June 2022
Volume 63, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2022
Exploration of Machine Learning-Enhanced Ellipsoid Zone Mapping and Radiomics-based Textural Features As Biomarkers for Risk of Geographic Atrophy Development in Dry AMD
Joseph Roy Abraham; Sudeshna Sil Kar; Hasan Cetin; Sunil K Srivastava; Anant Madabhushi; Justis P Ehlers
Author Affiliations & Notes
  • Joseph Roy Abraham
    Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio, United States
  • Sudeshna Sil Kar
    Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio, United States
    Case Western Reserve University, Cleveland, Ohio, United States
  • Hasan Cetin
    Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio, United States
  • Sunil K Srivastava
    Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio, United States
  • Anant Madabhushi
    Case Western Reserve University, Cleveland, Ohio, United States
  • Justis P Ehlers
    Cole Eye Institute, Cleveland Clinic, Cleveland, Ohio, United States
  • Footnotes
    Commercial Relationships   Joseph Abraham None; Sudeshna Sil Kar None; Hasan Cetin None; Sunil Srivastava Bausch and Lomb, Code C (Consultant/Contractor), Novartis, Code C (Consultant/Contractor), Carl Zeiss Meditec, Code C (Consultant/Contractor), Allergan, Code F (Financial Support), Bausch and Lomb, Code F (Financial Support), Leica, Code P (Patent); Anant Madabhushi Aiforia, Code C (Consultant/Contractor), Astrazeneca, Bristol Myers-Squibb, Philips. Equity: Inspirata, Elucid Bioimaging, Code F (Financial Support); Justis Ehlers Alcon, Allergan, Leica, Santen, Thrombogenics, Genentech, Novartis, Aerpio, Allegro, Regeneron, Adverum, Stealth, Roche, Zeiss, Code C (Consultant/Contractor), Alcon, Genentech, Regeneron, Boehringer-Ingelheim, Novartis, Aerpio, Thrombogenics, Code F (Financial Support), Leica, Code P (Patent)
  • Footnotes
    Support  National Institutes of Health/National Eye Institute, Bethesda, Mary-land, USA, K23-EY022947-01A1
Investigative Ophthalmology & Visual Science June 2022, Vol.63, 3020 – F0290. doi:
  • Views
  • Share
  • Tools
    • Alerts
      User Alerts
      You are adding an alert for:
      Exploration of Machine Learning-Enhanced Ellipsoid Zone Mapping and Radiomics-based Textural Features As Biomarkers for Risk of Geographic Atrophy Development in Dry AMD
      You will receive an email whenever this article is corrected, updated, or cited in the literature. You can manage this and all other alerts in My Account
      The alert will be sent to:
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation
      Citation

      Joseph Roy Abraham, Sudeshna Sil Kar, Hasan Cetin, Sunil K Srivastava, Anant Madabhushi, Justis P Ehlers; Exploration of Machine Learning-Enhanced Ellipsoid Zone Mapping and Radiomics-based Textural Features As Biomarkers for Risk of Geographic Atrophy Development in Dry AMD. Invest. Ophthalmol. Vis. Sci. 2022;63(7):3020 – F0290.

      Download citation file:

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

      ×
    • Get Permissions
  • Supplements
Abstract

Purpose : To evaluate compartmental SD-OCT imaging biomarkers using machine learning-enhanced outer retinal segmentation and textural radiomics that are associated with the development of geographic atrophy (GA).

Methods : This was a retrospective image analysis study that included 114 subjects with dry age-related macular degeneration (AMD) without GA at baseline with 5 years of clinical and SD-OCT follow-up. All SD-OCT scans were analyzed using a machine learning-enhanced multi-layer retinal segmentation platform that enabled quantitative ellipsoid zone (EZ) integrity assessment and EZ-RPE compartmental extraction. Eyes were categorized as either GA developers or non-GA developers based on development of GA on SD-OCT by year 5 of follow-up. Baseline quantitative features between the GA developers and non-developers were compared using traditional statistics and a radiomic classification model combined with a Random Forest algorithm. Radiomic features were identified using a Minimum Redundancy Maximum Relevance feature selection method, and the 10 topmost features were used to train a Random Forest classifier.

Results : At baseline, eyes that developed macular GA (n = 33) significantly decreased EZ integrity as measured by mean EZ-RPE central subfield thickness (28.5 vs 36.2µm p<0.001) and EZ-RPE volume (1.23 vs 1.30 mm3 p=0.002) compared to eyes that did not develop GA (n = 81). In addition, eyes that developed GA increased partial (5.4% vs 0.9%, p<0.001) and total EZ attenuation (2.6 vs 0.3%, p<0.001) compared to eyes that did not develop GA. Random Forest modeling utilizing identified radiomic features yielded a classifier performance of AUC=0.88±0.02 for GA development. Increased heterogeneity in the EZ-RPE compartment was associated with the development of GA.

Conclusions : The development of GA was significantly associated with reduced EZ integrity at baseline as well as radiomics-based EZ-RPE compartment textural features. These potential biomarkers could be key tools for risk-stratification of eyes for future development of GA and clinical trial enrichment.

This abstract was presented at the 2022 ARVO Annual Meeting, held in Denver, CO, May 1-4, 2022, and virtually.

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
Attribution-NonCommercial-NoDerivatives
Copyright © 2025 Association for Research in Vision and Ophthalmology.
×
×

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.

×
This site uses cookies. By continuing to use our website, you are agreeing to our privacy policy.Accept