Utilising a Visual Image Quality Metric to Optimise Spectacle Prescriptions for Eyes with Keratoconus

Abstract

Purpose:

To compare optical performance, visual performance and patient-perceived quality of vision with: (1) spectacles determined through subjective refraction and (2) spectacles determined using an objective optimisation method based on wavefront aberration data for eyes with keratoconus.

Methods:

Thirty-seven eyes (22 subjects) with keratoconus underwent both subjective refraction and uncorrected wavefront aberration measurement. Wavefront aberration data were used to objectively identify a sphero-cylindrical refraction that optimised the visual image quality metric visual Strehl ratio (VSX). The two refractions were assembled in trial frames and worn by the subject in a random order. High contrast visual acuity (VA), letter contrast sensitivity and the patient’s short-term subjective preference were recorded for each prescription.

Results:

Median magnitude of the dioptric difference (a measure of similarity between the subjective and objective refractions) was 2.77 D (range= 0.21 to 20.44 D, 1^st quartile=1.02 D, 3^rd quartile= 4.36 D). Sixty-eight percent of eyes had better VA with the objective refraction and 32% of eyes gained more than one line of visual acuity. Monocularly, objective refraction was preferred 68% of the time when looking at a distant acuity chart and 76% of the time when viewing a real-world dynamic scene.

Conclusions:

Objective refraction based on visual image quality derived from wavefront aberration data can be valuable in the determination of monocular spectacle refractions for individuals with keratoconus.

INTRODUCTION

Patients with keratoconus make use of spectacles in a number of settings. In the early stages of the disease – when higher-order aberrations (HOA) are modest and before rigid contact lenses are needed – spectacles can be a viable primary mode of correction.¹ In moderate to severe disease, where rigid contact lenses are the standard of care,^2,3 best clinical practice recommends prescribing a pair of spectacles for times when contact lenses are not being worn (e.g., upon waking; at the end of the day or when lens discomfort, allergies or infections necessitate lens removal). Given the frequent concurrence of corneal ectasia and ocular surface disease, there is also an additional population of eyes that are clinically ineligible for contact lenses and are therefore dependent on spectacles.⁴

The gold standard for determining a spectacle correction is subjective refraction, which presents a series of sphero-cylindrical lower-order aberration (LOA) correcting lenses in front of an eye whose optical imperfections consist of both LOA and HOA. The success of subjective refraction hinges upon the ability of a patient to make a perceptual judgment regarding the quality of vision through each of the lens options. In normal eyes, where the level of HOA is low, this process is generally successful and subjective responses from the patient lead the clinician to a correction that provides good optical and visual performance. However, subjective refraction in keratoconus can be uncertain and variable, where the significant visual consequence of elevated HOAs confounds the patient’s judgement when choosing which lens is better, and visual systems with keratoconus can be more strongly adapted to their habitual percept than normal eyes.⁵ Further, the test-retest variability for individuals with keratoconus is four to six times greater than test-retest levels seen in a typical myopic sample.⁶

Additionally, the starting point of subjective refraction can preclude the optimal sphero-cylindrical refraction for a given eye from ever being reached in keratoconus.^7,8 If the subjective refraction process encounters a local image quality maximum within the refractive space (an area of relatively high image quality surrounded by lower image quality) rather than the global maximum, then the two-alternative forced choice nature of subjective refraction will not allow the patient to escape that local maximum through their responses. In such a case, moving from a local maximum to the global maximum will require the patient to traverse areas of poorer image quality before reaching the global maximum. This conflicts with the methodology of subjective refraction where successive judgments aim to continually take the patient towards better visual quality.⁹

Objective methods of refraction have an appeal in that they do not rely on the responses of the patient and can evaluate large ranges of dioptric corrections, thereby being better equipped to find the global optimal correction. Therefore, this study tests whether an objective method of identifying a spectacle prescription that considers the combined influence and interactions of LOA and HOA yields visual benefit in eyes with keratoconus. This objective method leverages a metric of visual image quality, the visual Strehl ratio (VSX), that is calculated from wavefront error and captures the combined effects of LOA, HOA and diffraction in the context of neural contrast sensitivity as a function of spatial frequency.^10,11 Prior work has demonstrated that change in the base₁₀ logarithm of VSX (logVSX) is predictive of change in visual performance in eyes with keratoconus¹² as well as normal eyes.^12,13 Further, an objective optimisation of refraction using VSX has identified spectacle prescriptions that performed equivalent to or better than subjective refraction in normal eyes¹⁴ and equivalent to subjective refraction in eyes of individuals with Down syndrome.¹⁵

In this study, sphero-cylindrical corrections identified using this objective method were compared with subjective refraction dioptrically, in terms of VSX, as well as using performance measures of high contrast logMAR visual acuity (VA), letter contrast sensitivity (CS) and subjective patient preference in eyes with keratoconus. In the context of a clinical application, this study asked if the addition of an objective method to the conventional subjective procedures could provide a path to a larger percentage of keratoconus patients having the option of—and being more satisfied with— their spectacle corrections. The goal of the study was not to determine the ability of the metric-derived method to replace subjective refraction, but rather to compare the performance of the two methods with the view that the metric-derived method could augment and complement subjective refraction in challenging cases.

METHODS

Human Subject Recruitment and Protection

The study was reviewed and approved by the University of Houston Institutional Review Board and adhered to the tenets of the Declaration of Helsinki. Prior to data collection, the nature of the study and potential consequences were explained, and subjects reviewed and signed an informed consent document. Individuals were recruited from the University Eye Institute at the University of Houston, College of Optometry. Inclusion criteria included being over 18 years of age with a clinical diagnosis of keratoconus, free of central corneal scarring and any ocular disease other than keratoconus or systemic disease that may influence the visual performance measurements. All individuals that did not meet the inclusion criteria were excluded. Twenty-two individuals were enrolled, and 37 eyes from 20 subjects were included (two subjects did not return for the second study visit and three eyes of three individuals were excluded due to an inability of the wavefront sensor to capture the requisite aberration measurements).

Study Design

Visit 1: Baseline Clinical Measurements

Entering high-contrast visual acuity was recorded monocularly while the subject wore their habitual form of optical correction (Table 1). Corneal topography was recorded using a Pentacam HR Corneal Topographer (Oculus, pentacam.com/us/start/.html). Three measures of autorefraction were taken using a Grand Seiko open-field autorefractor (grandseiko.com/) while the subject fixated a distant target. The instrument-derived average of these three autorefractions was used as the starting point for subjective refraction, except when it was unable to be measured. When autorefractor data could not be recorded, subjective refraction began with the individual’s habitual spectacle refraction when available, or otherwise started from retinoscopy.

Table 1.

Entering optical correction modality, number of eyes with that form of correction and associated median (range) of logMAR visual acuity (VA). GP, gas permeable contact lenses.

Primary Modality of Habitual Refractive Correction	Number of Eyes	Median (range) of VA with Habitual Correction (logMAR unless otherwise stated)
Uncorrected	5	+1.00 (0.46 to counting fingers at 4 feet)
Spectacles	14	+0.14 (0.00 to 1.00)
Soft contact lenses	7	+0.04 (−0.08 to 1.10)
Corneal rigid GP lenses	2	+0.44 (0.22 to 0.66)
Hybrids or piggyback lenses	2	+0.10 (0.10)
Scleral lenses	7	+0.14 (−0.06 to 0.16)

A phoropter-based subjective refraction was performed by one of two clinicians with experience refracting individuals with keratoconus. Best clinical practices were followed by adapting step size during the sphere measurements based on the subject’s ability to detect a difference in the choices being offered. A ± 0.25 D Jackson cross cylinder (the most common step-size used clinically) was used to refine the axis and power of the cylindrical correction. The monocular endpoint providing best acuity was defined as the subjectively derived refraction. The sphero-cylindrical steps viewed during the subjective refraction were recorded, and used to determine whether the objective refraction fell within the refractive space explored subjectively.

Wavefront aberrations were recorded uncorrected using a COAS-HD aberrometer (J&J Vision, jjvision.com/). Accommodation was not relaxed with cycloplegia to limit pupil diameter to physiological pupil sizes, and minimise the influence of optical aberrations that increase with larger non-physiological pupil diameters. To record wavefront aberration data for distance viewing conditions, spectacle lenses were placed in front of the fellow eye during aberration measurement to allow the subject to fixate (as best as possible) a large, distant letter target. The measurement beam aligned to the measured eye was perceptually overlaid with the distant letter viewed by the fellow eye. Room lights were dimmed and time was given to allow the pupil to reach a maximal physiological diameter under low luminance. Three aberration measurements for each eye were averaged over the largest common pupil size and reported using a tenth-order normalised Zernike polynomial, which was seen as a good compromise between capturing the visually relevant aberrations, representing the ocular aberrations faithfully and not measuring a significant number of orders that contain little to no relevant information.¹⁶

Inter-visit Calculation of Objectively Derived Refraction

A custom MATLAB (Mathworks, mathworks.com/) algorithm mathematically applied sphero-cylindrical spectacle prescriptions at a 12 mm vertex distance to each eye’s uncorrected wavefront error. The method sequentially applied every combination of lenses in a standard phoropter: spherical powers from +16.00 to −20.00 D in 0.25 D steps, cylinder powers from 0 to −8.00 D in 0.25 D steps and the cylinder axis was adjusted in 2° steps through 180°. The result was 417,746 unique conditions for each eye, with the result of each condition representing the residual aberrations experienced by the eye with the theoretical sphero-cylindrical spectacle correction in place. The number 417,746 was arrived at by first considering the 145 unique sphere steps, 33 unique cylinder steps and 90 unique axis steps. This results in a total of 430,650 combinations, but duplicates must be removed anytime a zero cylinder value is encountered. A zero cylinder value only needs to be tested once. This results in 89 duplicates for each unique sphere power that is combined with a zero cylinder power, or a total of 145*89 = 12,905 duplicates. If these are removed from the 430,650, there are 417,745 conditions. Since the experiment only allows for even numbered axes, one additional condition (the subjects subjective refraction, which may include an odd axis value) must be tested as well, for a total of 417,746 total conditions.

Each spectacle condition was first vertexed to the pupil plane. The correction was transformed into power vectors (M, J₀, J₄₅) and then into Zernike terms C3, C4 and C5.^11,17 These Zernike terms were added to the wavefront error of the uncorrected eye (simulating the application of the optical correction) before being reduced to the single value metric VSX. The sphero-cylindrical lens that provided the best VSX value was identified and termed the optimised, objective, spectacle prescription for each eye. The VSX metric incorporated a neural contrast sensitivity weighting function of a normal eye, which has been shown¹⁸ capable of relatively ranking dioptric corrections for highly-aberrated eyes, as was done here. The optimised objective and subjective refractions were placed in two otherwise identical trial frames. To keep both the subject and experimenter masked as to the prescription under test, an additional member of the research team (not involved in visit 2 assessment of the subject), prepared the two trial frames.

Visit 2: Evaluation of Subjective and Objective Refractions

Visual performance was measured monocularly using a Metropsis Visual Testing System (Cambridge Research Systems, crsltd.com/) with both the subjectively and objectively derived refractions. The order of testing (right/left eye and subjective/objective refraction) was randomised. The background luminance of the Metropsis screen was 116 cd/m². Adjustments in room lighting were made to match as closely as possible the pupil size experienced during visual performance testing conditions, with the physiological pupil size quantified during wavefront evaluation in Visit 1.

Visual performance data were recorded for high contrast logMAR VA (tested three times and averaged) and letter contrast sensitivity at a fundamental spatial frequency of 4 cycles per degree (tested three times and averaged). VA was tested by presenting the subject with lines of 5 high contrast letters on the Metropsis monitor. Subsequent lines of logarithmically decreasing letter size consistent with ETDRS testing were presented until the termination criterion was reached. For the purposes of this study, changes in VA greater than one line on the logMAR chart were considered clinically significant. CS was tested by presenting groups of 5 letters starting at a log Weber contrast of 0.0, decreasing in log Weber steps of 0.1. Per-letter scoring was used and the termination criteria for both tests were five (not necessarily consecutive) incorrect letters over each individual trial.

Following the objective quantification of spectacle performance, the subject viewed a high contrast logMAR VA chart monocularly and was permitted to switch back and forth between the two corrections before reporting their visual quality preference on a seven-point Likert scale. This was repeated for the right eye and left eye. The subject was then moved out of the laboratory to a window to view an outdoor scene with moving, polychromatic objects at varying distances beyond six metres, and was asked to evaluate their visual quality preference in the same manner. Finally, subjects were asked to provide descriptive comments to convey the criteria they utilised to determine their preferences.

RESULTS:

Sample Characteristics:

Median age of study participants was 31.5 years (range 19–65 years). Distribution of disease severity of the enrolled eyes, as described by the CLEK³ grading scale based on the steepest keratometric meridian (“steeper K”), was: 8 mild, 20 moderate and 9 severely diseased eyes. Median steepest K reading was 47.5 dioptres (D) (range: 40.1 – 66.8 D) and the median pachymetry at the thinnest point of the cornea was 465 microns (range: 371–574 microns). Entering corrections worn by this cohort of subjects are shown in Table 1. Five eyes (three subjects) presented as habitually uncorrected to the study. All habitually uncorrected subjects had unsuccessfully attempted to wear contact lens corrections in the past.

Visual Image Quality with Uncorrected, Subjectively and Objectively Corrected Conditions:

For all 37 eyes where wavefront aberration data were successfully captured, logVSX values with the objective refraction were significantly better than logVSX values for the subjective refraction (−0.97 ± 0.35 vs. −1.65 ± 0.74 respectively, paired t-test, p<0.001) (Figure 1). As expected given the level of HOA, spectacle corrections did not return any sample eyes to the levels of logVSX observed in an optimally-corrected typically sighted sample (hatched box indicates mean ± 2SD for a normal 30–39-year-old eye with a 4mm pupil, which was the average age and pupil size for this sample).¹⁹

Figure 1.

Logarithm of the visual Strehl ratio (LogVSX) for the uncorrected, subjectively corrected and objectively corrected conditions. The box plot shows the 25th, 50th and 75% percentiles. Whiskers indicate the 10th and 90^th percentiles. The mean ± 2 standard deviation range for typically best-corrected logVSX for 30–39-year-old eyes with a 4 mm pupil (age and pupil matched for this sample) was −0.250 ± 0.104 (hatched box). Best mean visual image quality was provided by the objective correction. As anticipated for this population wearing a spectacle correction, no eye reached typical levels of logVSX. The figure omits one individual from the uncorrected column who had an uncorrected VSX value of 0.0, due to the complication of plotting log(0).

Difference in Refractive Prescription between Subjective and Objective Refractions:

One method which has been utilised in the literature to assess the difference between two refraction errors is the dioptric difference⁶ separating the two refractions, calculated as the Euclidean distance between refractions in power vector space¹⁷ (Figure 2). The median dioptric difference from the subjective to objective refraction was 2.77 D (range= 0.21 to 20.44 D, 1^st quartile=1.02 D, 3^rd quartile= 4.36 D).

Figure 2:

Euclidean dioptric difference between autorefraction (auto), subjective refraction (subj) and metric-derived objective refraction (obj) in power vector space. Larger differences indicates that the refractions were more different. The box plot shows 25th, 50th, and 75% percentiles. Whiskers indicate 10th and 90^th percentiles.

The dioptric differences displayed in Figure 2 distils the difference between the refractions into a single number. While simple, this technique masks the unique spherical and cylindrical contributions to that difference. In considering the differences between the subjective and objective refraction, it is important to consider whether the variability in the two approaches was predominantly from the spherical or cylindrical components. When broken into M and J (which is itself a vector of J₀ and J₄₅) the contributions from M (Q1: 0.625 D; M: 1.75 D; Q3: 4.00 D) and J (Q1: 0.53 D; M: 1.73 D Q3: 2.67 D) were shown to be similar in magnitude.

Visual Performance with Subjectively and Objectively Derived Refractions:

As shown in Figure 3A, 68% of eyes had better monocular VA with the objective refraction and 32% of eyes gained more than one line of acuity. Average VA for all eyes with the objective refraction was 0.36 logMAR. For reference, typical levels of VA for this age range is −0.14±0.07. All three subjects with only one eye included in the dataset benefitted from the use of the objective refraction. Twenty-three eyes (62%) had better CS with the objective refraction (Figure 3B).

Figure 3: Gain or loss in (A) logMAR visual acuity (VA) and (B) letter contrast sensitivity (CS).

All eyes are plotted; subjects (when both eyes were included, then the eyes are paired vertically on the chart) are ranked from left to right based on the most to least improvement in monocular logMAR VA or CS with the objective refraction. Repeatability of VA was all within 2.5 letters; error bars are not plotted to simplify the figure. Solid circles represent the eye with the most gain or least loss with the objective compared to the subjective refraction. Open circles represent that gain or loss for the fellow eye.

Patient Preference between Subjectively and Objectively Derived Refractions:

Monocularly, the objective refraction was preferred by 68% of eyes when viewing a distant acuity chart (Figure 4A) and by 76% when viewing a distant, outdoor, colourful dynamic scene (Figure 4B).

Figure 4: Monocular subject preference when (A) viewing a static indoor letter chart and (B) when viewing a dynamic outdoor scene.

Data were evaluated on a 7-point Likert scale. The objective refraction was preferred by 68% eyes when viewing a letter chart indoors and by 76% of eyes when viewing an outdoor, dynamic scene.

The words used to describe the factors that drove each subjects’ qualitative preference, which include both positive and negative descriptors, are presented in Table 2 based on the relative frequency of their use.

Table 2.

Qualitative descriptors reported by subjects as driving their preference either for (positive connotation) or against (negative connotation) either the objective or subjective refraction.

Positive Connotation		Negative Connotation
Word	Count	Word	Count
balanced	1	bent	1
better	1	blurry	1
comfortable	1	crowded	1
crisp	1	distant	1
magnified	1	strained	1
natural	1	streaked	1
straight	1	tilted	2
sharp	6	shadow	3
clear	11	distorted	4
		double	4

No limit was placed on the number of words an individual could provide. For descriptors with a positive connotation, “clear” appeared most frequently (11 times) followed by “sharp” (6 times). For descriptors with a negative connotation, “double” (4 times), “distorted” (4 times) and “shadow” (3 times) appeared most frequently. Thirteen words appeared only once.

DISCUSSION:

This study was designed to gauge the potential value of an objective method of spectacle refraction based on a visual image quality metric for eyes with keratoconus. By searching the entire dioptric correction space, the objective method offers an additional tool to identify a spectacle correction for an eye that more frequently provided benefit both subjectively and objectively. Because the search space for the objective refraction (the entire phoropter) included the subjective refraction within that space, the objective refraction could always equal or improve upon the logVSX value obtained by the subjective refraction. This is one advantage of the objective method, as it is able to compare, in this case, almost half a million refractions for an individual eye.

The cohort of subjects enrolled in this study included a higher percentage of spectacle wearers than has been reported for individuals with keratoconus: 16% in the CLEK study as compared to 38% in this sample.³ Because many of the subjects either presented wearing spectacles or had a pair of backup spectacles that were worn part-time, many subjects were already adapted to a sphero-cylindrical refraction, and because the visual system is strongly adapted to its habitual aberrations^21,22 we examined whether this could have factored into their subjective refraction, guiding it closer to their habitual refraction. However, this was not the case here. Of the seven subjects that entered the study wearing spectacles, a monocular preference was expressed for the objective refraction in 11 of 14 (78.6%) eyes when looking out a window, despite not having time to adapt to the new prescription.

The correction determined from the subjective refraction was most similar to the refractions determined through autorefraction, and greater differences existed between the subjective and metric-derived findings (Figure 2). This is likely because autorefraction was the starting point of the subjective refraction. Measuring visual performance with the autorefraction correction would have added to the already-lengthy experimental procedure and therefore was not performed.

Clinically, it would be convenient to have a method of easily identifying which individuals would benefit most from this additional step of a wavefront-derived objective refraction. However, no parameters including entering VA, corneal curvature, change in logVSX, pachymetry, any corneal topography parameters or the severity of the disease correlated with the change in VA or the subject’s preference. The low correlation of change in logVSX with VA and preference was due to the predictive models utilised here being good at predicting performance of a cohort, but less accurate in their prediction for an individual. This can be seen in Figures 1 and 3, where the group as a whole benefitted from the objective method, while some individuals continued to prefer subjectively-derived corrections.

Under the current optimisation paradigm, cylinder axes were incremented in 2° steps, regardless of cylinder power. It is known that as the cylinder power increases, then the tolerance to rotation decreases. Therefore, the algorithm could be optimised to include an axis step size that results in a uniform change in cylinder power, rather than a fixed number of degrees. This would reduce the computational intensity of the algorithm by reducing the unnecessarily close spacing at low cylinder powers, which likely provide little to no visual benefit. For instance, a sampling pattern that was homogeneous, with nearest-neighbour spacing of 0.125D could be employed, thereby reducing the number of locations searched with little loss in the ability to identify the optimal sphero-cylindrical correction.

Examination of individual cases provides an opportunity to understand the metric better, as well as its ability to predict the optimal refraction for an individual. As an example, one subject reported that severe doubling with the objective refraction drove their choice in preference towards the subjective refraction, explaining that the doubled letter was unacceptable, making that refraction intolerable. Because the VSX metric identified the dioptric correction that maximised retinal image contrast, it also made the ghost image much more prominent, of higher contrast and therefore more visually disturbing to the patient. Several additional factors may be at play in cases where the optimised refraction failed, including neural adaptation ^5,23 (our preference data were collected without allowing time for adaption).

In addition to cases where individuals performed better with the objective refraction, but preferred the subjective refraction, there were cases (Figure 3) where individuals performed worse with the objective refraction in both eyes. This is not unexpected, for the same reasons described above: the metric is good in aggregate, but exhibits variance across individual predictions.

Similar work by Jinabhai et al.²⁴ examined the benefit of an objective refraction for individuals with keratoconus. While the methods differed, the results of this study agreed with this prior work in finding that spectacles derived by subjective and objective means were quite different. However, the results of the studies differ in that they found improved visual performance with the subjective refraction and did not assess patient-perceived preference.

Several limitations exist in the present study. First, the objective refraction is dependent on the capture of good fidelity wavefront aberration measurements. When considering clinical deployment of this method, more highly aberrated eyes will be more difficult to measure on common clinical wavefront sensors designed for eyes with typical aberration levels. This clinical problem could be addressed by deploying wavefront sensors that are specifically designed to meet the dynamic range needs of a population with highly aberrated eyes. Additionally, because neither of these refractions were dispensed to the subjects (all visual performance tests were performed acutely), no adaptation time was provided, making it difficult to capture the real-world acceptability of the objective refraction.

VSX was optimised to find the “best” objective refraction. However, a change in VSX does not perfectly represent the change in perceived visual percept that is observed by a subject. For instance, examination of the negative connotation words in Table 2 shows that “doubling” was tied for the most common reason that the subject was unhappy with a refraction, and doubling was also a common complaint with VSX-optimised refractions. Future optimisation algorithms may incorporate one of several metrics where a change in the metric value has been shown to correlate with a change in VA.¹² Alternatively, perhaps several metrics could be combined such that the distinct attributes tested by the individual metric increase the predictive power of the combined metric. In this study, a phoropter was used to perform the subjective refractions. Additionally, the Jackson cross cylinder (JCC) was limited to ±0.25D. Use of a larger dioptric magnitude of JCC could have been used initially to explore a wider span of the dioptric space more easily. Larger JCCs could have been replaced with a smaller powered version as the refraction proceeded.

This study was completed in two visits due to the time needed in between to allow computation of the logVSX values for each of the 417,746 refractive combinations tested (~13 hours of calculation time). Future work should focus on reducing the search space for the objective refraction and speeding up computation time, thereby making the method clinically feasible.

Finally, this study did not consider balancing the eyes to work together binocularly, which is a critical step in prescribing a set of spectacles that a patient would accept for everyday wear. Keratoconus often presents asymmetrically, resulting in widely different refractions between eyes, which can be difficult to tolerate with spectacle lenses and must be accounted for clinically.²⁵ Given the current state of available technology, the objective, metric-derived method described here might be best used as a starting point for subjective refraction for individuals with keratoconus. Perhaps the optimisation process could be used objectively to identify several locations in the refractive space to explore using subjective refraction techniques.

CONCLUSIONS:

This study demonstrates that an objective, VSX-optimised wavefront refraction can provide monocular visual performance that is superior to that achieved with subjective refraction in eyes with keratoconus. The benefit was seen in terms of subject preference, visual acuity and contrast sensitivity. These findings support the long-term goal of adding a wavefront-derived objective refraction to current clinical procedures for individuals with highly aberrated eyes.

Key Points:

• This study was designed to gauge the potential value of an objective method of spectacle refraction based on a visual image quality metric for eyes with keratoconus.

• By searching the entire dioptric correction space with a visual image quality metric that is predictive of change in visual performance, the objective method identified the global maximum metric value and the associated monocular spectacle correction that, more-times-than-not, provided benefit and was preferred by patients.

• Given the current state of available technology, the objective, metric-derived method could be used as a starting point for subjective refraction and to identify whether multiple disparate dioptric areas of good visual quality exist for eyes with keratoconus.