Predictive modeling of cycloplegic refraction using non-cycloplegia ocular parameters with emphasis on lens-related features

Abstract Utilizing non-cycloplegia ocular parameters in the routine optometry examinations, this study focuses on developing a predictive model for ocular refraction after cycloplegia. The predictive model, incorporating lens-related features, outperformed a model without such features. Notably, the IOLcl group exhibited the highest predictive accuracy with r² of 0.964, MSE of 0.241, RMSE of 0.472, and MAE of 0.307. The physical model, validated through AS-OCT data, highlighted the anterior surface of the lens as a primary contributor to refraction changes after cycloplegia. Insights into optical mechanisms suggested a natural preference for anterior surface adjustments due to lower force requirements, enhancing the efficiency of the accommodation process. Overall, this predictive model has potential advantages to relieve outpatient congestion, avoid cycloplegia complications and widespread apply for vision screening of optometry.