True tomographic imaging requires the generation of an X, Y, and Z coordinate system and the measurement of true shape. The first commercially available elevation-based system was the PAR Corneal Topography System (PAR CTS) (PAR Technology, New Hartford, NY). The PAR CTS used a stereo-triangulation technique (rasterphotogrammetry) to make direct measurements on the anterior corneal surface. The PAR CTS used a grid pattern composed of horizontal and vertical lines projected onto the anterior surface at an angle (Fig. 7.2). In order to visualize the grid, the PAR system required a small amount of fluorescein in the tear film. From the known geometry of the grid and the change when projected onto the corneal surface the system was able to compute the X, Y, and Z coordinate in space. Because the system required a stained tear film to visualize the grid it was limited to measuring only the anterior corneal surface [8]. While, the PAR is no longer commercially available, it was the first system to utilize elevation data in a clinically useful form and had documented accuracy superior to the available Placido-based systems at that time [9].

The first elevation system with the capability to measure both anterior and posterior corneal surfaces utilized a scanning-slit technique of optical cross sectioning (Orbscan (formerly Orbtek) Bausch & Lomb, Rochester, NY). The ability to measure both corneal surfaces in space allowed for the generation of a corneal thickness map as corneal thickness is simply the spatial difference between the corresponding anterior and posterior corneal positions. A full corneal thickness map offers significant advantages over single point pachymetry (discussed in the following chapter). Numerous articles have demonstrated the limitations of the Orbscan, particularly in locating the posterior corneal surface after refractive surgery and subsequently underestimating the corneal thickness [10–13]. Currently, there are a number of different devices and different technologies that produced tomographic data. Although some differences exist between systems, they all display elevation data in a similar fashion that was first introduced with the PAR CTS in 1990 [8].

While we refer to standard anterior and posterior corneal tomographic maps as elevation maps, that is somewhat of a misnomer. Rarely does the clinician view elevation data in its raw form and most systems do not offer the actual raw elevation data. The reason is that the raw elevation data from normal eyes and markedly ectatic corneas look remarkably similar (Fig. 7.3). In order to make the maps clinically useful and to allow for a rapid visual inspection, the raw data is compared to some reference surface. The purpose of the reference surface is to magnify or amplify the surface differences that would otherwise not be appreciated to the naked eye. The so-called elevation maps depict how the corneal surface differs from a defined reference shape. While the appearance of the map will vary greatly depending on the reference surface used, all maps are generated using the raw elevation data. The reference surface will affect the appearance, but not the accuracy of the actual data [14].

The choice of the reference surface will often depend on the clinical situation, the population being evaluated, and the specific pathology you are screening for. For most applications the best fit sphere (BFS ) is the most qualitatively intuitive (easiest to read and understand) surface and the most commonly used. A BFS allows for the visualization of astigmatism as the flat meridian rises above the BFS , while the steep meridian drops below the BFS . The normal astigmatic pattern generated against a BFS is easily recognizable (Fig. 7.4). As opposed to a BFS , a best fit toric ellipsoid (BFTE ) will better fit or mask an astigmatic cornea. When screening for ectatic disease one is trying to identify an abnormal conical protrusion. A focalized protrusion will appear as an elevated area against the BFS (Fig. 7.5). Since the cornea is normally aspherical, steeper in the center and flatter toward the periphery, normal corneas will display a central positive elevation. The goal of screening is to allow for a rapid visual inspection to separate normal from abnormal. This task is made more difficult by the fact that the normal cornea is aspherical and displays, to a smaller degree, a positive elevation (“positive island of elevation”), similar to what is seen with ectatic disease. The BFS and the resultant elevation map will vary depending on how much of the cornea is utilized to construct the reference surface. If the entire cornea is used to construct the BFS then the normal asphericity of the cornea will be clearly demonstrated. As the area (optical zone) utilized to compute the BFS is decreased the BFS steepens as less of the flatter periphery is incorporated into the BFS . If only the central 3.0 mm of the cornea were to be utilized, the resultant BFS would be substantially steeper. It has been shown that taking the BFS from the central 8.0 mm optical zone steepens the BFS enough to effectively mask the normal asphericity (Fig. 7.6). Masking the normal asphericity makes screening for ectatic disease easier and allows for a rapid visual inspection looking for positive Island of elevation against the 8.0 mm derived BFS .

While a BFS derived from the central 8.0 mm zone allows for a rapid visual inspection, it should be understood that while the reference surface does not affect “accuracy,” it does affect quantitative data and published normal values are all reference surface specific [15, 16]. Normal elevation values will not only vary based on which reference surface is utilized, but will vary based on where or what part of the cornea the values are measured from. Elevation normal values have been reported at the apex, maximal value within the central 4.0 mm zone, and at the thinnest point [15–17]. The thinnest point has some advantages as in the keratoconic cornea , the thinnest point usually corresponds to the center of the cone and this is where we normally quote our recommended normal values. Additionally, there is some geographic and/or ethnic variation in normal values [16] as well as differences between myopic and hyperopic individuals (Table 7.1) [15].