Predicting the need for laser treatment in retinopathy of prematurity using computer-assisted quantitative vascular analysis.

Published

Journal Article

PURPOSE: To investigate whether parameters of retinal vascular dilation and/or tortuosity calculated by ROPtool software can help to predict, prior to diagnosis of plus disease, whether laser will be needed for treatment of retinopathy of prematurity (ROP). METHODS: Video indirect ophthalmoscopy recordings were obtained during routine ROP examinations. One posterior pole still image of one eye was selected from each examination for each infant. ROPtool was used to calculate vessel dilation, tortuosity, and the sum of adjusted indices (SAI, combining dilation and tortuosity) for each image. The following values were calculated for each vessel characteristic: maximum value from any one vessel across all examinations, largest increase per week in maximum value, highest mean value of all vessels from any one examination, and largest increase per week in mean value. These parameters were compared between infants who eventually received laser and those who did not. RESULTS: Medians for maximum tortuosity indices were, for the eventual laser group (n = 28), 8.92 tortuosity units and, for the no laser group (n = 56), 6.87 (P < 0.001). Medians for highest mean tortuosity indices were 4.95 tortuosity units for the eventual laser group and 3.66 for the no laser group (P < 0.001). Parameters involving dilation and SAI did not differ significantly between groups. In multivariable analysis, highest mean tortuosity was associated with need for laser (P = 0.003). CONCLUSIONS: ROPtool analysis of tortuosity parameters from indirect ophthalmoscopy images can help predict need for laser in ROP.

Full Text

Duke Authors

Cited Authors

  • Wu, KY; Wallace, DK; Freedman, SF

Published Date

  • April 2014

Published In

Volume / Issue

  • 18 / 2

Start / End Page

  • 114 - 119

PubMed ID

  • 24698605

Pubmed Central ID

  • 24698605

Electronic International Standard Serial Number (EISSN)

  • 1528-3933

Digital Object Identifier (DOI)

  • 10.1016/j.jaapos.2013.11.025

Language

  • eng

Conference Location

  • United States