Automatic Segmentation of the Central Epithelium Imaged With Three Optical Coherence Tomography Devices

Purpose: The aim of this study was to investigate the feasibility of automatic segmentation of the central corneal thickness (CCT) and epithelial thickness (ET) of the human cornea obtained with different spectral domain optical coherence tomography (OCT) instruments. Methods: Ten left eyes from 10 healthy subjects with a mean age of 22.5 +/- 1.5 years participated in this study. A custom-built ultra-high resolution OCT (UHR-OCT) with a 3-mu m axial resolution, ultralong scan depth OCT (UL-OCT) with a 7.5-mu m resolution, and commercial RTVue OCT with a 5-mu m resolution were used to image the cornea. An automated segmentation algorithm was developed to process the OCT images and yield the CCT and ET. The measurement was verified by a manual measurement. Results: The automatic algorithm successfully processed the central thickness of the cornea and corneal epithelium for all images. The average CCT obtained by the automatic segmentation algorithm was 528.1 +/- 22.4 mu m, 526.1 +/- 23.4 mu m, and 525.2 +/- 23.7 mu m for UHR-OCT, UL-OCT, and RTVue, respectively. The average ET was 53.2 +/- 2.0 mu m, 54.1 +/- 3.0 mu m, and 52.1 +/- 2.5 mu m for UHR-OCT, UL-OCT, and RTVue, respectively. These measurements were in agreement with those of the manual method for the CCT (all r>0.997, P<0.05) and for the ET (all r>0.71, P<0.05). Conclusions: The algorithm seemed to be feasible for automatically segmenting the CCT and ET in OCT images using these tested OCT devices. The segmented results were equivalent to that obtained with the manual method.