Abstract
Purpose :
Placido-based curvature maps are widely used for diagnosis of corneal ectasia, but they do not represent the true curvature of the human cornea. It is well known that they lead to frequent misdiagnosis of corneal disease. We have developed a quartic smoothing spline algorithm on non-rectangular domains to compute elevation maps for posterior and anterior corneal surfaces, pachymetry, and the true mean curvature. We applied the algorithm to classify keratoconus, and forme fruste keratoconus (FFK).
Methods :
The data collection was done with Texas Tech University Institutional Review Board permission using allowed protocols. Anterior and posterior corneal surfaces were constructed from data from an AS-OCT for three patients - one each with normal cornea, forme fruste keratoconus and keratoconus. We computed quartic smoothing splines for the anterior and posterior corneal surfaces. The true mean curvature (TMC) was computed for anterior surface. The difference between the corneal surface and the best-fit-sphere is the elevation map. The pachymetry map is the difference between the anterior and posterior corneal surfaces.
Results :
(I)Validation: (a) For a normal cornea, the elevation map of the anterior surface is within 2 μm to that from an ATLAS topographer (Fig-1). (b) The elevation map was within 33 μm for a sphere over a 12mm central region. The computed TMC for the sphere was within 0.3D from the correct value over a 3mm central region. (II) Diagnosis: The anterior elevation (AE) for forme fruste keratoconus (FFK) showed no significant difference with that of the normal cornea, but the pachymetry map showed a thin region of 250 μm (Fig-2). For the patient with keratoconus, high mean curvature was observed in a region that matches with largest deviation from the best-fit-sphere. The Pachymetry map showed a thin region of 350 μm.
Conclusions :
Pachymetry map should be used to diagnose forme fruste keratoconus since anterior elevation and curvature maps fail to diagnose this condition. The TMC can help to accurately locate the cone for keratoconus, and the cone location might not match the thin zone of the cornea.
This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.