September 2016
Volume 57, Issue 12
Open Access
ARVO Annual Meeting Abstract  |   September 2016
Fully Automated Identification of Lesion Activity in Neovascular AMD
Author Affiliations & Notes
  • Anat Loewenstein
    Ophthalmology, Tel Aviv Medical Center, Tel Aviv, Israel
  • Dafna Goldenberg
    Ophthalmology, Tel Aviv Medical Center, Tel Aviv, Israel
  • Usha Chakravarthy
    Queen's University, Center for Experimental Medicine, Belfast, Ireland
  • Graham Young
    Notal Vision, Tel Aviv, Israel
  • Moshe Havilio
    Notal Vision, Tel Aviv, Israel
  • Omer Rafaeli
    Notal Vision, Tel Aviv, Israel
  • Gidi Benyamini
    Notal Vision, Tel Aviv, Israel
  • Footnotes
    Commercial Relationships   Anat Loewenstein, Notal (C); Dafna Goldenberg, None; Usha Chakravarthy, Notal (C); Graham Young, None; Moshe Havilio, None; Omer Rafaeli, None; Gidi Benyamini, None
  • Footnotes
    Support  Notal
Investigative Ophthalmology & Visual Science September 2016, Vol.57, No Pagination Specified. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Anat Loewenstein, Dafna Goldenberg, Usha Chakravarthy, Graham Young, Moshe Havilio, Omer Rafaeli, Gidi Benyamini; Fully Automated Identification of Lesion Activity in Neovascular AMD. Invest. Ophthalmol. Vis. Sci. 201657(12):.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : Objective: The objective of the study was to evaluate the accuracy of the Notal OCT Analyzer (NOA) versus retina specialists (RS) in the automatic detection of fluid on optical coherence tomography (OCT).

Methods : Design: A prospective study of the performance of the NOA, which is compared to three retina specialists. The data source was OCT volume scans acquired in a retina clinic.
Participants: A random selection of anonymization OCT scans (Zeiss Cirrus, Carl Zeiss Meditec, Dublin CA) of 165 AMD patients attending a single tertiary referral center (Belfast HSC, UK).
Methods: OCT scans of AMD patients were exported. Each scan set was analyzed by the NOA, and by three independent RS for the presence of intra-retinal or sub-retinal fluid. NOA also ranked cross-sections of scans for likelihood of CNV activity allowing a second grading session by the three RS.

Outcome measures: NOA’s sensitivity and specificity versus the RS grading and NOA’s performance in ranking cross-section for activity.

Results : Results: 142 scan sets met the criteria for the primary analysis. On testing the RS grading versus the NOA, the accuracy was 91% (95% CI ±7%), sensitivity was 92% (±6%) and specificity was 91% (±6%). The graders’ accuracy when compared to majority of the other readers (including a 4th reader) was 93%. On average, the three readers could identify fluid in 95% of scans by just reviewing a single cross section with the highest NOA score, 99.5% of scans by viewing the top three cross sections.
the RS grading and NOA’s performance in ranking cross-section for activity.

Conclusions : Conclusions: Concordance between the NOA and the RS determination of lesion activity was extremely high. The level of discrepancy between RS and the NOA was similar to the NOA’s mismatches. Our results show that automated delineation of the retinal contours combined with interpretation of disease activity is feasible and has the potential to become a powerful tool in terms of its clinical applications.
pt;line-height:150%'>
the RS grading and NOA’s performance in ranking cross-section for activity.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.

×
×

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.

×