September 2016
Volume 57, Issue 12
Open Access
ARVO Annual Meeting Abstract  |   September 2016
Evaluation of three different cone spacing metrics in adaptive optics flood illuminated retinal images
Author Affiliations & Notes
  • Daniela Giannini
    Fondazione G.B. Bietti IRCCS, Rome, Italy
    Statistical Sciences, University of Rome "La Sapienza", Rome, Italy
  • Giuseppe Lombardo
    Istituto per i Processi Chimico-Fisici, Consiglio Nazionale delle Ricerche, Messina, Italy
    Vision Engineering Italy srl, Rome, Italy
  • Letizia Mariotti
    Applied Optics, School of Physics, NUIG, Galway, Ireland
  • Nicholas Devaney
    Applied Optics, School of Physics, NUIG, Galway, Ireland
  • sebastiano serrao
    Fondazione G.B. Bietti IRCCS, Rome, Italy
  • Maurizio Vichi
    Statistical Sciences, University of Rome "La Sapienza", Rome, Italy
  • Marco Lombardo
    Fondazione G.B. Bietti IRCCS, Rome, Italy
  • Footnotes
    Commercial Relationships   Daniela Giannini, None; Giuseppe Lombardo, None; Letizia Mariotti, None; Nicholas Devaney, None; sebastiano serrao, None; Maurizio Vichi, None; Marco Lombardo, None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science September 2016, Vol.57, 64. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Daniela Giannini, Giuseppe Lombardo, Letizia Mariotti, Nicholas Devaney, sebastiano serrao, Maurizio Vichi, Marco Lombardo; Evaluation of three different cone spacing metrics in adaptive optics flood illuminated retinal images. Invest. Ophthalmol. Vis. Sci. 2016;57(12):64.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To assess the reliability of three cone spacing metrics for evaluating the distribution of distances between cones in adaptive optics (AO) flood illuminated images of the cone mosaic.

Methods : The three cone spacing metrics used are: the density counting method, which assumes an ordered lattice of the mosaic to extract the Inter Cones Distance (ICD), the Nearest Neighbour Distance (NND), which provides the mean distance of 6 neighbouring cones for each cone in the mosaic, and the Density Recovery Profile (DRP), which extracts the effective radius (ER) from the autocorrelogram.
A simulated mosaic with different numbers of cones (decreasing from 36600 to 18300 cones/mm2 by steps of 10%) was created as a model of the healthy human cone mosaic in order to evaluate how cone density influences the spacing metrics. In addition, AO images of the parafoveal cone mosaic were acquired in nineteen healthy subjects and fourteen patients suffering from inherited and acquired retinal diseases. In real AO images, the spacing metrics were calculated using two different sampling window sizes (64x64 µm and 204x204 µm) at three retinal eccentricities (1.0, 1.5 and 2.5 degrees).

Results : In the simulated AO images, the ICD and NND values decreased with decreasing cone density and their differences were significant (p<0.05) when the cone number was ≥10% lower than normal value. In real AO images, the ICD and NND showed significant differences between controls and patients (ICD difference range: 1.27-0.48 µm; NND difference range: 0.58-0.15 µm, p<0.05) in both sampling windows at all retinal eccentricities. The agreement between the ICD and NND was high in controls (ICC>0.68) and low in pathologic cases (ICC<0.47). There was no agreement between ER and ICD (ICC<0.031) or NND (ICC<0.033). The mean percentage of hexagonal Voronois was higher in controls (≥44%) than in pathologic cases (≤44%).

Conclusions : The ICD and NND provided reliable and interchangeable estimates of the cone distribution in AO flood illuminated images acquired in healthy subjects. On the other hand, these metrics could not be used interchangeably when analyzing retinal diseases. As the cone density decreases, the number of six-sided Voronois decreases and the assumption of an ordered lattice, as made by ICD, is not valid. This makes ICD less reliable than NND when analyzing retinal diseases.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.

×
×

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.

×