Evaluating a Topography Software Program for Fitting RGPs

BY BETH A. SOPER, C.O.A., N.C.L.C., JOSEPH P. SHOVLIN, O.D.,
& EDWARD S. BENNETT, O.D., M.SC.
OCT. 1996

Computer videokeratography provides an alternative to trial lens fitting by calculating accurate RGP lens parameters.

This six-month, multi-site study evaluates a new contact lens fitting algorithm, Pro-Fit by EyeSys Technologies, which recommends lens parameters based on corneal topography data and user-defined selections for sagittal depth, overall lens diameter and back surface design. The program also provides simulated fluorescein patterns for analysis of computer-suggested lens prescriptions.

NOMOGRAMS VS. ALGORITHMS

Clinicians prescribe most RGPs using nomograms, which are formulas based upon central, and in some cases peripheral, keratometric readings. Central keratometric readings, numeric averages of four points measured near the center of the cornea, may omit critical information regarding the topography of the corneal periphery. This can lead to inappropriate lens designs or undetected cases of pathology and irregular corneas.

Computerized videokeratoscopy instruments measure 6,000 to 10,000 data points on the cornea. These measurements allow a better estimation of the base curve-to-cornea fitting relationship than does a simple numerical match of the posterior lens surface to the estimate of the anterior corneal curvature, as is done in the traditional keratometry method of fitting.

Pro-Fit is a software program that combines corneal topography data with the practitioner's established fitting protocol to determine the appropriate fit of an RGP lens. The practitioner selects a lens design, diameter and desired sagittal depth. The system then finds an exact base curve parameter by repetitively calculating the base curve of the optical zone until the desired sagittal depth is achieved between the front surface of the cornea and the center of the back surface of the contact lens (Fig. 1).

Simulated fluorescein patterns show how lenses with the computer-suggested parameters would fit, and practitioners can select from a database of lenses by various manufacturers or continue with a customized, computer-generated fit. Simulated fluorescein patterns can be evaluated at any position on the cornea using the pan and tilt lens positioning features of this program.

Figures 2 and 3 illustrate how a Pro-Fit simulated fluorescein pattern compares to the actual fluorescein pattern created when the suggested lens is placed on the patient's eye.

MATERIALS AND METHODS

We used the EyeSys 2000 Corneal Analysis System to obtain videokeratographic images of both corneas of 50 contact lens wearers. Participants had best corrected visual acuities of at least 20/20, less than three diopters of corneal astigmatism and regular, symmetric astigmatic topography patterns with corneal apices very near to center. All participants were subjected to slit lamp biomicroscopy, and normal external exams were indicated.

We edited images for limbal rings and other artifacts, to prevent erroneous data from appearing in the corneal power maps. We excluded 14 of the 100 eye images from the study because 12 contained unedited limbal rings, and two of the final contact lenses were manufactured improperly. [Author's note: The release version of Pro-Fit (version 3.10) contains an automatic limbus detection feature which will avoid the calculation errors that occurred in the 12 study subjects who were eliminated.] We viewed axial and tangential radius of curvature maps as well as profile displays to assess topography patterns, corneal astigmatism and location of corneal apices. We used only acceptable images for the fitting process.

The Pro-Fit contact lens fitting software allows the practitioner to select from five back surface designs for custom RGP lens fitting. To represent a cross-section of lenses typically ordered in an optometric practice, we evaluated four of these designs: a spherical base curve with two spherical peripheral curves (27 eyes); a spherical base curve with three spherical peripheral curves (13 eyes); an aspheric base curve with a spherical periphery (28 eyes); and an aspheric base curve with an aspheric periphery (18 eyes).

We ordered the lenses precisely as the program specified, in rigid gas permeable materials with a Dk of 70 or less and with a minimum central lens thickness of 0.12mm. RGP materials included SGP I (telefocon A, Permeable Technologies, Inc.), FluoroPerm 30 (paflufocon C, Paragon Vision Sciences), FluoroPerm 60 (paflufocon B, Paragon Vision Sciences) and Fluorex 700 (flusilfocon A, GT Labs). One laboratory, Firestone Optics, manufactured all the lenses using the Coburn 8001 SX automated numeric lathe.

Four independent clinical sites evaluated lens parameters produced using the pre-release version of Pro-Fit (version 3.04) for physical characteristics such as movement, centration, position, central apical clearance and fluorescein pattern. They measured visual acuity using a standard Snellen chart, and performed sphero-cylindrical overrefraction. We assigned a fitting protocol of either 10 or 15 microns central sagittal depth, representing average sagittal vaulting values for an alignment or parallel fitting technique. Seventy-one eyes were fit using a protocol for 10 microns central sagittal depth, and 15 eyes were fit using a protocol for 15 microns central sagittal depth. The clinician was able to choose total lens diameter.

Long-term follow-up was not indicated in this study, so dispensing of the lenses was left to the discretion of the clinicians, and subjective findings such as comfort and wearing time were not addressed.

RESULTS

Practitioners at the four sites said they would have dispensed 91.86 percent of the lenses (79 lenses) suggested by the software program to the subjects without modification to any lens parameters (Fig. 4). Of these lenses to be dispensed, 31.65 percent (25 eyes) obtained visual acuity of 20/15, 62.02 percent (49 eyes) saw 20/20, 5.06 percent (4 eyes) saw 20/25, and 1.27 percent (1 eye) had 20/30 vision (Fig. 5).

When asked if the simulated fluorescein pattern reflected the actual fluorescein pattern of the lens on the patient's cornea, practitioners deemed 84 out of 86 patterns as accurate simulations (Fig. 6). Practitioners also commented that some of the lenses would require extra blending of the peripheral curves for enhanced patient comfort and adaptation. CLS

References are available upon written request to the editors at Contact Lens Spectrum. To receive this information via fax, call 1-800-239-4684 and request Document #18. (Be sure to have a fax number ready.)

Beth Soper is the manager of professional services and clinical research at EyeSys Technologies, Inc.

Dr. Shovlin, a diplomate of the Cornea & Contact Lens Section of the AAO, is a clinical associate at Northeastern Eye Institute in Scranton, Pa.

Dr. Bennett is an associate professor of optometry at the University of Missouri - St. Louis, and executive director of the RGP Lens Institute.