June 2023
Volume 64, Issue 8
Open Access
ARVO Annual Meeting Abstract  |   June 2023
Defining the macular xanthophyll pigment (MXP) distribution pattern by automating centration in dual wavelength autofluorescence(2WAF): ALSTAR2 baseline cohort
Author Affiliations & Notes
  • Kenneth R Sloan
    Ophthalmology, The University of Alabama at Birmingham Heersink School of Medicine, Birmingham, Alabama, United States
  • Andreas Berlin
    The University of Alabama at Birmingham School of Public Health, Birmingham, Alabama, United States
  • Mark E. Clark
    Ophthalmology, The University of Alabama at Birmingham Heersink School of Medicine, Birmingham, Alabama, United States
  • Thomas A. Swain
    The University of Alabama at Birmingham School of Public Health, Birmingham, Alabama, United States
  • Gerald McGwin
    The University of Alabama at Birmingham School of Public Health, Birmingham, Alabama, United States
  • Jason Neal Crosson
    Ophthalmology, The University of Alabama at Birmingham Heersink School of Medicine, Birmingham, Alabama, United States
  • Cynthia Owsley
    Ophthalmology, The University of Alabama at Birmingham Heersink School of Medicine, Birmingham, Alabama, United States
  • Christine A. Curcio
    Ophthalmology, The University of Alabama at Birmingham Heersink School of Medicine, Birmingham, Alabama, United States
  • Footnotes
    Commercial Relationships   Kenneth Sloan None; Andreas Berlin Dr. Werner Jackstädt foundation, Code F (Financial Support); Mark Clark None; Thomas Swain None; Gerald McGwin None; Jason Crosson None; Cynthia Owsley NEI R01EY029595 , Code F (Financial Support); Christine Curcio NEI R01EY029595, NIH R01EY027948, Genentech/Hoffman LaRoche, Code F (Financial Support)
  • Footnotes
    Support  R01EY029595; R01EY027948
Investigative Ophthalmology & Visual Science June 2023, Vol.64, 5021. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Kenneth R Sloan, Andreas Berlin, Mark E. Clark, Thomas A. Swain, Gerald McGwin, Jason Neal Crosson, Cynthia Owsley, Christine A. Curcio; Defining the macular xanthophyll pigment (MXP) distribution pattern by automating centration in dual wavelength autofluorescence(2WAF): ALSTAR2 baseline cohort. Invest. Ophthalmol. Vis. Sci. 2023;64(8):5021.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose : The distribution of MXP can be determined objectively with 2WAF. To investigate the role of MXP in vision and age related macular degeneration (AMD) progression, we developed tools for centering the MXP distribution. This problem is nearly trivial for normal eyes when the MXP distribution has a clear central peak, but is difficult in the face of common deviations from this pattern.

Methods : Participants in the ALSTAR2 baseline cohort (NCT04112667) were assessed by the AREDS 9-step grading system, Optical coherence tomography (OCT), and 2WAF images (MPOD module; both Heidelberg Engineering) underwent a quality review. Custom FIJI plugins centered the MXP distribution by 6 different techniques: MANUAL (trained observer); FOVEA (rise of the external limiting membrane (ELM), previously determined manually by a trained observer in OCT); MAX (maximum in a central region); CENTROID (weighted average in a central region; CONTOUR (weighted average in the convex hull of a polygon delineating half-maximum intensities)l; HILLCLIMB (iterates CENTROID by adjusting the center and re-computing a centroid until convergence).

Results : Of 1064 eyes, 178 eyes were excluded due to lack of MPOD imaging, insufficient quality, or pathologies that distorted the fovea. 12 eyes were omitted because they produced clear outliers on one or more of the automatic methods, usually traced to image quality. HILLCLIMB and CONTOUR showed best overall agreement with the MANUAL ground truth. FOVEA, CENTROID, and MAX showed lowest overall agreement (see Table).

Conclusions : The automatic tools are reliable and robust. MANUAL centration was used as ground truth – but it is labor intensive and subjective. FOVEA is also subjective and not based on the MXP distribution; MAX is acceptable when there is a clear central peak, but fails for central valleys. CENTROID is heavily influenced by the initial guess. CONTOUR is more global and tries to avoid problems caused by pathology; HILLCLIMB produces the best agreement with MANUAL as measured by both the mean and maximum disagreement. Determination of the center of the MXP distribution is necessary, as it often does not align with the anatomical fovea. Our future studies will use HILLCLIMB and CONTOUR.

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

 

×
×

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.

×