July 2019
Volume 60, Issue 9
Open Access
ARVO Annual Meeting Abstract  |   July 2019
Deep Learning Models Predict Visual Function from Macula Thickness Map
Author Affiliations & Notes
  • Mark Christopher
    Ophthalmology, University of California, San Diego, La Jolla, California, United States
  • Akram Belghith
    Ophthalmology, University of California, San Diego, La Jolla, California, United States
  • Christopher Bowd
    Ophthalmology, University of California, San Diego, La Jolla, California, United States
  • Massimo Antonio Fazio
    Ophthalmology, University of Alabama, Birmingham, Birmingham, Alabama, United States
  • Michael Henry Goldbaum
    Ophthalmology, University of California, San Diego, La Jolla, California, United States
  • Robert N Weinreb
    Ophthalmology, University of California, San Diego, La Jolla, California, United States
  • Christopher A Girkin
    Ophthalmology, University of Alabama, Birmingham, Birmingham, Alabama, United States
  • Jeffrey M Liebmann
    Ophthamology, Columbia University Medical Center, New York, New York, United States
  • Linda M Zangwill
    Ophthalmology, University of California, San Diego, La Jolla, California, United States
  • Footnotes
    Commercial Relationships   Mark Christopher, None; Akram Belghith, None; Christopher Bowd, None; Massimo Fazio, None; Michael Goldbaum, None; Robert Weinreb, Aerie Pharmaceuticals (C), Carl Zeiss Meditec (F), Centervue (F), Genentech (F), Heidelberg Engineering (F), Konan Medical (F), National Eye Institute (F), Optos (F), Optovue (F), Research to Prevent Blindness (F); Christopher Girkin, EyeSight Foundation of Alabama (F), Heidelberg Engineering (F), National Eye Institute (F), Research to Prevent Blindness (F); Jeffrey Liebmann, Alcon (C), Allergan (C), Bausch & Lomb (C), Bausch & Lomb (F), Carl Zeiss Meditec (C), Carl Zeiss Meditec (F), Heidelberg Engineering (C), Heidelberg Engineering (F), National Eye Institute (F), Optovue (F), Reichert (C), Reichert Technologies (F), Research to Prevent Blindness (F), Valeant Pharmaceuticals (C); Linda Zangwill, Carl Zeiss Meditec (F), Heidelberg Engineering (F), National Eye Institute (F), Optovue (F), Topcon Medical Systems (F)
  • Footnotes
    Support  EI: EY11008, P30 EY022589, EY026590, EY022039, EY021818, EY023704, EY029058, T32 EY026590, R21 EY027945, Genentech Inc. Unrestricted grant from Research to Prevent Blindness (New York, NY)
Investigative Ophthalmology & Visual Science July 2019, Vol.60, 5600. doi:https://doi.org/
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Mark Christopher, Akram Belghith, Christopher Bowd, Massimo Antonio Fazio, Michael Henry Goldbaum, Robert N Weinreb, Christopher A Girkin, Jeffrey M Liebmann, Linda M Zangwill; Deep Learning Models Predict Visual Function from Macula Thickness Map. Invest. Ophthalmol. Vis. Sci. 2019;60(9):5600. doi: https://doi.org/.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To apply deep learning approaches to predict visual field (VF) mean deviation (MD) from multiple thickness maps derived from macular imaging.

Methods : Imaging and VF measurements were collected from a cohort of 394 healthy, suspect, and glaucoma participants and 703 eyes. The imaging consisted of 1448 Spectralis horizontal posterior pole (61 b-scans, 30°x25°) images. VF testing in the form of 10-2 (mean ± standard deviation MD = -3.9 ± 6.3 dB) and 24-2 (mean ± standard deviation MD = -4.1 ± 6.3 dB) tests was completed within 180 days of imaging. Participants were randomly divided into independent training (85%), validation (5%), and test (10%) sets. Using standard instrument segmentation, retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), and inner plexiform layer (IPL), ganglion cell – inner plexiform layer combined (GCIPL), ganglion cell complex (GCC), and whole retina (RET) thickness maps were generated Individual deep learning models were trained to predict both 10-2 and 24-2 MD from each layer. Combined deep learning models that use all layers simultaneously were also constructed. The Resnet50 architecture was used and transfer learning was applied by pre-training on a general image dataset (ImageNet). Models were evaluated using R2 to compare to predicted and true MD. As a baseline comparison, linear regression models that predicted MD from mean layer thicknesses were also evaluated.

Results : In predicting 10-2 MD, the individual deep learning models, the GCIPL (R2 = 0.84), GCL (R2 = 0.78), and IPL (R2 = 0.74) outperformed other layers. The combined deep learning model achieved an R2 = 0.87 (p < 0.001), outperforming mean layer thickness R2 = 0.60 (p < 0.001). In predicting 24-2 MD, the combined deep learning model achieved an R2 = 0.82 (p < 0.001) while the mean layer thickness prediction had an R2 = 0.62 (p < 0.001). Here, the RNFL (R2 = 0.73), GCIPL (R2 = 0.71), and IPL (R2 = 0.70) outperformed other layers. The full results are summarized in Table 1.

Conclusions : Applying deep learning models to macular structural measurements can predict both central and peripheral visual field loss in glaucoma with a high degree of accuracy. The results suggest that deep learning models can provide useful clinical information without relying on time-consuming and subjective visual field testing.

This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.

 

Table 1: Performance of macula layer thicknesses and deep learning models to predict MD of 10-2 and 24-2 visual fields.

Table 1: Performance of macula layer thicknesses and deep learning models to predict MD of 10-2 and 24-2 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.

×