This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
We appreciate the concerns of Dr Goren regarding our recent article in which we compare the sensitivity and specificity of three methods for the detection of keratoconus: a simple keratometry approach, a multipoint videokeratoscope method developed by Rabinowitz and McDonnell,1 and a more involved, expert classifier model.2 While the first two methods detected both clinical keratoconus and keratoconus suspect, the expert classifier system was trained specifically to recognize the topographic patterns on corneas previously diagnosed nonvideokeratographically with clinical keratoconus. Hence, this system was not designed to report keratoconus suspect but to provide an indication of the presence of clinical keratoconus for the ophthalmologist to verify. The purpose of the article was to point out the validity and utility of the expert system compared with less specific methods.The expert classifier system does not make the diagnosis of a clinical disease before its development. True, the system
Klyce SD, Smolek MK, Maeda N. Comparison of Methods for Detecting Keratoconus Using Videokeratography-Reply. Arch Ophthalmol. 1996;114(5):631-632. doi:10.1001/archopht.1996.01100130623030