Thanks for emailing that article!
"Optical Coherence Tomography — Past, Present, and Future."
Joel S. Schuman, MD, FACS, Chair of the Department of Ophthalmology at New York University's Langone Medical Center, Director, NYU Langone Eye Center, and Professor of Ophthalmology, Neuroscience and Physiology at NYU, delivered the Glaucoma 360 - New Horizons Forum keynote lecture in San Francisco on February 3, 2017.
The Drs. Henry and Frederick Sutro Memorial Lecture, the keynote talk at New Horizons, features key opinion leaders and innovators in the field of glaucoma.
Andrew Iwach, MD: This year’s featured lecturer is Dr. Joel Schuman, who many of you know well. His training is impeccable from Columbia University to Mount Sinai School of Medicine, Beth Israel Medical Center, to the Medical College of Virginia for Ophthalmology, and finally, to Harvard Medical School for his glaucoma research, and clinical fellowships.
He’s currently at NYU. He’s Professor and Chair of the Department of Ophthalmology, but, there’s more. He’s also Professor of Electrical and Computer Engineering. That solid background helps explain the successes he’s had in our field. He has received numerous awards for his accomplishments, in particular, with OCT technology. He has written 57 book chapters, over eight textbooks, over 500 peer-reviewed articles, two patents. He’s lectured worldwide, given over 16 named lectures.
Dr. Schuman will talk to us about OCT - Optical Coherence Tomography - Past, Present, and Future. If I could ask Dr. Schuman to come up, I’d like to present this plaque to you as a small token of our appreciation for your presentation today.
Joel Schuman, MD: Good morning. First of all, I want to express my gratitude for being asked to give today’s lecture. To give a lecture for the Glaucoma Research Foundation is a wonderful honor. To give the Drs. Henry and Frederick Sutro Memorial Lecture is a tremendous honor, and I am very grateful and humbled by this.
Let me give you my disclosures. I just want to mention a couple of things about the people that this lecture is named after, and this is just from some research that I did. I found out that Adolph Heinrich Joseph Sutro was the 24th Mayor of San Francisco; he died in 1898. He was a German-American Engineer - politician, and philanthropist, but engineer is the important thing here.
Actually, this is the type of picture that we usually see of philanthropists, but, the truth is, and many of you who are out there know this, this is another picture of Adolph Sutro. He was a miner, and that was how he generated his fortune. This is him working on the mines with his men.
His grandson was Frederick Sutro, and Frederick Sutro is the father of Henry Sutro. Interestingly, Frederick Sutro was an eye doctor. His son, Henry Sutro, is the person who gave the bequest that endowed this lecture. This is also a picture of Henry Sutro. Henry Sutro wasn’t just an oral surgeon, or a dentist, he was also an inventor. What you see there is a picture of Henry Sutro with a motor bike that he built, in order to deliver newspapers during the war. He was talking to a policeman who was interested in doing something similar for his own children.
Well, there are a number of people who have gone before me in giving the Sutro Lecture. The inaugural was by Bob Weinreb. Paul Lee gave the second Sutro Lecture. The third Sutro Lecturer is sitting right in front of me, in the audience here, and that’s Richard Lewis. Again, like to express my gratitude for being asked to give this talk.
Now, I’m going to talk about OCT, and I’ll talk about the invention of OCT, how it came to be where it is, and maybe a little about where it’s going in the future. OCT was really born in the laboratory of Jim Fujimoto, at Massachusetts Institute of Technology. The idea was to do a non-invasive tissue biopsy - to be able to get an image of tissue, just using light, that would be sufficient to make an appropriate diagnosis, or to determine whether or not disease was progressing.
This was done as a consequence of Jim’s work in high-speed lasers at MIT. That work that led up to OCT was predicated on work, earlier, at Bell Labs. At Bell Labs, they were looking at time-gating for ranging. The concept is that you can look at reflections from a certain point in space, by looking just at the reflections that are coming back at a specific time from something that you’re aiming at.
Here we have the AT&T logo on your left, and your right, it’s covered by some tape. By time-gated ranging, you could just look at the scattered light from that particular plane in space, and you can see the logo, even though it was masked by the tape. The idea was that you could look through objects, like skin.
I was working at the laser lab at Mass Eye & Ear. This is back in 1989. I’m the guy in the middle with the hair, and similar glasses, actually. I was working on a totally unrelated project, but I heard about this thing going on with optical ranging. They were measuring the cornea, and the thickness of the cornea, because photo-refractive laser surgery was being developed. The idea was you could measure the corneal thickness.
As it turned out, that project was failing. As many of you know, who are inventors, and entrepreneurs, in the audience, most projects fail. This project was failing, because it took too long to measure the thickness of the cornea, and, by the way, you didn’t need to do it, because you know how much tissue you took off with every excimer-laser pulse.
I realized that the wavelength of the light that was being used could reach the retina. I went in to talk to the head of the laser lab, who was Carmen Puliafito. Some of you may know Carmen Puliafito in the audience.
Carmen is a rather larger-than-life, colorful character, and I was a lowly fellow, going into his office, telling him about this idea, which I have to admit, he didn’t think was the greatest idea in the world; he told me that, but not in those words.
I asked if it was okay if I went over to MIT to see if we could get some signal from the retina. He said, “Sure, do whatever you want …” Again, not in those words.
I went over to MIT with a bag of calf eyes, and David Wong, who was an HST student, and MD PhD student, at the time, working with Jim Fujimoto … David Wong is now a cornea-refractive surgeon, who has led some of the largest glaucoma-imaging trials in the country, and really has continued to push this technology forward.
David, and I, and a few other people were in the lab. I took some calf eyes, and I cut them in half, and we put them under the OCT beam, and we went out to get some sodas, because it took a long time for a single OCT A-scan. We came back, and there was a signal, so we knew that we would be able to measure the retina with this technology. After that, it was a lot of hard work to get to the point that we were able to image in living beings, but that was the critical proof-of-principle moment.
David realized that you could take this A-scan, and you could scan transversely to make a B-scan. That seems very obvious to everyone in the audience, right now, but, at the time, no one had done it, so, it wasn’t obvious until it was actually done. That made us go from having just having an A-scan, with peaks and valleys, to actually having a tomogram image.
David was the first author on the publication, in Science, describing this procedure. This was published in 1991. The image that you’re looking at is of a human eye, and you’re looking at the retina in this human cadaver eye that’s been hemisected. That explains the subretinal fluid that you can see here. It took us a little while to figure out that there was subretinal fluid to explain this dark space, because this image didn’t come with the labels. Then, we were able to identify that section on microscopy, as well.
The idea of OCT is that you’re using low-coherence light - broad spectrum - and you are using an interferometer, in order to measure something in space. It’s a very simple system, and it’s the simplicity of it, I think, that gives it so much power. We’re able to get very high resolution because of the low-coherence length.
This was the first breadboard for OCT, and this was built by Eric Swanson, the man in the middle. Michael Hee, who is the person on the left, is a retina surgeon here in Northern California. He wrote all of the original software for the imaging, and for the image analysis, and segmentation.
Then, this was the prototype OCT unit that we used, at the New England Eye Center, to do the original human studies, in a variety of different areas. The person who licensed the technology was actually John Moore, who was president of Humphrey/Zeiss; that was the name of the company at the time.
John had the foresight to see the potential of this technology. Jay Wei was the program manager at Zeiss. Jay has since gone on to found OptoVue.
The first commercial system from Zeiss did not do very well, and neither did the second one. The third one is the one that really caught on. The third one could scan faster; it had a normative database. Also, there was a billing code that allowed ophthalmologists to get paid for the work that they were doing, and, I think, resulted in- or was one of the things that resulted in the adoption of the technology, because no matter how good the technology is, if it costs you money, and you’re not getting paid, it’s very hard to justify it, from a business sense. This was the most successful, or most rapidly adopted ophthalmic technology in history, and it drives clinical-decision-making in a variety of different areas of ophthalmology, and is continuing to evolve.
This gives you a sense of some ot the economics of OCT. It is an investment that your tax dollars were used for. The federal government, the NIH, made an investment in basic science that resulted in a technology that benefits individuals; it benefits the health of the American people, which is the mission of the NIH; and it also benefits the economy. I think that this a very good example of how the NIH can make a difference in the everyday lives of people, and make lives better.
What are we doing today? Today, most of us are using spectral-domain OCT. This is slightly different than time-domain, only in that you can scan much faster than you can with time-domain. OCT. You’re collecting all the information from a given A-scan simultaneously, instead of pixel by pixel. This is done by looking at the frequency of the light that’s returning, using then, using math to decode it.
Spectral-domain OCT allows us to scan quite quickly, and create three-dimensional images, and by using these volume scans, we can learn more about the tissue, we can get much more reproducible results, and we can really use this technology to guide therapy, including surgery.
In terms of glaucoma, we were interested in the structure/function relationship, and we looked cross-sectionally at the population to find out, at what point would there be a relationship between visual fields, and the structure of the retina? What we found was that there was a tipping point, kind of a cut off, at around 75 microns - nerve-fiber-layer thickness; mean nerve-fiber-layer thickness. Above that nerve-fiber-layer thickness, there was not much of a relationship between the two. Below that, the relationship was quite strong.
This was predicted by the work of Harry Quigley, and Al Summer, and others, who showed that you needed to lose a certain amount of nerve tissue, before you would be able to detect a visual-field defect. Their prediction was somewhat higher than what we found, but their data were based on using older types of visual fields.
We found that you needed to lose about 17 percent, before it was likely there would be a visual-field abnormality.
Let’s schematize that. This curve represents what I just showed you, and here, you see that when the nerve-fiber layer is thick, that you’re likely to measure abnormalities primarily in the structure, in OCT-measured nerve-fiber-layer thickness. Then, you get to a point where you start to be able to measure change, both in the nerve-fiber layer, and in the visual fields.
Then, finally, you get to a point where the nerve-fiber-layer-thickness measurements bottom out. There’s a floor effect, and below that, you’re not going to measure change in the nerve-fiber layer, but change may still be occurring in your patient, and you can be falsely assured that the patient is stable, when, in fact, the patient is progressing, if you only look at the nerve-fiber layer, and not the visual field, in late disease. Again, here is the tipping point.
That theory, or hypothesis, was based on cross-sectional data. We have longitudinal data, now, that show that this, in fact, is true in patients across the board. You’re looking at all of the trajectories of all of the patients in this study. What you see is a curve that looks very similar to the curve that I showed you for the cross-sectional data set of the population, but, this is longitudinal, showing you the actual trajectories of the patients.
Let me illustrate that for you in a couple of slides. Here’s a slide where you’re looking at the top, at the visual-field index, and, at the bottom, at the nerve-fiber-layer thickness. The visual-field-pattern deviation is showing above, on the right, and the OCT is showing on the bottom, on the right. What you see is it’s not until you reach the tipping point that you start to have visual-field abnormalities.
Here’s a patient where the visual field is already abnormal. You’re below the tipping point, and you’re seeing a decay in both the visual field, and in the OCT nerve-fiber-layer thickness. That’s in the red curves. In the blue curves, what you see is that the visual field continues to get worse, but, the nerve-fiber-layer thickness is flat.
That neve-fiber-layer thickness is flat, because you have already hit the floor of what the machine can measure. We need to be very careful, as clinicians, that we don’t assume that our patient’s stable when, in fact, they may be progressing, when they have thin nerve-fiber layer. That number is at about 50, or 55 microns - mean-fiber-layer thickness.
Then, here’s another patient, where the nerve-fiber layer is thinning, the visual field is getting worse, and then, you hit the floor, and the nerve-fiber-layer thickness is stable, but field continues to deteriorate.
There was a study that was put together by David Wong, called the Advanced Imaging Glaucoma Study. In that study, we looked at a number of factors associated with glaucoma detection, and also, measuring glaucoma progression.
What we found was that the OCT abnormality that predicts visual-field conversion in people who are suspects, or would have preperimetric glaucoma is actually in the macula, and not simply the retinal-nerve-fiber layer.
This is called the Ganglion Cell Complex, or the GCC. Specifically, it’s the focal loss-of-volume parameter. This is the GCC-FLV. That was the best predictor of glaucoma conversion in these patients with preperimetric, or glaucoma suspect. You had almost a two-year lag between the time that the disease conversion could be detected by OCT, so, progression by OCT, until there was a an abnormality on the visual field.
The GCC-FLV was also the best single predictor of progression in parametric glaucoma, and that’s what you see here. Then, the progression detection, using the GPA, and the trend analysis of the VFI - the nerve-fiber layer - and the GCC, you can see that the earliest detection of progression that was statistically significant occurred in the GCC.
The OCT detects progression earlier, among the glaucoma suspect, and preperimetric-glaucoma eyes, about two years earlier than you would detect it on visual field, or by the appearance of a visual-field abnormality. I just wanted to point that out. That doesn’t mean that the nerve-fiber layer is not a useful parameter - it’s an extremely useful parameter - but it may also be that the macula will provide us with additional information that will be quite valuable.
The last thing that I want to talk about is OCT angiography. OCT angiography is what is coming. It actually is here, in terms of people being able to do OCT angiography clinically. In terms of the actual quantification of the results, and what those results mean, I think we’re just at the tip of the iceberg, in terms of understanding OCT angiography. I think it is going to be quite valuable.
This is work, and I have to thank David Wong for many of these slides … This is new work in parametric glaucoma, and in controls. What you see is that the OCT angiography corresponds quite well to the area of abnormality in the Macular Ganglion Cell Complex, and in the visual field.
If you look at the statistics, what you find is that it’s the superficial vascular plexus that is significantly different, or is significantly lower in terms of vessel density than in the intermediate plexus, or the deep plexus, when you compare glaucoma to healthy eyes. Lower vascular density in people with glaucoma, in the superficial vascular plexus, which is just what you would expect, because of where damage occurs in glaucoma.
This gives you a sense of what the sensitivity, and specificity are for these detections, in terms of these areas under the AROC curve.
Then, if you want to look, instead, at the peripapillary retinal vascular density, you can do that, and you find very similar results, and, again, high correspondence between the abnormalities in the vascular density, and the visual-field abnormalities, as well as the retinal nerve-fiber-layer thickness.
Here, you have the radial peripapillary capillary plexus that is statistically different, as well as the superficial vasculature. These are the areas under the AROC curve for that. In terms of repeatability, you can see that the repeatability is quite good, at about two percent, or, actually better than two-percent coefficient variation.
To conclude, we have now advanced ocular-imaging technology that allows us to visualize, and quantify structures that we were not able to quantify before. We can use this for early diagnosis of disease, and also for early detection of change in disease. I’ve spoken specifically about glaucoma, but the same could be said for other diseases of the retina, and of the eye, in general.
What’s coming down the road? I think that we’ll see higher speeds. We’ll see better software. We’ll be able to do better, in terms of the biometrics, the parameters, that we’re evaluating, when we’re looking for disease, or progression. Hopefully, we’ll see what the market usually does, which is to push prices down when we have competition. I look forward to that.
I want to thank our collaborators, and, especially, Jim Fujimoto, with whom I’ve had a wonderful, and longstanding collaboration. The group at NYU, this is our Ophthalmic Imaging Research Laboratory Group, and I want to thank you very much.
Last reviewed on May 17, 2019