April 2014
Volume 55, Issue 13
Free
ARVO Annual Meeting Abstract  |   April 2014
Visualizing retinal vasculature using non-confocal adaptive optics scanning light ophthalmoscopy
Author Affiliations & Notes
  • Yusufu N B Sulai
    The Institute of Optics, University of Rochester, Rochester, NY
  • Drew H Scoles
    Biomedical Engineering, University of Rochester, Rochester, NY, NY
  • Alfredo Dubra
    Ophthalmology, Medical College of Wisconsin, Milwaukee, WI
    Biophysics, Medical College of Wisconsin, Milwaukee, WI
  • Footnotes
    Commercial Relationships Yusufu Sulai, None; Drew Scoles, None; Alfredo Dubra, Canon USA Inc. (C), US Patent 8,226,236 (P)
  • Footnotes
    Support None
Investigative Ophthalmology & Visual Science April 2014, Vol.55, 1656. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Yusufu N B Sulai, Drew H Scoles, Alfredo Dubra; Visualizing retinal vasculature using non-confocal adaptive optics scanning light ophthalmoscopy. Invest. Ophthalmol. Vis. Sci. 2014;55(13):1656.

      Download citation file:


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

      ×
  • Supplements
Abstract
 
Purpose
 

To explore non-invasive imaging of the retinal vascular structure and perfusion by non-confocal detection in adaptive optics scanning light ophthalmoscopy (AOSLO).

 
Methods
 

Five detection methods were tested by placing different spatial filters and/or light detectors in retinal conjugate planes: circular mask, annular mask, circular mask with filament, knife edge and split-detection. The dimensions and geometry of the detection apertures were varied and the effects of illumination pupil apodization, polarized detection and four different illumination wavelengths (500, 600, 680 and 790 nm) were also evaluated. A side by side comparison of all the detection schemes was performed at identical foci and retinal locations, using the signal-to-noise ratio (SNR) along capillary cross-sections as performance metrics, in both image sequence averages (structure maps) and standard deviations (perfusion maps).

 
Results
 

Detection apertures that include areas outside the confocal signal (1 Airy disk diameter) facilitate the visualization of capillary walls. In areas where the vascular structure is overwhelmed by the strong confocal signal from the nerve fiber layer, detection methods which reject the confocal signal allow visualization of all capillary beds. Of the four non-confocal detection methods investigated, split-detection (see figure 1) is superior in terms of contrast and SNR for both structural and motion contrast imaging at all retinal locations. Apodization of the illumination pupil and linearly polarized detection decrease the SNR of the split-detector images. Raw images with visible illumination show substantially noisier backgrounds in both reflectance and motion contrast than those collected using infrared light, potentially due to lower signal. Registered image averages with comparable SNR however, show that image contrast is indeed reduced when using visible wavelengths. The perfusion maps created using motion contrast derived from asymmetric non-confocal detection schemes, such as knife-edge and split-detection, have some predictable and repeatable artifacts (doubling of vessels).

 
Conclusions
 

Non-confocal AOSLO imaging can reveal the structure and perfusion of all retinal capillary beds non-invasively, including that serving the highly reflective nerve fiber layer.

 
 
Figure 1: AOSLO split-detection reflectance (top) and motion contrast (bottom) images in a human subject. Scale bar is 50 μm
 
Figure 1: AOSLO split-detection reflectance (top) and motion contrast (bottom) images in a human subject. Scale bar is 50 μm
 
Keywords: 552 imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • 596 microscopy: confocal/tunneling • 688 retina  
×
×

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.

×