June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Precisely Mapping Choroidal Tumour Margins
Author Affiliations & Notes
  • Roderick O'Day
    Centre for Eye Research Australia Ltd, East Melbourne, Victoria, Australia
    Ophthalmology - Department of Surgery, The University of Melbourne Faculty of Medicine Dentistry and Health Sciences, Melbourne, Victoria, Australia
  • Xavier Hadoux
    Centre for Eye Research Australia Ltd, East Melbourne, Victoria, Australia
  • Darvy Dang
    Centre for Eye Research Australia Ltd, East Melbourne, Victoria, Australia
  • Peter van Wijngaarden
    Centre for Eye Research Australia Ltd, East Melbourne, Victoria, Australia
    Ophthalmology - Department of Surgery, The University of Melbourne Faculty of Medicine Dentistry and Health Sciences, Melbourne, Victoria, Australia
  • Footnotes
    Commercial Relationships   Roderick O'Day None; Xavier Hadoux None; Darvy Dang None; Peter van Wijngaarden None
  • Footnotes
    Support  H&L Hecht Trust; Yulgilbar Foundation; University of Melbourne Hector Maclean Fellowship; CERA receives operational and infrastructure support from the Victorian Government.
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 456. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Roderick O'Day, Xavier Hadoux, Darvy Dang, Peter van Wijngaarden; Precisely Mapping Choroidal Tumour Margins. Invest. Ophthalmol. Vis. Sci. 2023;64(8):456.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : Assessing the malignant potential of choroidal tumours relies on accurate mapping of tumour margins. Standalone images taken using colour fundus photography (CFP), ultrawidefield scanning laser ophthalmoscopy (UWF SLO) or enhanced-depth imaging optical coherence tomography (EDI-OCT) are most commonly used to do this. Each of these methods has limitations: conventional EDI-OCT can only image tumours up to ~9mm in basal diameter and can be difficult to interpret without en face projections; tumour margin delineation with CFP or UWF SLO can be confounded by pigmentation and overlying retinal and pigment epithelial change. Our group has developed machine learning tool to co-register retinal images captured on different modalities, enabling assessment of the concordance between modalities in tumour margin mapping. Preliminary data using this technology is presented here.

Methods : Non-mydriatic retinal photography (Canon CR6-45NM), UWF SLO (Optos), EDI-OCT (Spectralis, Heidelberg) images of choroidal naevi from 9 people were co-registered using a software application developed by our team. Tumour margins were annotated independently for each image type by an ocular oncologist masked to patient identity. Dense raster EDI-OCT B-scans were annotated to delineate cross-sectional tumour measurements. These measurements were rendered as an en face projection for tumour margin and area determination.

Results : Mean tumour area was 21.9 mm2 (SD=16.6 mm2) based on EDI-OCT B-scan measurements. Tumour area measurements were, on average, larger for EDI-OCT (ground truth) than for CFP and UWF SLO: absolute difference 2.3 mm2 (SD = 4.7 mm2) and 2.7 mm2 (SD = 3.1 mm2), respectively. These differences were greater than 1 mm2 (pre-defined threshold of clinical relevance) in 4 out of 9 cases (45%). Figure 1 shows the discrepancy in tumour margin and area for a representative choroidal naevus between co-registered CFP, UWF SLO and OCT images.

Conclusions : There is discrepancy between tumour margin boundaries defined by EDI-OCT (the reference standard) and CFP as well as UWF SLO. Co-registered images with en-face projection of EDI-OCT B scans may allow more accurate delineation of tumour margins. Further studies are needed.

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

 

Figure 1.(A) Representative choroidal naevus on CFP, UWF SLO and IR/OCT imaging.(B) Manual annotation of the naevus for each imaging modality. *Annotation for the IR/OCT image is based on margins labelled on EDI-OCT B-scans (C) Comparison of lesion areas

Figure 1.(A) Representative choroidal naevus on CFP, UWF SLO and IR/OCT imaging.(B) Manual annotation of the naevus for each imaging modality. *Annotation for the IR/OCT image is based on margins labelled on EDI-OCT B-scans (C) Comparison of lesion areas

×
×

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.

×