Precise Thickness Measurements of Bowman's Layer, Epithelium, and Tear Film

Abstract

To visualize corneal microstructure such as tear film, epithelium, and Bowman's layer in three dimensions with spectral domain optical coherence tomography (SDOCT) exhibiting 1.3 μm axial resolution at 100,000 A-scans/s. This enables measurement of epithelial and Bowman layer thickness across an area of 8.4 mm × 8.4 mm and measuring the tear film thickness at the central cornea. We designed a high-performance SDOCT system, which uses a broad bandwidth TiSapph Laser and a high-speed complementary metal-oxide-semiconductor detector technology, providing a resolution in tissue of 1.3 μm and an acquisition speed of 100,000 A-scans/s. Such speed and resolution is a prerequisite if precise anatomy is to be determined. The high resolution gives access to corneal microstructure such as the epithelium layer as well as the boundaries of Bowman's layer and stroma. Even more interestingly, the tear film can be distinguished on the surface of the cornea. The Bowman's layer and epithelial thickness for both eyes of nine subjects have been measured out of which two subjects underwent photorefractive keratectomy treatment. Three-dimensional volumes of the human cornea have been recorded in vivo at an A-scan rate of 100,000 scans/s. Epithelial thickness was measured to be 55.8 ± 3.3 μm and Bowman's layer thickness 18.7 ± 2.5 μm in normal eyes. Epithelial thickness in the eyes after refractive surgery was measured to be 68.2 ± 5.0 μm. The Bowman layer was degenerated in these eyes. The average tear film thickness of four eyes was 5.1 ± 0.5 μm. Using a high-performance SDOCT system with high-imaging speed and ultrahigh resolution, we produced precise thickness maps of the epithelium and for the first time of the Bowman's layer. Such a system will give insight into high-fidelity three-dimensional corneal microstructure helping to precisely plan refractive surgery. It may furthermore yield new perspectives on studying and understanding tear film dynamics.