July 2019
Volume 60, Issue 9
Free
ARVO Annual Meeting Abstract  |   July 2019
Topographic Analysis of IS/OS Impairment and Correlation With Fluid Markers in Wet AMD Using Deep Learning
Author Affiliations & Notes
  • Sophie Klimscha
    Department of Ophthalmology and Optometry, Medical University Vienna, Vienna, Austria
  • Christoph Grechenig
    Department of Ophthalmology and Optometry, Medical University Vienna, Vienna, Austria
  • Amir Sadeghipour
    Department of Ophthalmology and Optometry, Medical University Vienna, Vienna, Austria
  • Eleonore Pablik
    Center for Medical Statistics, Informatics and Intelligent Systems, Medical University of Vienna , Austria
  • Sebastian M Waldstein
    Department of Ophthalmology and Optometry, Medical University Vienna, Vienna, Austria
  • Ursula Schmidt-Erfurth
    Department of Ophthalmology and Optometry, Medical University Vienna, Vienna, Austria
  • Footnotes
    Commercial Relationships   Sophie Klimscha, None; Christoph Grechenig, None; Amir Sadeghipour, None; Eleonore Pablik, None; Sebastian Waldstein, Bayer (F), Genentech (F), Novartis (C); Ursula Schmidt-Erfurth, Böhringer Ingelheim (C), Genentech (C), Novartis (C), Roche (C)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science July 2019, Vol.60, 1536. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Sophie Klimscha, Christoph Grechenig, Amir Sadeghipour, Eleonore Pablik, Sebastian M Waldstein, Ursula Schmidt-Erfurth; Topographic Analysis of IS/OS Impairment and Correlation With Fluid Markers in Wet AMD Using Deep Learning. Invest. Ophthalmol. Vis. Sci. 2019;60(9):1536.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : The extent of photoreceptor layer impairment during anti-VEGF therapy of wet AMD has been insufficiently characterized using semi-quantitative approaches. This large-scale post-hoc analysis of a prospective, multicenter trial aims to provide quantitative data on the course of IS/OS integrity. Furthermore, we analyze the spatial correspondence of impaired IS/OS integrity with fluid-related markers of AMD disease activity.

Methods : Volume scan SD-OCT data (Spectralis) of n=185 treatment naïve wet AMD eyes acquired at baseline (bsl), month 3 and month 12 were included into the analysis. Impaired IS/OS integrity was manually quantified in each of the 49 B-Scans comprising the volume and analyzed over time, reporting the median [IQR] / mean (SD) change of IS/OS integrity as percentage of the respective ETDRS field. Intraretinal cystoid fluid (IRC), subretinal fluid (SRF) and pigment epithelial detachment (PED) were automatically segmented in the full volumes using validated deep learning methods. To assess their topographic correlation with IS/OS impairment at the native stage, per-pixel chi-square testing between pairs of lesions was performed.

Results : IS/OS integrity decreased from 85% [70; 93] at bsl to 82% [66; 93] at month 3, and increased to 88% [70; 96] at month 12. Figure 1 shows results per ETDRS field. Improvement in IS/OS integrity was greater in the foveal area compared to the total scan from bsl to month 3 (+6% ±33 vs. -3% ±10, p<0.01), from month 3 to month 12 (+12% ±26 vs. +4% ±7, p<0.01), and from bsl to month 12 (+18 ±37 vs. +1 ±12, p<0.01). En-face chi-square maps revealed that IS/OS impairment was less likely in areas with SRF and vice versa (Figure 2). Moreover, we observed an association between the resolution of SRF and temporary IS/OS impairment.

Conclusions : Our study shows improvement of IS/OS integrity, predominantly in the fovea, during anti-VEGF therapy of wet AMD. IS/OS integrity was preserved in areas of SRF accumulation, and showed temporary changes upon SRF resolution.

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

 

IS/OS integrity relative to the respective ETDRS field at each timepoint.

IS/OS integrity relative to the respective ETDRS field at each timepoint.

 

En-face per pixel odds ratio maps show topographic relations between pathomorphologic lesions at bsl, which are statistically significant in chi-square testing. Positive correlations are color-coded in light to dark green; dissociation in yellow to red, as shown in log odds ratio scale below.

En-face per pixel odds ratio maps show topographic relations between pathomorphologic lesions at bsl, which are statistically significant in chi-square testing. Positive correlations are color-coded in light to dark green; dissociation in yellow to red, as shown in log odds ratio scale below.

×
×

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.

×