{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2024,8,22]],"date-time":"2024-08-22T00:44:32Z","timestamp":1724287472132},"reference-count":25,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2020,12,18]],"date-time":"2020-12-18T00:00:00Z","timestamp":1608249600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Big Data"],"abstract":"Purpose:<\/jats:bold> Our work introduces a highly accurate, safe, and sufficiently explicable machine-learning (artificial intelligence) model of intraocular lens power (IOL) translating into better post-surgical outcomes for patients with cataracts. We also demonstrate its improved predictive accuracy over previous formulas.<\/jats:p>Methods:<\/jats:bold> We collected retrospective eye measurement data on 5,331 eyes from 3,276 patients across multiple centers who received a lens implantation during cataract surgery. The dependent measure is the post-operative manifest spherical equivalent error from intended and the independent variables are the patient- and eye-specific characteristics. This dataset was split so that one subset was for formula construction and the other for validating our new formula. Data excluded fellow eyes, so as not to confound the prediction with bilateral eyes.<\/jats:p>Results:<\/jats:bold> Our formula is three times more precise than reported studies with a median absolute IOL error of 0.204 diopters (D). When converted to absolute predictive refraction errors on the cornea, the median error is 0.137 D which is close to the IOL manufacturer tolerance. These estimates are validated out-of-sample and thus are expected to reflect the future performance of our prediction formula, especially since our data were collected from a wide variety of patients, clinics, and manufacturers.<\/jats:p>Conclusion:<\/jats:bold> The increased precision of IOL power calculations has the potential to optimize patient positive refractive outcomes. Our model also provides uncertainty plots that can be used in tandem with the clinician\u2019s expertise and previous formula output, further enhancing the safety.<\/jats:p>Translational relavance:<\/jats:bold> Our new machine learning process has the potential to significantly improve patient IOL refractive outcomes safely.<\/jats:p>","DOI":"10.3389\/fdata.2020.572134","type":"journal-article","created":{"date-parts":[[2020,12,18]],"date-time":"2020-12-18T06:37:30Z","timestamp":1608273450000},"update-policy":"http:\/\/dx.doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":10,"title":["The Bayesian Additive Regression Trees Formula for Safe Machine Learning-Based Intraocular Lens Predictions"],"prefix":"10.3389","volume":"3","author":[{"given":"Gerald P.","family":"Clarke","sequence":"first","affiliation":[]},{"given":"Adam","family":"Kapelner","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2020,12,18]]},"reference":[{"key":"B1","doi-asserted-by":"publisher","first-page":"713","DOI":"10.1016\/s0886-3350(13)80339-2","article-title":"An improved universal theoretical formula for intraocular lens power prediction","volume":"19","author":"Barrett","year":"1993","journal-title":"J. Cataract Refract. Surg."},{"key":"B2","article-title":"Regression error characteristic curves","author":"Bij","year":"2003"},{"key":"B3","doi-asserted-by":"publisher","first-page":"1750","DOI":"10.1214\/14-aoas755","article-title":"Variable selection for BART: an application to gene regulation","volume":"8","author":"Bleich","year":"2014","journal-title":"Ann. Appl. Stat."},{"key":"B4","volume-title":"Empirical model-building and response surfaces","author":"Box","year":"2001"},{"key":"B5","doi-asserted-by":"publisher","first-page":"199","DOI":"10.1214\/ss\/1009213726","article-title":"Statistical modeling: the two cultures (with comments and a rejoinder by the author)","volume":"16","author":"Breiman","year":"2001","journal-title":"Stat. Sci."},{"key":"B6","doi-asserted-by":"publisher","first-page":"266","DOI":"10.1214\/09-aoas285","article-title":"BART: Bayesian additive regression trees","volume":"4","author":"Chipman","year":"2010","journal-title":"Ann. Appl. Stat."},{"key":"B7","doi-asserted-by":"publisher","first-page":"1585","DOI":"10.1016\/s0886-3350(97)80034-x","article-title":"Comparison of intraocular lens computations using a neural network versus the Holladay formula","volume":"23","author":"Clarke","year":"1997","journal-title":"J. Cataract Refract. Surg."},{"key":"B8","doi-asserted-by":"publisher","first-page":"1157","DOI":"10.1016\/j.jcrs.2016.06.029","article-title":"Comparison of 9 intraocular lens power calculation formulas","volume":"2","author":"Cooke","year":"2016","journal-title":"J. Cataract Refract. Surg."},{"key":"B9","volume-title":"General principles of software validation","year":"2020"},{"key":"B10","doi-asserted-by":"publisher","first-page":"362","DOI":"10.1016\/j.jcrs.2018.01.015","article-title":"Accuracy of 8 intraocular lens calculation formulas in relation to anterior chamber depth in patients with normal axial lengths","volume":"44","author":"G\u00f6kce","year":"2018","journal-title":"J. Cataract Refract. Surg."},{"key":"B11","doi-asserted-by":"crossref","DOI":"10.1201\/9780429258480","volume-title":"Bayesian data analysis","author":"Gelman","year":"2003"},{"key":"B12","doi-asserted-by":"publisher","first-page":"765","DOI":"10.1007\/s004170000188","article-title":"Comparison of immersion ultrasound biometry and partial coherence interferometry for intraocular lens calculation according to Haigis","volume":"238","author":"Haigis","year":"2000","journal-title":"Graefes. Arch. Clin. Exp. Ophthalmol."},{"key":"B13","doi-asserted-by":"crossref","DOI":"10.1007\/978-0-387-21606-5","volume-title":"The elements of statistical learning","author":"Hastie","year":"2001"},{"key":"B14","doi-asserted-by":"publisher","first-page":"700","DOI":"10.1016\/s0886-3350(13)80338-0","article-title":"The Hoffer Q formula: a comparison of theoretic and regression formulas","volume":"19","author":"Hoffer","year":"1993","journal-title":"J. Cataract Refract. Surg."},{"key":"B15","doi-asserted-by":"publisher","first-page":"17","DOI":"10.1016\/s0886-3350(88)80059-2","article-title":"A three-part system for refining intraocular lens power calculations","volume":"14","author":"Holladay","year":"1988","journal-title":"J. Cataract Refract. Surg."},{"key":"B16","doi-asserted-by":"publisher","first-page":"333","DOI":"10.1016\/j.jcrs.2016.12.021","article-title":"Accuracy of 3 new methods for intraocular lens power selection","volume":"43","author":"Kane","year":"2017","journal-title":"J. Cataract Refract. Surg."},{"key":"B17","doi-asserted-by":"publisher","first-page":"224","DOI":"10.1002\/cjs.11248","article-title":"Prediction with missing data via bayesian additive regression trees","volume":"43","author":"Kapelner","year":"2015","journal-title":"Can. J. Stat."},{"key":"B18","doi-asserted-by":"publisher","first-page":"1","DOI":"10.18637\/jss.v070.i04","article-title":"BartMachine: machine learning with bayesian additive regression trees","volume":"70","author":"Kapelner","year":"2016","journal-title":"J. Stat. Software"},{"key":"B19","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1111\/ceo.13435","article-title":"The eye in AI: artificial intelligence in ophthalmology","volume":"47","author":"Keel","year":"2018","journal-title":"Clin. Exp. Ophthalmol."},{"key":"B20","doi-asserted-by":"publisher","first-page":"368","DOI":"10.1016\/j.jcrs.2007.10.031","article-title":"Sources of error in intraocular lens power calculation","volume":"34","author":"Norrby","year":"2008","journal-title":"J. Cataract Refract. Surg."},{"key":"B21","doi-asserted-by":"publisher","first-page":"419","DOI":"10.1016\/j.jcrs.2005.12.139","article-title":"Prediction of the effective postoperative (intraocular lens) anterior chamber depth","volume":"32","author":"Olsen","year":"2006","journal-title":"J. Cataract Refract. Surg."},{"key":"B22","volume-title":"R: A language and environment for statistical computing [computer program]","year":"2018"},{"key":"B23","doi-asserted-by":"publisher","first-page":"785","DOI":"10.1016\/j.jcrs.2014.07.035","article-title":"Influence of corneal asphericity on the refractive outcome of intraocular lens implantation in cataract surgery","volume":"41","author":"Savini","year":"2015","journal-title":"J. Cataract Refract. Surg."},{"key":"B24","doi-asserted-by":"publisher","first-page":"2018","DOI":"10.1016\/j.jcrs.2011.05.042","article-title":"Optimizing intraocular lens power calculations in eyes with axial lengths above 25.0 mm","volume":"37","author":"Wang","year":"2011","journal-title":"J. Cataract Refract. Surg."},{"key":"B25","doi-asserted-by":"publisher","first-page":"394","DOI":"10.1016\/j.jcrs.2011.08.044","article-title":"Comparison of prediction error: labeled versus unlabeled intraocular lens manufacturing tolerance","volume":"38","author":"Zudans","year":"2012","journal-title":"J. Cataract Refract. Surg."}],"container-title":["Frontiers in Big Data"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2020.572134\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2020,12,18]],"date-time":"2020-12-18T06:37:36Z","timestamp":1608273456000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fdata.2020.572134\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,12,18]]},"references-count":25,"alternative-id":["10.3389\/fdata.2020.572134"],"URL":"https:\/\/doi.org\/10.3389\/fdata.2020.572134","relation":{},"ISSN":["2624-909X"],"issn-type":[{"value":"2624-909X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,12,18]]}}}