April 2010
Volume 51, Issue 13
Free
ARVO Annual Meeting Abstract  |   April 2010
Spectral Domain OCT Segmentation Accuracy in Monkeys
Author Affiliations & Notes
  • K. B. McIntyre
    Comparative Ophthalmic Research Labs,
    UW-Madison, Madison, Wisconsin
  • C. A. Rasmussen
    Ophthalmology and Visual Sciences,
    UW-Madison, Madison, Wisconsin
  • A. K. Goulding
    Comparative Ophthalmic Research Labs,
    UW-Madison, Madison, Wisconsin
  • V. Bantseev
    Covance Laboratories, Madison, Wisconsin
  • J. N. Ver Hoeve
    Comparative Ophthalmic Research Labs,
    UW-Madison, Madison, Wisconsin
  • P. L. Kaufman
    Ophthalmology and Visual Sciences,
    UW-Madison, Madison, Wisconsin
  • B. J. Christian
    Covance Laboratories, Madison, Wisconsin
  • T. M. Nork
    Comparative Ophthalmic Research Labs,
    UW-Madison, Madison, Wisconsin
  • Footnotes
    Commercial Relationships  K.B. McIntyre, None; C.A. Rasmussen, None; A.K. Goulding, None; V. Bantseev, None; J.N. Ver Hoeve, None; P.L. Kaufman, None; B.J. Christian, None; T.M. Nork, None.
  • Footnotes
    Support  EY02698, RPB, OPREF, P30 EY016665
Investigative Ophthalmology & Visual Science April 2010, Vol.51, 4401. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      K. B. McIntyre, C. A. Rasmussen, A. K. Goulding, V. Bantseev, J. N. Ver Hoeve, P. L. Kaufman, B. J. Christian, T. M. Nork; Spectral Domain OCT Segmentation Accuracy in Monkeys. Invest. Ophthalmol. Vis. Sci. 2010;51(13):4401.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose: : To evaluate scan segmentation accuracy and reproducibility using spectral domain optical coherence tomography (OCT) in monkeys; correlate retinal thickness values with fluorescein angiography (FA) leakage scores in a laser model of choroidal neovascularization (CNV) and compare time domain (Stratus®) with spectral domain (Cirrus®) results.

Methods: : Normal eyes from 40 cynomolgus monkeys and lasered eyes from 6 cynomolgus monkeys were studied. A Cirrus or Stratus OCT was used to acquire macular thickness cube scans or macular thickness map scans, respectively. The CNV laser model animals had FA performed on the same day as the OCT scans. Using the Cirrus software algorithm, segmentation lines were manually corrected in the set of CNV model monkeys.

Results: : Average thickness and volume values vs FA Grade 4 leakage from CNV model animals resulted in R2 values of 0.54 and 0.53.There was a significant, though small (5.4 µm), increase in thickness values post manual correction of segmentation errors, after which R2 was 0.66. Cirrus values (mean±sem) for central subfield thickness (µm), volume (mm3) and average thickness (µm) in 20 normal animals (40 eyes) were 232.7±3.88; 10.5±0.11; 291.2±3.02 respectively. Stratus central subfield thickness (µm), volume (mm3) in a gender matched group were 172.8±3.84; 7.2±0.09 respectively. Age did not significantly affect thickness values (p=0.50). Signal strength average values were 9.83±0.09 for the Cirrus and 9.95±0.03 for the Stratus. Cirrus intra-session COV for the 12 CNV model eyes ranged from 0.011 to 0.031; for 12 normal eyes 0.009 to 0.017; normal animals with the Stratus 0.004 to 0.013.

Conclusions: : Cirrus tended to underestimate CNV model thickness values. Manual segmentation of Cirrus scans resulted in a small increase in average thickness values and a mild improvement in correlation with FA grade. Stratus and Cirrus show comparable intra-session variability in normal eyes. Segmentation differences resulted in ~ 60µm average difference between Cirrus and Stratus values. Reproducible thickness values can be obtained in monkeys with both the Cirrus and Stratus instruments.

Keywords: imaging methods (CT, FA, ICG, MRI, OCT, RTA, SLO, ultrasound) • imaging/image analysis: non-clinical • retina 
×
×

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.

×