June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Patterns of Neuronal and Central Visual Field Loss in Optic Neuritis at Outcome Identified by Machine Learning
Author Affiliations & Notes
  • David Szanto
    Department of Neurology, Icahn School of Medicine at Mount Sinai, New York, New York, United States
  • Jui-Kai Wang
    Department of Ophthalmology and Visual Sciences, University of Iowa Hospitals and Clinics, Iowa City, Iowa, United States
  • Brian Woods
    Department of Ophthalmology, Cork University Hospital, Cork, Cork, Ireland
  • Tobias Elze
    Schepens Eye Research Institute of Massachusetts Eye and Ear, Boston, Massachusetts, United States
  • Mona K Garvin
    Department of Electrical and Computer Engineering, University of Iowa Center for the Prevention and Treatment of Visual Loss, Iowa City, Iowa, United States
  • Louis R Pasquale
    Department of Ophthalmology, Icahn School of Medicine at Mount Sinai Department of Medicine, New York, New York, United States
  • Randy H Kardon
    Department of Ophthalmology and Visual Sciences, University of Iowa Hospitals and Clinics, Iowa City, Iowa, United States
  • Joseph Branco
    New York Medical College, Valhalla, New York, United States
  • Mark J Kupersmith
    Departments of Neurology, Ophthalmology and Neurosurgery, Icahn School of Medicine at Mount Sinai, New York, New York, United States
  • Footnotes
    Commercial Relationships   David Szanto None; Jui-Kai Wang None; Brian Woods None; Tobias Elze Genentech Inc., Code F (Financial Support); Mona Garvin University of Iowa, Code P (Patent); Louis Pasquale Twenty Twenty Inc, Skye Biosciences, Eyenovia, Code C (Consultant/Contractor); Randy Kardon Fight for Sight, Department of Veterans Affairs Research Foundation, Code S (non-remunerative); Joseph Branco None; Mark Kupersmith None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 542. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      David Szanto, Jui-Kai Wang, Brian Woods, Tobias Elze, Mona K Garvin, Louis R Pasquale, Randy H Kardon, Joseph Branco, Mark J Kupersmith; Patterns of Neuronal and Central Visual Field Loss in Optic Neuritis at Outcome Identified by Machine Learning. Invest. Ophthalmol. Vis. Sci. 2023;64(8):542.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : Patterns of central visual field (VF) loss and retinal ganglion cell (RGC) thinning are reported in glaucoma but not for optic neuritis (ON), which almost always acutely affects the central VF. We hypothesized that machine learning methods to investigate the residual VF deficits and macula RGC thickness could reveal specific patterns of loss and enhance our understanding of the structure-function relationship in ON.

Methods : We applied archetypal analysis (AA) to 157 same-day pairings of 10-2 VFs and segmented optical coherence tomography (OCT) macula images collected from 88 eyes at least 90 days after acute ON attack, decomposing them into component VF and retinal thickness archetypes (total weight = 100%). We correlated archetypes (ATs) for total retinal thickness (TRT), inner retinal thickness (IRT), and macular ganglion cell-inner plexiform layer (GCIPL) thickness to ATs of visual field loss. We also performed a sub-analysis on VF and OCT patterns in eyes with a deficient outcome (MD <-5 dB), training on 50 ON VFs and testing on 26 separate eyes.

Results : AA identified seven VF loss patterns and 11 retinal thickness patterns for the three OCT models of all eyes. A total of 137 measurements (87%) had the normal VF AT as the dominant AT (weight ≥50%). Conversely, OCTs seldom decomposed into a single AT. The two most common GCIPL ATs each had 12 OCTs (8% and 8% of all GCIPL measurements, respectively) with weight ≥50%. We found GCIPL ATs best correlated with the significant VF metrics MD (r = .61) and the normal VF AT (r = .58). For eyes with poor vision (MD <-5 dB), Figure 1 describes a ten-AT-VF model showing three meaningful VF patterns corresponding to moderate diffuse vision loss (r = .79), major diffuse vision loss (r = .77), and central vision loss (r = .71), as well as strongly predicting MD (r = .77).

Conclusions : Except for eyes with poor VF outcomes, unlike glaucoma most eyes with acute ON had complete VF recovery, suggesting potential compensation due to overlapping receptor fields. AA identifies and correlates ON-specific patterns of retinal thickness and vision loss best when recovery is incomplete.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

 

Figure 1: The three dominant weight (DW) ATs for eyes with deficient VF outcome and the linear regression of all GCIPL ATs correlated with VF ATs are shown. AA trained on these VFs (n = 50) and transformed VFs with same-day OCT pairings (n = 26) showing AT decompositions.

Figure 1: The three dominant weight (DW) ATs for eyes with deficient VF outcome and the linear regression of all GCIPL ATs correlated with VF ATs are shown. AA trained on these VFs (n = 50) and transformed VFs with same-day OCT pairings (n = 26) showing AT decompositions.

×
×

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.

×