Purchase this article with an account.
Jigesh Baxi, William Calhoun, Hiroshi Ishikawa, Daniel X Hammer, Ilko Ilev, Joshua Pfefer, Gadi Wollstein, Anant Agrawal; Evaluation of OCT Optic Nerve Head Measurements Using a Phantom Eye. Invest. Ophthalmol. Vis. Sci. 2014;55(13):4809.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Glaucoma detection and progression monitoring with optical coherence tomography (OCT) has the potential to improve patient care. Diagnostic decisions may involve quantitative measurements of nerve fiber layer (NFL) thickness and optic nerve head (ONH) cup-to disc ratio (CDR). Numerous studies have highlighted retinal thickness and CDR measurement repeatability but few used a controlled test object. We have developed a robust phantom that mimics the morphology and optical properties of the ONH as a tool for evaluating OCT system performance and implemented the phantom in a comparison of CDR measurement accuracy and precision.
The phantom was constructed with layer-by-layer spin coating followed by laser micro-etching to modify surface topography for modeling ONH morphology. The phantom was measured with near-micron precision using a stylus profilometer to construct a 3D topographic map. We assumed the cup-rim boundary to be 150 μm above the Bruch’s membrane opening and 180 radial cross-sections centered on the ONH were used to determine the true phantom CDR. The phantom was imaged with three different clinical instruments using each device’s standard ONH scanning protocol and compared with profilometry.
All devices accurately segmented the ONH region. Figure 1 shows an OCT B-scan of the ONH phantom. As shown in Figure 2, none of the devices accurately matched the absolute CDR measured with profilometry, however, the measurements were within one standard deviation for device 1 and 2. The mean CDR was measured to be 0.34±0.01 by profilometry. Device 1, 2, and 3 measured a mean CDR of 0.29±0.05, 0.36±0.05, and 0.39±0.03 respectively.
We developed and tested an ONH mimicking phantom for evaluation of OCT CDR measurement accuracy. The low profilometry variance indicates ONH phantom radial uniformity allowing for precise analysis of CDR measurement accuracy. All devices successfully segmented and measured the ONH indicating the phantom’s effectiveness for use with clinical devices. Each device predicted different CDR values and showed inconsistencies between measurements. These variations may result from different cup-rim margins defined by each device. This phantom could prove to be a useful tool in determining critical differences between devices and lead to a standardization of OCT ONH measurements.
This PDF is available to Subscribers Only