Background: To find the correlations and differences between the nerve fiber layer parameters of a GDx (polarimetric retinal nerve fiber analyzer) in single high quality images and those in the mean image from three high quality images.
Methods: The results of scanning laser polarimetry were selected from 63 eyes of 63 patients (31 male and 32 female). Complete ophthalmic examinations included slit-lamp biomicroscopy, dilated pupil for vitreous and ocular fundus examination, applanation tonometry, and subjective or objective refraction. There were no significant ophthalmic disorders other than glaucoma and mild lens opacity. Nerve fiber layer thickness of each eye was measured with GDx and at least three high quality images (passing the software's quality criteria) were obtained. The best of these three high quality images was selected as a single high quality image (Group One). The mean image (Group Two) was calculated from the same three high quality images.
Results: We found that there was a high correlation in all nerve fiber layer parameters between Group One and Group Two. There was no significant difference in the seven relative values of nerve fiber layer parameters (Number, Ellipse Modulation, Symmetry, Superior Ratio, Inferior Ratio, Superior/Nasal and Maximal Modulation) between Group One and Group Two. Group One had significantly higher values compared with Group Two in the other seven absolute values of nerve fiber layer parameters (Average Thickness, Ellipse Average, Superior Average, Inferior Average, Superior Integral, Superior Maximal and Inferior Maximal).
Conclusion: It is reasonable to take a mean image from three good quality images from a cooperative patient. If only one high quality image can be obtained in repeated acquisition of GDx, the seven relative values of nerve fiber layer parameters in this high quality image can be used as a base-line image for detecting retinal nerve fiber layer defects and for determining changes of retinal nerve fiber layer thickness in sequential images of GDx.