August 2019
Volume 60, Issue 11
Open Access
ARVO Imaging in the Eye Conference Abstract  |   August 2019
Designing a Hyperspectral Autofluorescence (AF) Camera for Early Detection of Age-related Macular Degeneration (AMD)
Author Affiliations & Notes
  • R Theodore Theodore Smith
    Ophthalmology, Icahn School of Medicine of Mount Sinai, New York, New York, United States
  • Neel Dey
    Computer Science, New York University, New York, New York, United States
  • Thomas Ach
    Ophthalmology, University of Wurzburg, Wurzburg, Germany
  • Christine Curcio
    Ophthalmology, University of Alabama Birmingham, Birmingham, Alabama, United States
  • Footnotes
    Commercial Relationships   R Theodore Smith, None; Neel Dey, None; Thomas Ach, None; Christine Curcio, None
  • Footnotes
    Support  R01 EY027948, NEI
Investigative Ophthalmology & Visual Science August 2019, Vol.60, PB0185. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      R Theodore Theodore Smith, Neel Dey, Thomas Ach, Christine Curcio; Designing a Hyperspectral Autofluorescence (AF) Camera for Early Detection of Age-related Macular Degeneration (AMD). Invest. Ophthalmol. Vis. Sci. 2019;60(11):PB0185.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

Purpose : Optimize excitation wavelengths of a proposed clinical hyperspectral AF camera for early detection of AMD. Soft drusen and basal linear deposit (BLinD) are the lipid rich material of the Oil Spill on Bruch’s membrane (BrM) of early AMD. Drusen are focal and recognizable clinically. BLinD is thin, diffuse, and invisible clinically, even on high resolution optical coherence tomography (OCT), but is detectable ex vivo on hyperspectral AF imaging. Optimal spectral excitations, however, are yet unknown.

Methods : 20 tissues, retinal pigment epithelium (RPE)/BrM flatmounts from AMD donors underwent hyperspectral AF imaging with 4 excitation wavelengths (436, 450, 480 and 505 nm), and the resulting image cubes were simultaneously decomposed with non-negative tensor factorization (NTF), an extension of non-negative matrix factorization. NTF results are essentially unique and robust to spectral initializations, up to rotations and scale factors, and also deliver the excitation spectra for the fluorophore sources of the recovered emissions, that is, the most efficient wavelengths among those tested for excitation of given fluorophores. 2 basic emission spectra were recovered for a clear separation: a total RPE spectrum (lipofuscin (LF)) and a subRPE spectrum (drusen and BLinD).

Results : A composite emission spectrum for drusen and BlinD, the SDr spectrum, was consistently recovered with peak at 520 nm (Fig, upper left, blue spectrum.) The total LF spectrum was also consistent, peaking at about 570 nm, with secondary peaks or shoulders at 600 and 650 nm (Fig, upper left, red spectrum). SDr localized to drusen and subRPE deposits (Fig, lower left), the LF spectrum to the LF compartment in the RPE (Fig, lower right) with histopathologic sensitivity and specificity.
Remarkably, the excitation spectra for both SDr and LF peaked at 450 nm in all 20 samples (Fig, upper right, blue and red spectra). Thus, 450 nm not only optimally excited SDr, it was also more efficient than the classic 480 nm for exciting lipofuscin AF. Dual excitations at ~450 and ~480 nm were the most efficient and reliable pair for detection of early AMD lesions.

Conclusions : Classic 488 nm excitation, with additional 450 nm excitation. a hyperspectral AF detector, and suitable image analysis, should be capable of clinical detection and quantification of drusen and BLinD in early AMD.

This abstract was presented at the 2019 ARVO Imaging in the Eye Conference, held in Vancouver, Canada, April 26-27, 2019.

 

×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×