Geometric Modelling of the Human Cornea: A New Approach for the Study of Corneal Ectatic Disease. A Pilot Investigation

Abstract

The aim of this study was to describe the application of a new bioengineering graphical technique based on geometric custom modelling capable to detect and to discriminate small variations in the morphology of the corneal surface. Methods: A virtual 3D solid custom model of the cornea was obtained employing Computer Aided Geometric Design tools, using raw data from a discrete and finite set of spatial points representative of both sides of the corneal surface provided by a corneal topographer. Geometric reconstruction was performed using B-Spline functions, defining and calculating the representative geometric variables of the corneal morphology of patients under clinical diagnosis of keratoconus. Results: At least four variables could be used in order to classify corneal abnormalities related to keratoconus disease: anterior corneal surface area (ROC 0.853; p < 0.0001), posterior corneal surface area (ROC 0.813; p < 0.0001), anterior apex deviation (ROC 0.742; p < 0.0001) and posterior apex deviation (ROC 0.899; p < 0.0001). Conclusions: Custom geometric modelling enables an accurate characterization of the human cornea based on untreated raw data from the corneal topographer and the calculation of morphological variables of the cornea, which permits the clinical diagnosis of keratoconus disease.

Machine Learning Techniques in Keratoconus Classification: A Systematic Review

“…6) Geometric data: Correspond to information essentially describing the geometry of the anterior and posterior corneal surfaces to diagnose any pathology linked to an alteration in corneal morphology [55]. Among these data the total corneal volume, the anterior corneal surface, the posterior corneal surface, the total corneal surface, the deviation of the anterior apex and the deviation of the posterior apex [26].…”