Seeing blood vessels in the eye

When we gaze skyward at night, stars appear to twinkle because the light path passes through patches of warmer and colder air, and this bends the light rays. Astronomers use a technique called adaptive optics to detect and correct for these aberrations and still produce sharp images. And now scientists in the US have used the same trick to image non-invasively and in detail the blood vessels - and the movements of individual blood cells - through the retina. And some of what we thought we knew about blood flow in the eye turns out to need another look. From the University of Rochester, and speaking with Chris Smith, Aby Joseph…

Aby - What we have here is a new method to non-invasively visualise and measure single blood cells flowing in the smallest to largest vessels in the retina. If you look at the brain, one of the most metabolically active organs in the body, we've gotten a lot of insights into the functioning of the brain using functional MRI. However there are still some limitations to that; one of them being that the resolution such techniques can go to is about one millimetre, and it may miss changes happening at scales much lower than that, namely in the smallest vessels of the central nervous system. So what we've done here is we've borrowed a technique which was first pioneered in astronomy called adaptive optics. Our view of the stars above in the heavens gets distorted and blurred because of turbulence in the atmosphere, and mirrors which can change their shape on the ground adaptively correct for that turbulence and sharpen our view of the stars in the sky. And we've borrowed this technique to image the retina.

Chris - So how do you get from what's going on in the night sky to what's actually going on in the back of the eye?

Aby - Imperfections in the front of the eye cause the image to be blurred. So we use adaptive optics to get a high resolution of cells in a living eye. Additionally we used a very fast camera to now start visualising blood cells which obviously are moving, and combined this with the computational technique to now be able to measure speeds of blood cells from the smallest capillary to the largest vessels in the eye, giving us a complete picture of the interconnected network which becomes important in studying disease.

Chris - It sounds amazing. So talk me through then actually what the technique is. Because you're doing this initially on test animals aren't you? You're looking at mice to start with. But talk us through then what you actually do and how you obtain the images.

Aby - What we do is we use an infrared light source which is non-invasive and to which the eye is relatively insensitive. We scan a very fast beam across blood vessels in the eye and we're able to image back the light that is scattered by the red blood cells. And because we are imaging fast enough we're able to measure speeds of up to one metre per second in the eye, covering the entire possible range of speeds of blood cells from the smallest to the largest vessels.

Chris - So you have a light source sitting outside the eye which is beaming near-infrared through the pupil and the optics at the front of the eye. It's hitting the blood vessels which are actually behind the retina, bouncing back off those, and some of the reflected light comes back out the front of the eye where you can see it. And you're doing that by scanning in a series of lines a bit like a television would scan across the screen, in order to get a complete picture, many many times a second.

Aby - That's absolutely right, yes. 15,000 lines per second enables us to not miss anything that's happening.

Chris - How do you know that you're seeing the same bit of eye each time? Because obviously the eye might move a bit, your apparatus might vibrate a bit, and at the resolution you're looking - fractions of a millimetre - any tiny movement would blur things. So how do you match them up?

Aby - That's a great question. So it's partly done by the adaptive optics technique which adaptively corrects for changing imperfections at about 10 times a second. And additionally in post-processing we have what we call registration or correction algorithms which can correct for the eye drift or eye motion.

Chris - And so this gives you the architecture of the vessels, it gives you where the cells are going in those vessels, so that you can image right through from the smallest vessels to some of the largest in the retina. And that doesn't just mean the blood coming in, you can also therefore look at the veins that are taking the blood away as well presumably?

Aby - Exactly, yes. And that's what we think is one of the advances in this paper, where we can look at the complete vascular unit, which becomes important for example in diseases like diabetes where it's known that the disease may start at the smallest level of the capillary and have a cascading effect to larger and larger vessels. And this technique now enables us to study the complete system.

 Chris - Have you got any surprises emerging from this? Because physiologists have been studying how cells move along blood vessels for a really long time, and we've got various theories about what happens when blood flows through different sizes of blood vessels. When you actually now look at them rather than just go by what the theories say, were there any obvious differences?

Aby - That's a great question. And perhaps that's one of the unique selling points here. So what we find is, contrary to models which predict a certain shape to how the blood distribution is in a cross section of a vessel, we found that when we actually directly imaged them or visualised them that it's quite removed from what conventional models predict of how blood flow should look like at the smallest vessels. And this requires more study, but some of the first results indicate that we indeed need to directly measure some of these metrics, and not only rely on models because there can be a lot of surprises when it comes to these extremely tiny vessels.

Chris - Indeed but also big vessels. Because we're doing things like building stents to open up blocked arteries for example, and the way these are designed is based on our ideas about how blood flows through blood vessels, and how blood cells interact with the walls of blood vessels. And actually if those models are a bit wrong it argues that we may not be designing these instruments to go inside blood vessels as well as they could be designed.

Aby - Yes, certainly. And I couldn't agree more, and one of the advantages here is that although we're looking at the eye and the retina, because the retina is a part of the larger central nervous system which is a very special tissue with a very dedicated vascular network, what this really enables us to do is study what's happening in the nervous system  - as opposed to the skin which is more easily accessible - and see if there are differences in these models or assumptions, as you mentioned, of what's happening in blood vessels in the nervous system.