Fundamentals in Corneal Topography1

Different features used to characterize optical properties of the cornea must be introduced in advance:

The evolution of the quantitative assessment of the anterior corneal curvature has been progressive and its range of measurement has been extended from the four points a few millimeters apart measured by keratometers to a grid of thousands of points covering almost the entire cornea measured by computerized corneal topography.

Normal corneas have some interesting features:

Several studies have checked that the anterior corneal configuration tends to be prolate. This means negative asphericity values, ranging normal values from -0.26 to -0.11, depending on the study considered.

In a map with a color-coded power scale (corneal topographic map), we can see how the dioptric power declines towards the periphery (Fig. 1.4). Depending on how the dioptric power declines, we can recognise five different normal topographic patterns. The approximate distribution of keratographic patterns described in normal eyes includes the following:

The bow tie configuration represents the existence of corneal astigmatism. The localization of the steepest axis defines against-the-rule astigmatism (horizontal axis) (Fig. 1.9), with-the-rule astigmatism (vertical axis) (Fig. 1.10), or oblique astigmatism (near meridian angle 45º or 135º) (Fig. 1.11).

Scheiner in 1619 was the first in analysing corneal curvature by matching the image of a window frame reflected from a subject's cornea with that produced by one of his calibrated spheres.

Following this basis and using the ability of the anterior corneal surface to behave as a convex mirror, in 1854 Helmholtz created the first true keratometer, also called ophthalmometer. Before this, some authors (Mandell, Ramsden, Kohlrausch) had tried to developed an apparatus for determining corneal curvature, creating some kind of prototypes. The Helmholtz keratometer projects four points onto the cornea, creating a reflected image, which can be analysed and converted to corneal radius data by using an equation that considers distance from mire to cornea, image and mire size. In 1881, Javal and Schiotz improved Helmholtz laboratory apparatus and built an ophthalmometer designed for clinical use (Fig. 1.12). Although keratometers are still used commonly in clinical practice, it has important limitations:

The keratoscope let us examine a much greater area of the cornea. Clinical keratoscopy has been used since 1870, when Placido studied the corneal surface by observing the corneal reflections of a series of illuminated concentric rings (known as Placido's rings) (Fig. 1.13). In 1896 Allvar Gullstrand incorporated the disk in a self-developed instrument for examining photographs of the cornea by a means of a microscope and calculating corneal curvature by means of a numerical algorithm. Depending on the distance between the concentric rings, the examiner can evaluate qualitatively the corneal shape.

In the 1980s, photographs of the projected images were hand-digitalized and then analysed by a computer. However, the automation of process and the image capture by a digital camera became soon a reality. Advances in video and personal computer technology made it possible to instantaneously acquire and analyze high quality images. These improvements made quantitative keratoscopy 8practical for routine clinical use. The Corneal Modeling System (CMS) produced by Computed Anatomy became the first widely used computerized videokeratoscope.

In an effort to improve accuracy and avoid the assumptions necessary for computerized videokeratoscopy, new technologies (projection systems) have appeared commercially in the last years, such as rastereography, laser interferometry, Moiré interference and slit-scan systems (Fig. 1.14). These systems directly measure the true shape of the anterior corneal surface in terms of height, not from curvature. From these measurements slope, curvature and power can be calculated.

Some of the errors that examiner can make using a Placido-based system are the following: focusing errors (Fig. 1.15), alignment (Fig. 1.16) and fixation errors (that could induce wrong levels of astigmatisms), wrong calculation of the position of the center from the small central rings, increased inaccuracy toward the periphery due to the lack of accuracy of the preceding points. We can conclude that critical points for a precise measurement are accurate alignment, centring and focusing. These issues depend on the ability of the examiner to take a good measurement.

There is another potential source of errors. It is essential for obtaining a good exam that the tear film forms a smooth layer over the irregular corneal epithelium. Tear film break-up causes mistracking of the mires and artifacts in the corneal map that appears as irregularity areas or false irregular astigmatism. For example, a dry patch could be associated with an area of focal flattening in the corneal map (Fig. 1.17). In order to avoid disturbing of the tear film, corneal topography should be performed before giving dilating drops and taking intraocular pressures.

In addition, there are a variety of complex surfaces that are very difficult to characterize. Errors greater than 4 D may occur in very steep or flat corneas, keratoconus with local steepening, sharp transition zones after uncomplicated refractive surgeries, diffusely irregular surfaces after penetrating keratoplasty, and complex surfaces after decentered ablations in refractive surgery or central islands. These complex cases may be easily identified on the raw image of the Placido rings projected onto the cornea. When there are abnormalities in tear film break up, the raw image is also very valuable.

The presence of areas which are not analysed on the corneal maps may be due to the configuration of the nose or the position of the eyelids (Fig. 1.18). To avoid these artifacts, it is important to instruct the patient to keep the eyes wide open and to maintain a good position and fixation.

In the slit-scan systems, a slit is projected sequentially onto the cornea at different angles. This is the basis used by the Orbscan corneal topography system. A high-resolution video camera captures 40 light slits projected onto the cornea. The diffuse reflection is obtained from the cornea, iris, and lens. By triangulation, data is obtained from the anterior and posterior surfaces of the cornea and from other structures. These issues permits this instruments to calculate anterior chamber depth or full pachymetry of the cornea. The accuracy and repeatability of this instrument is reported to be below 10 µ and it is significantly dependent on many factors, such as 10movement of the patient's eye, stability of tear film, ability of patients to keep the eyes wide open, corneal transparency, and the presence of corneal abnormalities. One of the limitations of slit-scan system is the longer time of image acquisition in comparison with any of the other commercially available instruments. In the other hand, it brings the possibility to characterize the posterior corneal surface. This issue is important for the detection of incipient keratoconus conditions.

Before examining the color-coded map, it is critical to check the raw image of Placido rings projected on the cornea. It is important to confirm that the unprocessed data is reliable, as further analysis is based on these data (Fig. 1.20).

Depending on the kind of software associated to the videokeratocope, several other displays may be available as three dimensional maps (Fig. 1.23), relative maps, numerical views, profile views, etc.

There are some indexes that are specific of some kind of commercially available instruments. These indexes can be used in clinical practice as different mathematical estimates of the visual disturbance that can be expected to be caused by the amount of irregularities of the anterior corneal surface. Below we list some of them:

Common causes of corneal irregular astigmatism include dry eye, corneal scars, ectatic corneal degenerations, trauma, pterygium, and surgery. Some causes of corneal irregularity may be discovered just with careful inspection of the photokeratoscope view showing irregular mires as in poor quality of tear film, subtle corneal scar, or epithelial basement membrane degeneration. The latter is a risk factor for corneal ecstasies after refractive surgery.

Minor degrees of irregular astigmatism are not uncommon and would not contraindicate standard refractive surgery. However, significant amount of irregular astigmatism can limit surgical outcomes due to poor predictability. Surgeons should be especially aware when dealing with rigid contact lens wearers who expect final uncorrected visualacuity to be similar to their previous best corrected visual acuity, as well as in patients who state that their vision with rigid lenses is significantly better than with spectacles.

Contact lens wear can alter the shape of the cornea and induce corneal warpage (Fig. 1.26). Corneal warpage occurs more frequently in wearers of hard PMMA contact lenses or RGP contact lenses as a result of the mechanical pressure exerted by the lens, and corneal hypoxia.

There are at least four different forms of noticeable topographic changes that usually occur mixed with one or more of the other forms:

Although keratoconus can be detected by qualitative methods, early or subclinical keratoconus are usually missed. Modern corneal topographers usually include software to predict the presence of keratoconus by quantitative methods in a more sensitive and specific way. Special effort is made in seeking for early keratoconus or “forme fruste keratoconus”, considered a contraindication to LASIK due to possible evolution into clinically significant keratoconus within months or years after the surgery.

Rabinowitz and Mc Donnell developed the first numerical method for detecting keratoconus using only topographic data. They established three criteria to determine whether the patient has keratoconus or not:

However, using only these simple measurements for keratoconus diagnosis presents specificity problems.

To solve the specificity problem, the new strategy must be able to detect and consider the unique characteristics of keratoconus maps, such as local 16abnormal elevations. The Keratoconus Prediction Index, developed by Maeda et al, is calculated from the Differential Sector Index (DSI), the Opposite Sector Index (OSI), the Centre/Surround Index (CSI), the SAI, the Irregular Astigmatism Index (IAI), and the percent Analyzed Area (AA). This method partially overcomes the specificity limitation.

Maeda et al also developed the neural network model, based on artificial intelligence. It is a much more sophisticated method for classifying corneal topography and detecting different corneal topographic abnormalities; it employs indexes that were empirically found to capture specific characteristics of the different corneal pathologies, including keratoconus. Further modifications in neural network approach developed by Smolek and Klyce supposedly produce 100% accuracy, specificity and sensitivity in diagnosing keratoconus.

Newer approaches for detecting early keratoconus (Fig. 1.29) and increased risk of ectasia after refractive surgery include patterns and indices based on the anterior or posterior elevation best fit sphere maps, obtained from Orbscan ®. However, further studies are needed to assess the performance of the topography elevation criteria in detecting keratoconus. It is known that the cone of the keratoconic cornea appears as an area of increased elevation surrounded by concentric zones of decreasing elevation. Two indices have been proposed to screen keratoconic corneas based on the Orbscan system:

Keratoconic eyes exhibit significantly more ocular higher-order aberrations due to an increase of corneal higher-order aberrations compared to normal eyes. Coma-like aberrations were dominant compared with spherical-like aberrations in keratoconic eyes (Fig. 1.30), although both were significantly higher than in normal subjects. Wavefront sensing could also help to differentiate keratoconic eyes from normal eyes by analyzing the characteristics of the higher-order aberrations. The corneal wavefront aberration, which is the component of the total ocular wavefront aberration attributed to the cornea, can be derived from the corneal topographic height data.

To summarize, corneas with” forme fruste keratoconus” or other ectatic disorders have increased risk of post-LASIK ectasia. Clinicians may employ one of the several methods noted above. However, it is important to remember that there is no approach that guarantees correct discrimination of all normal and ectatic corneas. Therefore, if an ectatic disorder is suspected, LASIK will not be safely performed.

A very typical finding in these corneas is a topographic pattern with a polygonal shape (Fig. 1.32), which is seen in over a half of them. The polygons include squares, hexagons and octagons, depending on the number of incisions. The angles of the polygons correspond closely to the central ends of these incisions.18

The topographic pattern in myopic corrections shows an oblate profile, with flattening of the central cornea (Fig. 1.33). The treatment zone is usually well delineated by the close proximity of adjacent contours at its edge, especially in the next days following the procedure during the first period of overcorrection.

The intrastromal corneal ring inserts act as passive spacing elements that shorten the arc length of the anterior corneal surface and therefore flatten the cornea. Corneal topography demonstrate that general flattening of the central cornea occurs after placement of ring and ring segments, but the normal positive asphericity of the cornea is maintained. The ring itself can be identified in the topographic map as a peripheral annulus of local steepening, surrounding, as already mentioned, a flattened central cornea (Figs 1.42A and B).

According to the FDA, intrastromal corneal rings are intended for the reduction or elimination of mild myopia (-1.00 to -3.00 D spherical equivalent at the spectacle plane) and they may now be used to reduce irregular steepening (irregular astigmatism) caused by keratoconus.

In a recent study, Intacs (a model of intracorneal ring), provided better results in visual acuity and corneal topography quality and significantly reduced the spherical equivalent in eyes with keratoconus with relatively low mean K values (< or = 53 D) and a relatively low spherical equivalent. In contrast, when performing the procedure in advanced keratoconus (mean K-reading > or = 55 D), poor results can be anticipated.

Different patterns of corneal shape result after penetrating keratoplasty, the most frequently seen are:

Although we have mentioned penetrating keratoplasty, the same patterns can be seen in lamellar keratoplasty, such as in deep anterior lamellar keratoplasty. Deep endothelial lamellar keratoplasty does not generate postsurgical astigmatism (Fig. 1.44).

With the advent of computerized topographic systems, it is possible to distinguish two different types:

The typical associated topographic pattern is the presence of an inferior area of steepening usually in the temporal quadrant. In advanced cases, the dioptric power at the apex is at or above 55 D (Fig. 1.42A). The superior-nasal quadrant of the cornea is commonly the last to be affected, and thus an “island” of normal corneal topography often remains in this area even in the more advanced stages of the condition. In a small group of patients, the topographic alterations may be centred at the central cornea, showing an asymmetric bow tie configuration with the inferior loop normally larger than the superior loop (Fig. 1.45).

Keratoconic Corneas have three common characteristics that are not present in normal corneas:

The videokeratographic pattern has a classic butterfly appearance that results in a flattening of the vertical meridian and a marked against-the-rule irregular astigmatism (Fig. 1.28).

Other types of corneal pathology such as epithelial dysplasia and epithelial basement membrane disease can also be evaluated using corneal topography.