Quantitative Measurement of the Cornea by OCT

The cornea accounts for two thirds of the eye’s refractive power. The accurate measurement of the corneal shape and refractive power is essential for diagnostics and the planning of surgeries. Keratometry and corneal topography are clinically established measures for the quantitative description of the corneal shape and refractive power. One major application is the planning of cataract surgeries, where the natural lens gets replaced by an intraocular lens (IOL). Typically, different measurement modalities are combined to acquire all measures needed for accurate IOL selection.
Optical coherence tomography (OCT) potentially enables the three-dimensional measurement of all optically relevant structures of the eye at once – including the cornea. However, the use of OCT for corneal topography and keratometry is still limited. One limitation is the sensitivity of beam-scanning to eye motion. This is especially true for beam-scanning OCT, that relies on the sequential gathering of one-dimensional depth profiles. This sensitivity can be decreased by reducing the measurement duration, in particular by increasing the scan speed. Nevertheless, there is a trade-off between axial resolution, scan range, scan speed, signal-to-noise ratio (SNR) and the cost of the system; higher speed implies lower SNR and higher axial resolution implies shorter scan range. To take full advantage of OCT – measuring the full depth of the eye at once – one has to make compromises regarding the resolution and speed of the system.
In our work, we present solutions for OCT-based keratometry and topography, using a system with limited axial resolution and speed which is, in return, able to measure the full depth of the eye. The limited axial resolution asks for more extensive measurements to compensate for the increased uncertainties. The need for more extensive measurements combined with the limited speed makes the OCT measurements more sensitive to motion and asks for new scanning and motion compensation techniques.
In this PhD thesis, new methods for scanning, segmentation, motion compensation and reconstruction are presented. We propose new scanning techniques with two-dimensional scan trajectories, enabling robust reconstruction and accurate motion compensation with high temporal resolution. The motion compensation features model-based motion compensation in three dimensions. Because current segmentation methods do not apply to these new scanning techniques, we present a novel method for model-based segmentation. Further, we present methods for robust reconstruction, topography and accurate simulated keratometry (SimK) from OCT measurements.