Exudative lesions in age–related macular degeneration: comparison of fluorescein angiograms and polarimetry images

Abstract

Abstract: : Purpose:To perform a blind comparison of digital fluorescein angiograms (FA) and novel types of polarimetry images for the detection and localization of exudative lesions in patients with age–related macular degeneration (AMD). Methods: Two authors, who were not graders, selected 19 eyes with exudative AMD with good quality FA’s and neovascular membranes < 15 deg, the size of polarimetry images. Early and late phase FA’s and 3 types of images computed from polarimetry data were graded from a computer screen. Noninvasive polarimetry images were obtained from raw data using a GDx ( LDT) and in–house software (Matlab, Mathworks). In the GDx, returning light is collected by two detectors: a parallel polarized and a crossed detector. We computed a depolarized image (minimum of the crossed detector), a birefringence image (modulation of the crossed detector), and a confocal image (mean of both detectors). Two experienced ophthalmologists, blinded to the patients’ clinical status, graded each image separately and in random order. In each image, graders rated the probability of presence of choroidal neovascularization (CNV) and/or leakage (scale from 1–5). The grades for presence of CNV and fluid in were compared using nonparametric statistics (Kruskal–Wallis for all images, Mann–Whitney for paired comparisons). Results:The ratings for all the image types differed significantly for both the CNV and fluid scores (p < 0.0001 for both). For detection of CNV the rank is in the following order, (in decreasing probability of CNV): early phase FA > confocal image > depolarized > birefringent. For the fluid grading, the order was confocal image> late phase FA= depolarized > birefringent image. For CNV, the FA was significantly better than the next best image type (p<.0019). For fluid, the FA was not significantly better than the depolarized image and significantly worse than the confocal image (p > .7, p < .0114, respectively). Conclusions: Experienced clinicians prefer early phase FA for detection of CNV, when this is of good quality, but for the detection of the fluid component a noninvasive method was rated better.

Keywords: age–related macular degeneration • imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • choroid: neovascularization