Process and Outcomes of Fitting Corneoscleral Profilometry-Driven Scleral Lenses for Patients with Ocular Surface Disease

Abstract

Objectives:

To assess the feasibility of obtaining Cornea Scleral Profile (CSP) measurements using Scheimpflug imaging and report on the fitting process of free-form custom scleral lenses (SLs) for patients with ocular surface disease (OSD).

Methods:

This prospective study of patients fit with free-form SLs collected data on the following: demographics, indications for wear, corneal and scleral tomography, scan acquisition process, and SL fitting process.

Results:

CSP scans were acquired on 15 eyes of 9 patients. Mean scan time for right eyes was 10.7 minutes and 9.7 minutes for left eyes. A mean of 2.9 follow-up visits were required to complete SL fitting, with a mean of 2.1 lenses ordered. One eye did not tolerate lens wear, and one eye could not be fit using the CSP scan due to insufficient data. The initial lens ordered was dispensed at the first follow-up visit for 7 of the remaining 13 eyes, all of which were ultimately fit successfully in free-form lenses.

Conclusions:

In this study of profilometry-guided SL fitting for eyes with OSD and low magnitude corneal astigmatism, the number of lenses and follow-up visits required were similar to outcomes of previous studies that described the diagnostic approach to SL fitting. Additionally, imaging technology does not negate the need for skilled clinical observation while fitting SLs.

The first scleral lenses (SLs) fabricated from rigid gas permeable materials were described in 1983,^1,2 but interest in SL prescription, management, and research did not begin to expand significantly until the early 2000’s.^3,4 The earliest commercially available SLs were fit using diagnostic lenses with spherical landing zones (LZ) which frequently required considerable modification to improve comfort and provide physiologically acceptable fit against the conjunctiva. The repetitive process of evaluation and modification was time-consuming for both patients and practitioners.^5,6 Advances in ocular imaging technology and investigations characterizing the shape of the anterior ocular surface have facilitated a more evidence-based approach to SL design and fitting. Using the sMap3D corneoscleral topographer (Cedar Crest, NM), DeNaeyer et al. (2017) obtained scleral shape data from 140 eyes of prospective SL patients. About one-third of those eyes had rotationally symmetrical shapes, while 2/3 of eyes demonstrated irregular patterns consisting of asymmetric depressions and elevations.⁷ The asymmetric nature of the sclera has been confirmed using another corneoscleral topographer, the Eaglet-Eye Eye Surface Profiler (ESP; Houten, Netherlands).⁸ Expanding knowledge of ocular surface shape has prompted manufacturers to incorporate toric, bitangential,⁹ or quadrant-specific¹⁰ LZs into their diagnostic lenses.

Individual variations in scleral shape and the high percentage of eyes with irregular ocular contour continues to make an iterative fitting process challenging. Integrating anterior surface imaging technology with lens fabrication has been proposed as a way to reduce fitting time by eliminating the trial-and-error involved in the traditional approach to diagnostic SL fitting.¹¹ Using data obtained through a molded impression of the eye^6,12-14 or through corneoscleral profilometry^15,16, it is now possible to design SLs with customized LZs. The outcomes of using impression-based SLs for the management of advanced anterior segment disease have been previously reported.^6,12,17,18 The purpose of the current study is to assess the feasibility of obtaining corneoscleral profilometry measurements using the Cornea Scleral Profile (CSP) software on the Pentacam (OCULUS Optikgrate GmbH, Germany) tomographer and to report on the lens design and fitting process of free-form SLs.

Methods

This prospective study was approved by the Institutional Review Board of the University of Illinois at Chicago and all procedures were conducted in accordance with the requirements of the Declaration of Helsinki. Consecutive patients presenting for SL fitting at the Illinois Eye and Ear Infirmary were recruited to participate. Patient informed consent was acquired prior to collecting research data.

Instrument and Scan Acquisition

Corneoscleral tomography data was collected using the Pentacam, a Scheimpflug imaging-based tomographer. The CSP software on the instrument allows for the measurement of sagittal height of the anterior ocular surface over a chord of up to 18mm in diameter using 250 discrete images.^19,20 No fluorescein dye is required. The device automatically captures a tear film-independent image of the ocular surface when proper alignment is achieved. Prior to scan acquisition, 1 drop of preservative-free artificial tear was instilled to improve patient comfort. For patients who required bilateral SL fitting, the right eye was scanned first, followed by the left eye.

Scans are acquired in the following 5 positions: central, nasal, temporal, superior, and inferior. The first scan is the central corneal scan, followed by the 4 peripheral scans. After each scan is completed, the software provides feedback regarding the integrity of the data acquired. Potential acquisition errors include lid closure/blinking, unsteady fixation, incomplete coverage, vertex error, and 3D model deviation. The practitioner can repeat scans for each position and make appropriate adjustments to address the specific error. Each scan can be repeated before proceeding to the subsequent scan. The patient’s eyelids can be manipulated with a LidStick^® (Oculus Inc, Arlington WA, USA) when necessary to enhance data capture. The LidStick^® is a disposable, silicone-tipped applicator that allows for enhanced grip of the eyelid. Figure 1, a sample printout of the CSP report for one of the eyes scanned in this study is shown.

Figure 1.

Printout of the CSP report showing (A) scans in 25 directions, (B) cornea scleral elevation map, (C) coverage map for the data collected, (D) minimum and maximum sagittal heights for (E) specific chord lengths (can be adjusted).

Data Collected

The following information was collected from the CSP report for each eye: flattest simulated (SimKf) and steepest simulated (SimKs) keratometry readings, corneal astigmatism (the difference between SimKf and SimKs), horizontal visible iris diameter (HVID), and the maximum measurable anterior ocular surface diameter with a complete data set. Minimum (MinSH) and maximum sagittal heights (MaxSH) for chord lengths of 15, 16, and 17 mm were recorded and the scleral toricity (the difference between MaxSH and MinSH)²¹ at each chord length was calculated. In addition, information regarding the maximum scan coverage area with a complete data set captured by the CSP software was collected. The “Ring Diameter” (labeled as “E” in Figure 1) represents the chord length from which the data shown on the report was collected. The practitioner can manually adjust the value to determine the extent to which complete data was collected.

Lens Ordering

Following scan acquisition, over-refraction was performed over a diagnostic lens from the fitting set of the study lenses (BostonSight Scleral, BostonSight, Needham, MA) with known base curvature and power. Prior to initiating power determination, a slit lamp was used to confirm that central corneal clearance was present. Free-form lenses (BostonSight Smart 360, BostonSight, Needham, MA) were ordered for each patient using FitConnect, BostonSight’s online lens design and ordering platform. When using the Pentacam CSP software, the profilometry data is exported as a .csv file and imported into FitConnect. The practitioner selects lens diameter, base curve, corneal clearance and lens power. The lens diameter was selected between 16 and 19mm based on clinician preference taking into consideration scan coverage area and each patient’s clinical presentation (e.g., horizontal visible iris diameter, extent of conjunctival staining, degree of corneal ectasia, amount of scleral show, etc). In general, a larger diameter is often preferred for patient with ocular surface disease to provide protection over a larger area of the ocular surface.

Demographics, indications for SL prescription, data related to the scan acquisition process (duration of scan, number of clinicians required for image acquisition, need for lid holding, number of scans required per position, scan errors), and the lens fitting process (number of follow-up visits, number of lenses ordered, initial lens fitting characteristics, lens parameter changes made, lens successfully dispensed) were recorded. Profilometric characteristics of eyes included in this study were also recorded. All data was stored in REDCap (Research Electronic Data Capture) hosted at the University of Illinois at Chicago.^22,23

Data Analysis

Data were summarized with descriptive statistics including mean, standard deviation, and range reported.

Results

CSP imaging was completed for 9 participants (15 eyes). Six participants identified as females and 3 participants identified as males. The mean age of all participants was 53.0 ± 14.8 years (range 18 to 67). OSD was the general indication for SL prescription for all participants; specific conditions to be treated included keratoconjunctivitis sicca (5 participants, 9 eyes), neurotrophic keratitis (1 participant, 1 eye), ocular Graft-verus-host-disease (2 participants, 3 eyes), and neuropathic pain (1 participant, 1 eye). Additional clinical characteristics of patients fit into profilometry-guided lenses are available in Supplemental Appendix 1.

Scan Acquisition

A single clinician was able to retract eyelids and acquire imaging for all 15 eyes. Mean scan time for right eyes was 10.7 ± 6.5 minutes (range 4 to 19, n=6) and 9.7 ± 4.7 minutes (range 3 to 15, n=9) for left eyes. The number of scans required to achieve an adequate, error-free scan was highest in the superior quadrant for right eyes (2.7 ± 1.5 scans superior, 1.2 ± 0.4 central, 1.5 ± 0.8 nasal, 1.3 ± 0.5 temporal, and 1.2 ± 0.4 inferior). Left eyes required the highest number of scans in the inferior quadrant (2.4 ± 1.6 scans inferior, 1.3 ± 0.5 central, 1.9 ± 1.2 nasal, 1.1 ± 0.3 temporal, and 2.1 ± 1.7 superior). The most common scanning errors were related to lid closure/blinking and unsteady fixation. All eyes required manual lid retraction to obtain adequate scans in the superior and inferior positions.

Corneal and scleral tomography

Eyes (15 eyes) exhibited mean corneal astigmatism of 0.98 ± 0.88 D, an HVID of 11.6 ± 0.4 mm, and a mean maximum scan coverage area of 17.3 ± 0.3 mm (see Table 1). Mean scleral toricity was highest at the 17 mm chord (308.86 ± 175.40 um), compared to 208.64 ± 157.92 um at the 16 mm chord and 136.71 ± 132.96 um at 15 mm. Seven eyes exhibited scleral toricity in the oblique orientation, 4 eyes exhibited against-the-rule toricity, and 2 eyes exhibited with-the-rule toricity. In this series of eyes with low magnitude corneal astigmatism, there was no association between the orientation of corneal astigmatism and scleral toricity.

Table 1.

Tomography data from the Cornea Scleral Profile (CSP) report reported as mean ± SD.

Tomographical Characterisitcs of Eyes Scanned N=15 eyes; OD (n=6) and OS (n=9)
Corneal Astigmatism (in D)	1.0 ± 0.9
Horizontal Visible Iris Diameter (in mm)	11.6 ± 0.4
Maximum Scan Coverage Area (in mm)	17.3 ± 0.3
Scleral Toricity at 3 Chord Lengths (in um)	15 mm	16 mm	17 mm
	136.7 ± 133.0	208.6 ± 157.9	308.9 ± 175.4

Lens fitting

Fitting was completed using CSP technology to design free-form SLs for 7 of 9 participants (13 eyes). One participant (one eye, left eye) was unable to tolerate diagnostic SL wear during over-refraction and SL fitting was not pursued. A lens could not be ordered for one eye (left eye) of another participant due to an insufficient data set despite acquiring error-free scans; this eye was instead fit using a diagnostic lens.

The mean SL diameter ordered was 17.7 ± 0.8 mm for right eyes and 17.9 ± 0.7 mm for left eyes. A mean of 2.9 ± 1.3 total visits (range 2 to 5) over a period of 71.4 ± 37.5 total days (range 30 to 120) were required to complete SL fitting, with a mean of 2.1 ± 1.2 (1-3, n=6) lenses ordered for right eyes, and 2.1 ± 1.8 (1-4, n=7) lenses for left eyes (see Table 2). The initial lens ordered for 7 eyes (54%) provided adequate central corneal fluid reservoir, limbal clearance, SL LZ alignment, and visual acuity, and was therefore dispensed. For the 6 remaining eyes, the initial lens was not dispensed for the following reasons: poor LZ alignment (5/6), excessive central corneal fluid reservoir (4/6), and poor visual acuity requiring refractive power change (2/6). The total number of errors is greater than 6 because several lenses exhibited more than one inadequate fit characteristic. Ultimately, all 13 eyes were fit successfully in free-form lenses after lens modifications were made without repeat profilometry scanning. Additional details on scleral lens design and fitting, including changes made to first lens ordered and the number of lenses ordered, are available in Supplemental Appendix 2.

Table 2.

Free-form scleral lens (SL) fitting data and outcomes reported as mean ± SD.

Fitting Outcomes for Eyes Successfully Fit into Free-Form SLs N=13 eyes; OD (n=6) and OS (n=7)
SL Diameter (in mm)	OD	OS
	17.7 ± 0.8	17.9 ± 0.7
# of Lenses Ordered	OD	OS
	2.1 ± 1.2	2.1 ± 1.2
# of Visits	2.9 ± 1.3
Fitting Duration (in days)	71.4 ± 37.5

Habitual correction worn prior to free-form SL fitting was as follows: 1 eye wore a soft contact lens, 1 eye wore a quadrant-specific SL design, 10 eyes wore spectacles, and 3 eyes wore no correction. For right eyes, the presenting visual acuity improved from a mean of 0.08 to 0.03 logMAR with the final SL. For left eyes, the presenting mean visual acuity improved from 0.12 to 0.03 logMAR with the final SL.

Discussion

The iterative process of diagnostic SL fitting can be challenging and time-consuming for both patients and practitioners.^5,6 Theoretically, the fabrication of a custom lens based upon images of the ocular surface should provide an excellent fit with minimal adjustment, thus expediting the fitting process. Furthermore, exact alignment of the SL LZ with the ocular surface should provide an ideal fit with the first lens ordered. While CSP imaging was successfully acquired for a majority of patients in this series (14 of 15 total eyes) with ocular surface disease, ideal lens fit was not achieved with the first lens ordered for over half of the eyes included in this cohort.

Several previous studies have reported the numbers of lenses and visits needed to complete the process of fitting scleral lenses in clinical practice (see Table 3).^24-27 The mean number of lenses ordered per eye ranged from 1.5 to 4.5, and the mean number of visits needed to complete the fitting process ranged from 2.8 to 6.5. Most of these studies report data from a single practice, but the largest included survey-generated data on 419 eyes from 233 practices worldwide; mean number of lenses ordered in this study was 2.4/eye, and the mean number of visits was 3.8.²⁴ While the average number of lenses ordered and visits required to complete the fitting process using free-form lens design as reported in the present study suggests that profilometry may increase efficiency compared to some practices, differences in the fitting process were not remarkable for this study population of eyes with low magnitude corneal astigmatism.²⁸ Additionally, imaging technology does not negate the need for skilled clinical observation while fitting SLs.

Table 3.

Comparison of SL fitting outcomes reported by various practices

Comparison of Reported SL Fitting Outcomes
	Current Study (N=13 eyes)	Schornack et al. 2021 (N=419 eyes)	Scanzera el al. 2020 (N=133 patients)		Pecego et al. 2012 (N=107 eyes)	Schornack & Patel 2020 (N=30 eyes)
Study Type	Prospective	Survey	Retrospective		Retrospective	Retrospective
Indication for SL Prescription	OSD	OSD, CI, Uncomplicated Refractive Error	OSD (n=95 patients), CI (n=38 patients)		OSD, CI	Keratoconus
SL Fitting Approach	Free-Form	Not reported	Diagnostic		Diagnostic	Diagnostic
# of Lenses Ordered	2.1	2.4	Commercially available lens		3.2	1.5
			OSD	4.5
			CI	3.8
			PROSE
			OSD	6.7
			CI	6
# of Visits	2.9 follow up visits	3.8 visits	Commercially available lens		6.2 follow up visits	2.8 visits
			OSD	6.5 follow up visits
			CI	3.9 follow up visits
			PROSE
			OSD	7.4 follow up visits
			CI	7.5 follow up visits
Fitting Duration	1-4 months	Not reported	Not reported		3-17 months	3-32 months

Incorporating new skills and technology into clinical practice may initially be time-consuming. Macedo et al. (2019) reported that efficiency of the fitting process of SLs improved significantly as a novice practitioner gained experience. The mean number of lenses applied during the initial consultation dropped from 2.35 for the first 20 patients to 1.56 once the practitioner had completed 140 fits. The number of reorders placed during the fitting process also decreased from 0.95 to 0.25 during that time. This suggests that 1.95 lenses were required to complete the fitting process for the first 20 fits, while only 1.25 lenses were required following 140 fits.⁸

Obtaining high-quality, error-free CSP scans requires practice and experience. It is possible that poor peripheral scan coverage contributed to issues with LZ alignment in this study. Certain patient characteristics may make it difficult to obtain scans of optimal quality, including small palpebral apertures, severe tear film dysfunction, blepharospasm, ptosis, scleroderma, symblepharon, and history of aggressive blepharoplasty or tarsorrhaphy. Retracting lids sufficiently to obtain adequate images can be difficult in patients with deep periorbital sulci and may require positional modifications. Image acquisition can vary depending on examiner experience, patient’s orbital structure, and patient cooperation and mobility. Utilizing two individuals available to retract eyelids and obtain scans may be beneficial particularly when first incorporating new technology. In this series, the average time to capture scans in 5 different positions was 10 minutes for each eye. However, there are newer instruments that allow for data capture in a single snapshot, requiring less images and potentially less time. Additionally, the mean maximum chord length captured by CSP profilometry was less than the mean SL diameter ordered. Bandlitz et al. (2020) investigated the agreement and repeatability of Fourier-based profilometry (Eye Surface Profiler, ESP) and Scheimpflug imaging (Pentacam CSP) in normal eyes. The ESP requires scans in 3 positions of gaze. The study found that the maximum possible measurement zone diameter with ESP (16.4 ± 1.3 mm) was significantly greater than with Pentacam (14.8 ± 1.1 mm). The mean maximum scan coverage area (17.3 ± 0.26 mm) in the current study is higher than both of these reported values. However, a scleral profilometer that captures more data in the periphery where the SL LZs land could potentially have a beneficial effect on SL fitting outcomes.

The time required to obtain high-quality scans of the ocular surface may foil efforts to meaningfully reduce the amount of time required for initial evaluation at the onset of the image-guided lens fitting process. Accurate estimates of the amount of time required to complete an initial SL evaluation using either diagnostic or technology-based lenses have not been published. One might assume that the time needed to place one or more diagnostic lenses on the eye and allow the lens to settle on the eye to determine an appropriate initial lens design would be greater that the amount of time needed to capture images necessary to design and fabricate a lens. However, as noted previously, time required for image acquisition is not inconsequential. Even with image-based fitting, a diagnostic lens needs to be placed upon the eye to provide refractive data. The total amount of time required for initial evaluation (patient evaluation and education, image acquisition, and collection of refractive data) was not assessed as part of this study.

In addition to quantifying elements of the SL fitting process, the present study confirmed results from previous studies on characteristics of anterior surface contour in patients with ocular surface disease. Scleral toricity, defined as the greatest difference in scleral sagittal height between two perpendicular meridians, is known to increase as the distance from the limbus increases²¹. Previous studies suggested that a sclera with less than 100 um of toricity at a 15 mm chord length could be adequately fit in a spherical SL design.^13,29 Larger SLs that land beyond a 15mm chord length or eyes that exhibit scleral toricity greater than 200 um may require toric lens landing zone to ensure a stable lens fit and minimal decentration.^21,29 The average lens diameter ordered was 17.8 mm, which is larger than the average diameter of SLs fitted by a majority of practitioners.³⁰ The eyes included in this study exhibited a mean ± standard deviation scleral toricity of 309.0± 175.0 um at the 17mm chord. Both lens diameter and scleral toricity (greater than 200 microns) in the present sample suggest that toric or custom LZs would be required to provide optimal fit for these eyes. Despite this, the study did not demonstrate marked reduction in the number of lenses and visits required for free-form lens fitting over diagnostic fitting. Future studies including eyes that have been surgically altered (e.g.. glaucoma filtering procedures or scleral buckling) with highly irregular contours may provide more guidance on the role of CSP-assisted SL fitting, as previous studies have reported success with custom, impression-based SL fitting in eyes that previously failed SL wear^6,12,17.

Study limitations include a small sample size and lack of inclusion of patients with corneal irregularity (the most common indication for SL wear).³ Previous evaluations of scleral shape have suggested that eyes with corneal ectasia, particularly those with moderate to severe disease, exhibit more toricity and asymmetry compared to those with normal corneal profiles.^31,32 Therefore, it may be inaccurate to generalize the findings of this study to patients with corneal irregularity. Furthermore, this series utilized a single free-form design from a single manufacturer. Time required to acquire images using other profilometers may differ, and there are undoubtedly differences between lens designs from various manufacturers.

Conclusion

Theoretically, corneoscleral profilometry has the potential to streamline the SL fitting process by digitizing the ocular surface to create a truly customized fit. In this study of profilometry-guided SL fitting for eyes with OSD and low magnitude corneal astigmatism, the number of lenses and follow-up visits required were similar to previous studies that described the diagnostic approach to SL fitting. Future studies including more irregular ocular shapes may reveal greater differences between the two fitting approaches. Given the increasing opportunities to incorporate technology into SL design and fitting, additional investigations will also allow SL practioners to better understand the role of custom-SL designs in their practices. However, imaging technology does not negate the need for skilled clinical observation while fitting SLs.

Supplementary Material

Appendix 1

Supplemental Appendix 1. Clinical characteristics of patients fit into profilometry-guided lenses.

Appendix 2

Supplemental Appendix 2. Scleral lens design details: changes made to first lens ordered and number of lenses ordered.