June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
Evaluation of an OCT B-scan of interest tool in glaucomatous eyes
Author Affiliations & Notes
  • Gary C Lee
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Katherine Makedonsky
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Laura Tracewell
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Sophia Yu
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Mary K Durbin
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Noelleisha Graves
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Charles Wu
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Iryna Falkenstein
    Glaucoma Specialists of San Francisco, Oakland, California, United States
  • Todd Severin
    East Bay Eye Center, San Ramon, California, United States
  • Thomas Callan
    Carl Zeiss Meditec Inc, Dublin, California, United States
  • Footnotes
    Commercial Relationships   Gary Lee, Carl Zeiss Meditec, Inc (E); Katherine Makedonsky, Carl Zeiss Meditec, Inc (E); Laura Tracewell, Carl Zeiss Meditec, Inc (E); Sophia Yu, Carl Zeiss Meditec, Inc (E); Mary Durbin, Carl Zeiss Meditec, Inc (E); Noelleisha Graves, Carl Zeiss Meditec, Inc (E); Charles Wu, Carl Zeiss Meditec, Inc (C); Iryna Falkenstein, Carl Zeiss Meditec, Inc (C); Todd Severin, Carl Zeiss Meditec, Inc (C); Thomas Callan, Carl Zeiss Meditec, Inc (E)
  • Footnotes
    Support  None
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 996. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Gary C Lee, Katherine Makedonsky, Laura Tracewell, Sophia Yu, Mary K Durbin, Noelleisha Graves, Charles Wu, Iryna Falkenstein, Todd Severin, Thomas Callan; Evaluation of an OCT B-scan of interest tool in glaucomatous eyes. Invest. Ophthalmol. Vis. Sci. 2021;62(8):996.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : B-scans from optical coherence tomography (OCT) cube scans are commonly reviewed during clinical evaluation of the macula. A deep learning algorithm was previously trained to predict if a given B-scan might be flagged “of interest” based on ground truth labels from retinal specialists in healthy and eyes with retinal pathology or disease1. In this preliminary study, the performance of a B-scan of interest tool was compared to a clinical assessment of OCT cube data in healthy and glaucomatous eyes.

Methods : A B-scan of interest algorithm was previously trained and tested on 76,544 B-scans (598 subjects) and 25,600 B-scans (200 subjects), respectively2. Retrospective OCT data were analyzed from an ongoing visual field (VF) study including 19 eyes of 19 patients in both healthy and glaucoma groups, using CIRRUS™ 5000 HD-OCT (ZEISS, Dublin, CA) and HFA3 (ZEISS, Dublin, CA). At each visit, a qualified Macula 512x128 cube scan was used for inference. Mean Deviation (MD) of a SITA Standard 24-2 VF was used to indicate disease severity. A trained grader reviewed each cube and assessed if either a) <10% or b) >=10% of total B-scans of interest should be flagged.

Results : Mean age was 56.4 (standard deviation, SD: 7.7; range: 44.3 to 73.1) years and 73.4 (SD: 9.5; range 60.6 to 97.9) years for healthy and glaucomatous eyes (p<0.001). Mean VF MD was 0.66 (SD: 0.91; range: -0.62 to 2.62) dB and -6.39 (SD: 6.67; range: -23.16 to 1.63) dB in healthy and glaucomatous eyes (p<0.001). Observed grader and algorithm agreement (95% confidence interval) was 100.0 (82.4,100.0)% and 79.0 (54.4, 94.0)% in healthy and glaucomatous eyes (see Fig 1). B-scans from the four disagreement cases included the presence of unusual retinal curvature (two cases), unusual contrast in the vitreous, and general false positives with inference scores near the algorithm cut-off (see Fig 2).

Conclusions : The findings in this study suggest excellent agreement in healthy eyes and reasonable agreement in glaucomatous eyes for detecting the presence of pathologies typically associated with retinal disease. As a result, a B-scan of interest tool may be a useful workflow aid to identify retinal disease co-morbidities in glaucomatous eyes.

References
[1] Yu et al. IOVS 2020; 61(9): Abstract PB0085.
[2] Ren et al. IOVS 2020; 61(7): Abstract 1635.

This is a 2021 ARVO Annual Meeting abstract.

 

Figure 1. Observed agreement between algorithm and grader

Figure 1. Observed agreement between algorithm and grader

 

Figure 2. Examples of disagreement between algorithm and grader

Figure 2. Examples of disagreement between algorithm and grader

×
×

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.

×