A novel fitting algorithm for alignment curve radius estimation using corneal elevation data in orthokeratology lens trial

Comparison of trial lens and computer-aided fitting in orthokeratology: A multi-center, randomized, examiner-masked, controlled study

PURPOSE

To compare the efficacy and safety between traditional lens fitting and computer-aided fitting methods for orthokeratology (OrthoK) in the Chinese population.

METHODS

A multi-center, examiner-masked, randomized controlled study was conducted with a one-year follow-up period, enrolling 280 participants with spherical equivalent (SE) ranging from -0.5D to -4.0D. Participants were assigned to either the computer-aided orthokeratology fitting group (trial group) or the traditional lens fitting group (control group) using stratified randomization based on age (8 to 13 years, 13 to 18 years, and ≥ 18 years) to ensure a minimum of 30 cases in each sub-age group. Ocular examinations included visual acuity, objective and subjective refraction, corneal endothelial cell density, corneal topography, intraocular pressure, axial length, and ocular health assessment. Successful lens-correction was defined as the residual refraction with the OK lens, which should not exceed ± 0.5D, and/or an uncorrected visual acuity of no worse than 0.1 logMAR. Statistical analysis involves t-tests, analysis of variance, and Chi-squared tests.

RESULTS

215 subjects were included in the statistical analysis (109 in the trial group and 106 in the control group). In both groups, compared to baseline data, the uncorrected visual acuity (UCVA) improved significantly, with SE reduced and central corneal curvature flattened greatly after wearing OrthoK lens (P < 0.05 for all). Compared to the control group, the trial group exhibited a higher successful rate in correcting UCVA (93.6 % vs. 84.0 %, P = 0.03) and slightly better correction in refraction (77.1 % vs. 66.0 %, P = 0.07) at 1-month follow-up. However, no significant differences were observed in the axial length elongation, corneal changes, or the incidence of adverse events between the two groups.

CONCLUSION

These findings indicate the higher efficiency and slightly better performance in correcting myopia and improving UCVA of computer-aided lens fitting approach compared to the traditional one, but similar outcomes in controlling axial elongation.

Effects of orthokeratology on corneal reshaping and the delaying of axial eye growth in children

Purpose

To investigate the inhibition of myopia progression and axial elongation in children wearing orthokeratology (OK) lenses, as well as to evaluate the status of corneal reshaping, this study explores the relationship between changes in central corneal curvature (K-value) and e-value induced by OK lenses and axial elongation.

Methods

In this study, it is planned to select children aged 8-15 who wear orthokeratology lenses at the Pediatric Ophthalmology and Strabismus Clinic of the Second Affiliated Hospital of Dalian Medical University. All children will undergo slit lamp examination, visual acuity assessment, computerized refraction, intraocular pressure measurement, biometry, and corneal topography examination before lens wear and at 1 month, 3 months, and 6 months after lens wear in the pediatric ophthalmology clinic. Based on age (lower age group (8 < age ≤12 years); higher age group (12 < age ≤15 years)) and baseline equivalent spherical (SE) value (mild myopia group (-1.00 D < SE ≤ -3.25D); moderate myopia group (-3.25 D < SE ≤ -6.00 D)), four groups will be formed by pairing these factors. Suitable data will be selected according to inclusion and exclusion criteria, and different groups will be included. Data will be organized, and statistical analysis will be performed using SPSS software to obtain the results. The expected results will be discussed and analyzed.

Results

After wearing OK lenses, all four groups achieved good visual acuity at follow-up. At 6 months, there were no significant differences in visual acuity among the four groups (P = 0.149, >0.05). There were no significant differences in refractive error among the four groups (P = 0.066, >0.05). Baseline axial length differed significantly among the four groups (P = 0.000, <0.001), with the LM group having longer axial length than the LL group (P < 0.001, paired samples t-test), and the HM group having longer axial length than the HL group (P < 0.001, paired samples t-test). However, there were no significant differences in axial length change compared to baseline among the groups at 1 month, 3 months, and 6 months (P 1 = 0.053; P 3 = 0.557; P 6 = 0.329, >0.05). Significant differences were observed in corneal flat K-value change compared to baseline among the four groups at 1 month, 3 months, and 6 months (P 1 = 0.001, P 3 = 0.001, P 6 = 0.004, <0.05). There were no significant differences in e-value change among the groups at 1 and 3 months (P 1 = 0.205, P 3 = 0.252, >0.05), but significant differences were found in e-value change compared to baseline at 6 months (P 6 = 0.010, <0.05). Multiple regression analysis with changes in central corneal flat K-value and e-value as independent variables and axial elongation as the dependent variable showed a correlation between e-value change at 6 months and axial elongation (P = 0.004, <0.05), indicating a negative correlation.

Conclusion

Orthokeratology (OK) lenses effectively improve myopic children's vision by reshaping the cornea, leading to reduced central corneal curvature and flattening of its anterior surface. The effectiveness of OK lenses is not significantly affected by age or initial myopia severity. Children of varying ages and myopia levels experience similar levels of axial length control with OK lens wear. Changes in corneal shape due to OK lenses affect axial elongation, with greater changes in corneal morphology associated with smaller increases in axial length.

Machine learning models for orthokeratology lens fitting and axial length prediction

PURPOSE

In order to improve the efficiency of orthokeratology (OK) lens fitting and predict the axial length after 1 year of OK lens wear, machine learning models were proposed.

METHODS

Clinical data from 1302 myopic subjects were collected retrospectively, and two machine learning models were implemented. Demographic and corneal topographic data were collected as input variables. The output variables were the parameters of the OK lens and the axial length after 1 year. Eighty percent of input variables was used as the training set and the remaining 20% was used as the validation set. The first alignment curve (AC1) of the OK lenses, deduced using machine learning models and formula calculation, were compared. Multiple regression models (support vector machine, Gaussian process, decision tree and random forest) were used to predict the axial length after 1 year. In addition, we classified data based on lens brand, and carried out more detailed parameter fitting and analysis for spherical and toric OK lenses.

RESULTS

The OK lens fitting model showed higher (R2  = 0.93) and lower errors (mean absolute error [MAE] = 0.19, mean square error [MSE] = 0.09) when predicting AC1, compared with the formula calculation (R2  = 0.66, MAE = 0.44, MSE = 0.25). The machine learning model still had high R2 values ranging from 0.91 to 0.96 when considering the brand and design of the OK lenses. Further, the R2 value for the axial length prediction model was 0.94, which indicated that the machine learning model had high accuracy and good robustness.

CONCLUSION

The OK lens fitting model and the axial length prediction model played an important role in guiding OK lens fitting, with high accuracy and robustness in prediction performance.

Clinical tool to measure fluorescein patterns in orthokeratology

Background

Orthokeratology (ortho-k) is an overnight clinical contact lens wear technique to correct myopia and to reduce myopia progression wearing reverse-geometry rigid gas-permeable lenses. The lens fitting procedure in clinical practice usually requires subjective assessment of fluorescein pattern (fluorescein "bull's eye" pattern). The aim of this study was to develop a novel tool for fluorescein pattern measurements to reduce subjective practitioner dependency, especially in inexperienced practitioners, in ortho-k practice.

Methods

A new MATLAB customized algorithm to measure the horizontal width of the four main zones of ortho-k fluorescein patterns (central bearing, tear reservoir, mid-peripheral bearing and edge lift) was designed. The algorithm was tested on a small image database consisting of 26 ortho-k fluorescein pattern images of 13 volunteers fitted with reverse geometry lenses (Seefree, Conoptica-Hecht Contactlinsen). The agreement between two independent observers and the ImageJ measurements was determined.

Results

The new clinical tool provided similar measurements to ImageJ software for the central bearing (4.20 ± 0.74 and 4.27 ± 0.69 mm; P = 0.21), tear reservoir (1.69 ± 0.41 and 1.69 ± 0.45 mm; P = 0.69), mid-peripheral bearing (1.17 ± 0.11 and 1.13 ± 0.10 mm; P < 0.01) and edge lift (0.48 ± 0.06 and 0.48 ± 0.06 mm; P = 0.81) zones. Good agreement between the software (limits of agreement lower than ±0.55 mm) and inter-observer measurements (limits of agreement lower than ±0.66 mm) was found.

Conclusions

The proposed clinical tool for semiautomatic fluorescein pattern measurements in ortho-k could help to reduce practitioner dependency in fluorescein pattern assessment with future potential to introduce prediction algorithms or artificial intelligence methods in myopia control management.

Machine learning for predicting the treatment effect of orthokeratology in children

PURPOSE

Myopia treatment using orthokeratology (ortho-k) slows myopia progression. However, it is not equally effective in all patients. We aimed to predict the treatment effect of ortho-k using a machine-learning-assisted (ML) prediction model.

METHODS

Of the 119 patients who started ortho-k treatment between January 1, 2019, and January 1, 2022, 91 met the inclusion criteria and were included in the model. Ocular parameters and clinical characteristics were collected. A logistic regression model with least absolute shrinkage and selection operator regression was used to select factors associated with the treatment effect.

RESULTS

Age, baseline axial length, pupil diameter, lens wearing time, time spent outdoors, time spent on near work, white-to-white distance, anterior corneal flat keratometry, and posterior corneal astigmatism were selected in the model (aera under curve: 0.949). The decision curve analysis showed beneficial effects. The C-statistic of the predictive model was 0.821 (95% CI: 0.815, 0.827).

CONCLUSION

Ocular parameters and clinical characteristics were used to predict the treatment effect of ortho-k. This ML-assisted model may assist ophthalmologists in making clinical decisions for patients, improving myopia control, and predicting the clinical effect of ortho-k treatment via a retrospective non-intervention trial.

Machine learning algorithm improves accuracy of ortho-K lens fitting in vision shaping treatment

PURPOSE

To construct a machine learning (ML)-based model for estimating the alignment curve (AC) curvature in orthokeratology lens fitting for vision shaping treatment (VST), which can minimize the number of lens trials, improving efficiency while maintaining accuracy, with regards to its improvement over a previous calculation method.

METHODS

Data were retrospectively collected from the clinical case files of 1271 myopic subjects (1271 right eyes). The AC curvatures calculated with a previously published algorithm were used as the target data sets. Four kinds of machine learning algorithms were implemented in the experimental analyses to predict the targeted AC curvatures: robust linear regression models, support vector machine (SVM) regression models with linear kernel functions, bagging decision trees, and Gaussian processes. The previously published calculation method and the novel machine learning method were then compared to assess the final parameters of ordered lenses.

RESULTS

The linear SVM and Gaussian process machine learning models achieved the best performance. The input variables included sex, age, horizontal visible iris diameter (HVID), spherical refraction (SER), cylindrical refraction, eccentricity value (e value), flat K (K1) and steep K (K2) readings, anterior chamber depth (ACD), and axial length (AL). The R-squared values for the output AC1K1, AC1K2 and AC2K1 values were 0.91, 0.84, and 0.73, respectively. The previous calculation method and machine learning methods displayed excellent consistency, and the proposed methods performed best based on flat K reading and e values.

CONCLUSIONS

The ML model can provide practitioners with an efficient method for estimating the AC curvatures of VST lenses and reducing the probability of cross-infection originating from trial lenses, which is especially useful during pandemics, such as that for COVID-19.

Corneal Elevation, Power, and Astigmatism to Assess Toric Orthokeratology Lenses in Moderate-to-High Astigmats

OBJECTIVES

Fitting philosophies for toric orthokeratology are based on elevation or corneal astigmatism, but it is unclear which is more effective. The purpose of this analysis was to further understand corneal shape and the relationship between peripheral elevation and central astigmatism in moderate-to-high astigmats.

METHODS

Corneal tomography was measured three times on the right eyes of 25 moderate-to-high refractive myopic astigmatic adults. Corneal astigmatism and elevation were calculated at 4-, 6-, and 8-mm chords. Subjects were fitted with toric orthokeratology lenses following the manufacturer's guidelines based on elevation. Twenty subjects completed 10 days of wear. A masked examiner assessed movement and centration via slitlamp videos and quantified treatment zone and decentration from tangential power difference tomography maps. Correlations between variables were assessed.

RESULTS

Average corneal astigmatism was 2.20±0.70 DC and peripheral elevation was 50.88±18.92 μm and they were strongly correlated (4 mm R2=0.96, 6 mm R2=0.92, 8 mm R2=0.86, all P<0.001). Each diopter of astigmatism equated to 25 μm of elevation at an 8-mm chord. Via slitlamp, average treatment zone area was 12.73±4.62 mm2 and 13 lenses decentered. From tomography, average treatment zone area was 7.16±2.56 mm2 and 17 were decentered. Tomography treatment zone area was negatively correlated with central corneal astigmatism (R2=0.60) and elevation at an 8-mm chord (R2=0.64, both P<0.001).

CONCLUSIONS

For tomography images, central corneal astigmatism was highly correlated with peripheral elevation and may be a more expedient measure for clinical use. Treatment area decreased as corneal astigmatism and elevation increased.