Hyperspectral imaging of retinal microvascular anatomy

Abstract

Background: Hyperspectral image processing has been applied to many aspects of astronomical and earth science research. Furthermore, advances in computed tomographic imaging spectroscopy and diffraction grating design have allowed biological applications for non-invasive tissue analysis. Herein, we describe a hyperspectral computed tomographic imaging spectroscope (HCTIS) that provides high spatial, spectral and temporal resolution ideal for imaging biological tissue in vivo. Methods: We demonstrate proof-of-principle application of the HCTIS by imaging and mapping the microvascular anatomy of the retina of a model organism (rabbit) in vivo. The imaging procedure allows rapid and dense spectral sampling, is non-toxic, non-invasive, and easily adaptable to a commercially available fundus camera system. Results: HCTIS provides highly co-registered temporal, spatial and spectral data with resolution capable of reconstructing the fine vascular tree of the rabbit retina in vivo. Conclusions: We show that HCTIS allows for reliable and reproducible tissue classification and detection using signature discriminant analysis. Future applications of this system may provide promising diagnostic methods for diseases of many tissues.