April 2011
Volume 52, Issue 14
Free
ARVO Annual Meeting Abstract  |   April 2011
Mapping The Visual Field To Retinal Ganglion Cell Thickness Image
Author Affiliations & Notes
  • Xian Zhang
    Psychology, Columbia University, New York, New York
  • Ali S. Rasa
    Psychology, Columbia University, New York, New York
  • Gustavo V. De Moraes
    Ophthalmology, New York Univ School of Medicine, New York, New York
  • Donald C. Hood
    Psychology, Columbia University, New York, New York
  • Footnotes
    Commercial Relationships  Xian Zhang, Topcon, Inc. (C); Ali S. Rasa, None; Gustavo V. De Moraes, None; Donald C. Hood, Topcon, Inc. (F, C)
  • Footnotes
    Support  NIH/NEI Grant EY 02115
Investigative Ophthalmology & Visual Science April 2011, Vol.52, 5085. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Xian Zhang, Ali S. Rasa, Gustavo V. De Moraes, Donald C. Hood; Mapping The Visual Field To Retinal Ganglion Cell Thickness Image. Invest. Ophthalmol. Vis. Sci. 2011;52(14):5085.

      Download citation file:


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

      ×
  • Supplements
Abstract
 
Purpose:
 

To predict topographical loss on static automated perimetry from retinal ganglion cell (RGC) thickness measured with frequency domain optical coherence tomography (fdOCT), a multiple linear regression method was employed.

 
Methods:
 

151 eyes of 106 patients, glaucoma and glaucoma suspects, were tested with both 10-2 visual fields (VF) (Carl Zeiss, Inc) and fdOCT macular 3D scans (Topcon, Inc.). All had glaucomatous optic neuropathy. The mean total deviation of the VFs had a median (5%, 95%) of -4.96 (-18.31; 0.36) dB. For inclusion, VFs had to be reliable (indices <30%) and the fdOCT free of major algorithm failures as judged by visual inspection. The RGC+IPL (inner plexiform layer) thickness was determined by a segmentation algorithm previously described.[1] The linear regression method involved: 1. The 512x128 RGC+IPL thickness image was reduced to 24 principal components (PCs); 2. The RGC+IPL-to-VF map was defined as the product of the inverted 24x150 (PCs x number of eyes) matrix and the 68x150 (number of VF test points x number of eyes) matrix. 3. The predicted VF of each eye was obtained by multiplying the RGC+IPL thickness of that eye by the RGC+IPL-to-VF map obtained with data of the other 150 eyes. Probability plots for the observed VFs were determined using a -5 dB (about 5%) criterion. Probability plots for the predicted pattern deviation VFs were obtained by setting the cutoff such that the total number of abnormal points in the predicted VF was equal to number in the observed VF. For each patient, the percent of points agreeing on the predicted and observed VFs was calculated.

 
Results:
 

In general, the predicted VF showed good agreement with the actual visual field. The median agreement was 87% with 95% of the values > 51% (see Fig. for 2 median examples).

 
Conclusions:
 

The pattern of 10-2 VF loss can be predicted based upon the thickness of the RGC+IPL layer. This method provides a possible approach for combining VF and OCT information for improving the detection of glaucomatous defects. 1. Yang, Reisman et al, Opt. Exp, 2010.  

 
Keywords: imaging/image analysis: clinical • perimetry • visual fields 
×
×

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.

×