Custom algorithms to analyze the OCT 3D corneal images, and to correct for optical and fan distortion have been described previously.
25,27–29,33,34 In addition, we have previously presented dedicated image processing tools for automatic analysis of ICRS from OCT images. Processing tools included denoising algorithms, automatic identification of 3D structures (cornea, iris, ICRS), multilayer segmentation, and distortion corrections.
Figure 2 illustrates the image analysis in S#2 for pre and postoperative measurements (7, 30, and 90 days).
The pupil center (obtained from the automatically identified iris volume) was used as a reference in the analysis of pre- and postoperative measurements. Corneal elevation maps were reported within the optical zone defined by the ICRS and the natural pupil. The center of the implanted ICRS was obtained from the automatically identified ICRS volume, and its shift from the pupil center estimated for registration of pre- and postoperative measurements. The optical zone is defined by the 4-mm diameter circular zone inside the circumference defined by the ICRS radius, and centered at the pupil center.
29
Corneal shape was described in terms of the radii of curvature and asphericities from fittings to a sphere and a biconicoid (anterior cornea:
R [sphere]; Rx and Ry, Qx and Qy [biconicoid]; posterior surface:
R [sphere]), and corneal elevation maps were also fitted to Zernike polynomial expansions (sixth order) using the center of the optical ICRS zone described in the previous paragraph as the reference. This way, the same optical zone was analyzed in the pre- and postoperative conditions.
29
Topographic and thickness maps were displayed in a grid square of 100 × 100 points in the 4-mm of diameter optical ICRS zone in order to ensure quantitative analysis of the optical zone without the ICRS influence. Topographic maps were represented as the difference of corneal elevation data from the reference sphere, in the so called “height representation,” with warm colors representing points that are higher than the reference surface and cool colors representing points below the reference. Both anterior and posterior corneal surfaces were fitted by Zernike polynomial expansions (note that these are fits to surface elevations, not corneal wave aberrations). The symmetry of the corneal elevation maps was obtained using the Root Mean Square error (RMS) of the asymmetric terms of the corneal elevation Zernike expansion (RMS_asym). RMS_asym was therefore defined as the RMS for astigmatism, coma, trefoil, tetrafoil, pentafoil, and hexafoil terms (Z2 −2, Z2 2, Z3 −3, Z3 −1, Z3 1, Z3 3, Z4 −4, Z4 −2, Z4 2, Z4 4, Z5 −5, Z5 −3, Z5 −1, Z5 1, Z5 3, Z5 −5, Z6 −6, Z6 −4, Z6 −2 Z6 2, Z6 4, and Z6 6) of corneal height maps (not to be confused with wave aberration maps). RMS_asym was evaluated both including and excluding astigmatism coefficients.
Thickness maps were calculated from direct subtraction of the posterior corneal surface from the anterior corneal surface. RMS thickness maps were used to assess the regularity of the thickness distribution. RMS_thicknessmap is therefore defined as the deviation of the thickness maps from a uniform pachymetry across the cornea.
Corneal power (D) was calculated by using the paraxial formula with both the corneal (1.376) and the aqueous (1.336) refractive indices. All metrics were computed for a 4-mm optical ICRS zone.
3D ICRS positioning was described by the following parameters: 3D ICRS depth, defined as the distance between the center of mass of the ICRS and the anterior corneal surface, and ICRS tilt, defined as the angle between the ICRS axis (normal to the ICRS plane) and the pupillary axis (normal to the pupil plane). Positive tilts around
x-axis indicate a forward tilt of the nasal part (OD)/temporal part (OS) of the ICRS plane with respect to the pupil plane. Positive tilts around
y-axis indicate a forward shift of the inferior part of the ICRS plane with respect to the pupil plane.
Figure 3 illustrates the definitions of the pupilary plane (and axis), ICRS plane (and axis), and ICRS tilt (
Fig. 3a), as well as the tilt sign notations (
Figs. 3b,
3c).
The changes in corneal geometry and ICRS position were analyzed statistically using an ANOVA (general linear model for repeated measurements). Significant levels (ANOVA and pair-wise two-tailed comparison t-test) were set at P less than 0.05. The statistical tests were performed using SPSS software (SPSS, Inc., Chicago, IL).