Improving Accuracy of Refraction Derived form Wavefront Aberrometer Measurement

Abstract

Abstract: : Purpose:To evaluate the clinical efficacy of refraction provided by a grating–based wavefront aberrometer, (i) using only second order Zernikes and (ii) optimizing the refraction value using correlation analysis including wavefront aberrations up to 6^th order and patient’s parameters. Methods:We measured the refractive errors of 100 patients with ages ranging from 8 to 81 years using a grating–based wavefront system (Z–View, Ophthonix, Inc., San Diego, CA), and compare with the manifest refraction using a phoropter. Measurements were made for both right and left eyes without cycloplegia. All patients were independently examined, wavefront by a technician and manifest by an O.D. The O.D was masked to the patients’ spectacle prescriptions and to all wavefront results. Data were analyzed in terms of spherical equivalent (SE). The refraction from the wavefront result was derived using: (i) only the second order Zernikes, and (ii) all Zernike polynomials up to the 6^th order, and certain relevant patient parameters, including pupil size, and age. The manifest refraction was treated as the true refraction. Results:The difference between Z–View wavefront and manifest SE has an approximately normal distribution. The mean difference for 100 left and 100 right eyes was 0.065 D with the standard deviation of 0.4 D. Based only on second order Zernikes, a binomial statistical test showed that the 80.1% of cases in which the difference between the Z–View and manifest SE was within ± 0.5 D. A complex regression analysis was performed using the manifest SE as dependent variable, wavefront SE, pupil diameter, age, high order rms, total rms, and Zernike coefficients up to 6^th orders as independent variables. We also divided the patients into three groups (myopes: SE < –0.5 D, emmetropes: –0.5 D <= SE <= 0.5 D, and hyperopes: SE > 0.5 D), and derived equations that predicted with 93% probability that the difference between the Z–View and manifest SE is within ± 0.5 D. A paired T–test showed no significant difference between the Z–View and manifest refractions (–1.6 ± 3.1; –1.6 ± 2.8; t = 0.073; p = 0.94). Conclusions:In spite of some subjective variability within manifest refraction, refraction using only the second order Zernike polynomials of the Z–View aberrometer correlate well with the manifest refraction. Using regression model including higher order Zernikes and patient parameters, the derived equations improve the prediction of patient’s manifest refraction within ± 0.5 D with a 93% probability.

Keywords: refraction • refractive surgery • visual acuity