Imaging blood flow in the eye
The eye is a window to the human body and is important in non-invasive health monitoring. Abnormalities of the retinal blood vessels can help predict a patient's risk for diabetes, hypertension, stroke, and heart disease. The state of blood vessels in the retina also reflects the condition of vasculature in other organs, such as the kidney.
Ophthalmologists use a number of techniques for eye exams, but optical coherence tomography (OCT) stands out as the only one that enables non-invasive, high-resolution, cross-sectional imaging. It provides detailed morphology of the eye fundus and the dimensions and locations of blood vessels. However, assessment of blood flow could give better insight into the status of blood circulation. It is possible to expand OCT to measure blood flow. We are developing spectral OCT/Fourier domain OCT (SOCT/FDOCT) whose speed and sensitivity make it especially suited to performing functional analysis. Spectral OCT looks promising for accurate measurements of blood flow.
The idea of measuring blood flow using SOCT has been around since the technique's early days. The most common way to measure flow velocity is phase-resolved Doppler OCT,1 which is based on the linear relation between blood flow velocity and the OCT signal phase. Our group at Nicolaus Copernicus University proposed an alternative way of extracting Doppler information by looking at time-varying Fourier domain signals. We called this novel approach joint spectral and time domain OCT (STDOCT).2
The idea is to repeat spectral OCT measurements and to retrieve the Doppler shift from signal amplitude modulation in time. In practice, we register a highly oversampled signal and consider it as a set of 2D spectral fringes dependent on wavelength and time. By calculating the amplitude of a two dimensional Fourier transformation applied to such a signal, we generate an image of velocity distribution as a function of depth. Such numerical analysis is computationally efficient since we only use fast Fourier transforms (FFTs).
With the technique, we can reconstruct the structure of ocular tissue together with velocity maps—or, if we like, with blood velocity profiles. Figure 1 shows velocity analysis in retinal blood vessels. We have also demonstrated that this method can be more sensitive than phase-resolved techniques and more reliable under a low signal-to-noise ratio.2
The technique can be applied to the same SOCT data used for standard phase-resolved Doppler OCT with A-scans averaging. However, we need a minimum of three spectra to estimate the velocity value. Additionally, it strongly improves sensitivity to detect flows under low signal-to-noise ratio conditions with an increasing number of spectra. This property can be fully exploited in today's era of ultra-high-speed imaging, where large amounts of data can be collected in a couple of seconds. By using complementary metal oxide semiconductor cameras, we can advance the registration speed, making 3D blood flow imaging using STDOCT possible (see Figure 2). With 3D data, we can perform straightforward automatic segmentation of vessels together with velocity estimation, creating two sets of data, one quantitative and the other qualitative. Time requirements of both measurement (3s) and post-processing (3min) renders STDOCT the fastest 3D OCT technique to image blood vessels and blood velocity in the retina and choroid.
Physicians need reliable tools to exploit unique features of the eye in non-invasive diagnosis. STDOCT is a highly sensitive and computationally efficient method of blood flow velocity measurement. The next step is to conduct clinical studies to introduce the technique to widespread practice.
This project is operated within the Foundation for Polish Science START and Ventures Programme co-financed by the EU European Regional Development Fund and supported by EuroHORCs-European Science Foundation EURYI Award EURYI-01/2008-PL.
Anna Szkulmowska is a PhD student in the medical physics group in the Faculty of Physics, Astronomy and Informatics. Her main interest is application of spectral OCT in ophthalmology. She is focused on 3D techniques of retinal blood flow analysis.
Maciej Szkulmowski received his PhD in physics from NCU. He is a research assistant professor in the medical physics group in the Faculty of Physics, Astronomy and Informatics. His main research interests are numerical techniques in morphological and functional optical coherence tomography.
Andrzej Kowalczyk is a dean of the Institute of Physics and a founder and head of the medical physics group in the Faculty of Physics, Astronomy and Informatics. His scientific interests are biomedical imaging and application of fluorescence as an analytical method in medicine.
Maciej Wojtkowski received his PhD in physics from NCU. He is a research assistant professor in the medical physics group in the Faculty of Physics, Astronomy and Informatics. His research interest includes optical coherence tomography and low coherence interferometry applied to biomedical imaging.
