Investigative Ophthalmology & Visual Science Cover Image for Volume 63, Issue 7
June 2022
Volume 63, Issue 7
Open Access
ARVO Annual Meeting Abstract  |   June 2022
The RAP study, Report 6: Quantification of exudative biomarkers in neovascular age-related macular degeneration (nAMD) using deep learning, a type-based comparative analysis.
Author Affiliations & Notes
  • Bilal Haj Najeeb
    Medizinische Universitat Wien, Wien, Wien, Austria
  • Bianca S Gerendas
    Medizinische Universitat Wien, Wien, Wien, Austria
  • Hrvoje Bogunovic
    Medizinische Universitat Wien, Wien, Wien, Austria
  • Ursula Schmidt-Erfurth
    Medizinische Universitat Wien, Wien, Wien, Austria
  • Footnotes
    Commercial Relationships   Bilal Haj Najeeb RetInSight, Code C (Consultant/Contractor); Bianca S Gerendas Roche, Code C (Consultant/Contractor), Novartis, Code C (Consultant/Contractor), DXS, Code C (Consultant/Contractor); Hrvoje Bogunovic Apellis, Code C (Consultant/Contractor); Ursula Schmidt-Erfurth Roche, Code C (Consultant/Contractor), Novartis, Code C (Consultant/Contractor), Genentech, Code C (Consultant/Contractor), RetInSight, Code C (Consultant/Contractor), Boehringer, Code C (Consultant/Contractor)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2022, Vol.63, 3861. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Bilal Haj Najeeb, Bianca S Gerendas, Hrvoje Bogunovic, Ursula Schmidt-Erfurth; The RAP study, Report 6: Quantification of exudative biomarkers in neovascular age-related macular degeneration (nAMD) using deep learning, a type-based comparative analysis.. Invest. Ophthalmol. Vis. Sci. 2022;63(7):3861.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To quantify five exudative biomarkers in eyes with treatment-naïve macular neovascularization type 3 (MNV3) using deep learning and compare the outcomes with those of eyes with MNV 1 and MNV2.

Methods : 34528 optical coherence tomography scans of 457 eyes with nAMD consisting of 281 (250 unifocal, 31 multifocal) eyes with MNV3, 55 eyes with MNV2 and 121 (91 non-polypoidal, 30 polypoidal) eyes with MNV1 were included. The volumes of intraretinal (IRF) and subretinal fluid (SRF), pigment epithelial detachments (PED) and hyperreflective foci (HRF) in nanoliters (nl) and retinal thickness (RT) in micrometers (µm) were quantified by validated deep learning algorithms for each (sub)type of MNV.

Results : The analysis of MNV types shows that MNV3 had the highest mean (± standard deviation) of IRF volume in the central 1, 3 and 6mm: 67±61, 216±196, 291±289nl, and greatest mean of RT 418±104, 391±62, 356±49µm, followed by MNV2 with 31±54, 79±145, 105±216nl (IRF), and 380±80, 372±50, 338±39µm (RT), p<0.05. Also, MNV3 presented with the biggest mean of HRF volume 5±5, 32±24, 79±69nl, followed by MNV1 3±3, 22±19, 51±52nl, p<0.05. MNV3 revealed the lowest mean of SRF volume 15±35, 72±133, 215±382nl, whereas MNV1 showed the greatest mean of SRF volume 28±38, 168±193, 492±583nl, p<0.05, and the biggest mean of PED volume 84±94, 495±583, 667±855nl, p<0.05.
The analysis of MNV1 subtypes (polypoidal vs. non-polypoidal) shows that the polypoidal subtype had a significantly higher mean of PED volume in all areas, greater mean of RT in the 3 and 6mm areas, and bigger means of SRF and HRF volumes in the 6mm area only. The mean of IRF volume was not significantly different between both subtypes. In addition, the analysis of MNV3 subtypes (multifocal vs. unifocal) shows that the multifocal subtype revealed significantly higher means of IRF and HRF volumes in all areas and a greater mean of RT in the 3mm area only. No significant mean differences of SRF and PED volumes were observed.

Conclusions : Using deep learning to quantify exudative biomarkers in different MNV types provide clinicians with unprecedented quantitative and topographic details. These pathomorphological differences allow a better understanding of the clinical manifestation, natural course and prognosis of each (sub)type of MNV in nAMD.

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

×
×

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.

×