Prediction of Effective Lens Position Using Multiobjective Evolutionary Algorithm

Ray tracing intraocular lens calculation performance improved by AI-powered postoperative lens position prediction

AIMS

To assess whether incorporating a machine learning (ML) method for accurate prediction of postoperative anterior chamber depth (ACD) improves cataract surgery refraction prediction performance of a commonly used ray tracing power calculation suite (OKULIX).

METHODS AND ANALYSIS

A dataset of 4357 eyes of 4357 patients with cataract was gathered at the Kellogg Eye Center, University of Michigan. A previously developed machine learning (ML)-based method was used to predict the postoperative ACD based on preoperative biometry measured with the Lenstar LS900 optical biometer. Refraction predictions were computed with standard OKULIX postoperative ACD predictions and ML-based predictions of postoperative ACD. The performance of the ray tracing approach with and without ML-based ACD prediction was evaluated using mean absolute error (MAE) and median absolute error (MedAE) in refraction prediction as metrics.

RESULTS

Replacing the standard OKULIX postoperative ACD with the ML-predicted ACD resulted in statistically significant reductions in both MAE (1.7% after zeroing mean error) and MedAE (2.1% after zeroing mean error). ML-predicted ACD substantially improved performance in eyes with short and long axial lengths (p<0.01).

CONCLUSIONS

Using an ML-powered postoperative ACD prediction method improves the prediction accuracy of the OKULIX ray tracing suite by a clinically small but statistically significant amount, with the greatest effect seen in long eyes.

Laboratory and Clinical Experience With a Diffractive Trifocal Intraocular Lens Sutured to an Artificial Iris

PURPOSE

To determine in vitro, using a translational research approach before realizing the procedure in a patient with iatrogenic aphakia and partial aniridia, whether suturing a trifocal intraocular lens (IOL) to an artificial iris degrades the IOL's optical quality.

METHODS

Optical quality was analyzed by measuring the modulation transfer function (MTF) at a 3-mm aperture and at 50 and 100 lp/mm spatial frequencies. The FineVision Pod F GF IOL (PhysIOL) was assessed in two powers: two +20.00 diopters (D) (20A and 20B IOLs) and two +30.00 D (30A and 30B IOLs). The IOLs' decentration in relation to the artificial iris's center was evaluated. The laboratory results provided empirical evidence in the informed consent for surgical intervention in a patient with iatrogenic aphakia and iris defect in one eye. Clinical results were measured using the parameter of corrected distance visual acuity plus a patient self-assessment of the cosmetic appearance of the operated eye.

RESULTS

The 20A and 20B IOLs demonstrated a mean MTF reduction of up to 1.1%, whereas the 30A and 30B IOLs showed a decrease of up to 5.2% for both spatial frequencies. All lenses showed good centration levels. In the clinical case, the patient showed corrected distance visual acuity, distance-corrected near visual acuity, and distance-corrected intermediate visual acuity of 0.20, 0.20, and 0.22 logMAR, respectively. The patient was satisfied with the cosmetic outcome.

CONCLUSIONS

There was merely a slight reduction in trifocal IOL optical quality after it was sutured to an artificial iris. Clinically, the combined implantation of the artificial iris and FineVision IOL provided good functional and cosmetic outcomes. [J Refract Surg. 2022;38(1):61-68.].

Three-dimensional topographic changes of anterior chamber depth following phacoemulsification with intraocular lens implantation in cataract patients

PURPOSE

To investigate the three-dimensional topographic changes of anterior chamber depth (ACD) following cataract surgery.

METHODS

Seventy-eight eyes with age-related cataract undergoing phacoemulsification and intraocular lens (IOL) implantation were retrospectively enrolled. Participants were evaluated with Pentacam for ACD topography before and approximately four weeks after the surgery. The absolute changes of ACD (AACD) and the relative changes of ACD (RACD) topography were calculated, and three-dimensional topographic contours were plotted. The influence of age, gender, distance to corneal apex (DCA), temporal-nasal and superior-inferior on AACD and RACD was analyzed.

RESULTS

Both AACD and RACD were negatively correlated with the DCA (p < 0.001; p < 0.001) and positively correlated with the age at all DCA (p < 0.05 for all the analyses). Significantly greater AACD and RACD were observed in female subjects (p < 0.05, respectively, at all DCA). AACD was significantly larger in the temporal compared with the nasal region (p < 0.001) and at the superior compared with the inferior region (p < 0.001), but not RACD. Subgroup analysis indicated that the significant difference of the AACD between the temporal and nasal regions was manifested at the DCA of more than 6 mm (p < 0.001), and the difference between the superior and inferior regions was observed at 2 mm DCA for both AACD (p < 0.001) and RACD (p = 0.001).

CONCLUSIONS

We depicted the topographic changes of ACD following cataract surgery and found that it was significantly influenced by age, gender, DCA and quadrant location. The research provided the basis for including postoperative ACD topography prediction before cataract surgery in the future.

Changes in ocular parameters the crystalline lens after implantation of a collamer lens

CLINICAL RELEVANCE

Understanding changes in ocular anatomical parameters after intraocular lens implantation will allow a more accurate determination of dioptric power prior to surgery.

BACKGROUND

The crystalline lens position might change due to the implantation and removal of an implantable collamer lens (ICL) or toric implantable collamer lens (TICL). This study aimed to assess the effect of ICL implantation on position of the crystalline lens.

METHODS

This retrospective study was conducted on patients who underwent V4c ICL or V4c TICL implantation between March and September, 2018. Preoperative and post-operative (2 weeks, 3 months and 6 months) axial length, central corneal thickness, crystalline lens position, crystalline lens thickness and vault height were analysed. Multivariable linear regression was used to determine the variables associated with 6-month changes in lens position.

RESULTS

This study included 117 eyes of 117 patients. There were decreases in all vertical distance measures from the central corneal endothelium to the anterior and posterior crystalline lens capsule (all p > 0.05). The amount of reduction was related to the crystalline lens position before the operation and crystalline lens thickness after the operation (all p < 0.01). An error in anterior chamber depth and lens thickness may appear when the ICL/TICL is close to the crystalline lens.

CONCLUSION

Phakic intraocular lens implantation resulted in lens thickening and forward movement on day 1 post-operatively, which becomes stable within 6 months. Preoperative lens position and post-operative lens thickness were related to the amount of movement.

Comparison of various intraocular lens formulas using a new high-resolution swept-source optical coherence tomographer

PURPOSE

To compare vergence, artificial intelligence, and combined intraocular lens (IOL) calculation formulas using a new swept-source optical coherence tomographer (SS-OCT) and to analyze their performance based on manifest and estimated refractive outcomes of cataract surgery.

SETTING

Department of Ophthalmology, University of Pécs Medical School, Pécs, Hungary.

DESIGN

Retrospective data analysis.

METHODS

Optical biometry readings of patients who underwent uneventful cataract removal and implantation of a monofocal acrylic IOL were used to predict IOL power with Barrett Universal II (BUII), Haigis, Hoffer Q, Holladay 1, Radial Basis Function (RBF) 2.0, Kane, Ladas Super Formula, and SRK/T. All the implanted IOLs were calculated by using the Haigis formula. The arithmetic prediction error and median and mean absolute refractive errors for all formulas were computed. The percentage of eyes within ±0.25 diopters (D), ±0.50 D, and ±1.0 D of prediction error was calculated.

RESULTS

A total of 95 eyes of 95 patients with a mean age of 68.80 ± 7.57 years were included. There was a statistically significant difference in absolute prediction error across the 8 IOL calculation formulas (P < .0001). Haigis showed the lowest mean absolute error, and it differed significantly from the BUII, Hoffer Q, Holladay 1, Ladas, RBF 2.0, and SRK/T formulas (P < .05). In terms of eyes within ±0.25 D, ±0.50 D, and ±1.0 D of prediction error, the Haigis formula showed the overall best performance.

CONCLUSIONS

The results indicated that a recently developed SS-OCT provided accurate ocular biometry measurements before cataract surgery, and the Haigis formula incorporated in its software enabled precise calculation of IOL refractive power.

Determining the Theoretical Effective Lens Position of Thick Intraocular Lenses for Machine Learning-Based IOL Power Calculation and Simulation

Purpose

To describe a formula to back-calculate the theoretical position of the principal object plane of an intraocular lens (IOL), as well as the theoretical anatomic position in a thick lens eye model. A study was conducted to ascertain the impact of variations in design and IOL power, on the refractive outcomes of cataract surgery.

Methods

A schematic eye model was designed and manipulated to reflect changes in the anterior and posterior radii of an IOL, while keeping the central thickness and paraxial powers static. Modifications of the shape factor (X) of the IOL affects the thick lens estimated effective lens position (ELP). Corresponding postoperative spherical equivalent (SE) were computed for different IOL powers (-5 diopters [D], 5 D, 15 D, 25 D, and 35 D) with X ranging from -1 to +1 by 0.1.

Results

The impact of the thick lens estimated effective lens position shift on postoperative refraction was highly dependent on the optical power of the IOL and its thickness. Design modifications could theoretically induce postoperative refraction variations between approximately 0.50 and 3.0 D, for implant powers ranging from 15 D to 35 D.

Conclusions

This work could be of interest for researchers involved in the design of IOL power calculation formulas. The importance of IOL geometry in refractive outcomes, especially for short eyes, should challenge the fact that these data are not usually published by IOL manufacturers.

Translational Relevance

The back-calculation of the estimated effective lens position is central to intraocular lens calculation formulas, especially for artificial intelligence-based optical formulas, where the algorithm can be trained to predict this value.

AI-powered effective lens position prediction improves the accuracy of existing lens formulas

AIMS

To assess whether incorporating a machine learning (ML) method for accurate prediction of postoperative anterior chamber depth (ACD) improves the refraction prediction performance of existing intraocular lens (IOL) calculation formulas.

METHODS

A dataset of 4806 patients with cataract was gathered at the Kellogg Eye Center, University of Michigan, and split into a training set (80% of patients, 5761 eyes) and a testing set (20% of patients, 961 eyes). A previously developed ML-based method was used to predict the postoperative ACD based on preoperative biometry. This ML-based postoperative ACD was integrated into new effective lens position (ELP) predictions using regression models to rescale the ML output for each of four existing formulas (Haigis, Hoffer Q, Holladay and SRK/T). The performance of the formulas with ML-modified ELP was compared using a testing dataset. Performance was measured by the mean absolute error (MAE) in refraction prediction.

RESULTS

When the ELP was replaced with a linear combination of the original ELP and the ML-predicted ELP, the MAEs±SD (in Diopters) in the testing set were: 0.356±0.329 for Haigis, 0.352±0.319 for Hoffer Q, 0.371±0.336 for Holladay, and 0.361±0.331 for SRK/T which were significantly lower (p<0.05) than those of the original formulas: 0.373±0.328 for Haigis, 0.408±0.337 for Hoffer Q, 0.384±0.341 for Holladay and 0.394±0.351 for SRK/T.

CONCLUSION

Using a more accurately predicted postoperative ACD significantly improves the prediction accuracy of four existing IOL power formulas.

Gradient Boosting Decision Tree Algorithm for the Prediction of Postoperative Intraocular Lens Position in Cataract Surgery

Purpose

To develop a method for predicting postoperative anterior chamber depth (ACD) in cataract surgery patients based on preoperative biometry, demographics, and intraocular lens (IOL) power.

Methods

Patients who underwent cataract surgery and had both preoperative and postoperative biometry measurements were included. Patient demographics and IOL power were collected from the Sight Outcomes Research Collaborative (SOURCE) database. A gradient-boosting decision tree model was developed to predict the postoperative ACD. The mean absolute error (MAE) and median absolute error (MedAE) were used as evaluation metrics. The performance of the proposed method was compared with five existing formulas.

Results

In total, 847 patients were assigned randomly in a 4:1 ratio to a training/validation set (678 patients) and a testing set (169 patients). Using preoperative biometry and patient sex as predictors, the presented method achieved an MAE of 0.106 ± 0.098 (SD) on the testing set, and a MedAE of 0.082. MAE was significantly lower than that of the five existing methods (P < 0.01). When keratometry was excluded, our method attained an MAE of 0.123 ± 0.109, and a MedAE of 0.093. When IOL power was used as an additional predictor, our method achieved an MAE of 0.105 ± 0.091 and a MedAE of 0.080.

Conclusions

The presented machine learning method achieved greater accuracy than previously reported methods for the prediction of postoperative ACD.

Translational Relevance

Increasing accuracy of postoperative ACD prediction with the presented algorithm has the potential to improve refractive outcomes in cataract surgery.

AI-Powered Effective Lens Position Prediction Improves the Accuracy of Existing Lens Formulas

AIMS

To assess whether incorporating a machine learning (ML) method for accurate prediction of postoperative anterior chamber depth (ACD) improves the refraction prediction performance of existing intraocular lens (IOL) calculation formulas.

METHODS

A dataset of 4806 cataract patients were gathered at the Kellogg Eye Center, University of Michigan, and split into a training set (80% of patients, 5761 eyes) and a testing set (20% of patients, 961 eyes). A previously developed ML-based method was used to predict the postoperative ACD based on preoperative biometry. This ML-based postoperative ACD was integrated into new effective lens position (ELP) predictions using regression models to rescale the ML output for each of four existing formulas (Haigis, Hoffer Q, Holladay, and SRK/T). The performance of the formulas with ML-modified ELP was compared using a testing dataset. Performance was measured by the mean absolute error (MAE) in refraction prediction.

RESULTS

When the ELP was replaced with a linear combination of the original ELP and the ML-predicted ELP, the MAEs ± SD (in Diopters) in the testing set were: 0.356 ± 0.329 for Haigis, 0.352 ± 0.319 for Hoffer Q, 0.371 ± 0.336 for Holladay, and 0.361 ± 0.331 for SRK/T which were significantly lower than those of the original formulas: 0.373 ± 0.328 for Haigis, 0.408 ± 0.337 for Hoffer Q, 0.384 ± 0.341 for Holladay, and 0.394 ± 0.351 for SRK/T.

CONCLUSION

Using a more accurately predicted postoperative ACD significantly improves the prediction accuracy of four existing IOL power formulas.