June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Characterization of Baseline “EZ At-Risk” Burden Using Deep Learning Feature Extraction in the GATHER1 Phase 2/3 Clinical Trial.
Author Affiliations & Notes
  • Sari Yordi
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Gagan Kalra
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Hasan Cetin
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Jon Whitney
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Conor Anthony McConville
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Yavuz Cakir
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Victoria Eileen Whitmore
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Jamie Reese
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Julie Clark
    IVERIC bio, Cranbury, New Jersey, United States
  • Liansheng Zhu
    IVERIC bio, Cranbury, New Jersey, United States
  • Sunil K Srivastava
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Justis P Ehlers
    Cleveland Clinic Cole Eye Institute, Cleveland, Ohio, United States
  • Footnotes
    Commercial Relationships   Sari Yordi Betty J. Powers Retina Research Fellowship, Code S (non-remunerative); Gagan Kalra None; Hasan Cetin None; Jon Whitney None; Conor McConville None; Yavuz Cakir None; Victoria Whitmore None; Jamie Reese None; Julie Clark IVERIC bio, Code E (Employment); Liansheng Zhu IVERIC bio, Code E (Employment); Sunil Srivastava Bausch and Lomb, Adverum, Novartis, and Regeneron, Code C (Consultant/Contractor), Regeneron, Allergan, and Gilead, Code F (Financial Support), Leica, Code P (Patent); Justis Ehlers Aerpio, Alcon, Allegro, Allergan, Genentech/Roche, Novartis, Thrombogenics/Oxurion, Leica, Zeiss, Regeneron, Santen, Stealth, Adverum, Iveric BIO, Apellis, Boehringer-Ingelheim, RegenxBIO, Code C (Consultant/Contractor), Aerpio, Alcon, Thrombogenics/Oxurion, Regeneron, Genentech, Novartis, Allergan, Boehringer-Ingelheim, Iveric Bio, Adverum, Code F (Financial Support), Leica, Code P (Patent)
  • Footnotes
    Support  P30EY025585(BA-A), Research to Prevent Blindness (RPB) Challenge Grant, Cleveland Eye Bank Foundation Grant
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 231. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Sari Yordi, Gagan Kalra, Hasan Cetin, Jon Whitney, Conor Anthony McConville, Yavuz Cakir, Victoria Eileen Whitmore, Jamie Reese, Julie Clark, Liansheng Zhu, Sunil K Srivastava, Justis P Ehlers; Characterization of Baseline “EZ At-Risk” Burden Using Deep Learning Feature Extraction in the GATHER1 Phase 2/3 Clinical Trial.. Invest. Ophthalmol. Vis. Sci. 2023;64(8):231.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To evaluate the feasibility of automated characterization of Ellipsoid Zone (EZ) At-Risk and characterize the baseline EZ At-Risk feature spectrum across the study eyes in the GATHER1 AMD Trial.

Methods : The GATHER1 study was a prospective, randomized, double-masked, phase 2/3 trial that evaluated avacincaptad pegol compared with sham in patients with geographic atrophy (GA) secondary to age-related macular degeneration (AMD).

The EZ At-Risk automated feature assessment tool is a deep learning model that was developed to identify areas of the outer retina, particularly the EZ, that are abnormal and at-risk for progressive degenerative changes. This model trained on over 100,000 SD-OCT independent B-scans that were annotated for thinning and abnormalities consistent with photoreceptor outer segment thinning of 15 microns or less (i.e., EZ-RPE thickness). The annotations excluded areas of EZ loss overlying GA with the intent to have the model primarily focus in areas “at-risk” rather than areas that were already atrophic.

SD-OCT scans from the GATHER1 study were imported into the ML analysis platform to evaluate both EZ At-Risk and to provide enhanced multi-layer segmentation to identify areas of GA. All baseline scans were included for analysis that were of sufficient quality for analysis and segmentation. EZ At-Risk was evaluated by the model for each B-scan and overall percentage of the macular cube en face area that included EZ At-Risk was determined. In addition, the EZ At-Risk Index was calculated based on the percentage area of EZ At-Risk divided by the macular area not occupied by GA.

Results : At baseline, 194 subjects (194 eyes) were included in this analysis. Within the whole cohort, EZ At-Risk ranged from 6.3% to 79.5% per scan. The mean and median EZ At-Risk for the study were 26.8% and 23.6%, respectively. The EZ At-Risk Index ranged from 0.07 to 1.01 per scan, with a mean and median of 0.25 and 0.31, respectively.

Conclusions : EZ At-Risk was able to be successfully calculated and characterized in a Phase 2/3 Clinical Trial and demonstrated a wide range of baseline values. Ongoing analyses for GA growth risk stratification and the impact of EZ At-Risk on treatment response will continue to inform the role of this biomarker in AMD management.

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

×
×

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.

×