Polarimetric analysis of the human cornea measured by polarization-sensitive optical coherence tomography

J Biomed Opt. 2010 Sep-Oct;15(5):056004. doi: 10.1117/1.3486540.

Abstract

Corneal polarimetry measurement has been the object of several papers. The results of techniques like polarization-sensitive optical coherence tomography (PS-OCT), scanning laser polarimetry, or polarization microscopy are contradictory. Some studies propose a biaxial-like birefringence pattern, while others postulate that birefringence grows at corneal periphery. Several theoretical approaches were proposed for the interpretation of these measurements, but they usually lack accuracy and an adequate consideration of the nonnormal incidence on the tissue. We analyze corneal polarization effects measured by PS-OCT. In vivo and in vitro PS-OCT images of the human cornea are acquired. PS-OCT measurements are apparently not in agreement with the biaxial-like birefringence pattern. We present a polarimetric model of the human cornea based on the extended Jones matrix formalism applied to multilayered systems. We also apply the Poincaré equivalence theorem to extract optic axis orientation and birefringence. The results show that for a fibrils orientation pattern composed by alternating circular and radial fibrils, the birefringence is biaxial-like at the corneal center, and there is an almost circularly symmetric high-birefringence area at corneal periphery. The model could be useful for diagnosis of corneal diseases or corneal compensation in retinal polarimetric imaging.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Birefringence
  • Cornea / anatomy & histology*
  • Cornea / physiology*
  • Corneal Diseases / diagnosis
  • Humans
  • Image Interpretation, Computer-Assisted
  • Microscopy, Polarization / methods
  • Models, Biological
  • Optical Phenomena
  • Refraction, Ocular
  • Scanning Laser Polarimetry
  • Tomography, Optical Coherence / methods*
  • Tomography, Optical Coherence / statistics & numerical data