Integrated Ocular Response Analyzer Waveform Score as a Biomechanical Index of Keratoconus Disease Severity

Abstract

Purpose: : To evaluate a new biomechanical measure to grade keratoconus (KC) disease severity.

Methods: : The Ocular Response Analyzer (Reichert Inc.) was used to gather biomechanical data on 110 patients (211 eyes) identified with KC. Eyes having undergone surgery were excluded from analysis. Degree of KC disease was determined using two methods; a clinical ratings score (CRS) of normal, suspect, mild, moderate and severe as determined by topography analysis (normal, atypical or characteristic KC steepening as well as flat keratometry reading) and the Keratoconus Severity Score (KSS) proposed by McMahon et. al. The KSS relies on calculation of average corneal power, RMS wavefront error (6^th to 27^th terms) of the first corneal surface as determined using VOL-CT software (Sarver & Associates), classification of topography mapping as well as slit lamp signs. A KSS of 0 (normal) through 5 (most severe disease) was assigned to all eyes. A group of 360 pre-LASIK eyes were chosen as a "normal" control group. Waveform numerical scores (WS) derived from 24 parameters of ORA corneal deformation signals were assigned by a neural network. WS were compared to numerical scores designated by KSS and CRS for all groups.

Results: : For KC patients, mean age was 35.0 years (SD 10.7; range 13 to 67 years). Mean flat keratometry was 46.08D (SD 5.92; range 37.92 to 74.83) and mean corneal power was 47.80D (range 40.52 to 77.03). Mean RMS higher order wavefront error of the first corneal surface was 2.25microns (SD 3.97; range 0.03 to 30.9).Correlations of the KSS and CRS with WS were 0.50 and 0.52, p<0.0001 respectively excluding pre-LASIK eyes. Including normal pre-LASIK eyes, categories normal, suspect, mild and moderate were significantly separated, p<0.0002 while moderate and severe failed to reach significance, p=0.20. For KSS groups, only groups 3 to 4 (equivalent to mild and moderate disease) were significantly separated.

Conclusions: : Neural network analysis of ORA signals can provide an index of keratoconus severity that agrees well with clinical observations based on topographic assessment and may be clinically more relevant than KSS evaluation. Development of a KC disease index would be useful in etiologic studies as well as evaluation of efficacy of surgical interventions.

Keywords: keratoconus • clinical (human) or epidemiologic studies: systems/equipment/techniques • cornea: clinical science