Purchase this article with an account.
S.R. Russell, A.H. Durukan, H.T. Davis, S.C. Nemeth, G. Zamora, P. Fournier, B.K. Taylor, J.M. Reinhardt, M.D. Abramoff; Spectral Imaging Demonstrates Principal Wavelength Components That Differentiate Stargardt Disease, Age–Related Macular Degeneration (AMD) and Non–AMD Drusen–Associated Disorders . Invest. Ophthalmol. Vis. Sci. 2006;47(13):5808.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
To evaluate the reflectance spectral components that differentate the fundi of Stargardt disease, age–related macular degeneration (AMD), non–AMD drusen–associated disorders and age–matched normal subjects (AN)
Thirty–five subjects (15 with fundus disorders and 20 normals) were evaluated with a high spectral resolution (hyperspectral) fundus imager. Each hyperspectral image consisted of 512 interferograms in a linear array, each with a footprint 150 by 500 µm. Interferograms were converted to spectra by Fourier transform, yielding 249 wavelength amplitudes between 500 and 700 nm. Simultaneous monochromatic fundus images were obtained that provided registration to fundus structures. For analysis of funduscopically pathologic regions in Stargardt disease and drusen–related disorders, interferograms were selected that corresponded to the center of white lesions. Principal spectral components (PSC) with greatest correlation to disease status were determined by a partial least squares (PLS) method. Source of variation and PLS were verified utilizing a nested analysis of variance within subjects, between subjects and between disease states.
Each interferogram demonstrated an intra–class correlation of 91 to 95%. For Stargardt disease, funduscopically unaffected macula demonstrated a significant spectral difference from AN (P<0.05). PLS demonstrated PSC with maxima at 681 and 651 nm, and minima at 640 and 659 nm. For non–AMD drusen disorders, funduscopically unaffected macular regions also showed statistically significant differences from AN (P<0.004) with a PSC maximum at 670 nm. Comparison of white lesions of Stargardt disease and non–AMD drusen, and between AMD and non–AMD drusen, demonstrated statistically significant spectral differences (P<0.0001 and P<0.0003, respectively).
Spectral imaging demonstrates structure– and lesion– associated spectral signatures that can reliable differentiate disease states, even from funduscopically unaffected macular regions. Phenotypic spectral differences that separate drusen of AMD from those of non–AMD disorders may be related to age, stage or disorder.
This PDF is available to Subscribers Only