lens. If the vault height is too high after surgery, it may cause high intraocular pressure and angle-closure glaucoma; if the vault height is too low after surgery, it may cause subcapsular opacity of the lens and lens rotation. Therefore, correctly predicting the postoperative vault height before surgery has important value and significance. Currently, the prediction of the vault height mostly relies on doctors' experience and the NK Formula and KS Formula in the optical coherence tomography system. However, these two methods have relatively large errors in predicting the postoperative vault height. Summary of the Invention

[0004] In view of the technical problem of corresponding errors existing in the prior art, the present invention provides a vault height prediction method and system based on multi-coordination of parametric models.

[0005] The technical solution adopted by the present invention is as follows: A vault height prediction method based on multi-coordination of parametric models specifically includes the following steps:

[0006] Step 1: Select and collect patients who meet the sample criteria and relevant parameters of their implantable phakic intraocular lenses.

[0007] Step 2: Use statistical software to perform correlation matrix and multiple linear regression analysis on the parameters described in Step 1 to establish a new model Ma.

[0008] Step 3: Establish a comparison model library, which contains common formulas or models Mb used to predict the vault height.

[0009] Step 4: Compare the model Ma with Mb to predict the vault height value and the actual vault height value. Use analysis of variance, pairwise comparison using the LSD method, and consistency test using the Bland-Altman method to verify the effectiveness and accuracy of the model Ma; set the verified model as Mn and add it to the model library described in Step 3.

[0010] Step 5: Repeat the above model N times until the latest model M is obtained. 现 ;

[0011] Step 6: Apply the preoperative relevant parameter values to model M 现 , and predict and analyze to obtain the postoperative arch height value.

[0012] Furthermore, the patient information includes age, diopter, spherical lens parameter, cylindrical lens parameter, intraocular pressure, anterior chamber volume, anterior chamber depth, anterior chamber angle, and pupil diameter.

[0013] Furthermore, the passing of the verification described in step 4 is based on whether the deviation and variance values obtained by using the analysis of variance, LSD method, and Bland - Altman comparison simultaneously fall within the preset range.

[0014] An arch height prediction system based on multi - coordination of parameter models, comprising:

[0015] A human - machine interaction module, used for inputting data, formulas or models, and instructions;

[0016] A storage module, used for storing data, formulas or models;

[0017] A data analysis module, used for retrieving data, formulas or models from the storage module and running instructions;

[0018] A control module, used for controlling the human - machine interaction module, storage module, and data analysis module;

[0019] An output module, used for reading data or information.

[0020] Furthermore, the carrier of the module is a hardware device or software embedded in the hardware device.

[0021] Furthermore, the hardware device includes a mobile communication device, a computer or a single - chip microcomputer.

[0022] The beneficial effects of the present invention are:

[0023] The prediction method and system proposed in this solution have good prediction effects and relatively accurate results. Description of the drawings:

[0024] Figure 1 is the flow chart of the present invention;

[0025] Figure 2 is the correlation matrix analysis chart of the present invention;

[0026] Figure 3 is the regression analysis training data in the present invention;

[0027] Figure 4 This is the normal P-P plot of the standardized residuals of the multiple linear regression formula for the vault height after implantation of phakic posterior chamber intraocular lens in myopic patients in the present invention;

[0028] Figure 5 This is the Bland-Altman agreement detection plot in the present invention;

[0029] Figure 6 This is the Bland-Altman agreement detection plot of the NK formula model in the present invention;

[0030] Figure 7 This is the Bland-Altman agreement detection plot of the KS formula model in the present invention;

[0031] Figure 8 This is the pairwise comparison plot of the LSD method in the present invention;

[0032] Figure 9 This is the plot showing that there are statistically significant differences in the vault height predicted by the KS model in the present invention.

[0033] Among them, Figure 1 The content within the trapezoidal frame is a process that can be cycled and repeatedly executed.

[0034] Figure 3 In, age refers to age, SE refers to spherical equivalent, W-T-W refers to white-to-white diameter (corneal diameter), ACA refers to anterior chamber angle, ACD refers to anterior chamber depth, ACV refers to anterior chamber volume, and size ICL refers to ICL size. Detailed implementation manners

[0035] Next, in combination with the embodiments of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0036] Refer to Figure 1 - 7 , to solve the problems existing in the background technology, the present application proposes the following technical solution: a vault height prediction method based on multi-coordination of parameter models, specifically including the following steps:

[0037] Step 1: Select and collect patients meeting the sample criteria and the relevant parameters of their phakic intraocular lenses;

[0038] Step 2: Use statistical software to perform correlation matrix and multiple linear regression analysis on the parameters described in Step 1 to establish a new model Ma;

[0039] Step 3: Establish a comparison model library, which contains common formulas or models Mb for predicting the vault height;

[0040] Step 4: Compare model Ma with Mb to predict the vault height value and the actual vault height value. Use analysis of variance, pairwise comparison using the LSD method, and consistency test using the Bland - Altman method to verify the effectiveness and accuracy of model Ma; Set the verified model as Mn and add it to the model library described in Step 3;

[0041] Step 5: Repeat the above model N times until the latest model M is obtained 现 ;

[0042] Step 6: Apply the preoperative relevant parameter values to model M 现 , and predict and analyze to obtain the postoperative vault height value.

[0043] Patients who underwent posterior chamber intraocular lens implantation for phakic eyes aged 18 - 45 years were included. Inclusion criteria: Having an obvious willingness to remove glasses, the myopia degree has been stable in the past 2 years, with a change of no more than -0.50D; The myopic spherical lens degree is between -3.00D and -18.00D, and the cylindrical lens degree is between -1.00D and -4.00D. Exclusion criteria: Active ocular inflammation or infection; Glaucoma, cataract; Psychiatric diseases such as depression. We collected data such as anterior chamber parameters, refractive power, age, vault height, and lens size of 208 eyes of patients 6 months after surgery, and obtained the model through statistical software calculation; Collected the vault height of 47 eyes 6 months after surgery for model verification. The main relevant factors were obtained by correlation matrix analysis, and the model was obtained by multiple linear regression equation analysis.

[0044] Use analysis of variance, pairwise comparison using the LSD method (Least Significant Difference method) for comparison with the existing model, and Bland - Altman for consistency test (Note: The Bland - Altman plot is a visual display method for consistency measurement. After calculating the relevant measurement data, it is displayed as scatter points. If the scatter points are within the confidence interval {generally within 1.96 standard deviations of the difference}, it means that the data has a good consistency level).

[0045] Among them, the correlation matrix, also called the correlation coefficient matrix, is composed of the correlation coefficients between the columns of the matrix.

[0046] Then perform multiple linear regression analysis. In regression analysis, if there are two or more independent variables, it is called multiple regression. In fact, a phenomenon is often related to multiple factors. Predicting or estimating the dependent variable by the optimal combination of multiple independent variables is more effective and more in line with reality than using only one independent variable for prediction or estimation, so as to be able to predict the vault height more accurately.

[0047] Among them, the data for analysis of variance needs to meet the requirements of independence, normality, and homoscedasticity. Therefore, before performing analysis of variance, it is necessary to test the normality and homoscedasticity of the data.

[0048] In further designs, patient information includes age, diopter, spherical lens parameter, cylindrical lens parameter, intraocular pressure, anterior chamber volume, anterior chamber depth, anterior chamber angle, pupil diameter, and so on. Multiple other pieces of information beneficial for prediction can also be set.

[0049] It also includes performing consistency test through Bland - Altman, integrating the multiple predicted elevation information in step S103, then taking out a close elevation value and excluding other values. At the same time, ROC curve accuracy test can also be performed.

[0050] Based on the above analysis, in the present invention, a large amount of patient information and the crystal parameters of the implanted intraocular lens are obtained, using correlation matrix and multiple linear regression analysis, then performing analysis of variance, comparing the predicted elevation value and the actual elevation value through the models in current literature, using analysis of variance, pairwise comparison using the LSD method, and using the Bland - Altman method for consistency test, so as to verify the effectiveness and accuracy of the model, and thus establish a prediction model; then outputting the elevation information after implanting the crystal, and the prediction model analyzes to obtain the final elevation prediction information. Therefore, the prediction model has a good prediction effect and the results are relatively accurate.

[0051] In this embodiment, the criteria for patient information are as follows:

[0052] Parameter Average value (range) Eye (n) 208 Age (years) 25.6±4.72(18～40) Spherical lens -8.73±3.41(－3.00～－24.00) Cylindrical lens -1.09±0.75(0.00～－4.50) Equivalent spherical lens -8.77±3.61(－3.00～－24.00) Intraocular pressure (mmHg) 15.49±2.55(9.3～22.0) White - to - white (mm) 11.5±0.37(10.6～12.6) Anterior chamber volume (mm3) 203.41±27.67(141～278) Anterior chamber depth (mm) 3.22±0.21(2.80～3.75) Anterior chamber angle (°) 40.04±4.98(26.3～53.4) Pupil (mm) 3.18±0.62(2.12～5.39) ICL size (mm) 12.7±0.29(12.1～13.7) Toric ICL / ICL 76 / 132 12.1:12.6:13.2:13.7 14:133:57:4

[0053] Experimental example:

[0054] By including patients with phakic posterior chamber intraocular lens implantation aged 18 - 45 years.

[0055] Inclusion criteria: Having an obvious willingness to remove glasses, the myopia degree being stable in the past two years with a change of no more than -0.50D; the myopia spherical lens degree being between -3.00D and -18.00D, and the cylindrical lens degree being between -1.00D and -4.00D.

[0056] Exclusion criteria: Active ocular inflammation or infection; glaucoma, cataract; mental diseases such as depression. We collected data such as anterior chamber parameters, diopter, age, elevation, and lens size of 210 eyes of patients 6 months after surgery, and obtained the model through calculation in statistical software; collected the elevation of 47 eyes 6 months after surgery for model verification.

[0057] The main relevant factors are obtained through correlation matrix analysis, and the model is obtained through multiple linear regression equation analysis. Analysis of variance is used, and pairwise comparisons are made with the existing model using the LSD method (Least Significant Difference method), and Bland-Altman is used for consistency testing {Note: The Bland-Altman plot is a visual display method for consistency measurement. After calculating the relevant measurement data, it is displayed as scatter points. If the scatter points are within the confidence interval (generally within 1.96 standard deviations of the difference), it indicates that the data has a good level of consistency}.

[0058] Arch height

[0059] =-9.391*age - 8.787*SE + 5.573*ACA + 201.152*size - 67.607*WTW - 1.821*ACV + 509.775*ACD - 2579.972

[0060] Figure 4 It is the normal P-P plot of the standardized residuals of the multiple linear regression formula for the arch height after phakic posterior chamber intraocular lens implantation in 208 myopic eyes. It shows that the expected cumulative probability and the measured cumulative probability of this regression formula have good consistency, indicating that this formula has good prediction accuracy.

[0061] In this embodiment, it further includes an arch height prediction system based on multi-coordination of parametric models;

[0062] Specifically, it includes:

[0063] A human-computer interaction module, used for inputting data, formulas or models, and instructions;

[0064] A storage module, used for storing data, formulas or models;

[0065] A data analysis module, used for retrieving data, formulas or models from the storage module and running instructions;

[0066] A control module, used for controlling the human-computer interaction module, the storage module and the data analysis module;

[0067] An output module, used for reading data or information.

[0068] A control module, used for controlling the human-computer interaction module, the storage module and the data analysis module; the control module can be a circuit including at least one processor, or a circuit including at least one single-chip microcomputer, or a combined form of multiple circuits or chips, as long as the corresponding functions can be realized. It can be understood that for those skilled in the art, the control module can also be a common circuit composed of amplifiers, comparators, triodes, MOS tubes, etc. to realize the corresponding functions in a pure hardware manner.

[0069] The system can be executed on a computer installed with SPSS Statistics 25 and GraphPad Prism 9 software. Then, the human-computer interaction module is the computer's mouse, keyboard, etc.; the storage module is the computer hard drive; the data analysis module is the computer system running software, the control module is the computer processor, and the output module is the monitor.

[0070] Among them, referring to Figure 5 , the arch height information of the reference experimental example is as follows:

[0071] =-9.391*age - 8.787*SE + 5.573*ACA + 201.152*size - 67.607*WTW - 1.821*ACV + 509.775*ACD - 2579.972

[0072] In summary, referring to Figure 6 , so it shows that the prediction model has a good prediction effect and the results are relatively accurate.

[0073] Therefore, the present invention obtains a large amount of patient information and the crystal parameters of the implanted intraocular lens, uses the correlation matrix and multiple linear regression analysis, and then conducts variance analysis. By comparing the predicted arch height value and the actual arch height value through the models in the current literature, using variance analysis, pairwise comparison using the LSD method, and using the Bland-Altman method for consistency testing, the effectiveness and accuracy of the model are verified, and then a prediction model is established; then the arch height information after implanting the crystal is output, and the prediction model analyzes to obtain the final arch height prediction information. Therefore, the prediction model has a good prediction effect and the results are relatively accurate.

[0074] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for predicting the arch height based on multi - coordination of parametric models, characterized in that: Specifically, it includes the following steps: Step 1: Select and collect several groups of relevant parameters of patients meeting the sample criteria and their implantable phakic intraocular lenses, with no less than 30 cases in a group; Step 2: Use statistical software to perform correlation matrix and multiple linear regression analysis on the parameter groups described in Step 1, and establish a new model Ma; Step 3: Establish a comparison model library, which contains common formulas or models Mb for predicting vault height; Step 4: Compare the predicted vault height values and actual vault height values of model Ma and Mb. Use analysis of variance, the LSD method for pairwise comparison, and the Bland - Altman method for consistency test to verify the effectiveness and accuracy of model Ma; Set the verified model as Mn and add it to the model library described in Step 3; Step 5: Repeat the above model N times until the latest model M is obtained 现 ; Step 6: Apply the preoperative relevant parameter values to the model M 现 , and predict and analyze to obtain the postoperative arch height value.

2. The method for predicting the arch height based on multi-coordination of parameter models according to claim 1, wherein: The relevant parameters described in Step 1 include age, diopter, spherical lens parameter, cylindrical lens parameter, intraocular pressure, anterior chamber volume, anterior chamber depth, anterior chamber angle, and pupil diameter.

3. A method for predicting the arch height based on multi - coordination of parameter models according to claim 1, characterized in that: The statistical software described in Step 2 is SPSS Statistics 25 and / or GraphPad Prism 9.

4. A method for predicting the arch height based on multi - coordination of parametric models according to claim 1, characterized in that: The model Mb described in Step 3 includes the commonly used NK and KS models, as well as several Mn models 4 described in Step 4.

5. A method for predicting the arch height based on multi-coordination of parametric models according to claim 1, characterized in that: The verification described in Step 4 is based on whether the deviation and variance values obtained by using the analysis of variance, LSD method, and Bland - Altman comparison simultaneously fall within the preset range.

6. An arch height prediction system based on multi-coordination of parameter models, characterized in that: It includes: A human - machine interaction module for inputting data, formulas or models, and instructions; A storage module for storing data, formulas or models; A data analysis module for retrieving data, formulas or models from the storage module and running instructions; A control module for controlling the human - machine interaction module, storage module, and data analysis module; An output module for reading data or information.

7. A height prediction system for an arch based on multi - coordination of parametric models according to claim 6, characterized in that: The carrier of the module is a hardware device or software embedded in the hardware device.

8. A height prediction system for an arch based on multi-coordination of a parameter model according to claim 7, characterized in that, The hardware device includes a mobile communication device, a computer, or a single - chip microcomputer.

Images