American Journal of Ophthalmology
Volume 141, Issue 1 , January 2006, Page 1
Kuang-Mon Ashley Tuan OD, PhDa, Dimitri Chernyak PhDa and Sandy T. Feldman MD, MSb, ,
a) VISX Inc, Santa Clara, California
b) ClearView Eye and Laser Medical Center, San Diego, California.
Purpose: To evaluate the accuracy of the diagnostic capabilities of optical metrics generated from wavefront measurements in relationship to post–laser-assisted in situ keratomileusis (LASIK) visual complaints as expressed and drawn by patients.
Design: Retrospective analysis and observational case series.
Methods: Patient wavefront data from an investigational device exemption study for wavefront-guided ablations were used to derive normative modulation transfer function (MTF), encircled energy (EE), and Strehl ratio. These optical metrics and their point-spread functions (PSF) were compared with data from five postoperative patients with night vision complaints. Patients were asked to draw their symptoms, which were elicited by testing with a Fenthoff muscle light, while using their best-corrected distance vision.
Results: The MTF, EE, and Strehl ratio of most patients were markedly different from those of the averages of 208 normal myopic eyes before and after LASIK surgery. The spatial extent of the PSF correlated positively with the severity of the visual complaints. Wavefront-derived PSFs were markedly similar to the patients’ drawings.
Conclusions: The results of this study demonstrate the diagnostic capability of the wavefront system in predicting visual symptoms and complaints of patients with high-order aberrations. Objective visual metrics from patients with night vision complaints were different from those of normal myopic eyes that had undergone LASIK procedures.
Excerpts: If the human eye were a perfect optical system, the appearance of a point source of light would be limited only by diffraction effects, and the source would appear as a single point to the observer. In addition to spherocylindrical components, the optical system of the human eye generates other complex optical aberrations, which contribute to the distortion of retinal images and determine the quality of the image formed on the retina. The optical imperfections of the eye cause rays of light traveling from a point source through the eye’s optics to intercept the retina at different locations, thereby blurring the appearance of the point. The distorted appearance of the point on the retina is called a point-spread function (PSF).
The recent development of wavefront aberrometers for use in ophthalmology has given clinicians an objective measurement method for optical aberrations other than sphere and cylinder. The wavefront maps generated by such devices represent deviations from the ideal diffraction-limited optical systems and enable the physician to precisely diagnose visual impairment. In addition to guiding customized refractive surgery, wavefront sensors can be used to evaluate the source of visual complaints. The appearance of the PSF can be computed directly from the wavefront measurement and corroborated by the patient with a simple drawing.
In recent years, much emphasis has been placed on the use of PSF and PSF-based metrics to objectively evaluate the quality of vision (Gross E, Wavefront Congress 2004, Abstract).3 and 4 Whereas the wavefront map (which is analytically defined by methods that include the Zernike coefficients) describes the aberrations in the pupillary plane, the PSF describes aberrations in the retinal plane. Retina-based descriptions of aberrations have been shown to correlate strongly with subjective visual metrics, such as visual acuity and contrast sensitivity.
In addition to correlating with subjective measures of visual performance, the PSFs are valuable tools for understanding visual symptoms, such as ghost images, which occur with multiple peaks of the PSF, or halos and starbursts, which may also be inferred from the appearance of the PSF. The purpose of this study is to determine whether quantitative analysis of PSFs is helpful to the clinician’s understanding of nighttime visual complaints.
Methods: This study was a retrospective analysis and observational case series. Two groups of patients were included in this study. The first group of eyes were nonsymptomatic and were used to establish normative values. Data from this group were gathered from 208 eyes from the six clinical sites in the United States that participated in the Institutional Review Board–approved investigative device exemption study for the CustomVue wavefront-guided laser vision correction procedure (VISX Inc, Santa Clara, California, USA). The second group of patients, all of whom had night complaints after laser-assisted in situ keratomileusis (LASIK), was recruited from the practice of ClearView Eye Center, San Diego, California, USA.
The purpose of this part of the study was to see whether sketches of the point source of light made by patients with night vision complaints would be reasonably similar to the polychromatic PSF plot created by the WaveScan aberrometer, and to determine whether their optical metrics were different from the normal population. Inclusion criteria included a complaint of glare, halo, starburst at night, or a poorer quality of vision. Patients were also required to be willing to sketch their symptoms.
WaveScan version 3.5 software was used to convert the wavefront data into PSF. The PSF was computed by using multiple wavelengths to accurately simulate the appearance of white light and to account for chromatic aberrations of the eye. The Stiles-Crawford effect of the first kind, which depicts the probability of photon absorption by photoreceptors as a function of ray position from the pupil center, was also taken into account. Additionally, absorption spectra of the three retinal photoreceptor types were used to correctly weight different wavelengths in the PSF computations. Finally, spacing of the retinal receptor mosaic was used to determine the maximum spatial frequency that is used in the rendition of the PSF.
An important issue in assessing retinal image quality is a realistic computation of the PSF. The above method enables the instrument to create realistic renditions of the appearance of a white light point source on the image plane, which is the polychromatic PSF. An optical system with little aberration will generate a compact and high-intensity point source on the image plane. Therefore, the brain would be expected to perceive sharp images and resolve small details. A highly aberrated eye will form a PSF that spreads out and has lower contrast within the retinal plane. In this case, the brain would perceive a dim and blurry image resulting in low resolution and poor visual performance.
To effectively compare the visual impact of different PSFs, it is necessary to quantify PSF or PSF plots. To date, there is no consensus in the vision science community as to a standardized method for quantifying the visual impact of measured wavefront aberration. However, Strehl ratio, encircled energy (EE), and modulation transfer function (MTF) are the commonly used optical metrics believed to provide useful information on perceptual image quality and therefore predict visual performance.3, 4, 7 and 8 For this study, the PSF was broken down into two parts for visual evaluation: circle of blur; and the length of the vertical and horizontal strikes from the center. Optical metrics such as Strehl ratio, EE, and MTF were derived from the PSF8 with the intention of quantifying the visual impact of the PSF.
Strehl ratio is defined as the ratio of the peak intensities of the aberrated PSF and the diffraction-limited PSF. EE is the two-dimensional integral of the PSF. EE represents the proportion within a given radius of the image center and the total energy from a point source. EE was calculated at 1, 3, 5, 8, and 10 min of arc. MTF is the module of Fourier transform of the PSF. MTF represents contrast information, and it varies with spatial frequencies. MTF was calculated from 3 cycles per degree (cpd) to 30 cpd for this study. An MTF of 1 indicates that 100% contrast was maintained after the light passed through the optical system.
The normative value of the optical metrics described above was used for comparison with patients with postoperative visual symptoms. Lower-order wavefront information was deleted from the calculation to remove optical degradation arising from refractive error.
Results: All patients were healthy men ranging in age from 31 to 43. Patient demographic information is listed in Table 1. Their drawings and corresponding PSFs are presented in Figure 1. The postoperative follow-up period ranged from 6 months to 3 years. Postoperative uncorrected vision ranged from 20/30 to 20/20 and was correctable in each eye to 20/20 or better. However, patients reported visual symptoms at follow-up visits. Unfortunately, not all of the patients’ preoperative visual information was available.
Discussion: It is well known that patients with nighttime vision complaints may nevertheless have excellent uncorrected acuity. The clinician does not have many clinical tests outside of questionnaires to assess the extent of symptoms such as glare, halo, and starburst. Additionally, literature on clinical methods with which to demonstrate night complaints or quantify the visual disturbance is scarce. In this study, we have demonstrated that optical metrics derived from polychromatic PSF are a reliable quantitative method for the validation of night vision complaints.
The PSF diagram represents the appearance of one small spot of light after it passes through an optical system. The authors found the WaveScan-generated PSF to be consistently similar to the subjective visual experience of the study subjects despite differences in the artistic abilities of the patients. The polychromatic PSF plots derived from the wavefront measurements bear marked similarities to the drawings of the patients. The PSF plots show diffraction effects and some subthreshold light. Diffraction-based spatial frequencies are missing from drawings because they are not perceived. Additionally, the spacing of the photoreceptors limits visual acuity and spatial resolution of the retina. Occupation, personality, and level of detail orientation also appear to help determine the way patients perceive their visual quality. Therefore, patients with similar PSFs can have different reactions to their quality of vision. Nevertheless, the similarity between the PSF maps of subjects who have visual symptoms and their drawings are evident.
Optical metrics generated from the PSF were used to create a quantitative description of a PSF diagram. Optical metrics apply summations from different areas and orientations, and the result is a directionless quantitative number. The authors chose to use three metrics: Strehl ratio, EE, which signifies the intensity of the brightness of the image, and MTF, which represents contrast information. The study patients were compared with the normative values of EE and MTF generated from a cohort that had never undergone refractive surgery. Most optical metrics in the study patients were significantly worse than the normative values.
Lowered EE and MTF in the left eye of patient 4 are marginally significant, and there is a high amount of negative spherical aberration. The PSF maps are more vertically distributed. After averaging in all directions, the effect on overall optical metrics was reduced, which may explain why patient 4 had both positive visual symptoms but was within the normal limits of EE and MTF. However, patient 4’s PSF map shows an asymmetrical pattern in which the vertical dimension spreads outside the normal range. This extensive vertical spread of light accounts for the patient’s visual symptoms and correlates with his drawings.
After examining the visual functions of seven patients who underwent LASIK procedures, Holladay and coauthors hypothesized that corneas with an oblate shape or positive spherical aberration contributed to the decrease of visual functions. In our study patients, patients 1 and 4 had significant negative spherical aberration or prolate cornea (−0.41 μm and −0.25 μm, respectively) but also had visual symptoms. Spherical aberration was one of the few “abnormal” optical metrics for patient 4. This observation suggests that negative spherical aberration is not necessarily associated with visual satisfaction. On the contrary, when spherical aberration is present in sufficient magnitude, it could cause visual dissatisfaction.
In our study patients, some patients’ uncorrected visual acuity nevertheless experienced persistent visual symptoms, even with their best-corrected vision.
Polychromatic PSF plots generated from VISX’s polychromatic wavefront algorithms showed high correlation with the patients’ subjective visual experiences.
In conclusion, this study shows that wavefront systems are capable of predicting nighttime visual symptoms and complaints of patients whose visual acuities are otherwise good. Polychromatic PSF and the optical metrics derived from it are valuable diagnostic tools for predicting quality of vision and for evaluating the optical quality of an eye.