June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Deep-learning-aided vascular tortuosity analysis in eyes with idiopathic epiretinal membrane
Author Affiliations & Notes
  • KEITA BABA
    Aichi Ika Daigaku, Nagakute, Aichi, Japan
  • Footnotes
    Commercial Relationships   KEITA BABA None
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 272. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      KEITA BABA; Deep-learning-aided vascular tortuosity analysis in eyes with idiopathic epiretinal membrane. Invest. Ophthalmol. Vis. Sci. 2023;64(8):272.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : Vascular tortuosity is one of the biomarkers of epiretinal membrane (ERM). We analyzed vascular tortuosity before and after surgery in eyes with ERM using a deep-learning approach.

Methods : A retrospective case-control study included eyes with idiopathic ERM. For a control group, eyes with idiopathic macular hole (MH) were enrolled. The eyes with other retinal diseases were excluded. All eyes underwent a 6x6-mm spectral domain (SD) OCT angiography (OCTA) macular scan (Avanti, Optovue, or Ciruss-5000, Zeiss) before and after the vitrectomy with the internal limiting membrane (ILM) peeling. A previously developed and validated deep-learning network by Gao et al. (Ophthalmology Science, 2022) was utilized to extract and classify arteries and veins in the scan area. The output vessels were skeletonized and used to calculate vascular tortuosity. The tortuosity value was mapped to the binarized artery-vein OCTA map (Figure).

Results : Twenty-four eyes with ERM and 10 eyes with MH were analyzed. Before the surgery, vascular tortuosity was significantly greater in eyes with ERM than in eyes with MH (1.14 ± 0.028 vs. 1.11 ± 0.020, P = 0.0095 by t-test). After the surgery, vascular tortuosity in eyes with ERM significantly decreased to 1.12 ± 0.019 (P = 0.0017 by paired t-test), while no change was observed in eyes with MH (P = 0.75). When arteries and veins were separately analyzed in eyes with ERM, there was no difference in preoperative vascular tortuosity between arteries and veins (P = 0.38 by paired t-test). In contrast, postoperative vascular tortuosity in arteries was significantly lower than in veins (P = 0.019).

Conclusions : A deep-learning method can reveal the vascular tortuosity changes between artery and vein after surgery in the eyes with ERM, suggesting it will become a useful tool to evaluate ERM surgery in releasing retinal traction.

This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.

 

Fig 1. Preoperative and postoperative vessel tortuosity map in eyes with epiretinal membrane (ERM). Deep-learning network analyzes the vessel tortuosity in 6x6-mm OCT angiography images. Preoperative mean venous tortuosity was 1.16, arterial tortuosity was 1.20, and total tortuosity was 1.18. Postoperative mean venous tortuosity was 1.15, arterial tortuosity was 1.13, and total tortuosity was 1.14.

Fig 1. Preoperative and postoperative vessel tortuosity map in eyes with epiretinal membrane (ERM). Deep-learning network analyzes the vessel tortuosity in 6x6-mm OCT angiography images. Preoperative mean venous tortuosity was 1.16, arterial tortuosity was 1.20, and total tortuosity was 1.18. Postoperative mean venous tortuosity was 1.15, arterial tortuosity was 1.13, and total tortuosity was 1.14.

×
×

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.

×