Abstract
Purpose::
Atypical birefringence patterns (ABP) are artifactous retinal nerve fiber layer thickness patterns occasionally seen with scanning laser polarimetry (SLP). In this study we set out to identify which clinical correlations might be associated with the occurrence of these atypical patterns.
Methods::
Sixty-one subjects (glaucoma patients, glaucoma suspects and normal individuals) underwent a full ophthalmic eye examination, standard visual field (VF) test and GDx variable cornea compensation (GDx-VCC) SLP examination. One eye per patient was selected for this study. The ABP magnitude was measured in two independent ways: using a built-in support vector machine analysis (typical scan score, TSS), and by a masked experienced observer. We correlated the magnitude of ABP with: gender, age, refractive state of the eye, corneal polarization axis, corneal polarization magnitude, GDx-VCC global parameters (TSNIT and NFI), and the amount of damage seen in the VF, as evident from: the glaucoma hemifield test (GHT), mean deviation (MD) and pattern standard deviation (PSD).
Results::
Of the 61 study eyes, 27 (44%) showed an ABP, based on a TSS cut-off <82.5). A very high correlation was found between the TSS score and the masked experienced observer (r-square = 0.80; p-value < 0.001). The following were the only clinical parameters found, on bivariate analysis, to be significantly correlated with the presence of an ABP: age (r-square = 0.086; p-value = 0.02); corneal polarization magnitude (r-square = 0.069; p-value = 0.04); TSNIT (r-square = 0.16; p-value < 0.001).
Conclusions::
A correlation was not found between the presence and magnitude of ABP and most clinical parameters evaluated in this study. A low, but statistically significant, correlation was found for age and corneal polarization magnitude (r-square = 0.086 and 0.069, respectively). A low-medium correlation was found for TSNIT (r-square = 0.16), however, we speculate that this might represent a confounding effect, rather than an underlying association. We conclude that none of the clinical parameters investigated in this study appear to be strongly correlated with the presence of an ABP on SLP scans performed using the commercially available GDx-VCC.
Keywords: imaging/image analysis: clinical • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound)