Dr. Dan Reinstein: What I'm going to show you right now is an area that has been quite under the radar screen for the last 20 years, mainly because there was only literally one device, I mean singularly, one machine that was able to measure epithelium in three dimensions with one micron precision. And it was an ultrasound system that we built at Cornell 10--20 years ago and developed into a scanner
So, a whole science of how--or what's the point of measuring epithelium in the cornea arose the day after Cal Roberts [sp] came in when I first measured it when I was a pre-resident doing research. I said to him, "Would it be of any use that we can measure epithelium?" And he said, "No." And I thought, "That's it. That's what I have to do for the rest of my career. That's what I have to do if he thinks it's of no use."
So, I’m going to run through, first of all, the fact that there is now a commercially available machine that can measure epithelial thickness in three dimensions. And that's the RTVue, as you know.
My financial interests, of course, are because I have some interests with the Artemis technology. I am a consultant for Carl Zeiss Meditec for their refractive lasers. And I receive a speaking honorarium for this evening, which I would have done even without their honorarium because I'm very honored to be asked to speak.
As you've seen, the--using an average--averaging technique, you can actually get a fairly good signal to noise ratio in order to detect what are really transparent boundaries within the cornea. But, there is just enough optical disturbance to be able to distinguish the interface between the basal epithelial cell and Bowman's and the back of Bowman's and the rest of the stroma, which looks all speckly, and thus the correlation
Of course, what you have to remember, though, is that whenever we're going to make a measurement of the epithelium with OCT, we are also including the tear film, because that interface between those two is a little bit too small in terms of a signal to noise ratio to be able to separate it out from the measurement. So, when we--so, I will be saying epithelial measurements, but I mean epithelium plus the tear film.
Of course, when you're measuring with ultrasound, the tear film is not included because the eye is immersed in saline. And so, you're measuring directly to the tissue level. So, that's a slight difference.
With the machine delivered even a few years ago, I was using this in my clinic and we were measuring epithelium with a manual caliper. And I’m sure many of you have tried this. You can measure at any location that you choose and you get a epithelial thickness and a stromal thickness underneath that.
Each click, each pixel you move up would change the measurement by four microns. So, basically you've got a four micron resolution. Of course, the resolution of light is in angstroms. So, somewhere in between there there's a lot more gain to be had
The new RTVue system that is able to measure epithelial thickness uses a scanning algorithm that works like this. Basically it scans eight meridians very quickly, and then it goes and does it again five times. And what that enables it to do is basically to take five scans from each plane and use the averaging techniques to give you these beautiful sharp images, and then from there an automatic algorithm goes through.
And by detecting the A-scans along the way, it looks for the first peak and the second peak. And it knows then that that is the epithelial thickness measurement, and it builds up maps. So, you can click on your corneal thickness map up here, or you can click on the second button and that now becomes the epithelial thickness map.
Now, the color scaling isn't great right now because you can't--this all looks like one color. But, actually, if you can see there, that's 57 microns. That's 50. That's 54. And this gradient of epithelial thickness is something which we published using ultrasound many years ago showing that the epithelial thickness profile of the normal cornea is actually thicker inferiorly and slightly nasal than superiorly. So, that matches.
If we look at the repeatability studies that have been done so far internally by the company, this is a six eye study, two normals, two post LASIK, and two keratoconic eyes. They measured several locations in three scans of the same cornea. And if you look at the results, these are all the results, they had two observers doing it manually and then the algorithm.
So, if you just condense this down, on average basically the algorithm is able to get about a 1.5 micron repeatability from one scan to the next, and about a 1.8 in the vertical. So, let's say about 1.5 microns repeatability, which is very decent
And if you compare this to previous publications using optical coherence tomography to measure epithelial thickness, this actually compares very favorably with all the other systems that have been done. Interestingly, our manual study that we did in-house got very similar results to the company's internal study measuring epithelial manually too.
But, anyway, we're down to about a 1.3 micron standard deviation for these measurements. Why is that important? Well, it's important because the epithelial thickness varies micronically from eye to eye, and very small changes in epithelial thickness profile can cause refractive changes.
As we know, in a six millimeter zone, if you take six microns of tissue from the center in an arc, you have actually corrected the cornea by half a diopter. And so, these are very, very sensitive changes, and you need to be able to measure these things to the micron if you're going to be--want to use the epithelial thickness measurements for refractive purposes.
Here is another normal example here. There's another example. Remember I've put these circles here because these maps pertain to the central six millimeter zone.
Here is three maps of the right eye and three maps of the left eye of a patient scanned on Thursday right when we got this thing to work in our clinic, the mapping thing, because we just had this installed. And here are the Pentacam maps for these same eyes. You see it was quite an--a high astigmatic eye. And I did an Artemis at the same time.
So, we were able to do a sample of one study where we measured the epithelial thickness profile with ultrasound and with the OCT. And if I just block off here so that we are looking at the same areas, this is the central six millimeter area, you can see that this pattern of it being thicker inferiorly and nasally is held and demonstrated by the OCT system, and the values at those locations are quite similar.
The average thickness of the epithelium in the human cornea, which we measured without tear film, was--is--we've done this twice. We did this in 2004 and again in 2008. It was--we came up with the same number. It's 53.5 microns, basically, with a standard deviation of about four to five microns.
So, if you look at the tear film inclusion studies of all the optical coherence studies before, they tend to get, of course, thicker measurements. And if you look at contact methods such as confocal microscopy, they get thinner measurements because they're probably compressing the epithelium a little bit or compressing the mucin layer maybe a little bit.
So, now let's look at some examples of abnormal corneas. Well, here is a cornea that has had LASIK for a minus two. Now, the average epithelium is 53, but in this case it's gone up to 60 in the center. All around here, this central zone, the epithelium has thickened. Why? Because of myopic ablation.
Here's a minus eight ablation showing epithelial thickening to 70 microns, right? And here--David Huang [sp] just said he's going to present some stuff on Monday--on Tuesday morning about keratoconus screening. But, here we've got the corneal thickness and the epithelial thickness map of a keratoconic case. And you see where the cone is sticking out, the epithelium becomes thinner. And here is another example where that happens.
Now, that is something which brings me to then the second part of my talk, which is really what I was asked to talk about, which is why--what's the point of measuring epithelium, and how can we use epithelial thickness mapping as a tool for corneal refractive surgery? And I can tell you that this is the dawn of a new era, because being able to measure epithelial thickness profiles I put at number two after topography in terms of diagnostics in corneal refractory surgery. And you'll see now in a minute why.
Our very first maps that we did back in the early '90s is an ophthalmology publication where we were measuring actually only the central three millimeter zone. Later in '96 when I went back to Cornell after residency, we expanded this to 10 millimeters, published that in 2000
But, here is an example of how--an epithelial thickness profile before and after a minus five ablation, and this is a difference map. So, you see here the epithelium has changed by 20 microns in the center, but centrifugally less as we go out towards the periphery. So, the epithelium here is actually like a lens, so it has power.
So, there's a power shift that is due to the epithelium that was not expected by the ablation. In fact, the only reason why refractive surgery works is because we overcome the epithelial compensatory function. But, that's what nomograms are for. That's why you have to keep on adjusting nomograms. That's why theoretical calculations for ablation profiles don't work, because the epithelium comes and tries to reverse some of what we're doing.
The diagnosis of keratoconus utilizing epithelial data is something which occurred to me very early on because I was noticing that the epithelium was thin on keratoconics. And so, after publishing the normal cornea and the keratoconic cornea, we wrote a paper describing a method of using epithelial thickness profiles in addition to, of course, topography so that we could get down to the problem, the problem being, and this has been going on for the last 10 years, that when refractive surgeons see a cornea that's slightly abnormal, they want to do LASIK--they want to do refractive surgery on it.
Instead of calling it, "I think it's keratoconus," they say, "It's forme fruste keratoconus." As soon as they say that, it gives them permission to give a consent form to the patient that says, "Look, I think you don't have keratoconus, but--this might be slightly higher risk, but if I do PRK you're less likely to get ectasia. So, would you sign the form so I can do your PRK?"
And I find that to be a rather gray zone of medicine, frankly, because we're operating without really knowing what is happening. I would much rather have a diagnostic technique that will tell us if it's keratoconus, in which case you do no surgery, or if it's not keratoconus, and then you do LASIK. You don't pussyfoot around with PRK.
Here's from the keratoconus publication. You can see that as the keratoconus cases become more and more severe so the size of the area of thinning increases. The difference between the thickest and the thinnest point of the epithelial profile increases, something which is not necessarily obvious from just looking at the topographies. So, now you have subsurface information. We're now doing--effectively, the epithelial map is like a proxy of the stromal surface topography
Here is the map of normals. This is 110 eyes that were--right and left eyes were mirrored so that there's nasal and temporal. And you can see the relationship as the eyelid would be blinking. You could see the eyelid blinking here and the angle of the eyelid hitting, chafing the superior epithelium, keeping thicker epithelium here which is protected by the lower lid.
And this is the average keratoconic, and you can see that there are marked differences in the pattern. So, we thought, "Can we exploit that?" Well, the answer is what we--where we really need to exploit it is where the topography doesn't give us the answer. And if you look here at this model, here we've got a stromal layer and an epithelial layer.
There is going to be a point on the early stages of deformation where the stroma is going to deform, but the epithelium is going to absorb that stromal deformity and there is going to be no change on the topography of the cornea. There is going to be a point when that is happening. It is not until the epithelium has ceased to be able to compensate fully for the stromal deformation that you will then see it as a bulge on topography and say, "It's forme fruste keratoconus."
So, armed with this knowledge, we can see that even in cases where the machine says it's completely normal and the topography looks completely normal, even if I showed these to Steve Sklite--Steve Klyce [sp], he would say that they're normal. In fact, it turns out that all of these were keratoconic cases that were very early and they all have patterns that are characteristic of a subsurface cone, which coincides with a back surface--eccentric back surface, etc., etc.
Likewise--so, that's--so, those are six cases of ectasia that I avoided. And likewise, if you've got a case where your heart sinks because the machine writes this nasty red subclinical keratoconus but you think the topography looks okay, by proving that it wasn't keratoconus because there is no doughnut pattern and there is a totally normal epithelial thickness pattern, you can actually go ahead and do LASIK on these cases.
And does that matter? I mean, does it--is it an important problem? Well, in our clinic, we looked at 1,500 consecutive cases that came through our clinic. We scanned with topography and tomography looking at the back surface, and we found that 9.2 percent of these cases were cases that we didn't want to--we weren't comfortable with, we would not want to do LASIK on.
Those eyes were then scanned and we--for epithelial thickness mapping. Out of those cases, 84 percent of them were found to be not keratoconus and therefore had surgery. And we've published the one year results and presented the two year results of LASIK with a Hansatome in these cases, therefore showing that we were probably right.
The interesting thing is that by using this method, our rejection rate, which would have been 9.2 percent from odd topography, went down therefore to 1.4 percent, which means that you're increasing your surgical volume by 7 to 8 percent just because you can operate on people who you otherwise would have felt restricted on because of the appearance of the topography or because of the digital diagnosis given by the topographer.
So, that's a--that's keratoconus screening. Here's another application, and that is in hyperopic LASIK. We've published a paper where we studied this quite carefully, looking at how the epithelium thickens where the tissue was removed in the periphery and forms a doughnut right over the diameter of the treatment zone. So, here we treated in a seven millimeter zone, and here we have a seven millimeter zone thickening doughnut here
But, of course we also get thinning of the epithelium where the cornea was steepened. And as you know, most people talk about not being able to steepen the cornea above 49 because, if it's too steep, you can sort of fall over. The epithelium does the same thing and there--and one tries to calculate what the post-op Ks are going to be to decide whether you can do a hyperopic LASIK.
Well, it turns out that a much more relevant method of determining how much hyperopic treatment can be done is to actually do it by measuring the epithelial thickness, much more relevant because, as it turn out, yes, the thickest epithelium thickens the more you attempt to correct. And the thinnest point of the epithelium thins the more you correct
But, after the surgery, the correlation between the post-op K and the thinnest epithelium is not logical. For example, here is a patient who had a keratometry of less than 42 post-op, and yet his epithelium was already 26 microns, which, you know, I can tell you is too low to do an enhancement on. But, the average surgeon would have done an enhancement and rendered this patient with an apical syndrome because his epithelium was already maxed out. On the other hand, at a Ks of 50 here, there is someone with an epithelium of 44, ample room to do a plus two enhancement.
So, epithelial thickness is actually the point, not the keratometry. And this leads us to understanding complicated corneas and how--it is an understanding of how the epithelium compensates for stromal surface irregularities that's going to get us out of the reason why refractive surgery is essentially a failed business, because there's no business where you would spend billions of R&D research monies from Wall Street and then have less than 1 percent market penetrants for your product. That's LASIK. That's LASIK right now.
And one of the reasons is people don't want to get it because they're worried what would happen if something goes wrong. And the thing is that if you know what's going on anatomically then there's no problem, because if you know the anatomy of the cornea, you're going to know exactly how to fix it. And so far, the problem is we've only ever had the anatomy of the surface of the cornea.
And any--these three yellow lines are the same shape but, look at the surface below, can be completely different. And it was--interestingly, Vogt, in his old textbook of slit lamp exam, said corneal stromal defects are filled with surface epithelial cells. I said--in the '90s I wrote in a paper, "Irregular astigmatism, irregular epithelium." If an eye presents with irregular topography, by definition the epithelium has reached its maximum compensatory function. Therefore, topography and wavefront measurements are inaccurate means of describing the irregularity of the stromal surface.
So, wavefront, the great promise of wavefront-guided surgery, turns out is not the solution for complex corneas at all. And here is a number of papers that you can look up in the literature showing how the epithelium can change. This is the normal, as you see. Here is a keratoconic. Here is a myopic LASIK. Here is RK. Here is hyperopic, as we saw earlier, orthokeratology, ectasia.
Here is an example of a nasal--a short nasal flap with an ablation that was performed. There is excess flattening just inside the hinge. There is steepening here, but you can see how irregular the epithelial thickness profile is. The epithelium is filling the trough formed by the double ablation here. There is no way that this topography represents what needs to be done to the stromal surface.
Here is an RK. That's what's underneath. That's the stromal surface. That's what the topography looks like. You know, no way a wavefront-guided treatment is going to be of any help there.
Here is multiple procedures, ALK followed by PRK followed by LASIK, giving you multiple optical zones. That is not evident on topography at all, but you've got the anatomy delineated here.
And here is a tracker that was locked inferiorly, and then a hyperopic ablation was done. So, here you have an epithelial hyperopic compensation, but off center in a patient that we had to repair.
I'll give you an example of how this works and how powerful the epithelium is in understanding corneal refractive surgery. This patient had RK and trapezoid cuts, rather like a Zorro job here, back in 1990--'82. He did nothing with the other eye, and he presented with plus 650, minus eight, best corrected 21--20/50. This was the epithelial profile over that and this was his topography, which really doesn't tell you at all why he was 20/50. You can see that
Now, if we think of the stroma here and the epithelial compensation there, the epithelium was incapable of completing the compensation, therefore it is still tilted. And so, the surface topography is represented by that surface, but it's marred by the epithelial thickness changes. If we were to plot the stromal surface topography, now we have a true sense of what is going on optically in his cornea.
The superior to inferior difference is six diopters on the surface but nine diopters subsurface. So, these big differences are why topography or wavefront-guided treatments don't work well in highly aberrated corneas. It's because we're measuring the wrong thing.
In this case, the topography-guided treatment would want to straighten out the obliqueness of the surface. But, that amount of tissue removed from the stroma results in this stromal surface, which may only partially be compensated by epithelium with remaining irregularity and an unfixed patient.
What we did was to do a transepithelial PTK guided by these epithelial maps. And as you see here, the white areas is where the epithelium has broken through and has already started to take stroma. We calculated the stromal ablation based on our digital subtraction. And--this is pre-op, and we realized that it looked like a hyperopic cylinder ablation. We calculated the equivalent refraction that we would have expected to see had we done this, and it said it was going to be a plus three diopter cyl.
We PTK'ed down to 45 microns, stopped, took a picture just to see if it looked like what we were doing was what we were doing, and it did, carried on. And here we are nine months after this PTK, pre-op, post-op, and the difference map showing that we had indeed corrected three diopters--3.5 diopters of cyl just as predicted.
That was in the epithelium. The epithelium had cylinder, therefore his refraction was not represented by his stromal surface. And as you see, his epithelium, which was very irregular to start with, was now much more regularized. It changed by this amount.
And that means that, by understanding these relationships, we were then able to--because the epithelium was now regular, then we were able to do a topography-guided treatment and render the patient more regular, and then a third topography-guided treatment, which rendered this patient 20/20 plano with no problems at all, seeing great. And that was the initial refraction with that that's corrected. Now, that couldn't have been done if we didn't know the epithelial thickness profile or the epithelial power.
Now, why is epithelial power important to the average ophthalmologist, not just the sort of nerdy LASIK fixer-upper guy? Well, it turns out that, in my belief, cataract surgery calculations post LASIK are going to become exceedingly accurate once we measure epithelial power.
It's obvious stuff. You're measuring through epithelium and stroma. And of course, you know, for years, as David pointed out, we've only ever used some approximation for the curvature of the front surface and some refractive index that doesn't exist to--called the K, and now we're using back surface.
And in fact, you may not know this, but in 1999 I published a paper on total optical power calculations of the Orbscan, and found that the four millimeter zone correlated with the refractive change better than anything else. And we published it. It's in the JCRS. A few years later, Jack found the same thing and published it for Pentacam
But, you see, that's not enough. And it doesn't matter if you do--if you're doing ray tracing or--it doesn't matter how fancy you get. That's not enough because of this. A normal cornea has that epithelial thickness profile. But, after refractive surgery, the epithelium might have produced a minus 125 shift in the refraction or a plus 2.3 diopter shift in the refraction.
And so, these contact lenses of epithelium sitting on the stroma and the refractive index difference between epithelium on the stroma, which is 1.401 versus 1.377, or 78 at the front of stroma, these actually make a difference on the order of a diopter in some cases.
And so, we--there is no reason why we can't make post LASIK IOL calculations extremely accurate, but we need to know the epithelial power and the stromal power separately. Then it's just an optical bench. One lens, two lens, and three lens, the other one that you're taking out, and the length of the eye. That's it.
I'll finish with this last example of a patient who had had PRK for hyperopia but had regressed. And I didn't want to do another PRK because she'd developed a little bit of peripheral reticular haze. And as you know, that produces cylinder and it's a mess.
But, because I had a femtosecond laser that was very accurate and I had an epithelial map, I decided to measure the thickness of the epithelium at the thickest point on each side and plan for a flap that goes just under the thickest point. And we did that in both eyes, and we were therefore safely able to avoid a buttonhole through this extremely thick epithelium of 99 microns.
So, as you can see, a knowledge of the epithelial thickness profile is going to be--is going to open up a whole new world of the way that we look at corneal refractive surgery, which, at the moment, is more like a--sort of like a, "Let's do it. Oh, it didn't work, let's do a little bit more," science. And now we can turn it into an actual anatomical problem that we can solve by anatomical surgery
Thank you.