June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
AI-based quantification of photoreceptor maintenance in the treatment of Geographic Atrophy Secondary to AMD in the FILLY trial
Author Affiliations & Notes
  • Ursula Schmidt-Erfurth
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Julia Mai
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Gregor Sebastian Reiter
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Sophie Riedl
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Dmitrii Lachinov
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Wolf-Dieter Vogl
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Hrvoje Bogunovic
    Department of Ophthalmology, Medizinische Universitat Wien, Wien, Wien, Austria
  • Footnotes
    Commercial Relationships   Ursula Schmidt-Erfurth, Apellis (F), Genentech (C), Heidelberg (C), Novartis (C), RetInSight (C), Roche (C); Julia Mai, None; Gregor Reiter, None; Sophie Riedl, None; Dmitrii Lachinov, None; Wolf-Dieter Vogl, None; Hrvoje Bogunovic, Apellis (F)
  • Footnotes
    Support  Research Grant from Apellis
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 236. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Ursula Schmidt-Erfurth, Julia Mai, Gregor Sebastian Reiter, Sophie Riedl, Dmitrii Lachinov, Wolf-Dieter Vogl, Hrvoje Bogunovic; AI-based quantification of photoreceptor maintenance in the treatment of Geographic Atrophy Secondary to AMD in the FILLY trial. Invest. Ophthalmol. Vis. Sci. 2021;62(8):236.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To quantify the therapeutic effect of pegcetacoplan (APL2) on inhibition of photoreceptor (PR) loss on conventional OCT images using deep learning in GA.

Methods : SD-OCT scans (Spectralis) in a raster of 512x49x496 voxels were obtained at monthly intervals during the phase II FILLY trial of APL2, an investigational complement C3 inhibition, in patients with geographic atrophy (GA) secondary to AMD. Patients with complete outer retinal atrophy (cRORA) were randomized into a sham group (SM), a group receiving monthly (AM) and another group receiving bimonthly (AEOM) intravitreal injections of APL2. Using expert manual annotation at an A-scan level, a deep learning segmentation algorithm was developed based on a convolutional neural network (CNN). Fully automated delineation of the area presenting with complete photoreceptor (PR) loss was performed at baseline, at 2, 6 and 12 months. The associated retinal pigment epithelial (RPE) loss was measured via a previously established algorithm.

Results : A total of 31,752 B-Scans of 648 volumes of 162 study eyes (SM:56, AM:54, AEOM:54) were evaluated from baseline to month 12. The square-root lesion area based on PR loss was 0.108, 0.111, 0.106 mm at baseline in the SM, AEOM and AM groups. While the mean PR defect size increased to 0.110, 0.112, 0.116 mm over months 2,6,12 in the SM group, the AM group presented a significantly smaller lesion extension with 0.105, 0.106, 0.110 mm, respectively over time, which was also seen as a trend for the AEOM group (Table 1). Moreover, the PR lesion sizes were found to consistently exceed the underlying borders of RPE loss by 73, 64, 61, and 60 % at the subsequent intervals.

Conclusions : Deep-learning based algorithms are able to distinctly and reliably quantify the extension and growth of photoreceptor loss and appear to represent ideal tools to evaluate disease activity as well as therapeutic efficacy of slowing GA progression. Based on standard OCT imaging, automated monitoring of RPE and PR loss and maintenance should be considered for an optimized management of patients with GA.

This is a 2021 ARVO Annual Meeting abstract.

 

 

×
×

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.

×