July 2018
Volume 59, Issue 9
Open Access
ARVO Annual Meeting Abstract  |   July 2018
The relationship between macular vessel density and ganglion cell/inner plexiformlayer thickness and their combinational index using artificial neural network
Author Affiliations & Notes
  • Jiwoong Lee
    Ophthalmology, Pusan National University Hospital , Busan, Korea (the Republic of)
  • Keunheung Park
    Ophthalmology, Pusan National University Hospital , Busan, Korea (the Republic of)
  • Jinmi Kim
    Biostatistics, Clinical Trial Center, Biomedical Research Institute, Pusan National University Hospital, Busan, Korea (the Republic of)
Investigative Ophthalmology & Visual Science July 2018, Vol.59, 5072. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jiwoong Lee, Keunheung Park, Jinmi Kim; The relationship between macular vessel density and ganglion cell/inner plexiformlayer thickness and their combinational index using artificial neural network. Invest. Ophthalmol. Vis. Sci. 2018;59(9):5072.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To evaluate the relationship between macular vessel density and ganglion cell-inner plexiform layer thickness (GCIPLT) and to compare diagnostic performance of them and to make a new clinically useful single parameter with artificial neural network which utilize both macular vessel density and GCIPLT.

Methods : A total of 173 subjects were included and divided into major 2 groups, 100 for validation group and 73 for neural net training. Macular GCIPLT and vessel density were measured with Spectralis optical coherence tomography (SD-OCT) and Topcon swept-source OCT (SS-OCT) respectively. Linear, quadratic and exponential regression models were used to investigate relationship between macular vessel density and GCIPLT. Multilayer perceptron with 1 hidden layer was used to determine single combined parameter and the diagnostic performance was compared with macular vessel density and GCIPLT.

Results : Pearson’s correlation coefficients showed significant correlation between macular vessel density and GCIPLT and inferior sectors generally showed stronger correlation than corresponding superior sectors. Both macular vessel density and GCIPLT were not significantly correlated with central visual field in normal subjects but in early and advanced glaucoma patients, they were significantly correlated. The diagnostic power of macular GCIPLT was much better than macular vessel density in most sectors. However, when macular vessel density was incorporated into macular GCIPLT with the aid of neural network, the combined parameter showed significantly enhanced diagnostic performance. Especially, in differentiating normal and early glaucoma, artificial neural network showed significant better AUROC than both macular vessel density and GCIPLT throughout all sectors.

Conclusions : Diagnostic performance of macular vessel density was much lower than macular GCIPLT. However, when it was incorporated into macular GCIPLT using artificial neural network, a single combined parameter showed better performance than macular GCIPLT alone especially in differentiating normal and early glaucoma.

This is an abstract that was submitted for the 2018 ARVO Annual Meeting, held in Honolulu, Hawaii, April 29 - May 3, 2018.

×
×

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.

×