Fourier analysis of corneal Scheimpflug imaging: clinical use in keratoconus

To evaluate the clinical use of Fourier analysis of videokeratography data in the diagnosis and follow-up of keratoconus (KC).

We conducted a chart review of consecutive patients presented to our cornea clinic. A team of two experienced cornea specialists divided the patients into three groups: normal cornea, forme fruste KC (FFKC), and clinical KC. The exclusion criteria were a history of previous ocular surgery, any accompanying corneal pathology other than KC, high myopia (>6.00 diopters), amblyopia, pregnancy, breastfeeding, or any current autoimmune disease. The data of Fourier series harmonic analysis were evaluated for their diagnostic capacity using the receiver operating characteristic (ROC) curve. A binary logistic regression analysis was also conducted to construct a diagnostic model. A total of 259 eyes showed progression in the clinical KC group and underwent a combination of accelerated corneal collagen cross-linking and topography-guided customized treatment with an excimer laser.

The study included 1262 eyes (618 normal, 530 KC, and 114 FFKC) of 1262 patients. We observed that maximum decentration (MaxDec) was almost as good as maximum keratometry (Kmax) in detecting progressive KC. The area under the curve (AUC) was 0.95 for KC [95% confidence interval (CI): 0.93-0.96] and 0.84 for FFKC (95%CI: 0.79-0.88). Higher predictive accuracy was obtained using a model combining the spherical component, MaxDec, irregularity, and regular astigmatism in the center of the cornea (AUC: 0.97; sensitivity: 89%, and specificity: 96%).

Decentration, Kmax, and posterior radii of curvatures from a 3.0-mm optical zone centered on the thinnest point of the cornea provide the highest accuracy with low reproducibility of Kmax.