April 2009
Volume 50, Issue 13
ARVO Annual Meeting Abstract  |   April 2009
Heterogeneity of Gene Expression (GEP) Signature in Melanocytic Uveal Tumors Sampled at Two Sites by FNAB
Author Affiliations & Notes
  • J. J. Augsburger
    Ophthalmology, University of Cincinnati, Cincinnati, Ohio
  • Z. M. S. Corrêa
    Ophthalmology, University of Cincinnati, Cincinnati, Ohio
  • C. C. Simões
    Ophthalmology, University of Cincinnati, Cincinnati, Ohio
  • J. W. Harbour
    Ophthalmology, Washington University School of Medicine, St Louis, Missouri
  • Footnotes
    Commercial Relationships  J.J. Augsburger, None; Z.M.S. Corrêa, None; C.C. Simões, None; J.W. Harbour, None.
  • Footnotes
    Support  NEI grant EY02687,NCI grant R01CA125970, Kling Family Foundation,Tumori Foundation,Barnes-Jewish Hosp Foundation, Horncrest Foundation,Research to Prevent Blindness,Quest for Vision Fund-U Cincinnati
Investigative Ophthalmology & Visual Science April 2009, Vol.50, 6183. doi:
  • Views
  • Share
  • Tools
    • Alerts
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      J. J. Augsburger, Z. M. S. Corrêa, C. C. Simões, J. W. Harbour; Heterogeneity of Gene Expression (GEP) Signature in Melanocytic Uveal Tumors Sampled at Two Sites by FNAB. Invest. Ophthalmol. Vis. Sci. 2009;50(13):6183.

      Download citation file:

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

  • Supplements

Purpose: : To determine the frequency of genetic heterogeneity of melanocytic uveal tumors (defined by discordant GEP signatures of specimens obtained from different tumor sites) sampled in 2 distinct sites by fine needle aspiration biopsy (FNAB).

Methods: : Samples from 39 melanocytic uveal tumors obtained by FNAB from 2 different sites within each tumor were analyzed by GEP for the presence of Class 1 (favorable) or Class 2 (unfavorable) signature. GEP class assignment was performed using a weighted voting (WV) learning algorithm calibrated against a well-characterized dataset of 30 tumors with known GEP signature. Confidence regarding each classification was expressed as a value between -1 (complete confidence in Class 1 assignment) through 0 (no confidence in assignment to either class) to +1 (complete confidence in Class 2 assignment). Cases were arbitrarily regarded by WV confidence ≥ ± 0.5 as high confidence, those with WV confidence < ± 0.5 but ≥ ± 0.2 as medium confidence, and those with WV confidence < ± 0.2 as low confidence.

Results: : Tumor thickness ranged from 1.8 to 12.0 mm (mean = 5.9 mm, SD = 2.8 mm). Pre-FNAB clinical diagnosis was choroidal melanoma in 26 cases, ciliary body melanoma in 9, and choroidal nevus versus melanoma in 4. Twenty-one tumors were categorized as Class 1 at each of the two sampled tumor sites, 13 were categorized as Class 2 at each sampled site, and 3 (7.7%) were discordant (i.e., Class 1 at one site and Class 2 at the other site, both with at least medium confidence). One tumor that exhibited medium Class 1 WV confidence at one site and low Class 2 WV confidence at the other was categorized as Class 1, and another tumor categorized as indeterminate (i.e., WV confidence was extremely low at each site). Both of these were regarded as concordant. Two of 12 tumors with thickness ≤ 3.5 mm, 1 of 15 tumors with thickness > 3.5 but ≤ 7 mm, and none of the 12 tumors with thickness > 7 mm exhibited discordant GEP results at the 2 sites. The homogeneous tumors mean thickness was 6.0 mm (SD = 2.9 mm) while that of heterogeneous tumors was 4.4 mm (SD = 1.8 mm). This difference was not statistically significant (P = 0.3, t test).

Keywords: melanoma • gene/expression • oncology 

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.