Enhanced IOL Calculation Based on Machine Learning Algorithms
Purpose
This retrospective cross-sectional study was conducted at the Centro de Estudos Renato Ambrosio (CEORA) from Brasilia Ophthalmology Hospital (HOB; Brasília, Federal District, Brazil) to develop an enhanced method for intraocular lens (IOL) calculation using artificial intelligence techniques based on ocular biometry and corneal and anterior segment tomography (CAST) data.
Methods
A total of 200 eyes from 200 patients implanted with an Acrysof IQ intraocular lens (IOL) were selected. Manifest refraction data from 1 month postoperative was considered along with the IOL implanted to calculate the ideal IOL for each case. Preoperative ocular biometry data from IOL Master (Zeiss-Meditec) and Lenstar (Haag-Streit AG) and CAST from the Oculus Pentacam were retrieved.
Results
Artificial intelligence (AI) methods based on machine learning algorithms (MLA) were developed to enhance IOL calculation by theBrAIn (Brazilian Study Group of Artificial Intelligence and Corneal Analysis). Result details will be updated by January 31. Different AI approaches provided significantly higher correlations with the ‘ideal IOL’ than currently available formulas.
Conclusion
Artificial intelligence methods provide the ability to significantly improve refractive outcomes of cataract surgery through enhanced IOL power calculations. Future validation studies are necessary.