Purpose : To investigate a method for discriminating changes in the pattern of corneal collagen fibrils to detect keratoconus (KC).

Methods : Patients with KC and healthy controls were included in this prospective study. Corneal tomography (Pentacam) was taken as well as interferometric analysis of diffractive and polarizing effects related to the birefringent properties of corneal collagen fibrils (Lumaxis device, Phronema S.r.l., Bari, Italy). Three scans of each eye per patient were acquired. Custom software was developed to improve the visualisation of the representative cross by extracting the histogram equalised luminance to provide a numerical outcome to characterise the We then fit a set of ellipses to the image in a semi-automatic approach, aiming to capture the boundaries of the iris, pupil, each four quadrants of the cross, and the upper and lower eyelids. Combining this information, we are able to obtain cross (called cross parameter). A correlation between cross outcome and Kmax was calculated.

Results : A total of sixty eyes of sixty patients were included: 30 healthy subjects and 30 patients with KC. The control group included 60 eyes, mean age of 38.0 ± 13.2 years. The KC group comprised 60 eyes, mean age of 39.2 ± 12.9 years. Maximum keratometry value were 59.4 ± 11.2 D in the KC group and 45.3 ± .8 D in the healthy group. The cross parameter was 0.41 ± 0.028 in the KC group and 0.24 ± 0.055 in the healthy group. Using only measures taken from the cross resulting from the Lumaxis scan, we were able to distinguish between KC and healthy eyes.

Conclusions : The interferometric analysis of diffractive and polarizing effects related to the birefringent properties provide new morphological information at corneal fibrils level. A different corneal pattern of corneal collagen fibrils can be recognised in keratoconus patients compared to healthy patients.