June 2021
Volume 62, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2021
Modelling structure-function relationship in patients undergoing screening for hydroxychloroquine toxicity
Author Affiliations & Notes
  • Gopal Jayakar
    Unit on Clinical Investigation of Retinal Disease, National Eye Institute, Bethesda, Maryland, United States
  • Tharindu De Silva
    Unit on Clinical Investigation of Retinal Disease, National Eye Institute, Bethesda, Maryland, United States
  • Catherine A Cukras
    Unit on Clinical Investigation of Retinal Disease, National Eye Institute, Bethesda, Maryland, United States
  • Footnotes
    Commercial Relationships   Gopal Jayakar, None; Tharindu De Silva, None; Catherine Cukras, None
  • Footnotes
    Support  NEI Intramural Research Support
Investigative Ophthalmology & Visual Science June 2021, Vol.62, 2193. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Gopal Jayakar, Tharindu De Silva, Catherine A Cukras; Modelling structure-function relationship in patients undergoing screening for hydroxychloroquine toxicity. Invest. Ophthalmol. Vis. Sci. 2021;62(8):2193.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : Purpose: The American Academy of Ophthalmology issued recommendations on annual screening of patients using hydroxychloroquine (HCQ) including functional testing with a Humphrey 10-2 visual field and structural evaluation using spectral domain optical coherence tomography (SD-OCT). This study investigates the association between retina microstructure and visual function in a cohort with long-term HCQ use.

Methods : Methods: The case-control study included 87 participants (173 eyes, mean age 60, 93% female, 55 without toxicity, 32 with toxicity) and SD-OCT imaging (Heidelberg Spectralis) and Humphrey 10-2 visual field (VF) testing were obtained during each visit. Quantitative metrics computed using OCT included total and outer retina thickness, minimum intensity (along A-scans within total and outer retina), and ellipsoid zone (EZ) loss. Enface 2D maps of OCT metrics were sampled in 68 circular regions (288 µm diameter) corresponding to VF loci. First, correlations were analyzed between each OCT metric and corresponding VF sensitivity by locus using a linear model. Second, ability to predict VF sensitivity from the combination of OCT metrics were analyzed using a multi-variate, non-linear random forest regression using leave-one-out cross-validation.

Results : Results: In linear regression, the strongest relationship with VF sensitivity in the parafoveal ring was found with EZ loss, with R2=0.60. Linear analyses using total and outer retinal thickness revealed R2=0.57 and 0.40, respectively, while total and outer retinal minimum intensity yielded R2=0.10 and 0.22, respectively. Using the multivariate model that included all five OCT metrics, the VF prediction markedly improved with R2 = 0.79 and measured root mean squared error of 3.5 dB. Feature importance analysis identified EZ loss as the most relevant predictor.

Conclusions : Conclusions: Multiple OCT-derived quantitative metrics exhibited high correlation with VF sensitivity with the multivariate model achieving a high-degree of predictability of visual function from combined structural metrics. The results indicate a confluence of features available in SD-OCT and VF demonstrating the potential predictability of function from structure. With further analysis, this data could provide evidence to streamline the early detection of toxicity.

This is a 2021 ARVO Annual Meeting abstract.

×
×

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.

×