Update on Intraocular Lens Formulas and Calculations

A comparison of white-to-white measurements obtained by Anterion AS-OCT versus three optical devices in refractive surgery candidates

BACKGROUND AND OBJECTIVES

To assess the agreement between Anterion AS-OCT and three optical devices in measuring the white-to-white (WTW) diameter in candidates for refractive surgery.

METHODS

In this cross-sectional prospective study, 129 right eyes of 129 normal individuals underwent consecutive scans with the Anterion, the Pentacam AXL, the IOLMaster 700, and the Orbscan IIz. Mean difference (MD), 95% limits of agreement (LoA), and concordance correlation coefficient (CCC) were calculated to assess agreement and interchangeability.

RESULTS

The mean age of participants was 30.4 ± 5.9 (range: 21-47) years. The mean WTW distance measured by the different instruments was as follows: 12.00 ± 0.42 mm (range, 10.57 to 13.05) using the Anterion, 11.87 ± 0.34 (11.10 to 12.80) using the Pentacam, 12.12 ± 0.44 (11.00 to 13.30) using the IOLMaster, and 11.73 ± 0.37 (11.00 to 13.10) using the Orbscan. The MD and 95% LoA for Anterion vs. Pentacam, IOLMaster, or Orbscan were [0.11; -0.31 to 0.54 mm], [-0.13; -0.93 to 0.66], and [0.25; -0.28 to 0.78], respectively. The corresponding CCCs were 0.803, 0.514, and 0.631.

CONCLUSIONS

This study found weak agreement between Anterion and Pentacam AXL, IOLMaster 700, and Orbscan IIz devices regarding WTW distance measurements in refractive surgery candidates. Therefore, it is not recommended to use Anterion's WTW measurements interchangeably with the other three devices, particularly for phakic intraocular lens sizing.

Comparison of Hill-RBF 3.0 with Barrett Universal II, SRK/T, Hoffer Q, Haigis, and Holladay 1 to predict the accuracy of post-cataract surgery refractive outcomes in Indian eyes

PURPOSE

To compare Hill-RBF 3.0 with Barrett Universal II (BU II), SRK/T, Hoffer Q, Haigis, and Holladay 1 in predicting the accuracy of post-cataract surgery refractive outcomes in Indian eyes.

METHODS

In this prospective, comparative, observational study, consecutive patients with uncomplicated age-related cataracts undergoing uneventful phacoemulsification with posterior chamber intraocular lens (IOL) implantation were included. The mean absolute errors (MAEs) and median absolute errors were used to determine the accuracy of predicted postoperative target refractions.

RESULTS

A total of 219 eyes of 173 patients were enrolled. Based on the axial lengths (AL), the patients were classified into: AL <22 mm (short), 22-24.5 mm (normal), and >24.5 mm (long). BU II exhibited the lowest MAE for normal ALs (0.2683 ± 0.2790 D) as well as for the entire population (0.2764 ± 0.2764 D). For the short ALs, Hill RBF 3.0 exhibited the lowest MAE (0.3268 ± 0.3268 D), while for the long ALs, SRK/T showed the lowest MAE (0.2823 ± 0.2642 D). BU II exhibited the highest percentage of eyes of 57.5%, 95.4%, and 98.6% within ±0.25, ±0.75, and ±1.0 D of postoperative target refractions respectively, whereas Hill RBF 3.0 had the highest percentages of eyes (88.1%) within ±0.5 D of postoperative target refraction.

CONCLUSION

Hill-RBF 3.0 exhibited the least MAE for patients with short ALs, while BU II showed the least MAE for normal ALs as well as for the entire population and SRK/T for long ALs. This study is likely to aid surgeons in selecting the most appropriate IOL power formula, which thereby improves the refractive outcomes with utmost accuracy.

A Multi Comparison of 8 Different Intraocular Lens Biometry Formulae, Including a Machine Learning Thin Lens Formula (MM) and an Inbuilt Anterior Segment Optical Coherence Tomography Ray Tracing Formula

A comparison of the accuracy of intraocular lens (IOL) power calculation formulae, including SRK/T, HofferQ, Holladay 1, Haigis, MM, Barrett Universal II (BUII), Emmetropia Verifying Optical (EVO), and AS-OCT ray tracing, was performed. One hundred eyes implanted with either the Rayone EMV RAO200E (Rayner Intraocular Lenses Limited, Worthing, UK) or the Artis Symbiose (Cristalens Industrie, Lannion, France) IOL were included. Biometry was obtained using IOLMaster 700 (Carl Zeiss Meditec AG, Jena, Germany) and MS-39 AS-OCT (CSO, Firenze, Italy). Mean (MAE) and median (MedAE) absolute errors and percentage of eyes within ±0.25D, ±0.50D, ±0.75D, and ±1.00D of the target were compared, with ±0.75D considered a key metric. The highest percentage within ±0.75D was found with MM (96%) followed by the Haigis (94%) for the enhanced monofocal IOL. SRK/T (94%) had the highest percentage within ±0.75D, followed by Holladay 1, MM, BUII, and ray tracing (all 90%) for the multifocal IOL. No statistically significant difference in MAE was found with both IOLs. EVO showed the lowest MAE for the enhanced monofocal and ray tracing for the multifocal IOL. EVO and ray tracing showed the lowest MedAE for the two respective IOLs. A similar performance with high accuracy across formulae was found. MM and ray tracing appear to have similar accuracy to the well-established formulae and displayed a high percentage of eyes within ±0.75D.

Distribution and associated factors of keratometry and corneal astigmatism in an elderly population

To determine the distribution of keratometry and corneal astigmatism (CA) and their association with demographic factors, systemic parameters, anthropometric measures, ocular biometric indices, and refractive errors in people aged 60 years and above. In this cross-sectional study, 160 clusters were randomly selected from Tehran city (Iran) using the multi-stage cluster sampling method. All participants underwent optometric examinations including testing uncorrected and best-corrected distance visual acuity, non-cycloplegic autorefraction, and subjective refraction. Pentacam imaging for all participants was carried out using Pentacam AXL. Keratometry and CA were reported based on Pentacam's data. The average, standard deviation (SD) and 95% confidence interval (CI) of flat keratometry (Kf), steep keratometry (Ks), mean keratometry (mean K), and CA were 44.02 ± 1.58 D (95% CI 43.94-44.1), 44.86 ± 1.67 D (95% CI 44.78-44.94), 44.44 ± 1.58 D (95% CI 44.36-44.52), and 0.84 ± 0.74 D (95% CI 0.81-0.87), respectively. The 95% and 99% percentiles of mean K were 47.1 and 48.6 D, respectively. According to the multiple generalized estimating equation model, the mean K was significantly higher in males, in myopes, and in those with higher systolic blood pressure. Moreover, the mean K was inversely related to the axial length, height, anterior chamber depth (ACD), corneal diameter, and central corneal thickness (CCT). The prevalence of various types of CA based on a cut-off > 0.50 D was as follows; with-the-rule: 32.5% (95% CI 30.6-34.4), against-the-rule: 18.2% (95% CI 16.7-19.7), and oblique: 10.0% (95% CI 9.1-11.0). The present study investigated the normal distribution of keratometry and CA in individuals ≥ 60 years, and results can be used in clinical matters, especially in intraocular lens power calculation. Sex, systolic blood pressure, height, and some biometric components such as ACD, corneal diameter, and CCT were significantly related to keratometry and should be considered.

Effect of capsular tension ring implantation on intraocular lens calculation formula selection for long axial myopia

PURPOSE

The study investigated the effect of capsular tension ring (CTR) implantation on postoperative refractive stability and accuracy of intraocular lens (IOL) formulas for axial length (AL) ≥ 27.0 mm patients.

METHODS

Prospective case series. The eyes of patients underwent phacoemulsification extraction combined with IOL implantation were classified as CTR implantation (A-CTR) and without CTR implantation (B-CON) groups. Refractive outcome and anterior chamber depth (ACD) were recorded at 1 week, 1 month, and 3 months post-operation. Prediction refractive error (PE) and absolute refractive error (AE) of each formula were calculated.

RESULTS

A total of 89 eyes (63 patients) were included and randomized into the CTR (A-CTR) and control groups (B-CON). Comparison of refraction at different postoperative times of the CTR group showed no statistical difference (all P > 0.05). The ACD in the A-CTR group gradually deepened, and that in the B-CON group gradually shallowed (all P > 0.05). The formulas' AE showed statistically significant differences in CTR and CON groups (P < 0.001). The PE of Hill-RBF 2.0 and EVO formulas in the A-CTR group were more hyperopic than that in the B-CON group (all P > 0.05), the other five formulas were more myopic in A-CTR group than that in the B-CON group (all P > 0.05).

CONCLUSION

Patients with 13 mm diameter CTR implantation tended to have stable refraction at 1 week post-surgery and 1 month for those without it. CTR of the 13 mm diameter had no effect on the selection of formulas. Additionally, it is found that Kane and EVO formulas were more accurate for patients with AL ≥ 27.0 mm.

Accuracy of intraocular lens power calculation formulae for the Yamane technique of secondary fixation: a systematic review and meta-analysis

OBJECTIVE

This systematic review and meta-analysis aims to assess the refractive outcomes of the Yamane technique for intrascleral fixation of intraocular lenses (SF-IOL) and compare the predictive ability of the various intraocular lens power calculation formulae commonly used in conjunction with the technique.

METHODS

A literature search was conducted in the Medline, Scopus, and Cochrane Library databases for articles published from January 2014 to May 2023. Studies that met the predetermined inclusion criteria were included and subjected to analysis. The primary outcome evaluated was the refractive predictive error, defined as the difference between predicted refraction and post-operative manifest refraction.

RESULTS

Eleven studies met the inclusion criteria, with a cumulative sample size of 615 patients (mean age: 66.6 years). Various IOL formulae were used, with SRK/T being the most frequently adopted formula. The overall mean refractive predictive error for all formulae combined was -0.02 D, which was not statistically significant (p = 0.99). Subgroup analysis for individual formulae also showed no significant difference from predicted error for any formula (p > 0.05).

CONCLUSION

The Yamane technique for SF-IOL shows promising refractive outcomes, and the choice of IOL power calculation formula should be tailored based on patient characteristics and surgeon preference. No formula demonstrated superior predictive ability over others. Further research is needed to develop formulae specifically for eyes with secondary aphakia and poor capsular support.

Accuracy of using the axial length of the fellow eye for IOL calculation in retinal detachment eyes undergoing silicone oil removal

PURPOSE

Evaluate whether the axial length of the fellow eye can be used to calculate the intraocular lens (IOL) in eyes with retinal detachment.

DESIGN

Retrospective, consecutive case series.

METHODS

Our study was conducted at the Goethe University and included patients who underwent silicone oil (SO) removal combined with phacoemulsification and IOL implantation. Preoperative examinations included biometry (IOLMaster 700, Carl Zeiss). We measured axial length (AL) of operated eye (OE) or fellow eye (FE) and compared mean prediction error and mean and median absolute prediction error (MedAE) using four formulas and AL of the OE (Barrett Universal II (BUII)-OE). Additionally, we compared the number of eyes within ±0.50, ±1.00 and ±2.00 dioptre (D) from target refraction.

RESULTS

In total, 77 eyes of 77 patients met our inclusion criteria. MedAE was lowest for the BUII-OE (0.42 D) compared with Kane-FE (1.08 D), BUII-FE (1.02 D) and Radial Basis Function 3.0 (RBF3.0)-FE (1.03 D). This was highly significant (p<0.001). The same accounts for the number of eyes within ±0.50 D of the target refraction with the BUII-OE (44 eyes, 57%) outperforming the RBF3.0-FE (20 eyes, 25.9%), Kane-FE and BUII-FE formula (21 eyes, 27.2%) each.

CONCLUSION

Our results show a statistically and clinically highly relevant reduction of IOL power predictability when using the AL of the FE for IOL calculation. Using the AL of the SO filled eye after initial vitrectomy results in significantly better postoperative refractive results. A two-step procedure using the AL of the OE after reattachment of the retina is highly recommended.

Influence of Lens Thickness on Accuracy of Kane, Hill-RBF 3.0, Barrett Universal II, Emmetropia Verifying Optical, and Pearl-DGS Formulas in Eyes with Nonhigh Myopia and High Myopia

PURPOSE

To investigate the influence of lens thickness (LT) on accuracy of Kane, Hill-RBF 3.0 Barrett Universal II (BUII), Emmetropia Verifying Optical (EVO), and Pearl-DGS formulas in eyes with different axial lengths (AL).

METHODS

The prospective cohort study was conducted at Eye and ENT Hospital of Fudan University. Patients who had uneventful cataract surgery between March 2021 and July 2023 were recruited. Manifest refraction was conducted two-month post-surgery. Eyes were divided into 4 groups based on AL: short (<22mm), medium (22-24.5 mm), medium long (24.5-26mm) and very long (≥26mm). In each AL group, eyes were then divided into 3 subgroups based on the LT measured with IOLmaster700: thin (<4.5 mm), medium (4.5-5.0 mm), and thick (≥ 5 mm). The influence of LT on accuracy of Kane, Hill-RBF 3.0, BUII, EVO, and Pearl-DGS formulas were investigated in each AL group.

RESULTS

A total of 327 eyes from 327 patients were analyzed, with 64, 102, 73 and 88 eyes in each AL group, respectively. In eyes with AL < 24.5 mm, myopic PE was significantly associated with greater LT using all the 5 formulas (all p < 0.05). Backward stepwise multivariate regression analyses revealed that LT was an important influencing factor for PE in all 5 formulas, particularly in eyes with AL <24.5 mm. In eyes with AL <24.5 mm and LT > 5.0 mm, PE of all 5 formulas calculated with the optional parameter LT were more myopic than those calculated without LT.

CONCLUSIONS

Thicker LT was associated with more myopic PE among eyes with AL <24.5 mm when using all 5 formulas. Further optimization of current formulas is necessary, especially for eyes with short AL and thick LT.

Current Concepts and Recent Updates of Optical Biometry- A Comprehensive Review

One of the most recent advancements in the field of cataract surgery is optical biometry. With the advent of optical biometry ocular measurements are now simpler, quicker, and more precise. The devices have made intraocular lens (IOL) power calculations easier in difficult situations too, such as in cases with extremes of axial lengths, silicone filled eyes, cataract surgery in post-keratoplasty eyes, post Laser-Assisted in Situ Keratomileusis (LASIK) eyes, etc. The gold standard for IOL power calculation in the present day is by the use of optical biometry devices. The anatomical measurements by these devices are highly precise and because of these measurements and the incorporation of various IOL power calculation formulas the optical biometry devices give the accurate power and the post-operative visual outcome is highly satisfactory among the patients. The growing use of these devices has made cataract the most commonly performed refractive surgical procedure nowadays. In the current scenario, optical biometry has widespread acceptance in almost all countries and has many advantages over ultrasound or immersion biometry. Cataract surgeons can obtain easy and reliable measurements from these devices. Refractive surprises have also decreased considerably with their use. This article will comprehensively review the principles of the various optical biometry devices, the parameters used in each of the devices, the advantages and disadvantages, and add more like what all this article will add.

Retrospective assessment of accuracy of nine intraocular lens power calculation formulae in eyes with axial myopia

PURPOSE

To compare the accuracy of nine conventional and newer-generation formulae in calculating intraocular lens power in eyes with axial myopia.

SETTING

Tertiary eye care center, Bengaluru, India.

DESIGN

Retrospective cross-sectional, comparative study conducted in India.

METHODS

Patients undergoing uneventful phacoemulsification in eyes with axial length >26 mm were included. Preoperative biometry was done using Lenstar LS 900 (Haag-Streit AG, Switzerland). Single eye of patients undergoing bilateral implantation was randomly selected. Optimized lens constants were used to calculate the predicted postoperative refraction of each formula, which was then compared with the actual refractive outcomes to give the prediction errors, following which subgroup analysis was performed. The Kane formula, Barrett universal II, Emmetropia Verifying Optical (EVO) 2.0, Hill Radial Basis Function (Hill RBF) 3.0, Olsen formula, along with Wang Koch-adjusted four formulae, that is, Sanders Retzlaff Kraff/Theoretical (SRK/T), Holladay 1, Haigis, and Hoffer Q formula, were compared for intraocular lens power calculations.

RESULTS

One hundred and sixty-five eyes that fulfilled all the inclusion criteria were studied. Hill RBF 3.0 had the lowest mean and median absolute prediction errors (0.355 and 0.275, respectively) compared to all formulas. In subgroup analysis (26-28, >28-30, and >30 mm), significant difference was seen only in extremely long eyes (>30 mm). The Hill RBF 3.0 formula generated the maximum percentage of eyes with refractive errors within ±0.25, ±0.5, ±0.75, and ±1 D (46%, 76.2%, 89.9%, and 95.8%, respectively).

CONCLUSION

This is the first study evaluating all the formulas exclusively in the myopic eyes. Hill RBF 3 was found to be superior in accuracy to all other formulas.

Complications of foldable intraocular lenses requiring explantation or secondary intervention: 2022 survey with update of long-term trends

PURPOSE

To assess the complications that resulted in the explantation or secondary intervention with foldable intraocular lenses (IOLs).

SETTING

University setting, Salt Lake City, Utah.

DESIGN

Survey study.

METHODS

For the 25th consecutive year, surgeons were surveyed regarding complications associated with foldable IOLs requiring explantation or secondary intervention over the 2022 calendar year. These forms were made available online using the ASCRS and ESCRS websites and a fax-on-demand service. Surgeons completed 1 survey for each foldable IOL requiring explantation or secondary intervention. Further analysis determined complication trends related to specific IOL styles, materials, and types over the past 16 years (2007 to 2022).

RESULTS

103 completed surveys were returned in 2022 contributing to a total of 1627 tabulated surveys since 2007. In the 2022 survey, dislocation/decentration continued to be the most common complication overall. Glare/optical aberrations was a common complication associated with multifocal IOLs continuing a 16-year trend. In addition, hydrophilic acrylic IOLs as well as some silicone lenses in eyes with asteroid hyalosis demonstrated calcification as the most common complication necessitating explantation.

CONCLUSIONS

Dislocation/decentration remains the leading cause of explantation in most IOL types. Glare/optical aberrations continue to be an associated complication of multifocal IOLs suggesting this ongoing issue has yet to be resolved with this type of IOL. In addition, calcification of hydrophilic acrylic lenses and silicone lenses is a rare event but continues to occur.

A Review of Intraocular Lens Power Calculation Formulas Based on Artificial Intelligence

PURPOSE

The proper selection of an intraocular lens power calculation formula is an essential aspect of cataract surgery. This study evaluated the accuracy of artificial intelligence-based formulas.

DESIGN

Systematic review.

METHODS

This review comprises articles evaluating the exactness of artificial intelligence-based formulas published from 2017 to July 2023. The papers were identified by a literature search of various databases (Pubmed/MEDLINE, Google Scholar, Crossref, Cochrane Library, Web of Science, and SciELO) using the terms "IOL formulas", "FullMonte", "Ladas", "Hill-RBF", "PEARL-DGS", "Kane", "Karmona", "Hoffer QST", and "Nallasamy". In total, 25 peer-reviewed articles in English with the maximum sample and the largest number of compared formulas were examined.

RESULTS

The scores of the mean absolute error and percentage of patients within ±0.5 D and ±1.0 D were used to estimate the exactness of the formulas. In most studies the Kane formula obtained the smallest mean absolute error and the highest percentage of patients within ±0.5 D and ±1.0 D. Second place was typically achieved by the PEARL DGS formula. The limitations of the studies were also discussed.

CONCLUSIONS

Kane seems to be the most accurate artificial intelligence-based formula. PEARL DGS also gives very good results. Hoffer QST, Karmona, and Nallasamy are the newest, and need further evaluation.

Network Meta-analysis of Intraocular Lens Power Calculation Formula Accuracy in 1016 Eyes With Long Axial Length

PURPOSE

To systematically review the literature and quantitatively synthesize the currently available evidence to compare the accuracy of different intraocular lens calculation formulas in eyes with long axial length (AL).

DESIGN

Network meta-analysis.

METHODS

PubMed, Embase, Web of Science, and the Cochrane Library were systematically searched for studies published between January 2000 and June 2022. Included were prospective or retrospective clinical studies reporting the following outcomes in cataract patients with long AL (ie, ≥26 mm): percentage of eyes with a prediction error (PE) within ±0.25, ±0.50, and ±1.00 diopters (D). Network meta-analysis was conducted using R software (version 4.2.1).

RESULTS

Ten prospective or retrospective clinical studies, including 1016 eyes and 11 calculation formulas, were identified. A traditional meta-analysis showed that for the percentage of eyes with PE within ±0.25 and ±0.50 D, the Olsen, Kane, and Emmetropia Verifying Optical (EVO) all had insignificantly higher percentages compared with others. Considering the percentage of eyes with PE within ±1.00 D, the original and modified Wang-Koch adjustment formulas for Holladay 1 (H1-WK and H1-MWK) and EVO formulas showed superiority, but the difference was insignificant. This network meta-analysis revealed that compared with the widely used Barrett Universal II (BUII) formula, the Olsen, Kane, and EVO formulas had higher percentages of eyes with PE within ±0.25, ±0.50, and ±1.00 D (all odds ratios >1 but P >.05). Based on the surface under the cumulative ranking area (SUCRA) values for the percentage of eyes with PE within ±0.25 D, the Olsen (96.4%), Kane (77.5%), and EVO (75.9%) formulas had the highest probability of being in the top 3 of the 11 formulas.

CONCLUSIONS

The Olsen, Kane, and EVO formulas may perform better than others in calculating IOL power in eyes with long AL. Nevertheless, there is still considerable uncertainty in this regard and the accuracy of these formulas in highly myopic eyes should be confirmed in studies based on large multicenter registries.

Refractive surgical correction and treatment of keratoconus

Keratoconus is an ectatic corneal disorder that causes severe vision loss. Surgical options allow us to correct, partially or totally, the induced refractive error. Intracorneal ring segments (ICRS) implantation represents a minimally invasive surgical option that improves visual acuity, with a high success rate and a low overall complication rate. Corneal allogenic ICRS consists of ring segments derived from allogenic eye bank-processed donor corneas. Selective topography-guided transepithelial photorefractive or phototherapeutic keratectomy combined with CXL is another way in selected cases to improve spectacles corrected distance visual acuity. The microphotoablative remodeling of the central corneal profile is generally planned by optimizing the optical zones and minimizing tissue consumption. Phakic intraocular lens (PIOL) implant is considered in patients with stable disease and acceptable anatomical requirements. The two types of pIOLs, depending on their implantation inside the eye, are anterior chamber-pIOLs, which fixate to the anterior surface of the iris by using a polymethomethacrolate claw at the two haptics, and posterior chamber-pIOLs. In patients with both cataracts and keratoconus, the correct IOL power is difficult to obtain due to the irregular corneal shape and K values. Toric IOL is recommended, but carefully judging the topography and the possible need of subsequent keratoplasties.

Intraocular lens power calculation formulas: a scientometric analysis

Background

The most accurate method of intraocular lens (IOL) power calculation in cataract surgery has not been determined, and further studies are needed to reach a consensus. The aim of this study was to assess publications related to IOL power calculation formulas, mapping their yearly trends, most productive authors, top publishing countries and institutions, and areas of specialization for IOL power formulas.

Methods

We conducted a comprehensive analysis of research articles published on the topic of IOL power calculation formulas. Using PubMed, we employed appropriate search terms and filtered the results for the period of January 1, 1946, to June 28, 2023. Data were analyzed using CiteSpace, VOSviewer, and Microsoft Excel programs. The visual representations of the collected data through the use of figures was provided to demonstrate the aspects of IOL power calculation research.

Results

We retrieved 5475 documents in the initial search. Analysis of these documents revealed an increase in the number of publications, from one publication in 1946 to 201 publications in 2023. The top three countries contributing to these publications were the United States, China, and Japan, collectively accounting for over 27% of the total articles. However, the two institutions with the highest contributions were located in the United Kingdom and Hungary, neither of which was among the top 10 countries in overall contributions. Overall 15 326 authors contributed to publications pertaining to IOL power calculation formulas. Among these authors, the most prolific contributors included Achim Langenbucher from Saarland University (Germany), Giacomo Savini from G.B. Bietti Foundation I.R.C.C.S. (Italy), and Kenneth J Hoffer from the University of California (United States). Saarland University emerged as the most productive institution, contributing equally to two distinct departments: the Dr. Rolf M. Schwiete Center for Limbal Stem Cell Research and Congenital Aniridia, as well as the Department of Experimental Ophthalmology. The School of Physical Science at the Open University in the United Kingdom engaged in partnership with various institutions including Eye & Laser Clinic Castrop Rauxel in Germany and Johannes Kepler University Linz in Austria. Among the top 10 keywords found in the publications were "cataract", "cataract surgery", and "intraocular lens".

Conclusions

This study represents the first scientometric analysis of publications related to IOL power calculation formulas. The study offers valuable insights into the geographic distribution, contributing authors, and emphasis of research on the IOL power calculation formulas. Further cooperation is essential to pinpoint the most suitable formula and to address gaps in our current understanding.

Artificial Intelligence in Ophthalmic Surgery: Current Applications and Expectations

Artificial Intelligence (AI) has found rapidly growing applications in ophthalmology, achieving robust recognition and classification in most kind of ocular diseases. Ophthalmic surgery is one of the most delicate microsurgery, requiring high fineness and stability of surgeons. The massive demand of the AI assist ophthalmic surgery will constitute an important factor in boosting accelerate precision medicine. In clinical practice, it is instrumental to update and review the considerable evidence of the current AI technologies utilized in the investigation of ophthalmic surgery involved in both the progression and innovation of precision medicine. Bibliographic databases including PubMed and Google Scholar were searched using keywords such as "ophthalmic surgery", "surgical selection", "candidate screening", and "robot-assisted surgery" to find articles about AI technology published from 2018 to 2023. In addition to the Editorials and letters to the editor, all types of approaches are considered. In this paper, we will provide an up-to-date review of artificial intelligence in eye surgery, with a specific focus on its application to candidate screening, surgery selection, postoperative prediction, and real-time intraoperative guidance.

The Repeatability of Axial Length Measurements Using a Scheimpflug-based System

Purpose

To assess the repeatability of Pentacam AXL as a Scheimpflug-based system or measuring axial length according to the age, sex, lens type, axial length value, and type of cataract.

Methods

The present study was conducted using multistage cluster sampling in Tehran, Iran. Ocular biometry was performed, using the Pentacam AXL, by an experienced optometrist on all the participants. The axial length (AL) measurements were taken thrice, with a gap of 10 minutes. To evaluate the repeatability, the intraclass correlation coefficient (ICC) and the repeatability coefficient (RC) were calculated. To determine the significant difference in the repeatability index among study variables, the tolerance index (TI) was calculated.

Results

In this report, 897 eyes of 677 individuals aged between 20 and 91 years (mean ± SD: 64.90 ± 13.62 years) were reported. The ICC of the axial length measurements was 0.981 for all cases. Based on the within-subject standard deviation, the RC was 0.401. The ICC was 0.976 and 0.985 in men and women, respectively. The TI showed better RC of measurements among females. The ICC decreased from 0.999 in participants under 40 years to 0.973 in individuals over 60 years of age. The TI showed a decrease in RC with advancing age. The RC was worse in eyes with nuclear cataracts; the RC was also worse in the first quartile of the signal-to-noise ratio (SNR) compared to the other SNR quartiles.

Conclusion

The Scheimpflug-based systemPentacam AXL had high repeatability in measuring axial length. Some variables such as male gender, older age, and nuclear cataract were associated with reduced repeatability of the measurements. A higher SNR was associated with better repeatability of the axial length measurements.

Guidelines for the application of artificial intelligence in the diagnosis of anterior segment diseases (2023)

The landscape of ophthalmology has observed monumental shifts with the advent of artificial intelligence (AI) technologies. This article is devoted to elaborating on the nuanced application of AI in the diagnostic realm of anterior segment eye diseases, an area ripe with potential yet complex in its imaging characteristics. Historically, AI's entrenchment in ophthalmology was predominantly rooted in the posterior segment. However, the evolution of machine learning paradigms, particularly with the advent of deep learning methodologies, has reframed the focus. When combined with the exponential surge in available electronic image data pertaining to the anterior segment, AI's role in diagnosing corneal, conjunctival, lens, and eyelid pathologies has been solidified and has emerged from the realm of theoretical to practical. In light of this transformative potential, collaborations between the Ophthalmic Imaging and Intelligent Medicine Subcommittee of the China Medical Education Association and the Ophthalmology Committee of the International Translational Medicine Association have been instrumental. These eminent bodies mobilized a consortium of experts to dissect and assimilate advancements from both national and international quarters. Their mandate was not limited to AI's application in anterior segment pathologies like the cornea, conjunctiva, lens, and eyelids, but also ventured into deciphering the existing impediments and envisioning future trajectories. After iterative deliberations, the consensus synthesized herein serves as a touchstone, assisting ophthalmologists in optimally integrating AI into their diagnostic decisions and bolstering clinical research. Through this guideline, we aspire to offer a comprehensive framework, ensuring that clinical decisions are not merely informed but transformed by AI. By building upon existing literature yet maintaining the highest standards of originality, this document stands as a testament to both innovation and academic integrity, in line with the ethos of renowned journals such as Ophthalmology.

Comparing the Accuracy of the Kane, Barrett Universal II, Hill-Radial Basis Function, Emmetropia Verifying Optical, and Ladas Super Formula Intraocular Lens Power Calculation Formulas

Purpose

To assess the accuracy of five new-generation intraocular lens (IOL) power formulas: Barrett Universal II (BUII), Emmetropia Verifying Optical (EVO) Formula, Hill-Radial Basis Function (Hill-RBF), Kane Formula, and Ladas Super Formula (LSF).

Patients and Methods

This is a retrospective single-surgeon study from a refractive clinic and clinical research center in Draper, UT, USA. The primary outcome measures were mean absolute error (MAE) and median absolute error (MedAE). Secondary outcome measures were the standard deviation (SD) of each formula's refractive prediction errors (RPE) and the percentage of eyes within ±0.50D. Refractive predictions were compared to the postoperative spherical equivalent to determine the RPE for each formula. RPEs were optimized, and MAE, MedAE, SD of the AME, and percent of eyes achieving RPEs within the specified ranges of ±0.125 D, ±0.25 D, ±0.50 D, ±0.75 D, ±1.0 D were calculated. Subgroup analysis between different axial lengths was attempted but yielded insufficient statistical power to draw meaningful conclusions.

Results

A total of 103 eyes of 103 patients were included in our study after applying inclusion and exclusion criteria to 606 eyes from 2019 to 2021. Formulas ranked in ascending order by MAE were Kane, EVO, BUII, Hill-RBF, and LSF. The ascending rankings of MedAE were Kane, BUII, Hill-RBF, EVO, Ladas. Kane had a significantly lower MAE than Hill-RBF (p<0.001). EVO had the lowest SD of AMEs and the highest percentage of eyes within ±0.50 D. According to heteroscedastic testing, EVO also had a statistically significant lower SD than Hill-RBF.

Conclusion

Kane was the most accurate formula in terms of MAE and MedAE. EVO and BUII achieved marginally higher MAEs than Kane, suggesting these three formulas are comparable in performance. With the exception EVO and Hill-RBF, the heteroscedastic test yielded no significant differences in SD between the formulas. Although there were multiple statistically significant differences between the formulas in terms of MAE, MedAE, and SD, these differences may not be appreciable clinically. Lastly, there were no statistically significant differences in the percent of eyes with RPEs within ±0.50 D, suggesting similar clinical performance between formulas.

Accuracy of Traditional and Modern Formulas for Intraocular Lens Power Calculation After Radial Keratotomy Using Standard Keratometry

Purpose

To compare the accuracy of multiple traditional and modern intraocular lens (IOL) power calculation formulas in post-radial keratotomy (RK) patients undergoing cataract surgery.

Methods

This retrospective case series included 50 eyes with prior RK who underwent routine phacoemulsification surgery with single-piece acrylic IOL implantation (A constant = 118.8). Outcomes of multiple formulas were calculated. Included formulas were SRK/T, Holladay 1, Holladay 2, Haigis, Barrett True-K, Haigis and Barrett True-K (target refraction of 0.50 D), Barrett Universal II, Kane, PEARL-DGS, Shammas no history, DK SRK/T, DK SRK/T (target refraction of 0.50 D), Double K (DK) Holladay 1, and DK Holladay 1 (target refraction of 0.50 D). Averages of multiple combinations of best-performing single formulas were calculated. Primary outcome is mean absolute error (MAE).

Results

Haigis (with -0.50 D target refraction) and DK SRK/T showed the lowest mean and median absolute errors (MedAE) followed by Haigis, Barrett True-K, and Barrett True-K (with -0.50 D target refraction). Combinations of 3, 4, or 5 of best performing single formulas yielded good results with >60% of cases within +0.50 D of intended refraction and MAE around 0.50 D. The best performing formulas with flatter K readings were PEARL-DGS and Haigis (with additional -0.50 D target refraction) with MAE of 0.72 + 0.71 D and 0.70 + 0.70 D, respectively, followed by Barrett True-K (with intended -0.50 D target refraction) with MAE of 0.75 + 0.63 D.

Conclusion

Using an average of three or more Haigis (with -0.50 D target refraction), the Barrett True-K, DK Holladay 1, and DK SRK/T formulas showed better outcomes than using a single formula for IOLMaster 700 standard K readings. The PEARL-DGS formula showed better accuracy in eyes with flatter K readings (<38 D).

A Novel Role for Corneal Pachymetry in Planning Cataract Surgery by Determining Changes in Spherical Equivalent Resulting from a Previous LASIK Treatment

Objectives

To provide a metric to differentiate between hyperopic and myopic ablation of a prior LASIK treatment based on the corneal pachymetry profile after laser vision correction (LVC).

Methods

Pachymetry data were retrospectively recovered from patients who had previous LASIK for refractive purposes between 2019 and 2020. Patients with any corneal disorder were excluded. Ablation spherical equivalent was predicted from the central to semiperipheral corneal thickness (CPT) ratio, both values were provided by using the Pentacam user interface software (UI), and values were computed from extracted raw pachymetry data.

Results

Data of 157 eyes of 81 patients were collected, of which data were analysed for 73 eyes of 73 patients to avoid concurrence of measurements in both eyes per subject (42% female; mean age 40.9; SD 12.8). The CPT ratio cutoff for distinction between myopic and hyperopic LASIK was 0.86 for Pentacam UI data. Sensitivity and specificity were 0.7 and 0.95, respectively. Accuracy increased with computation of the CPT ratio based on extracted raw data with sensitivity and specificity of 0.87 and 0.99, respectively. There was a marked linear correlation between the CPT ratio and the ablation spherical equivalent (R2 = 0.93).

Conclusions

CPT ratio cutoffs can correctly classify if a cornea previously had a hyperopic versus myopic LASIK surgery and estimate the ablation spherical equivalent of such treatment. This could prove useful for increased accuracy of intraocular lens (IOL) calculations for patients with no historical data of their prior LVC surgery at the time of cataract surgery planning.

Visual Outcomes of Cataract Surgery in Patients With Keratoconus Using Toric and Non-toric Lenses

PURPOSE

To compare the accuracy and outcomes of different intraocular lens (IOL) power calculation formulas in eyes with keratoconus undergoing cataract surgery with toric and non-toric IOLs.

METHODS

This was a consecutive retrospective case series study including patients from the Cornea Service at the Department of Ophthalmology and Visual Sciences at the University of British Columbia, Vancouver, Canada, from 2000 to 2020. Keratoconus was diagnosed based on corneal topography and clinician opinion. Patients who underwent topography-guided photorefractive keratectomy, intracorneal ring segments implantation, or corneal transplant were excluded. The manifest spherical equivalent, prediction errors, and median absolute errors were calculated. Descriptive statistics were expressed as mean ± standard deviation.

RESULTS

There were 160 eyes from 101 patients; 136 eyes received non-toric lenses and 24 eyes received toric lenses. Most patients had mild disease (< 48.00 diopters [D]) when stratified by steep keratometry values. Patients with severe disease (> 53.00 D) were significantly more hyperopic following surgery (P < .05). The Barrett Universal II (0.26 D, inter-quartile range [IQR] = 0.4), Holladay 2 (0.31, IQR = 1.2), and SRK/T (0.42, IQR = 0.86) formulas had the lowest median absolute error. The postoperative prediction error following toric lens insertion was not significantly different than following non-toric lens insertion, and the mean absolute astigmatism was significantly reduced with toric lenses.

CONCLUSIONS

The Barrett Universal II, Holladay 2, and SRK/T were the most accurate IOL power calculation formulas in patients with keratoconus undergoing cataract surgery. Hyperopic surprise was increased in severe keratoconus. Toric IOLs may be considered in patients with mild keratoconus. [J Refract Surg. 2023;39(5):319-325.].

Technical failure rates for biometry between swept-source and older-generation optical coherence methods: a review and meta-analysis

PURPOSE

Precise ocular measurements are fundamental for achieving excellent target refraction following both cataract surgery and refractive lens exchange. Biometry devices with swept-source optical coherence tomography (SS-OCT) employ longer wavelengths (1055-1300 nm) in order to have better penetration through opaque lenses than those with partial coherence interferometry (PCI) or low-coherence optical reflectometry (LCOR) methods. However, to date a pooled analysis showing the technical failure rate (TFR) between the methods has not been published. The aim of this study was to compare the TFR in SS-OCT and in PCI/LCOR biometry.

METHODS

PubMed and Scopus were used to search the medical literature as of Feb 1, 2022. The following keywords were used in various combinations: optical biometry, partial coherence interferometry, low-coherence optical reflectometry, swept-source optical coherence tomography. Only clinical studies referring to patients undergoing routine cataract surgery, and employing at least two (PCI or LCOR vs. SS-OCT) optical methods for optical biometry in the same cohort of patients were included.

RESULTS

Fourteen studies were included in the final analysis, which presented results of 2,459 eyes of at least 1,853 patients. The overall TFR of all included studies was 5.47% (95% confidence interval [CI]: 3.66-8.08%; overall I² = 91.49%). The TFR was significantly different among the three methods (p < 0.001): 15.72% for PCI (95% CI: 10.73-22.46%; I² = 99.62%), 6.88% for LCOR (95% CI: 3.26-13.92%; I² = 86.44%), and 1.51% for SS-OCT (95% CI: 0.94-2.41%; I² = 24.64%). The pooled TFR for infrared methods (PCI and LCOR) was 11.12% (95% CI: 8.45-14.52%; I² = 78.28%), and was also significantly different to that of SS-OCT: 1.51% (95% CI: 0.94-2.41%; I² = 24.64%; p < 0.001).

CONCLUSIONS

A meta-analysis of the TFR of different biometry methods highlighted that SS-OCT biometry resulted in significantly decreased TFR compared to PCI/LCOR devices.

Analysis of accuracy of twelve intraocular lens power calculation formulas for eyes with axial hyperopia

PURPOSE

The purpose was to compare twelve intraocular lens power calculation formulas for eyes smaller than 22.0 mm in terms of absolute error (AE), the percentage of postoperative emmetropia, and agreement interval in Bland-Altman analysis.

METHODS

The data of hyperopic patients who underwent uneventful phacoemulsification between January 2016 and July 2021 were reviewed. Intraocular lens power was calculated using Holladay 1, SRK/T, Hoffer Q, Holladay 2, Haigis, Barrett Universal II, Hill-RBF, Ladas, Kane, Emmetropia Verifying Optical (EVO), Pearl-DGS, and K6 formulas. Three months after phacoemulsification, refraction was measured, and the mean AE was calculated. The percentage of patients with full visual acuity (VA) without any correction, with ± 0.25D, ±0.5D, ±0.75D, and limits of agreement for each formula was established.

RESULTS

Seventy-two patients, whose ocular axial length (AL) ranged between 20.02 mm and 21.98 mm, were included. The Kane formula achieved the lowest mean AE of 0.09 ± 0.09 just before EVO (0.12 ± 0.09), Hill-RBF (0.17 ± 0.12), and Hoffer Q formulas (0.19 ± 0.16). In addition, with the Kane formula, the percentage of patients with full VA without any correction (80.6%) was the highest ahead of EVO and Hoffer Q formulas (51.5% and 50.0%, respectively). Finally, Kane, EVO, and Hill-RBF obtained the lowest agreement interval (0.4923, 0.5815, and 0.7740, respectively).

CONCLUSION

The Kane formula is recommended for intraocular lens power calculation for eyeballs with the AL smaller than 22.0 mm. The EVO formula gives very promising results in regarding the accuracy of intraocular lens power for hyperopic eyes.

Analysis of accuracy of twelve intraocular lens power calculation formulas for eyes with axial myopia

PURPOSE

The aim of this study is to compare twelve intraocular lens power calculation formulas for eyes longer than 25.0 mm in terms of absolute error (AE), the percentage of postoperative emmetropia, and agreement interval in Bland-Altman analysis.

MATERIALS AND METHODS

Data of myopic patients who underwent uneventful phacoemulsification between January 2016 and July 2021 was reviewed. Intraocular lens power was calculated using Holladay 1, SRK/T, Hoffer Q, Holladay 2, Haigis, Barrett Universal II, Hill-RBF, Ladas, Kane, EVO, Pearl-DGS, and K6 formulas. Three months after phacoemulsification, refraction was measured, and mean AE was calculated. The percentage of patients with full visual acuity (VA) without any correction, with ± 0.25D, ±0.5D, ±0.75D, and limits of agreement for each formula were established.

RESULTS

Ninety-one patients, whose ocular axial length ranged between 25.03 mm and 28.91 mm, were included in the study. The Barrett Universal II formula achieved the lowest mean AE of 0.11 ± 0.11 (P < 0.001) just before Kane (0.13 ± 0.09; P < 0.001 except vs. Haigis and Holladay 2) and SRK/T formulas (0.18 ± 0.12). In addition, the Barrett Universal II formula had the highest percentage of patients with full VA without any correction (72.5%) followed by Kane and Holladay 2 formulas (56.0% and 49.5%, respectively). Finally, Barrett Universal II, Kane, and Haigis formulas obtained the lowest agreement interval (0.5725, 0.6088, and 0.8307, respectively).

CONCLUSION

The Barrett Universal II formula is recommended for intraocular lens power calculation for eyeballs with the axial length longer than 25.0 mm. The Kane formula also gives very promising results in regarding the accuracy of intraocular lens power for myopic eyes.

Recent developments in the intraocular lens formulae: An update

BACKGROUND

The precision of refractive outcomes after uneventful cataract surgery largely depends on the biometry and intraocular lens (IOL) formula used for selecting the IOL. To improve the accuracy of post-op refractive outcomes, several new IOL power calculation formulae have come up. This review would aim to summarise the differences among the new formulae in their performance among normal and variable ocular biometry conditions like short and long axial lengths.

METHODS

A literature review was performed by searching the PubMed and Cochrane databases from 2016 to 2021, identified 483 articles, of which 51 were included in the review.

RESULTS

We identified 15 new IOL formulas (including updates on older formulas) of which, only 8 newer formulas (BUII, Hill-RBF 2.0, Kane, Pearl DGS, LSF AI, Naesar 2, EVO 2.0 and VRF) met the eligibility criteria. They were compared according to the reported median absolute error, mean absolute error and percentage of eyes within 0.5D.

CONCLUSION

The Kane formula and Barrett Universal-II formula performed better than other formulas over the entire axial length (AL) spectrum. In the long eye (AL > 26.0 mm) sub-group, the Kane formula was the most accurate, while in the short eye (AL < 22.0 mm) sub-group, both Kane and EVO 2.0 formulas fared better than other formulas.

Optical Biometry and IOL Calculation in a Commercially Available Optical Coherence Tomography Device and Comparison With Pentacam AXL

PURPOSE

Optical devices are the gold standard for ocular biometry; however, they are unable to obtain high-quality optical coherence tomography (OCT) images. The current study aimed to evaluate ocular measurements and intraocular lens (IOL) calculation used in an anterior/posterior segment OCT device and to compare the results with those of a validated biometer.

DESIGN

Prospective evaluation of a diagnostic tool.

METHODS

This study enrolled healthy subjects at the Hygeia Clinic, Gdańsk, Poland, between October 2021 and November 2021. All individuals had ocular biometry measured with a validated biometer (Pentacam AXL) and with a new module of an anterior/posterior segment OCT device (Revo 80, Optopol Technologies). All IOL calculations were performed for the right eye with keratometric values from the Pentacam for one IOL: the Alcon AcrySof IQ SN60WF, with plano target setting.

RESULTS

The mean age of the 144 participants was 25.23 ± 7.15 years. The axial length measured with Revo was longer than with Pentacam AXL (24.08 ± 1.13 vs 23.98 ± 1.13; P < .0001), a 0.10 ± 0.04 mm difference. This translated into a significantly lower IOL power to achieve emmetropia for all formulas (-0.34 ± 0.15, -0.32 ± 0.13, -0.34 ± 0.19, and -0.30 ± 0.15 for the Hoffer Q, Holladay I, Haigis, and SRK/T formulas, respectively). The study showed high agreement between the devices: nearly 90% of eyes were within ±0.50 diopters for all of the analyzed formulas (r > 0.99).

CONCLUSIONS

The present study demonstrates that the results of IOL calculation with the OCT biometer have a very strong correlation with those obtained with the Pentacam AXL; however, axial length measurements and calculated IOL power cannot be considered interchangeable.

A Comprehensive Review of Pediatric Glaucoma Following Cataract Surgery and Progress in Treatment

Glaucoma following cataract surgery (GFCS) remains a serious postoperative complication of pediatric cataract surgery. Various risk factors, including age at lensectomy, intraocular lens implantation, posterior capsule status, associated ocular/systemic anomaly, additional intraocular surgery, and a family history of congenital cataract and GFCS, have been reported. However, the optimal surgical approach remains unclear. This review evaluates the diagnostic criteria, classification, risk factors, mechanism, and surgical management, especially the efficacy of minimally invasive glaucoma surgery, in GFCS, and aims to propose an optimal clinical management strategy for GFCS. The results of our review indicate that ab interno trabeculotomy (goniotomy) may be the most appropriate first-line treatment for GFCS.

Intraocular Lens Calculation Using 8 Formulas in Silicone Oil-Filled Eyes Undergoing Silicone Oil Removal and Phacoemulsification After Retinal Detachment

PURPOSE

To evaluate formulas for intraocular lens (IOL) calculation in silicone oil (SO)-filled eyes.

DESIGN

Retrospective, consecutive case series.

METHODS

We conducted our study at the Department of Ophthalmology, Goethe University, Frankfurt, Germany, and included SO-filled eyes that received SO removal combined with phacoemulsification and IOL implantation. Preoperative assessments included biometry (IOLMaster 700; Carl Zeiss Meditec). To evaluate the measurements, we compared the mean prediction error, and mean and median absolute prediction error of 8 different formulas.

RESULTS

A total of 90 eyes matched our inclusion criteria. The median absolute error was lowest in the Barrett Universal II formula (0.43 diopters [D] ± 0.75) followed by Kane (0.44 D ± 0.75), Hill-radial basis function (0.47 D ± 0.74), Holladay II (0.47 D ± 0.77), Sanders Retzlaff Kraff/theoretical (0.51 D ± 0.74), Holladay I (0.51 D ± 0.76), and Haigis and Hoffer Q (0.52 D ± 0.74 each). Regarding eyes within ±0.5 D Barrett Universal II (57.8%, 52 eyes) performed best, again followed by Kane (56.7%, 51 eyes) and Hill-radial basis function (54.4%, 49 eyes).

CONCLUSION

Using modern formulas for IOL calculation in oil-filled eyes improves predictability but still not as good as in unoperated eyes. This issue is created by the change in refractive index due to the SO fill and therefore a lower precision of axial length measurement and effective lens position prediction.

Accuracy of newer generation intraocular lens power calculation formulas in pediatric cataract patients

PURPOSE

To evaluate the accuracy of newer generation intraocular lens (IOL) power calculation formulas (EVO 2.0 and Kane) with established formulas (Barrett Universal II, Haigis and SRK/T) in pediatric cataract patients.

METHODS

Retrospective study. We enrolled 110 eyes (110 patients) in Eye Hospital of Wenzhou Medical University. All patients underwent uneventful cataract surgery and implanted with posterior chamber IOL in the bag. We calculate the mean prediction errors (PE) and percentage within 1 diopter (D) at 1 month to assess the accuracy, and percentage > 2D was defined as prediction accident. Then, we performed subgroup analysis according to age and axial length (AL).

RESULTS

The mean age and AL were 37.45 ± 23.28 months and 21.16 ± 1.29 mm. The mean PE for all patients was as follows: Barrett (- 0.30), EVO (0.18), Haigis (- 0.74), Kane (- 0.36), and SRK/T (0.58), p < 0.001. In addition, EVO and SRK/T formulas were relatively accurate in patients younger than 24 months and with AL ≤ 21 mm, while EVO got lower prediction accident rate than SRK/T (3/41 vs 8/41, 4/52 vs 5/52). Moreover, Barrett, EVO, and Kane formulas achieved better accuracy and lower prediction accident rate in patients older than 24 months and with AL > 21 mm (both > 51/69 and 43/58, and < 3/69 and 3/58).

CONCLUSIONS

In patients older than 24 months and with AL > 21 mm, Barrett, EVO, and Kane formulas were relatively accurate, while in patients younger than 24 months and with AL ≤ 21 mm, EVO was more accurate, followed by SRK/T formula.

A formula to improve the reliability of optical axial length measurement in IOL power calculation

To verify the influence of axial length (AL) variations after cataract surgery in IOL power calculation. Patients underwent ophthalmic evaluation before surgery, including optical biometry with IOLMaster 500. Same exams were repeated 2 months after surgery: AL of operated eye was evaluated using two modes (pseudophakic/aphakic options). Mean Keratometry and AL changes were analyzed. Furthermore, refractive prediction error (PE) was back-calculated with Barrett Universal-II, Hoffer-Q, Holladay-1 and SRK/T formulas. To eliminate any systematic error, the mean error (ME) was zeroed-out for each formula. MEs and median absolute errors (MedAEs) of PEs were analyzed. Two-hundred-one operated eyes of 201 patients and 201 opposite eyes were evaluated. In operated eyes, mean AL difference was - 0.11 ± 0.07 mm (p < 0.001) with pseudophakic option and 0.00 ± 0.07 mm (p = 0.922) with aphakic option. There were not-statistically significant differences between MedAE of PEs calculated after zeroing-out the ME with different ALs (p > 0.05). Instead, only MEs of PEs obtained with postoperative ALs-pseudophakic option were not-statistically different from zero (p > 0.05). AL measurement change after cataract surgery is probably due to a systematic error in optical biometer in case of phakic eyes. A correction factor applied to preoperative AL could eliminate any systematic error in IOL power calculation without modifying the lens constant.

A systemic review and network meta-analysis of accuracy of intraocular lens power calculation formulas in primary angle-closure conditions

BACKGROUND

For primary angle-closure and angle-closure glaucoma, the fact that refractive error sometimes deviates from predictions after intraocular lens (IOL) implantation is familiar to cataract surgeons. Since controversy remains in the accuracy of IOL power calculation formulas, both traditional and network meta-analysis on formula accuracy were conducted in patients with primary angle-closure conditions.

METHODS

A comprehensive literature search was conducted through Aug 2022, focusing on studies on intraocular lens power calculation in primary angle-closure (PAC) and primary angle-closure glaucoma (PACG). A systemic review and network meta-analysis was performed. Quality of studies were assessed. Primary outcomes were the mean absolute errors (MAE) and the percentages of eyes with a prediction error within ±0.50 diopiters (D) or ±1.00 D (% ±0.50/1.00 D) by different formulas.

RESULTS

Six retrospective studies involving 419 eyes and 8 formulas (Barrett Universal II, Kane, SRK/T, Hoffer Q, Haigis, Holladay I, RBF 3.0 and LSF) were included. SRK/T was used as a reference as it had been investigated in all the studies included. Direct comparison showed that none of the involved formula outperformed or was defeated by SRK/T significantly in terms of either MAE or % ±0.50/1.00 D (all P>0.05). Network comparison and ranking possibilities disclosed BUII, Kane, RBF 3.0 with statistically insignificant advantage. No significant publication bias was detected by network funnel plot.

CONCLUSIONS

No absolute advantage was disclosed among the formulas involved in this study for PAC/PACG eyes. Further carefully designed studies are warranted to evaluate IOL calculation formulae in this target population.

TRAIL REGISTRATION

Registration: PROSEPRO ID: CRD42022326541.

Accuracy of Six Intraocular Lens Power Calculations in Eyes with Axial Lengths Greater than 28.0 mm

The purpose of this study was to compare the accuracy of several intraocular (IOL) lens power calculation formulas in long eyes. This was a single-site retrospective consecutive case series that reviewed patients with axial lengths (AL) > 28.0 mm who underwent phacoemulsification. The Wang−Koch (WK) adjustment and Cooke-modified axial length (CMAL) adjustment were applied to Holladay 1 and SRK/T. The median absolute error (MedAE) and the percentage of eyes with prediction errors ±0.25 diopters (D), ±0.50 D, ±0.75 D, and ±1.00 D were used to analyze the formula’s accuracy. This study comprised a total of 35 eyes from 25 patients. The Kane formula had the lowest MedAE of all the formulas, but all were comparable except Holladay 1, which had a significantly lower prediction accuracy with either AL adjustment. The SRK/T formula with the CMAL adjustment had the highest accuracy in predicting the formula outcome within ±0.50 D. The newer formulas (BU-II, EVO, Hill-RBF version 3.0, and Kane) were all equally predictable in long eyes. The SRK/T formula with the CMAL adjustment was comparable to these newer formulas with better outcomes than the WK adjustment. The Holladay 1 with either AL adjustment had the lowest predictive accuracy.

Evaluation of Refractive Accuracy of ORA and the Factors Impacting Residual Astigmatism in Patients Implanted with Trifocal IOLs During Cataract Surgery: A Retrospective Observational Study

Purpose

Patients and Methods

Results

Conclusion

Evaluation of IOL power calculation with the Kane formula for pediatric cataract surgery

PURPOSE

To assess the accuracy of the Kane formula for intraocular lens (IOL) power calculation in the pediatric population.

METHODS

The charts of pediatric patients who underwent cataract surgery with in-the-bag IOL implantation with one of two IOL models (SA60AT or MA60AC) between 2012 and 2018 in The Hospital for Sick Children, Toronto, Ontario, CanFada, were retrospectively reviewed. The accuracy of IOL power calculation with the Kane formula was evaluated in comparison with the Barrett Universal II (BUII), Haigis, Hoffer Q, Holladay 1, and Sanders-Retzlaff-Kraff Theoretical (SRK/T) formulas.

RESULTS

Sixty-two eyes of 62 patients aged 6.2 (IQR 3.2-9.2) years were included. The SD values of the prediction error obtained by Kane (1.38) were comparable with those by BUII (1.34), Hoffer Q (1.37), SRK/T (1.40), Holaday 1 (1.41), and Haigis (1.50), all p > 0.05. A significant difference was observed between the Hoffer Q and Haigis formulas (p = 0.039). No differences in the median and mean absolute errors were found between the Kane formula (0.54 D and 0.91 ± 1.04 D) and BUII (0.50 D and 0.88 ± 1.00 D), Hoffer Q (0.48 D and 0.88 ± 1.05 D), SRK/T (0.72 D and 0.97 ± 1.00 D), Holladay 1 (0.63 D and 0.94 ± 1.05 D), and Haigis (0.57 D and 0.98 ± 1.13 D), p = 0.099.

CONCLUSION

This is the first study to investigate the Kane formula in pediatric cataract surgery. Our results place the Kane among the noteworthy IOL power calculation formulas in this age group, offering an additional means for improving IOL calculation in pediatric cataract surgery. The heteroscedastic statistical method was first implemented to evaluate formulas' predictability in children.

The repeatability and agreement of biometric measurements by dual Scheimpflug device with integrated optical biometer

To determine the repeatability of biometric measurements by dual Scheimpflug Devices with Integrated Optical Biometers and its agreement with partial coherence interferometry according to the axial length (AL), and the presence of cataracts. The present population-based cross-sectional study was conducted on the geriatric population in Tehran. For participants, imaging was performed by dual Scheimpflug Devices with Integrated Optical Biometers (Galilei G6) and partial coherence interferometry (IOL Master 500). All measurements were performed by one person. In both normal and cataractous eyes, the ICC values were above 0.99 for three measurements of AL, intraocular lens (IOL) power target, anterior chamber depth (ACD), central corneal thickness (CCT), flat and steep keratometry readings, and mean total corneal power (MTCP). The repeatability coefficient for the AL measurements was 0.003 and 0.002 in eyes with and without cataracts, respectively. The mean difference of AL between IOL Master 500 and Galilei G6 in normal and cataractous eyes was 0.015 and -0.003 mm, respectively. The 95% limits of agreement (LoA) of AL between these two devices were -0.09 to 0.12 mm in normal and -0.09 to 0.08 mm in cataractous eyes. The 95% LoA of ACD between the two devices was -0.13 to 0.36 mm and -0.10 to 0.31 mm in eyes without and with cataracts, respectively. The 95% LoA of steep K between the two devices was -0.63 to 0.32 and -1.04 to 0.89 diopter in normal and cataractous eyes, respectively. The results of the present study indicate the high repeatability of Galilei G6 in ocular biometric measurements. Galilei biometric measurements, had a very high agreement with the IOL Master 500.

Refractive Outcomes of Non-Toric and Toric Intraocular Lenses in Mild, Moderate and Advanced Keratoconus: A Systematic Review and Meta-Analysis

Background: To perform a systematic review and meta-analysis of the refractive outcomes of non-toric and toric intraocular lenses (IOLs) in keratoconus (KC) using different IOL power calculation formulas. Methods: A systematic search was conducted to identify studies that report on refractive outcomes of different IOL power calculation formulas in KC patients undergoing cataract surgery. Inclusion criteria were primary posterior chamber non-toric and toric monofocal intraocular lens implantation, data on the degree of KC, explicit mention of the formula used for each stage of KC, and the number of eyes in each category. We calculated and compared the absolute and mean prediction errors, percentage of eyes within 0.5 D and 1 D from target, and the weighted absolute prediction errors of IOL formulas, all were given for KC degrees I–III. Results: The bibliographic search yielded 582 studies published between 1996 and 2020, 14 of which (in total 456 eyes) met the criteria: three studies on non-toric IOL (98 eyes), eight studies on toric IOLs (98 eyes) and three studies of unknown separation between non-toric and toric IOLs (260 eyes). The lowest absolute prediction error (APE) for mild, moderate, and advanced KC was seen with Kane’s IOL power formula with keratoconus adjustment. The APE for the top five IOL power formulas ranged 0.49–0.73 diopters (D) for mild (83–94%) of eyes within 1 D from the target), 1.08–1.21 D for moderate (51–57% within 1 D), and 1.44–2.86 D for advanced KC (12–48% within 1 D). Conclusions: Cataract surgery in eyes with mild-to-moderate KC generally achieves satisfactory postoperative refractive results. In patients with advanced KC, a minority of the eyes achieved spherical equivalent refraction within 1 D from the target. The Kane’s formula with keratoconus adjustment showed the best results in all KC stages.

Applications of Artificial Intelligence in Myopia: Current and Future Directions

With the continuous development of computer technology, big data acquisition and imaging methods, the application of artificial intelligence (AI) in medical fields is expanding. The use of machine learning and deep learning in the diagnosis and treatment of ophthalmic diseases is becoming more widespread. As one of the main causes of visual impairment, myopia has a high global prevalence. Early screening or diagnosis of myopia, combined with other effective therapeutic interventions, is very important to maintain a patient's visual function and quality of life. Through the training of fundus photography, optical coherence tomography, and slit lamp images and through platforms provided by telemedicine, AI shows great application potential in the detection, diagnosis, progression prediction and treatment of myopia. In addition, AI models and wearable devices based on other forms of data also perform well in the behavioral intervention of myopia patients. Admittedly, there are still some challenges in the practical application of AI in myopia, such as the standardization of datasets; acceptance attitudes of users; and ethical, legal and regulatory issues. This paper reviews the clinical application status, potential challenges and future directions of AI in myopia and proposes that the establishment of an AI-integrated telemedicine platform will be a new direction for myopia management in the post-COVID-19 period.

Prediction of the axial lens position after cataract surgery using deep learning algorithms and multilinear regression

BACKGROUND

The prediction of anatomical axial intraocular lens position (ALP) is one of the major challenges in cataract surgery. The purpose of this study was to develop and test prediction algorithms for ALP based on deep learning strategies.

METHODS

We evaluated a large data set of 1345 biometric measurements from the IOLMaster 700 before and after cataract surgery. The target parameter was the intraocular lens (IOL) equator plane at half the distance between anterior and posterior apex. The relevant input parameters from preoperative biometry were extracted using a principal component analysis. A selection of neural network algorithms was tested using a 5-fold cross-validation procedure to avoid overfitting. The results were then compared with a traditional multilinear regression in terms of root mean squared prediction error (RMSE).

RESULTS

Corneal radius of curvature, axial length, anterior chamber depth, corneal thickness, lens thickness and patient age were identified as effective predictive parameters, whereas pupil size, horizontal corneal diameter and Chang-Waring chord did not enhance the model. From the tested algorithms, the Gaussian prediction regression and the Support Vector Machine algorithms performed best (RMSE = 0.2805 and 0.2731 mm), outperforming the multilinear prediction model (0.3379 mm). The mean absolute prediction error yielded 0.1998, 0.1948 and 0.2415 mm for the respective models.

CONCLUSION

Modern prediction techniques may have the potential to outperform traditional multilinear regression techniques as they can deal easily with nonlinearities between input and output parameters. However, in all cases a cross-validation is mandatory to avoid overfitting and misinterpretation of the results.

Advancements in intraocular lens power calculation formulas

PURPOSE OF REVIEW

We review recent studies comparing intraocular lens (IOL) formulas with an emphasis on selection of the highest performing formulas based on patient axial length, age, and history of previous corneal refractive surgery.

RECENT FINDINGS

The Barrett Universal II formula based on a theoretical model has consistently demonstrated high accuracy. The Olsen four-factor formula using ray tracing optics and the Hill-RBF calculator using artificial intelligence have also demonstrated good prediction results after being updated. Notably, the Kane formula, incorporating artificial intelligence, has overall shown the best performance for all axial lengths. Although newly developed and updated IOL formulas have improved refractive prediction in patients with short or long axial length eyes or prior history of corneal refractive surgery, these challenging cases still require special consideration. The Barrett True-K formula has shown accurate results regardless of preoperative data in eyes with previous myopic refractive surgery.

SUMMARY

Advancements in optical biometry and IOL calculation formulas continue to improve refractive outcomes. The clinician can optimize refractive outcomes in the majority of patients with the use of formulas that have shown consistent results and accuracy in several large studies.

Comparison of Kane, Hill-RBF 2.0, Barrett Universal II, and Emmetropia Verifying Optical Formulas in Eyes With Extreme Myopia

PURPOSE

To compare the accuracy of the Kane, Hill-RBF 2.0, Barrett Universal II (BUII), and Emmetropia Verifying Optical (EVO) formulas in calculating intraocular lens power in extremely myopic eyes.

METHODS

A total of 1,054 highly myopic eyes were included and divided into three groups according to axial length: control (⩾ 26 to < 28 mm), long (⩾ 28 to < 30 mm), and extreme axial length (⩾ 30 mm) groups. Prediction accuracies of the four formulas were compared and factors influencing the refractive errors were evaluated.

RESULTS

The Hill-RBF 2.0 formula generated the largest percentage of eyes with refractive errors within ±0.50 and ±1.00 D (71.44% and 94.59%, respectively, compared to 63.38% and 92.31% for the Kane, 61.76% and 94.02% for the BUII, and 59.01% and 87.57% for the EVO formulas; P < .001). The mean absolute errors of the Kane, Hill-RBF 2.0, BUII, and EVO formulas were 0.46 ± 0.38, 0.40 ± 0.39, 0.44 ± 0.30, and 0.58 ± 0.68 D (P < .001). In the long axial length group, the Hill-RBF 2.0 formula had the smallest MAE (all P < .001), whereas the extreme axial length group only had a smaller MAE than the Kane and EVO formulas (both P < .001). The accuracy of the Kane and Hill-RBF 2.0 formulas was affected by corneal curvature and A-constant; the accuracy of the BUII and EVO formulas was affected by corneal curvature, axial length, and A-constant.

CONCLUSIONS

The Hill-RBF 2.0 formula outperformed all three other formulas in eyes with axial lengths ⩾ 28 to < 30 mm, and outperformed the Kane and EVO formulas in eyes with axial lengths of 30 mm or greater. [J Refract Surg. 2021;37(10):680-685.].

Artificial intelligence applications and cataract management: A systematic review

Artificial intelligence (AI)-based applications exhibit the potential to improve the quality and efficiency of patient care in different fields, including cataract management. A systematic review of the different applications of AI-based software on all aspects of a cataract patient's management, from diagnosis to follow-up, was carried out in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. All selected articles were analyzed to assess the level of evidence according to the Oxford Centre for Evidence-Based Medicine 2011 guidelines, and the quality of evidence according to the Grading of Recommendations Assessment, Development and Evaluation system. Of the articles analyzed, 49 met the inclusion criteria. No data synthesis was possible for the heterogeneity of available data and the design of the available studies. The AI-driven diagnosis seemed to be comparable and, in selected cases, to even exceed the accuracy of experienced clinicians in classifying disease, supporting the operating room scheduling, and intraoperative and postoperative management of complications. Considering the heterogeneity of data analyzed, however, further randomized controlled trials to assess the efficacy and safety of AI application in the management of cataract should be highly warranted.

Intraocular lens power calculations in eyes with pseudoexfoliation syndrome

To compare refractive outcomes after cataract surgery in pseudoexfoliation syndrome (PEX) and control eyes and to investigate the accuracy of 3 intraocular lens (IOL) calculation formulas in these eyes. In this prospective comparative study 42 eyes (PEX group) and 38 eyes (control group) of 80 patients were included. The follow-up was 3 months. The refractive prediction error (RPE), mean absolute error (MAE), median absolute error (MedAE) and the percentages of eyes within ± 0.25 D, ± 0.5 D, ± 1.0 D and ± 2.0 D of prediction error were calculated. Three IOL calculation formulas (SRK/T, Barrett Universal II and Hill-RBF) were evaluated. PEX produced statistically significantly higher mean absolute errors and lower percentages of eyes within ± 0.5 D than control eyes in all investigated IOL calculation formulas. There were no statistically significant differences in the median absolute errors between the 3 formulas in either PEX or control eyes. Refractive outcomes after cataract surgery are statistically significantly worse in PEX than in control eyes. All three IOL calculation formulas produced similar results in both PEX and control eyes.Trial registration: ClinicalTrials.gov registration number NCT04783909.

Implications of the Relationship Between Refractive Error and Biometry in the Pathogenesis of Primary Angle Closure

Purpose

The purpose of this study was to investigate the relationship between refractive error and ocular biometry and its implication in the pathogenesis of primary angle closure (PAC).

Methods

We have retrospectively recruited 119 PAC eyes and 388 non-PAC eyes with an axial length (AL) of ≤25.0 mm and a spherical equivalent (SE) of ≥−6.0 diopters (D). Stepwise multiple regression was performed for keratometry value (K), AL, anterior chamber depth (ACD), and SE.

Results

PAC eyes were more likely to be in women and have a higher IOP and shorter AL than non-PAC eyes. In a multiple regression analysis, SE was not associated with PAC. The associations between AL and SE or AL and ACD were not different in PAC eyes compared with non-PAC eyes. However, the cornea was flatter in PAC eyes (β = −0.448, P < 0.001), and a flatter cornea was associated with more hyperopic refraction (β = −0.454, P < 0.001) and shallower ACD (β = 0.073, P < 0.001) in PAC eyes. ACD was not associated with SE in non-PAC eyes, but shallower ACD was associated with greater myopic refraction in PAC eyes (β = 1.117, P = 0.006).

Conclusions

PAC eyes seem to have flatter cornea compared with non-PAC eyes. A shallower ACD seems to be associated with greater myopic refraction in PAC eyes, but not in non-PAC eyes.

Artificial Intelligence in Cornea, Refractive Surgery, and Cataract: Basic Principles, Clinical Applications, and Future Directions

ABSTRACT

Corneal diseases, uncorrected refractive errors, and cataract represent the major causes of blindness globally. The number of refractive surgeries, either cornea- or lens-based, is also on the rise as the demand for perfect vision continues to increase. With the recent advancement and potential promises of artificial intelligence (AI) technologies demonstrated in the realm of ophthalmology, particularly retinal diseases and glaucoma, AI researchers and clinicians are now channeling their focus toward the less explored ophthalmic areas related to the anterior segment of the eye. Conditions that rely on anterior segment imaging modalities, including slit-lamp photography, anterior segment optical coherence tomography, corneal tomography, in vivo confocal microscopy and/or optical biometers, are the most commonly explored areas. These include infectious keratitis, keratoconus, corneal grafts, ocular surface pathologies, preoperative screening before refractive surgery, intraocular lens calculation, and automated refraction, among others. In this review, we aimed to provide a comprehensive update on the utilization of AI in anterior segment diseases, with particular emphasis on the recent advancement in the past few years. In addition, we demystify some of the basic principles and terminologies related to AI, particularly machine learning and deep learning, to help improve the understanding, research and clinical implementation of these AI technologies among the ophthalmologists and vision scientists. As we march toward the era of digital health, guidelines such as CONSORT-AI, SPIRIT-AI, and STARD-AI will play crucial roles in guiding and standardizing the conduct and reporting of AI-related trials, ultimately promoting their potential for clinical translation.

Correlating Kane formula with existing intraocular lens formulae for corneal curvatures and axial lengths

BACKGROUND:

To evaluate the predictability of the Kane formula in estimating postoperative refractive outcome with various corneal curvatures and axial lengths (ALs) besides comparing with existing intraocular lens (IOL) formulae.

MATERIALS AND METHODS:

A prospective cross-sectional study was carried out among patients having uneventful cataract surgery at an eye hospital. A total of 50 eyes were considered for the study. The corresponding A-constant for the model of IOL implanted into the patient's eye was taken along with the actual power of IOL implanted and corresponding predicted power for the IOL power inserted were taken for all the chosen formulae and was termed as "Adjusted Predicted Refractive Power." This was compared with the actual refractive outcome and the absolute error (AE) was measured. The eyes were separated into groups in terms of corneal curvature as flat (<42D), medium (42D–46D), and steep (>46D) corneas. In terms of AL, it was grouped as short (≤22 mm), medium (>22.0–<24.0 mm), and long (>24.0 mm) eyes.

RESULTS:

The study included 50 eyes and the mean AE for all the selected formulae were calculated for each group. Over the entire corneal curvature range, none of the formulae showed any significance when compared with the Kane formula (P > 0.05). In short AL, SRK-T formula had a statistical significance over the Kane formula (P = 0.043), whereas no other group had any significance over the Kane formula in AL groups.

CONCLUSION:

The study shows, all formulae (SRK-T, Holladay1, Hoffer Q, Hill RBF, Barrett Universal II, Kane) are interchangeable to predict the IOL power for any of the corneal curvature and ALs.

Correlating Kane formula with existing intraocular lens formulae for corneal curvatures and axial lengths

BACKGROUND

To evaluate the predictability of the Kane formula in estimating postoperative refractive outcome with various corneal curvatures and axial lengths (ALs) besides comparing with existing intraocular lens (IOL) formulae.

MATERIALS AND METHODS

A prospective cross-sectional study was carried out among patients having uneventful cataract surgery at an eye hospital. A total of 50 eyes were considered for the study. The corresponding A-constant for the model of IOL implanted into the patient's eye was taken along with the actual power of IOL implanted and corresponding predicted power for the IOL power inserted were taken for all the chosen formulae and was termed as "Adjusted Predicted Refractive Power." This was compared with the actual refractive outcome and the absolute error (AE) was measured. The eyes were separated into groups in terms of corneal curvature as flat (<42D), medium (42D-46D), and steep (>46D) corneas. In terms of AL, it was grouped as short (≤22 mm), medium (>22.0-<24.0 mm), and long (>24.0 mm) eyes.

RESULTS

The study included 50 eyes and the mean AE for all the selected formulae were calculated for each group. Over the entire corneal curvature range, none of the formulae showed any significance when compared with the Kane formula (P > 0.05). In short AL, SRK-T formula had a statistical significance over the Kane formula (P = 0.043), whereas no other group had any significance over the Kane formula in AL groups.

CONCLUSION

The study shows, all formulae (SRK-T, Holladay1, Hoffer Q, Hill RBF, Barrett Universal II, Kane) are interchangeable to predict the IOL power for any of the corneal curvature and ALs.

Changes in the Ocular Parameters of Patients with Graves' Disease after Antithyroid Drug Treatment

Background and Objectives: To find the differences in ocular axial length, keratometric measurements, and intraocular lens (IOL) power in patients with Graves’ disease (GD) after treatment with a thionamide antithyroid drug (ATD), methimazole. Materials and Methods: The medical charts of 28 patients (4 males and 24 females; mean age: 47.2 ± 21.2 years) were studied. Each patient was examined twice using an IOL Master Device and keratometry at the first visit (before ATD treatment) and after 1 month of ATD treatment. The IOL power was calculated for each patient using the Hoffer Q, SRK-2, and SRK/T formulas according to axial length. Results: After 1 month, the axial length increased (right and left eyes: p < 0.001 and p = 0.05, respectively). Based on keratometry, changes in the horizontal and vertical optical power [in diopters (D)] were not statistically significant. However, the IOL power changed after 1 month of ATD treatment in 64.3% of the patients. In 14 patients (50%), there was a 0.5–1.0 D IOL power decrease in single eyes; in two patients (7.1%), an IOL power decrease of 0.5–1.0 D in both eyes; and in two patients (7.1%), a 0.5 D IOL power increase in single eyes. The calculated IOL power values were lower after ATD treatment (right and left eyes, p = 0.010 and p = 0.018, respectively). Conclusions: The IOL power changed in 64.3% of GD patients after ATD treatment. Therefore, avoiding cataract surgery at the early stage of ATD treatment would be appropriate for selecting a more accurate IOL power.

Current Challenges in the Postoperative Management of Cataract Surgery

Collapse

Artificial Intelligence, Machine Learning and Calculation of Intraocular Lens Power

BACKGROUND AND PURPOSE

In the last decade, artificial intelligence and machine learning algorithms have been more and more established for the screening and detection of diseases and pathologies, as well as for describing interactions between measures where classical methods are too complex or fail. The purpose of this paper is to model the measured postoperative position of an intraocular lens implant after cataract surgery, based on preoperatively assessed biometric effect sizes using techniques of machine learning.

PATIENTS AND METHODS

In this study, we enrolled 249 eyes of patients who underwent elective cataract surgery at Augenklinik Castrop-Rauxel. Eyes were measured preoperatively with the IOLMaster 700 (Carl Zeiss Meditec), as well as preoperatively and postoperatively with the Casia 2 OCT (Tomey). Based on preoperative effect sizes axial length, corneal thickness, internal anterior chamber depth, thickness of the crystalline lens, mean corneal radius and corneal diameter a selection of 17 machine learning algorithms were tested for prediction performance for calculation of internal anterior chamber depth (AQD_post) and axial position of equatorial plane of the lens in the pseudophakic eye (LEQ_post).

RESULTS

The 17 machine learning algorithms (out of 4 families) varied in root mean squared/mean absolute prediction error between 0.187/0.139 mm and 0.255/0.204 mm (AQD_post) and 0.183/0.135 mm and 0.253/0.206 mm (LEQ_post), using 5-fold cross validation techniques. The Gaussian Process Regression Model using an exponential kernel showed the best performance in terms of root mean squared error for prediction of AQDpost and LEQpost. If the entire dataset is used (without splitting for training and validation data), comparison of a simple multivariate linear regression model vs. the algorithm with the best performance showed a root mean squared prediction error for AQD_post/LEQ_post with 0.188/0.187 mm vs. the best performance Gaussian Process Regression Model with 0.166/0.159 mm.

CONCLUSION

In this paper we wanted to show the principles of supervised machine learning applied to prediction of the measured physical postoperative axial position of the intraocular lenses. Based on our limited data pool and the algorithms used in our setting, the benefit of machine learning algorithms seems to be limited compared to a standard multivariate regression model.