Purpose
Phase-variance optical coherence tomography (pvOCT) angiography is a non-invasive three dimensional imaging technique that allows for visualization of microvascular features in the retina and choroidal layers. Due to the special acquisition and processing needs of pvOCT angiography, the optimal parameters for a sampling density and a beam spot size remain unclear, requiring the development of quantitative approaches for judging OCT angiogram image quality.
Methods
Normal subjects were imaged with pvOCT using a custom-built, fiber-based Fourier-domain OCT system. Two collimator lenses with different focal lengths were used in the sample arm to create different beam sizes at the pupil and resulting spot sizes on the retina. A range of two-dimensional scanning patterns were used to explore the changes in image quality and detection of microvascular features from differing sampling densities for each collimator lens. The total acquisition times required to image a large field of view were also evaluated for each scanning pattern. For each spot size and scan type, OCT volumetric data were processed by the phase-variance method and the resulting angiogram was subjected to quantitative assessment of its signal to noise, contrast to noise, and feature detection with the assistance of Fourier domain analysis tools.
Results
Of the large range of parameters explored, the figure depicts the results from two different retinal spot sizes when scanning at an isotropic 4µm spacing, illustrating the different capabilities to capture microvascular features. Qualitative comparison of the 1.5mm x 1.5mm area angiograms (3 degree nasal retina with in 4.5s) acquired from large (a) and small (b) retinal spot sizes suggests the small spot size is superior. Quantitative assessment through Fourier analysis (c) of the red dashed areas supports this impression by the greater spectral densities, offering a quantitative measure that can be optimized.
Conclusions
Angiograms generated with pvOCT are critically dependent on the employed scanning parameters of a sampling density and a retinal spot size. Quantitative image analysis using Fourier domain and other analyses offer not only a clear path to optimal sampling and optics, but also a means to judge image quality and to assure angiogram quality from pvOCT.
Keywords: 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) •
549 image processing •
550 imaging/image analysis: clinical