N. Maeda et al., AUTOMATED KERATOCONUS SCREENING WITH CORNEAL TOPOGRAPHY ANALYSIS, Investigative ophthalmology & visual science, 35(6), 1994, pp. 2749-2757
Purpose. Although visual inspection of corneal topography maps by trai
ned experts can be powerful, this method is inherently subjective. Qua
ntitative classification methods that can detect and classify abnormal
topographic patterns would be useful. An automated system was develop
ed to differentiate keratoconus patterns from other conditions using c
omputer-assisted videokeratoscopy. Methods. This system combined a cla
ssification tree with a linear discriminant function derived from disc
riminant analysis of eight indices obtained from TMS-1 videokeratoscop
e data. One hundred corneas with a variety of diagnoses (keratoconus,
normal, keratoplasty, epikeratophakia, excimer laser photorefractive k
eratectomy, radial keratotomy, contact lens-induced warpage, and other
s) were used for training, and a validation set of 100 additional corn
eas was used to evaluate the results. Results. In the training set, al
l 22 cases of clinically diagnosed keratoconus were detected with thre
e false-positive cases (sensitivity 100%, specificity 96%, and accurac
y 97%). With the validation set, 25 out of 28 keratoconus cases were d
etected with one false-positive case, which was a transplanted cornea
(sensitivity 89%, specificity 99%, and accuracy 96%). Conclusions. Thi
s system can be used as a screening procedure to distinguish clinical
keratoconus from other corneal topographies. This quantitative classif
ication method may also aid in refining the clinical interpretation of
topographic maps.