Investigative Ophthalmology & Visual Science Cover Image for Volume 61, Issue 9
July 2020
Volume 61, Issue 9
Free
ARVO Imaging in the Eye Conference Abstract  |   July 2020
3D characterization of retinal microvasculature in OCT-Angiography images via Reeb analysis
Author Affiliations & Notes
  • Jiong Zhang
    Keck School of Medicine, University of Southern California, Los Angeles, California, United States
  • Amir kashani
    Keck School of Medicine, University of Southern California, Los Angeles, California, United States
  • Yonggang Shi
    Keck School of Medicine, University of Southern California, Los Angeles, California, United States
  • Footnotes
    Commercial Relationships   Jiong Zhang, None; Amir kashani, Carl Zeiss Meditec (R), Carl Zeiss Meditec (F); Yonggang Shi, None
  • Footnotes
    Support  This work was supported by NIH grants UH3NS100614, R21EY027879, U01EY025864, K08EY027006, P41EB015922, P30EY029220, Research Equipment from Carl Zeiss Meditec and Unrestricted Departmental Grant from Research to Prevent Blindness.
Investigative Ophthalmology & Visual Science July 2020, Vol.61, PB0050. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Jiong Zhang, Amir kashani, Yonggang Shi; 3D characterization of retinal microvasculature in OCT-Angiography images via Reeb analysis. Invest. Ophthalmol. Vis. Sci. 2020;61(9):PB0050.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : We have set up a high-quality 3D surface-based representation for the visualization and topology preservation of retinal OCTA microvasculature. To promote diseases studies, we aim to design novel techniques to extract more complete shape-based 3D metrics for microvascular characterization. Significant geometric and topological vessel biomarkers that can reflect the intrinsic microvascular changes need to be explored for analyzing relevant diseases. Statistical analysis based on those novel metrics will be performed on a clinical diabetic retinopathy (DR) dataset to show the benefits of using 3D surface-based approach.

Methods : We propose a novel microvascular shape modeling and analysis framework via Reeb graph and its analysis to explore rich 3D vessel geometry and topology. The Reeb graph is intuitively a graph of level contours constructed based on a dedicated feature function defined on surfaces. It is established to provide a full microvascular topology description and thus the intrinsic Reeb analysis can be performed to quantify local geometric and topological changes. In this work, a novel mesh geodesic distance feature is considered as the function for constructing Reeb graph. Afterwards, several valuable features such as 3D vessel skeleton length, bifurcations, ending points, vessel calibers, and capillary loops are extracted to characterize the 3D retinal vascular networks.

Results : The proposed Reeb analysis framework has been evaluated on a DR dataset including 20 normal controls and 20 proliferative diabetic retinopathy subjects to find the statistical significance of its extracted 3D metrics at differentiating different DR stages. Fig. 1 shows examples of several 3D Reeb-based vessel metrics. High statistical significance (with p<0.05) has been found on ten 3D microvascular metrics for DR discrimination. The mean caliber on large vessels and the ending points show much higher statistical power (with p<0.001) at differentiating the two DR stages.

Conclusions : The proposed Reeb analysis framework shows its effectiveness in extracting useful 3D metrics to characterize microvascular variations. It provides great potential of exploring 3D vessel geometry and topology to support subsequent disease diagnosis.

This is a 2020 Imaging in the Eye Conference abstract.

 

Figure 1. Examples of OCTA vessel metrics extracted via Reeb analysis.

Figure 1. Examples of OCTA vessel metrics extracted via Reeb analysis.

×
×

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.

×