June 2017
Volume 58, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2017
Improved Visualization and Quantification of OCT Angiography Data using a Novel 3D Projection Artifacts Removal Algorithm
Author Affiliations & Notes
  • Yi-Sing Hsiao
    Optovue, Inc., Fremont, California, United States
  • Yulia Wolfson
    Optovue, Inc., Fremont, California, United States
  • Jing Tian
    Optovue, Inc., Fremont, California, United States
  • Xingwei Wang
    Optovue, Inc., Fremont, California, United States
  • Susan Luh
    Optovue, Inc., Fremont, California, United States
  • Ben K Jang
    Optovue, Inc., Fremont, California, United States
  • Qienyuan Zhou
    Optovue, Inc., Fremont, California, United States
  • Footnotes
    Commercial Relationships   Yi-Sing Hsiao, Optovue, Inc. (E); Yulia Wolfson, Optovue, Inc. (E); Jing Tian, Optovue, Inc. (E); Xingwei Wang, Optovue, Inc. (E); Susan Luh, Optovue, Inc. (E); Ben Jang, Optovue, Inc. (E); Qienyuan Zhou, Optovue, Inc. (E)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2017, Vol.58, 5998. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Yi-Sing Hsiao, Yulia Wolfson, Jing Tian, Xingwei Wang, Susan Luh, Ben K Jang, Qienyuan Zhou; Improved Visualization and Quantification of OCT Angiography Data using a Novel 3D Projection Artifacts Removal Algorithm. Invest. Ophthalmol. Vis. Sci. 2017;58(8):5998.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : To evaluate the performance of applying a novel 3D projection artifacts removal (PAR) algorithm to OCT Angiography (OCTA) volume data acquired with AngioVue® (Optovue, Inc., Fremont, CA).

Methods : AngioVue Retina 3×3 mm fovea-centered scans from 10 normal, 10 diabetic retinopathy (DR), and 10 age-related macular degeneration (AMD) subjects with choroidal neovascularization (CNV) were included. Analyses were performed on 1) OCTA B-scans; and 2) OCTA en face images generated from superficial plexus (ILM–IPL), deep plexus (IPL–OPL), outer retina (OPL–BM) and choriocapillaris (30 micron slab under BM) layers. 2D correlation coefficient was used to measure the similarity between superficial and deep plexus in all subjects. Area vessel density analysis was applied to deep plexus for normal and DR subjects. To evaluate CNV, grading of outer retinal slab images generated with and without PAR was performed by two independent human graders to evaluate identification of CNV (yes/no) and to estimate level of difficulty to identify CNV extent/boundary (easy/difficult).

Results : Following PAR, B-scans display less tailing, and projection artifacts are largely reduced in the outer retinal layer (Fig. 1). 2D correlation coefficient between superficial and deep plexus without PAR is 0.67 ± 0.05 (N = 30) and it lowers to 0.20 ± 0.12 with PAR, indicating that the two layers have more distinct capillary networks. With PAR, deep plexus area vessel density in the entire 3×3 mm image is 45.27 ± 2.92 % in normal and 41.60 ± 9.18 % in DR subjects. Following PAR application, CNV vessels remain connected and intact, and the CNV extent becomes easier to identify on the outer retinal slab (Fig. 2).

Conclusions : The novel 3D PAR algorithm applied to AngioVue® data has the potential to improve visualization for identification of pathologies, and may facilitate quantification of separate retinal vascular networks.

This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.

 

Fig 1. Comparison of (A) OCTA en face images and (B) B-scans from original OCTA and post-3D PAR volume of a normal eye. B-scan location is indicated by the dashed line in the upper left image in (A).

Fig 1. Comparison of (A) OCTA en face images and (B) B-scans from original OCTA and post-3D PAR volume of a normal eye. B-scan location is indicated by the dashed line in the upper left image in (A).

 

Fig 2. (A) Comparison of the outer retina in CNV cases, without (original) and with 3D PAR. Bar plots of (B) number of “Yes” cases when grader 1 and 2 (G1, G2) attempt to identify CNV from en face images without and with PAR, and (C) number of “Easy” cases when identifying the CNV extent.

Fig 2. (A) Comparison of the outer retina in CNV cases, without (original) and with 3D PAR. Bar plots of (B) number of “Yes” cases when grader 1 and 2 (G1, G2) attempt to identify CNV from en face images without and with PAR, and (C) number of “Easy” cases when identifying the CNV extent.

×
×

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.

×