AIM: To evaluate the potential of the Support Vector Machine Regression model (SVM-RM) and Multilayer Neural Network Ensemble model (MLNN-EM) to improve the intraocular lens (IOL) power calculation for clinical workflow. BACKGROUND: Current IOL power calculation methods are limited in their accuracy with the possibility of decreased accuracy especially in eyes with an unusual ocular dimension. In case of an improperly calculated power of the IOL in cataract or refractive lens replacement surgery there is a risk of re-operation or further refractive correction. This may create potential complications and discomfort for the patient. METHODS: A dataset containing information about 2,194 eyes was obtained using data mining process from the Electronic Health Record (EHR) system database of the Gemini Eye Clinic. The dataset was optimized and split into the selection set (used in the design for models and training), and the verification set (used in the evaluation). The set of mean prediction errors (PEs) and the distribution of predicted refractive errors were evaluated for both models and clinical results (CR). RESULTS: Both models performed significantly better for the majority of the evaluated parameters compared with the CR. There was no significant difference between both evaluated models. In the ±0.50 D PE category both SVM-RM and MLNN-EM were slightly better than the Barrett Universal II formula, which is often presented as the most accurate calculation formula. CONCLUSION: In comparison to the current clinical method, both SVM-RM and MLNN-EM have achieved significantly better results in IOL calculations and therefore have a strong potential to improve clinical cataract refractive outcomes.

Department of Circuit Theory Faculty of Electrical Engineering Czech Technical University Prague Prague Czech Republic

Department of Ophthalmology 3rd Faculty of Medicine Charles University Prague Czech Republic

Research and Development Department Gemini Eye Clinic Zlin Czech Republic

See more in PubMed

Abell RG, Vote BJ. Cost-effectiveness of femtosecond laser-assisted cataract surgery versus phacoemulsification cataract surgery. Ophthalmology. 2014;121(1):10–16. doi: 10.1016/j.ophtha.2013.07.056. PubMed DOI

Abulafia A, Barrett GD, Rotenberg M, Kleinmann G, Levy A, Reitblat O, Koch DD, Wang L, Assia EI. Intraocular lens power calculation for eyes with an axial length greater than 26.0 mm: comparison of formulas and methods. Journal of Cataract & Refractive Surgery. 2015;41(3):548–556. doi: 10.1016/j.jcrs.2014.06.033. PubMed DOI

Anastassiou GA. Multivariate hyperbolic tangent neural network approximation. Computers & Mathematics with Applications. 2011;61(4):809–821. doi: 10.1016/j.camwa.2010.12.029. DOI

Armstrong RA. Statistical guidelines for the analysis of data obtained from one or both eyes. Ophthalmic and Physiological Optics. 2013;33(1):7–14. doi: 10.1111/opo.12009. PubMed DOI

Astbury N, Ramamurthy B. How to avoid mistakes in biometry. Community Eye Health Journal. 2006;19(60):70–71. PubMed PMC

Chen Y-A, Hirnschall N, Findl O. Evaluation of 2 new optical biometry devices and comparison with the current gold standard biometer. Journal of Cataract & Refractive Surgery. 2011;37(3):513–517. doi: 10.1016/j.jcrs.2010.10.041. PubMed DOI

Cicchetti DV. Guidelines, criteria, and rules of thumb for evaluating normed and standardized assessment instruments in psychology. Psychological Assessment. 1994;6(4):284–290. doi: 10.1037/1040-3590.6.4.284. DOI

Clarke GP, Burmeister J. Comparison of intraocular lens computations using a neural network versus the Holladay formula. Journal of Cataract & Refractive Surgery. 1997;23(10):1585–1589. doi: 10.1016/S0886-3350(97)80034-X. PubMed DOI

Conrad-Hengerer I, Al Sheikh M, Hengerer FH, Schultz T, Dick HB. Comparison of visual recovery and refractive stability between femtosecond laser–assisted cataract surgery and standard phacoemulsification: six-month follow-up. Journal of Cataract & Refractive Surgery. 2015;41(7):1356–1364. doi: 10.1016/j.jcrs.2014.10.044. PubMed DOI

Cooke DL, Cooke TL. Comparison of 9 intraocular lens power calculation formulas. Journal of Cataract & Refractive Surgery. 2016;42(8):1157–1164. doi: 10.1016/j.jcrs.2016.06.029. PubMed DOI

Dixon WJ, Mood AM. The statistical sign test. Journal of the American Statistical Association. 1946;41(236):557–566. doi: 10.1080/01621459.1946.10501898. PubMed DOI

Ferrari S, Stengel RF. Smooth function approximation using neural networks. IEEE Transactions on Neural Networks. 2005;16(1):24–38. doi: 10.1109/TNN.2004.836233. PubMed DOI

Frampton G, Harris P, Cooper K, Lotery A, Shepherd J. The clinical effectiveness and cost-effectiveness of second-eye cataract surgery: a systematic review and economic evaluation. Health Technology Assessment. 2014;18(68):1–206. doi: 10.3310/hta18680. PubMed DOI PMC

Gale RP, Saldana M, Johnston RL, Zuberbuhler B, McKibbin M. Benchmark standards for refractive outcomes after NHS cataract surgery. Eye. 2009;23(1):149–152. doi: 10.1038/sj.eye.6702954. PubMed DOI

Gatinel D. Calculation of implant—theoretical formula. 2018. https://www.gatinel.com/en/recherche-formation/biometrie-oculaire-calcul-dimplant/calcul-dimplant-formule-theorique/ [13 February 2019]. https://www.gatinel.com/en/recherche-formation/biometrie-oculaire-calcul-dimplant/calcul-dimplant-formule-theorique/

Girosi F. Some extensions of radial basis functions and their applications in artificial intelligence. Computers & Mathematics with Applications. 1992;24(12):61–80. doi: 10.1016/0898-1221(92)90172-E. DOI

Gökce SE, Montes De Oca I, Cooke DL, Wang L, Koch DD, Al-Mohtaseb Z. Accuracy of 8 intraocular lens calculation formulas in relation to anterior chamber depth in patients with normal axial lengths. Journal of Cataract & Refractive Surgery. 2018;44(3):362–368. doi: 10.1016/j.jcrs.2018.01.015. PubMed DOI

Haag-Streit AG, Koeniz, Switzerland . Hill-RBF method. Koeniz, Switzerland: White Paper; 2017.

Haigis W. Challenges and approaches in modern biometry and IOL calculation. Saudi Journal of Ophthalmology. 2012;26(1):7–12. doi: 10.1016/j.sjopt.2011.11.007. PubMed DOI PMC

Herbrich R. Support vector learning for ordinal regression. 9th International Conference on Artificial Neural Networks: ICANN’99; Edinburgh: IEEE; 1999. pp. 97–102.

Hill W. Hill-RBF Calculator. 2018. www.rbfcalculator.com. [9 December 2018]. www.rbfcalculator.com

Hoffer KJ. Biometry of 7,500 cataractous eyes. American Journal of Ophthalmology. 1980;90(3):360–368. doi: 10.1016/S0002-9394(14)74917-7. PubMed DOI

Jap D, Stöttinger M, Bhasin S. Support vector regression. Proceedings of the Fourth Workshop on Hardware and Architectural Support for Security and Privacy—HASP’15; New York, USA. 2015.

Jiang Y, Li M, Zhou Z-H. Mining extremely small data sets with application to software reuse. Software: Practice and Experience. 2009;39(4):423–440. doi: 10.1002/spe.905. DOI

Kaiser J. Dealing with missing values in data. Journal of Systems Integration. 2014;5:42–51. doi: 10.20470/jsi.v5i1.178. DOI

Kane JX, Van Heerden A, Atik A, Petsoglou C. Intraocular lens power formula accuracy: comparison of 7 formulas. Journal of Cataract & Refractive Surgery. 2016;42(10):1490–1500. doi: 10.1016/j.jcrs.2016.07.021. PubMed DOI

Kane JX, Van Heerden A, Atik A, Petsoglou C. Accuracy of 3 new methods for intraocular lens power selection. Journal of Cataract & Refractive Surgery. 2017;43(3):333–339. doi: 10.1016/j.jcrs.2016.12.021. PubMed DOI

Koch DD, Hill W, Abulafia A, Wang L. Pursuing perfection in intraocular lens calculations: I. Logical approach for classifying IOL calculation formulas. Journal of Cataract & Refractive Surgery. 2017;43(6):717–718. doi: 10.1016/j.jcrs.2017.06.006. PubMed DOI

Kononenko I, Kukar M. Machine learning and data mining. Sawston: Woodhead Publishing Limited; 2007.

Koo TK, Li MY. A guideline of selecting and reporting intraclass correlation coefficients for reliability research. Journal of Chiropractic Medicine. 2016;15(2):155–163. doi: 10.1016/j.jcm.2016.02.012. PubMed DOI PMC

Kourentzes N, Barrow DK, Crone SF. Neural network ensemble operators for time series forecasting. Expert Systems with Applications. 2014;41(9):4235–4244. doi: 10.1016/j.eswa.2013.12.011. DOI

Kurban T, Beşdok E. A comparison of RBF neural network training algorithms for inertial sensor based terrain classification. Sensors. 2009;9(8):6312–6329. doi: 10.3390/s90806312. PubMed DOI PMC

Lampariello F, Sciandrone M. Efficient training of RBF neural networks for pattern recognition. IEEE Transactions on Neural Networks. 2001;12(5):1235–1242. doi: 10.1109/72.950152. PubMed DOI

Le NQK, Ho Q-T, Ou Y-Y. Incorporating deep learning with convolutional neural networks and position specific scoring matrices for identifying electron transport proteins. Journal of Computational Chemistry. 2017;38(23):2000–2006. doi: 10.1002/jcc.24842. PubMed DOI

Le NQK, Ho Q-T, Ou Y-Y. Classifying the molecular functions of Rab GTPases in membrane trafficking using deep convolutional neural networks. Analytical Biochemistry. 2018;555:33–41. doi: 10.1016/j.ab.2018.06.011. PubMed DOI

Le NQK, Nguyen V-N. SNARE-CNN: a 2D convolutional neural network architecture to identify SNARE proteins from high-throughput sequencing data. PeerJ Computer Science. 2019;5:e177. doi: 10.7717/peerj-cs.177. PubMed DOI PMC

Le NQK, Ou Y-Y. Incorporating efficient radial basis function networks and significant amino acid pairs for predicting GTP binding sites in transport proteins. BMC Bioinformatics. 2016a;17(S19):501. doi: 10.1186/s12859-016-1369-y. PubMed DOI PMC

Le NQK, Ou Y-Y. Prediction of FAD binding sites in electron transport proteins according to efficient radial basis function networks and significant amino acid pairs. BMC Bioinformatics. 2016b;17(1):298. doi: 10.1186/s12859-016-1163-x. PubMed DOI PMC

Le NQK, Yapp EKY, Ho Q-T, Nagasundaram N, Ou Y-Y, Yeh H-Y. iEnhancer-5Step: identifying enhancers using hidden information of DNA sequences via Chou’s 5-step rule and word embedding. Analytical Biochemistry. 2019;571:53–61. doi: 10.1016/j.ab.2019.02.017. PubMed DOI

Lee TH, Sung MS, Cui L, Li Y, Yoon KC. Factors affecting the accuracy of intraocular lens power calculation with lenstar. Chonnam Medical Journal. 2015;51(2):91–96. doi: 10.4068/cmj.2015.51.2.91. PubMed DOI PMC

Leys C, Ley C, Klein O, Bernard P, Licata L. Detecting outliers: do not use standard deviation around the mean, use absolute deviation around the median. Journal of Experimental Social Psychology. 2013;49(4):764–766. doi: 10.1016/j.jesp.2013.03.013. DOI

Linebarger EJ, Hardten DR, Shah GK, Lindstrom RL. Phacoemulsification and modern cataract surgery. Survey of Ophthalmology. 1999;44(2):123–147. doi: 10.1016/S0039-6257(99)00085-5. PubMed DOI

Mahdavi S, Holladay J. IOLMaster® 500 and integration of the Holladay 2 formula for intraocular lens calculations. European Ophthalmic Review. 2011;5(2):134–135. doi: 10.17925/EOR.2011.05.02.134. DOI

MathWorks . Fit a support vector machine regression model. Natick: MathWorks; 2017a.

MathWorks . Function fitting neural network—MATLAB fitnet—MathWorks Benelux. Natick: MathWorks; 2017b.

MathWorks . Levenberg–Marquardt backpropagation. Natick: MathWorks; 2017c.

MathWorks . Matlab documentation. Natick: MathWorks; 2019.

Melles RB, Holladay JT, Chang WJ. Accuracy of intraocular lens calculation formulas. Ophthalmology. 2018;125(2):169–178. doi: 10.1016/j.ophtha.2017.08.027. PubMed DOI

Mercier É, Even A, Mirisola E, Wolfersberger D, Sciamanna M. Numerical study of extreme events in a laser diode with phase-conjugate optical feedback. Physical Review E. 2015;91(4):042914. doi: 10.1103/PhysRevE.91.042914. PubMed DOI

Mongillo M. Choosing basis functions and shape parameters for radial basis function methods. SIAM Undergraduate Research Online. 2011;4:190–209. doi: 10.1137/11S010840. DOI

Nguyen D, Widrow B. Improving the learning speed of 2-layer neural networks by choosing initial values of the adaptive weights. 1990 IJCNN International Joint Conference on Neural Networks; San Diego, CA, USA. 1990.

Norrby S. Sources of error in intraocular lens power calculation. Journal of Cataract & Refractive Surgery. 2008;34(3):368–376. doi: 10.1016/j.jcrs.2007.10.031. PubMed DOI

Olsen T. Prediction of the effective postoperative (intraocular lens) anterior chamber depth. Journal of Cataract & Refractive Surgery. 2006;32:419–424. doi: 10.1016/j.jcrs.2005.12.139. PubMed DOI

Olsen T. Calculation of intraocular lens power: a review. Acta Ophthalmologica Scandinavica. 2007;85(5):472–485. doi: 10.1111/j.1755-3768.2007.00879.x. PubMed DOI

Olson M, Wyner A, Berk R. Modern neural networks generalize on small data sets. Neural Information Processing Systems (NIPS); Montréal, Canada. 2018.

Park J, Sandberg IW. Universal approximation using radial-basis-function networks. Neural Computation. 1991;3(2):246–257. doi: 10.1162/neco.1991.3.2.246. PubMed DOI

Pascolini D, Mariotti SP. Global estimates of visual impairment: 2010. British Journal of Ophthalmology. 2012;96(5):614–618. doi: 10.1136/bjophthalmol-2011-300539. PubMed DOI

Ranganathan A. The Levenberg–Marquardt algorithm. Tutoral on LM Algorithm. 2004. https://www.semanticscholar.org/paper/The-Levenberg-Marquardt-Algorithm-Ranganathan/1e0078d36080d288dc240877fdb33f54ef5028c6#references https://www.semanticscholar.org/paper/The-Levenberg-Marquardt-Algorithm-Ranganathan/1e0078d36080d288dc240877fdb33f54ef5028c6#references

Retzlaff JA, Sanders DR, Kraff MC. Development of the SRK/T intraocular lens implant power calculation formula. Journal of Cataract & Refractive Surgery. 1990;16(3):333–340. doi: 10.1016/S0886-3350(13)80705-5. PubMed DOI

Roberts TV, Hodge C, Sutton G, Lawless M, Contributors to the Vision Eye Institute IOL outcomes registry Comparison of Hill-radial basis function, Barrett Universal and current third generation formulas for the calculation of intraocular lens power during cataract surgery. Clinical & Experimental Ophthalmology. 2018;46(3):240–246. doi: 10.1111/ceo.13034. PubMed DOI

Romero Reyes IV, Fedyushkina IV, Skvortsov VS, Filimonov DA. Prediction of progesterone receptor inhibition by high-performance neural network algorithm. International Journal of Mathematical Models and Methods in Applied Sciences. 2013;7:303–310.

Ross KA, Jensen CS, Snodgrass R, Dyreson CE, Jensen CS, Snodgrass R, Skiadopoulos S, Sirangelo C, Larsgaard ML, Grahne G, Kifer D, Jacobsen H-A, Hinterberger H, Deutsch A, Nash A, Wada K, Aalst WMP, Dyreson C, Mitra P, Witten IH, Liu B, Aggarwal CC, Özsu MT, Ogbuji C, Patel C, Weng C, Patel C, Weng C, Wright A, Shabo (Shvo) A, Russler D, Rocha RA, Russler D, Lussier YA, Chen JL, Russler D, Zaki MJ, Corral A, Vassilakopoulos M, Gunopulos D, Wolfram D, Venkatasubramanian S, Gunopulos D, Vazirgiannis M, Davidson I, Sarawagi S, Peyton L, Hinterberger H, Speegle G, Vianu V, Van Gucht D, Etzion O, Etzion O, Curbera F, Ericsson A, Berndtsson M, Mellin J, Aalst WMP, Gray PMD, Trajcevski G, Wolfson O, Scheuermann P, Dorai C, Weiner M, Borgida A, Mylopoulos J, Vossen G, Reuter A, Grahne G, Tannen V, Elnikety S, Fekete A, Bertossi L, Geerts F, Geerts F, Fan W, Westerveld T, Jacobsen H-A, Gurrin C, Westerveld T, Etzion O, Kekäläinen J, Arvola P, Junkkari M, Wada K, Mouratidis K, Yu JX, Yao Y, Gehrke J, Babu S, Reuter A, Palmer N, Leung CK-S, Aalst WMP, Carroll MW, Gokhale A, Ouzzani M, Medjahed B, Elmagarmid AK, Manegold S, Cormode G, Mankovskii S, Zhang D, Härder T, Gao W, Niu C, Li Q, Yang Y, Refaeilzadeh P, Tang L, Liu H, Pedersen TB, Morfonios K, Ioannidis Y, Böhlen MH, Jensen CS, Snodgrass RT, Chen L. Cross-Validation. In: Liu L, Özsu MT, editors. Encyclopedia of Database Systems. Boston: Springer; 2009. pp. 532–538.

Shajari M, Kolb CM, Petermann K, Böhm M, Herzog M, De’Lorenzo N, Schönbrunn S, Kohnen T. Comparison of 9 modern intraocular lens power calculation formulas for a quadrifocal intraocular lens. Journal of Cataract & Refractive Surgery. 2018;44(8):942–948. doi: 10.1016/j.jcrs.2018.05.021. PubMed DOI

Shammas HJ, Shammas MC. Measuring the cataractous lens. Journal of Cataract & Refractive Surgery. 2015;41(9):1875–1879. doi: 10.1016/j.jcrs.2015.10.036. PubMed DOI

Shrivastava AK, Behera P, Kumar B, Nanda S. Precision of intraocular lens power prediction in eyes shorter than 22 mm: an analysis of 6 formulas. Journal of Cataract & Refractive Surgery. 2018;44(11):1317–1320. doi: 10.1016/j.jcrs.2018.07.023. PubMed DOI

Smola AJ, Schölkopf B. A tutorial on support vector regression. Statistics and Computing. 2004;14(3):199–222. doi: 10.1023/B:STCO.0000035301.49549.88. DOI

Snyder EM. Hill-RBF Calculator in clinical practice. 2019. https://crstodayeurope.com/articles/new-frontiers-in-iol-prediction-for-improved-refractive-outcomes/hill-rbf-calculator-in-clinical-practice/ [22 January 2019]. https://crstodayeurope.com/articles/new-frontiers-in-iol-prediction-for-improved-refractive-outcomes/hill-rbf-calculator-in-clinical-practice/

The American Society of Cataract and Refractive Surgery ASCRS Announces Hill-RBF Calculator for Cataract Surgery IOL Power Calculations. 2018. http://ascrs.org/about-ascrs/news-about/ascrs-announces-hill-rbf-calculator-cataract-surgery-iol-power-calculations. [5 December 2018]. http://ascrs.org/about-ascrs/news-about/ascrs-announces-hill-rbf-calculator-cataract-surgery-iol-power-calculations

Thulasi P, Khandelwal SS, Randleman JB. Intraocular lens alignment methods. Current Opinion in Ophthalmology. 2016;27(1):65–75. doi: 10.1097/ICU.0000000000000225. PubMed DOI

Trafalis TB, Ince H. Support vector machine for regression and applications to financial forecasting. Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks. IJCNN 2000. Neural Computing: New Challenges and Perspectives for the New Millennium; Como, Italy. Piscataway: IEEE; 2000.

Tuckova J. Selected applications of the artificial neural networks at the signal processing. Prague: Nakladatelství ČVUT; 2009.

Wang B, Gong NZ. Stealing hyperparameters in machine learning. 2018 IEEE Symposium on Security and Privacy (SP); San Francisco, CA, USA. Piscataway: IEEE; 2018. pp. 36–52.

Wang L, Koch DD, Hill W, Abulafia A. Pursuing perfection in intraocular lens calculations: III. Criteria for analyzing outcomes. Journal of Cataract & Refractive Surgery. 2017a;43(8):999–1002. doi: 10.1016/j.jcrs.2017.08.003. PubMed DOI

Wang W, Yan W, Fotis K, Prasad NM, Lansingh VC, Taylor HR, Finger RP, Facciolo D, He M. Cataract surgical rate and socioeconomics: a global study. Investigative Opthalmology & Visual Science. 2017b;57(14):5872–5881. doi: 10.1167/iovs.16-19894. PubMed DOI

Westfall PH, Troendle JF, Pennello G. Multiple McNemar Tests. Biometrics. 2010;66(4):1185–1191. doi: 10.1111/j.1541-0420.2010.01408.x. PubMed DOI PMC

Wu Y, Wang H, Zhang B, Du K-L. Using radial basis function networks for function approximation and classification. ISRN Applied Mathematics. 2012;2012:1–34. doi: 10.5402/2012/324194. DOI

Wu C-H, Wei C-C, Su D-C, Chang M-H, Ho J-M. Travel time prediction with support vector regression. Proceedings of the 2003 IEEE International Conference on Intelligent Transportation Systems; Shanghai, China. Shanghai: IEEE; 2003. pp. 1438–1442.

Yamaguchi T, Negishi K, Tsubota K. Functional visual acuity measurement in cataract and intraocular lens implantation. Current Opinion in Ophthalmology. 2011;22(1):31–36. doi: 10.1097/ICU.0b013e3283414f36. PubMed DOI

Yu P-S, Chen S-T, Chang I-F. Support vector regression for real-time flood stage forecasting. Journal of Hydrology. 2006;328(3–4):704–716. doi: 10.1016/j.jhydrol.2006.01.021. DOI

Zeng Z-Q, Yu H-B, Xu H-R, Xie Y-Q, Gao J. Fast training Support Vector Machines using parallel sequential minimal optimization. 2008 3rd International Conference on Intelligent System and Knowledge Engineering; Xiamen, China. Piscataway: IEEE; 2008. pp. 997–1001.