EyeWorld Asia-Pacific December 2013 Issue

32 EWAP rEfrActivE December 2013 How - from page 31 software’s ability to remove the color coding of the sub-D values to reduce inadvertent medico-legal concerns. The power of a multivariate regression analysis as the final value is the best predictor. The individual values by definition are poorer predictors of the disease and should not be used individually in the screening process. The reason that five (of nine) of the most heavily weighted values are shown on the BAD is to help the clinician explain the final result value, not predict ectasia. It is also important to remember that even the best predictor BAD Final D value is not 100% accurate. Pentacam BAD Final D is a multivariate regression. Each univariate variable is weighted according to its own individual predictive value. Each variable is compared to a large known database of normal patients and patients with keratoconus then plotted on ROC curve plots showing sensitivity vs. specificity allowing comparison of individual variables’ predictive nature. So Final D is not a percentage predictor of ectasia or a standard deviation but a predictor of true/ false positives vs. true/false negatives based on a normative database population. I recommend additional factors to consider when trying to diagnose keratoconus (see tables). Ultimately, the question that needs to be answered is not “Does this patient have keratoconus?” but “Can this patient’s cornea safely tolerate laser vision correction?” In order for us to better predict if a patient’s cornea can “survive” LVC, I think we need to look closer at corneal biomechanics. Most of the current biomechanical studies in progress involve the identification of “at risk or keratoconus suspect.” There are no large scale studies in publication on LVC effects on biomechanical changes of the cornea. I believe every cornea is at risk if the “wrong” collagen fibrils are compromised. A first step would include collecting biomechanical data before and after we alter corneas with laser vision correction, which could lead to a better understanding of this dynamic process, ultimately allowing us to perform more surgery and offer a better safety profile for all our patients. Therefore if topography and tomography can only show keratoconus once it begins, the addition of corneal biomechanical analysis should further improve our ability to predict a patient’s relative risk of ectasia based on corneal thickness, viscosity, IOP, and tissue removed by surgery. The near future for laser vision correction patient screening is likely to include a combination of advanced tomography and corneal biomechanical analysis to push our ability to provide the safest form of refractive surgery for our patients. EWAP Editors’ note: Dr. Tullo is vice president of clinical services, TLC Laser Eye Centers, Princeton, NJ, USA. Dr. Tullo has financial interests with Oculus. contact information tullo: William.Tullo@tlcvision.com