An Engineering-based Methodology to Characterize the In Vivo Nonlinear Biomechanical Properties of the Cornea with Application to Glaucoma Subjects

Abstract

 

Purpose

 

To characterize the in-vivo nonlinear biomechanical properties of normal and glaucoma corneas using a robust inverse finite element approach.

 

Methods

 

The corneas of 12 subjects (3 normal, 3 ocular hypertensive, 2 angle-closure, and 4 open-angle glaucoma; IOP=14.9±3.1 mmHg; Age=61±16 years; CCT=560±42 μm) were deformed (air jet) and their transverse cross-sections simultaneously imaged using the Corvis ST Tonometer (Oculus, Wetzlar, Germany). Since the Corvis cannot directly derive ‘true-engineering’ biomechanical properties, we propose a novel methodology that uses Corvis images. Briefly, the corneal geometry of each subject was digitally reconstructed and meshed using 8-node hexahedrons. A Veronda-Westmann constitutive model that incorporated stretch-induced stiffening of the collagen fibrils was introduced. Stiffness parameters were varied until model corneal displacements (affected by IOP and air jet loading) matched those derived experimentally. This was performed using an inverse finite element approach driven by a global optimization algorithm (differential evolution). Such a methodology was able to derive stress & strain (air jet induced), and a unique set of biomechanical properties for each subject’s cornea.

 

Results

 

In all cases, our models matched the Corvis data well (Figure). On average, corneas exhibited 9.05±2.29% maximum effective strain and 65.5±10.9 kPa maximum effective stress as a result of air jet loading. Corneas had an average initial stiffness of 0.08 MPa (95^th percentile: 0.11 MPa) that increased to 0.31 MPa (95^th percentile: 1.02 MPa) at 5% strain, indicating nonlinear stiffening behavior. Corneal bulk moduli were on average 5.49 MPa (95^th percentile: 17.6 MPa). No differences in mechanical properties between groups could be reported due to the small sample size.

 

Conclusions

 

Our novel methodology can estimate ‘true’ in-vivo corneal biomechanical properties that are likely more relevant than surrogate parameters provided either by the Corvis or the Ocular Response Analyzer. Our ultimate goal is to identify whether corneal biomechanics could serve as a biomarker for glaucoma.