Biomechanically Informed Patient-Specific in Silico Models for Laser Refractive Surgery

Abstract Purpose: Corneal biomechanics plays a key role in the planning and outcomes of laser refractive surgery. This study presents a validated methodology for simulating patient-specific refractive treatments, focusing on the optomechanical effects of the three most commonly performed procedures: PRK, LASIK, and SMILE. Methods: For the first time, patient-specific mechanical properties of the cornea were incorporated into finite element simulations. These properties were estimated using an artificial neural network trained on in silico data from fluid–structure interaction models of non-contact tonometry. The tool takes as input corneal deformation images acquired with the Corvis ST device, intraocular pressure (IOP), and corneal geometry obtained from Pentacam imaging. IOP is estimated independently of corneal geometry and mechanical properties using a novel algorithm developed in prior studies. The methodology was tested on a cohort of 58 eyes from 29 patients who underwent one of the three procedures. Results: By integrating patient-specific geometry, IOP, and biomechanical characterization, the proposed framework successfully simulated postoperative corneal responses, yielding a mean dioptric error of +0. 40 0. 30\, D + 0. 40 ± 0. 30 D relative to clinical outcomes. Among the three procedures, SMILE produced the highest mechanical impact on the corneal model. Conclusion: This study introduces a personalized, biomechanically informed approach to simulate corneal behavior following refractive surgery. The proposed framework enhances surgical planning and improves prediction of postoperative refractive stability, offering a step toward personalized refractive correction and safer, more predictable clinical outcomes.