Research Review

Detecting Keratoconus Early

Research Review

Detecting Keratoconus Early



Detecting subclinical keratoconus continues to be challenging in the clinical environment. Screening tools to detect early disease are pivotal for preventing complications after corneal refractive surgery, for genetic studies, and to assess candidacy for cross-linking. Most practicing clinicians do not have access to the cutting edge tools and instruments that are typically reserved for referral practices or for the research environment. Therefore, a simple, accessible index to screen for subclinical keratoconus in the absence of clinical disease is desired in the clinical setting.

New Detection Methods

With advancements in ocular surface imaging, new diagnostic methods have promising ability to diagnose early keratoconus. For example, epithelial thickness variables from anterior segment optical coherence tomography (OCT) have been used for keratoconus detection (Li et al, 2012). In polarization sensitive OCT, the observed birefringence patterns from normal corneas are used as standard patterns for comparisons with pathologic changes in keratoconus, in which the pattern is heavily distorted due to stacked collagen fibril lamellae of different orientations.

Other sophisticated detection methods include corneal biomechanical metrics (i.e., the corneal deformation response to air pressure or hysteresis), which has been shown to improve the detection of keratoconus in combination with traditional keratometric and pachymetric assessments (Wolffsohn et al, 2012). Additionally, the evaluation of total and corneal wavefront higher-order aberrations has been suggested as a better method to detect early very mild forms of ectasia compared to placido-based programs alone (Saad and Gatinel, 2012).

Topography: Tried and True

However, corneal topography remains a trusted method to identify keratoconus suspects (Maeda et al, 1995; Rabinowitz and Rasheed, 1999), and it is readily available to practicing clinicians.

The first disease detection topographic algorithms were based solely on anterior surface topography (Rabinowitz, 1995; Li et al, 2009; Maeda et al, 1994). In fact, Ramos-Lopez et al (2013) confirmed that placido-based corneal indices can discriminate between subtle forms of keratoconus.

While the good sensitivity and specificity of placido-based topographic algorithms for screening and grading keratoconus are well documented, their intrinsic limitations are that they provide information only on the anterior corneal surface. The scanning slit-based and rotating Scheimpflug camera–based systems, which provide curvature and elevation of both anterior and posterior corneal surfaces as well as corneal thickness, have been suggested as better methods to discriminate early disease from normals.

Posterior Elevation Systems Mean posterior elevation and/or corneal thickness are sensitive indices to assist in the screening of keratoconus (Ambrosio et al, 2006; Ambrosio et al, 2011; Auffarth et al, 2000; Lim et al, 2007). Systems capable of posterior elevation have been available for some time. Much is published on the slit-based Orbscan II (Bausch + Lomb) system regarding its good reliability for measuring the posterior corneal surface (Maldonado et al, 2006; Rabsilber et al, 2005; Seitz et al, 2001). Posterior surface elevation using the Scheimpflug camera–based Pentacam (Oculus) also demonstrates good reliability for pachymetry measurements and posterior corneal curvature (Chen and Lam, 2007 and 2009; Jain et al, 2007; Lackner et al, 2005).

What is “Normal?”

Normative data for posterior surface elevation using the Orbscan II default settings states that the average amount of maximum posterior elevation compared to the best fit sphere is about 21μm to 28μm ±7μm (Lim et al, 2007; Rao et al, 2002; Sonmez et al, 2007; Wei et al, 2006). Similarly, for Orbscan II, in Wei et al’s (2006) study of 140 normal eyes, the maximum posterior elevation was never greater than 46μm, while in Rao et al’s (2002) study of 50 normal eyes, the posterior elevation was never greater than 40μm.


Figure 1. A clinically normal cornea with posterior elevation >50μm above the best-fit sphere.

In Pentacam-derived indices, normative values for posterior elevation are lower compared to the Orbscan counterparts. Feng et al (2011) published international values of corneal elevation in normals using the Pentacam Eye Scanner. Posterior elevations were measured at the apex and thinnest point in 555 patients. Upper limits of normal for collective international data were 7.5μm and 13.5μm at the posterior apex and posterior thinnest point, respectively. These results suggest that a posterior elevation index (which varies by instrument) that is greater than the instrument’s upper limits in normals can raise a suspicion of early keratoconus even in the absence of clinical signs.

Can a Single Elevation-Based Index Detect Keratoconus?

Several studies have suggested that posterior elevation indices alone are good at discriminating between normal eyes and clinically apparent keratoconus, but delineating between normals and keratoconus suspect eyes is much more difficult (Lim et al, 2007; Rao et al, 2002; Mihaltz et al, 2009). Kollbaum (2006) presented data delineating normals from clinically apparent keratoconus, and posterior elevation was found to be the best-performing single Orbscan II index to detect keratoconus when the highest elevation on the posterior surface was >50μm above the best-fit sphere (Figure 1). This value yielded sensitivity of 0.93, specificity of 0.99, and accuracy of 0.96.

Posterior elevation measurements clearly differ between Orbscan II and Pentacam systems as seen by the normative databases described earlier. In confirmed keratoconus patients, a difference of almost 14μm has been recorded, with lower mean posterior elevation in Pentacam (Quisling et al, 2006). Studies on Pentacam to determine optimal cutpoints for detection of subclinical keratoconus found a mean posterior elevation greater than 29μm to 32μm (in the central 5mm cornea) as being predictive (de Sanctis et al, 2008). Furthermore, a study using the Galilei (Ziemer Ophthalmic Systems AG) Scheimpflug-based analyzer to determine optimal elevation cutpoints for discriminating subclinical keratoconus from normals found that posterior elevation (in the central 7mm zone) >50.5μm was predictive.


Posterior corneal elevation as a single predictive index is readily available to clinicians who have scanning slit-based and Scheimpflug camera–based systems. This index delineates normals from clinical keratoconus, but alone its efficacy is lower for subclinical keratoconus. Some studies support that abnormal posterior corneal elevation may be a suitable standalone screening tool to detect subclinical keratoconus when anterior topography and clinical exam reveals an otherwise normal cornea. Clinicians should understand the research done on their respective instruments compared to their instrument’s default settings prior to applying such information in practice. CLS

For references, please visit and click on document #209.

Dr. Szczotka-Flynn is a professor in the Departments of Ophthalmology & Visual Sciences and Epidemiology & Biostatistics at Case Western Reserve University and director of the Contact Lens Service at University Hospitals Case Medical Center Eye Institute. She is also the director of the Coordinating for the CPTS Study. She receives research funding from Alcon and Vistakon.