Validating Corneal Topography Maps
Learn how to recognize an invalid topography map and what steps to take to obtain an accurate image.
By Randy Kojima
Mr. Kojima is the director of technical affairs for Precision Technology Services in Vancouver, BC. You can reach him at email@example.com.
In the hands of many practitioners, a corneal topographer is an indispensable instrument for analysis, diagnosis and especially for contact lens design. But like all instruments, it has its own specific shortcomings. For instance, corneal topographers reflect light off the anterior surface of the cornea. If the tear film or epithelium is inconsistent because of dryness, staining or scarring, the topographer won't accurately measure and calculate the true shape of the cornea.
You can easily resolve a dry eye prior to topography capture by instilling artificial tears to "even" or "smooth" the fluid film layer. It's a little trickier to correct for a disrupted epithelium that limits the accuracy of the instrument in reading the shape of the cornea.
I'll explain how you can determine whether you have an invalid topography map and how to obtain accurate captures.
What to Look for
Most corneal topographers, both new and old, are based on placido disk technology. When the patient looks down the axis of the instrument, a series of rings reflect off the corneal surface back to the camera at the vortex of the disk or cone. The topographer treats the cornea as a convex mirror and in fact measures the surface tear film rather than the epithelium itself.
However, tear film, disrupted epithelium or corneal scarring can cause a topographical error called "ring jam," which occurs when the mirror surface (tear film or cornea) is inconsistent, causing the placido rings to break or intersect. Ring jam can result in analysis miscalculation or "extrapolation" error. This type of assumption or guess results in completely invalid corneal topography interpretation.
Figure 1. Accurate capture.
Figure 2. Poor capture exhibiting a central "divot" or depression in the corneal curvature.
Figure 3. The photokeratoscope image reveals ring distortion (ring jam), which caused the unusual appearing map in Figure 2.
Compare the two maps taken on the same eye within seconds of each other (Figures 1 and 2) with the Medmont E300 corneal topographer (Medmont International Pty, Ltd./Precision Technology Services). Note the significant difference in the interpretation of shape and curvature. Figure 1 shows a normal corneal topography capture on a mildly astigmatic patient. Figure 2 is the same eye with a central divot or depression in the corneal curvature. When you take a closer look at the photokeratoscope image on the latter capture (Figure 3), you can see that two of the central placido rings have distorted and intersected each other. This ring jam correlates precisely with the position of the "divot" (flat spot) in Figure 2. Because Figure 1 shows a normal appearing topography capture and the biomicroscopic evaluation of the cornea reveals this to be a healthy cornea, we must deduce that the ring jam on the successive map was an isolated tear film inconsistency.
Topography users must capture images while the tear film is even and consistent to achieve accurate corneal topography interpretation. If patients exhibit tear break-up or dryness, which is evident in the ring reflection (ring jam), then instill artificial tears and retake captures. Before saving any map, always review the photokeratoscope image for even ring reflection and the absence of ring jam. Additionally, it's best to plan on taking at least three images and if two are similar, then use one of those two.
Irregular Cornea Topography Captures
What can you do when the corneal surface is disrupted by staining or scarring? This problem presents itself most prominently with keratoconus patients. Corneal staining and worse, scarring, eliminates clear reflection of the rings.
For example, Figure 4 shows two consecutive topography captures of the same right eye taken within seconds of each other. Note the significant difference in both shape and curvature (12.80D difference between apical curvature readings). Which is the valid interpretation, or is either one an accurate determination of the true shape of the cornea?
To help determine which capture may be valid, a cursory glance at the photokeratoscope image (Figure 5) shows significant distortion and breaking of the rings, resulting in extrapolation error. This error is the cause for our two captures (taken seconds apart) appearing so dissimilar.
So what is the true shape of the cornea? If determining a contact lens course of action, knowing the actual shape of the cornea, or at the very least having a better idea of the topography and curvature, would help with the initial diagnostic lens selection (both type of lens and parameters). With such significantly different topography readings, what could assist in attaining a more accurate map?
Consider placing a low power soft lens on such stained or scarred corneas to smooth out the surface. This will eliminate most, if not all, of the ring jam as well as reduce topographical guestimation (extrapolation error). The result is a more accurate interpretation of the contour of the anterior surface. The soft lens will "blur out" curvature readings and finite shape changes, but the contour of the cornea will be far clearer. For example, a junction between scarred and healthy epithelium may be less distinct, but elevations and depressions will be more accurately interpreted from quality ring reflection. Also, the topography may be better represented with the soft lens on, but the radii will be longer by the approximate thickness of the lens.
Figure 4. Note the difference in appearance of the maps. Additionally, repeatability of these captures on the same eye is low with apical curvature readings separated by 12.80D!
Figure 5. The photokeratoscope image reveals significant ring jam, which caused our dissimilar appearing maps. This is due partly to tear breakup, but mostly to corneal scarring.
Figure 6. The soft contact lens "smoothes" out the disrupted and stained epithelium, allowing for quality ring reflection. Note the absence of ring jam compared with the Figure 5 captures.
Figure 7. Achieving quality ring reflection has resulted in both reproducible captures on this same eye and the likelihood of accurate interpretation of contour and shape.
Figure 6 shows the same right eye with two successive captures over a CIBA Vision O2 Optix diagnostic lens. Note the total absence of ring jam, which will provide us with a better interpretation of corneal contour and shape as evidenced by the Figure 7 analysis.
Figure 8. Approximately 13 of the 32 rings on this topographer's placido disk have significant ring jam superior to the apex. This severely hinders the topographer's ability to determine shape and curvature.
There is symmetry between captures (repeatability of data) and the corneal curvatures are very similar, which is also an indication of the accuracy and validity of the maps.
The previous case allowed us to determine that a stained and distorted epithelium had a 'peak' in curvature. In other words, the steepest reading (and possibly the highest elevation) was at the point of greatest staining. This is predictive for a keratoconus patient who has a small area of steep ectasia such as this. However, do all keratoconus patients with heavy staining have the steepest point at the most severe area of staining? The actual curvature could be difficult to determine with significant ring jam.
The following case exhibits again the difficulty that topographers have in determining shape on severely distorted corneas. Because of the heavy central staining on this keratoconus patient, the topographer is unable to achieve quality ring reflection as evidenced by the photokeratoscope image in Figure 8.
The topographer's calculation of this capture resulted in the Figure 9 analysis. Note the small area of color contour displayed at the apex which showed a massive 33D+ difference from the steepest reading to the flattest (shape scale). This is an obvious example of topographical error resulting from heavy ring jam caused by severe staining. What approximate curvature are we dealing with? Is this an elevated or depressed area of contour? More importantly, we don't have a clear basis point from which to calculate an initial diagnostic lens for keratoconus contact lens dispensing.
The poor quality of our ring reflection has limited our access to a clear picture of the shape and curvature of this cornea. A better perspective is possible by placing a soft contact lens on eye to eliminate the disruption in the epithelial surface.
Figure 10 shows the same cornea with an O2 Optix (plano) lens on eye. Previously, more than a third of the placido ring data was invalid. With a smoother surface from which to reflect light, you can see that little ring jam is present.
Figure 9. Determination of curvature was only possible on a small area of this capture. As it's based on poor ring reflection, we assume it to be completely invalid map.
Figure 10. The soft lens has smoothed out this very disrupted epithelium, creating quality ring reflection with an almost total absence of ring jam.
By achieving quality ring reflection, we're able to interpret the shape and curvature of the cornea to a high level of accuracy. Without the soft contact lens, the topographer could calculate only a small area of the apex of this cornea, and the data was completely invalid because of the significant ring jam. With the smoother surface created by our soft contact lens, we now have an almost complete picture of the shape and curvature of this eye.
Additionally, by capturing two independent readings, we can use the repeatability as a gauge of accuracy. Note in Figure 11 the similarity between captures. We can assume with a level of confidence that we're looking at valid topography captures because both ring reflection and symmetry between captures are high.
In analyzing the above case, we can see a significant depression in the curvature of both maps located 1mm (one grid box) inferior to the apex. This cornea is characterized as an invaginated cone. The cause is likely an extended period of time with a GP lens bearing on the cone. The peak of the ectasia has now collapsed, leaving a depression where there once was an elevation. Any GP lens placed on this eye would likely have a large bubble locked behind the lens in this depression. Without our soft lens to aid in the capture and interpretation, we wouldn't know the approximate curvature of this cornea. We also wouldn't see the depression or invagination that determines what type of lens to diagnostically fit for this patient. Additionally, knowing about the depression and predicting bubble formation behind any corneal GP lens would tell us that a scleral lens may be the best lens design for this patient.
Tips for Consistent Results
Corneal topography is an invaluable instrument in every eyecare practice, especially for contact lens specialists. But for some of the toughest corneas in which you require the topographer the most, dry eye, staining and scarring can limit the accuracy of the information. For such corneas, instill artificial tears to reduce or eliminate tear breakup and to even out the tear film for quality ring reflection.
For corneas that have very disrupted surfaces in which ring jam and distortion can't be avoided, capture over a thin, low power soft contact lens to help determine shape and curvature. We need to further explore the altering of curvature values in capturing over soft contact lenses, but for very distorted epitheliums it can provide much more anterior shape information in cases when little is possible. CLS
Figure 11. With quality ring reflection, we're now able to interpret the shape of this cornea. Both captures over the soft lens (on this left eye) exhibit similar maps, which is a good indication that we have valid interpretations of shape.