May 2004
Volume 45, Issue 13
Free
ARVO Annual Meeting Abstract  |   May 2004
Predicting Visual Field Loss in Ocular Hypertensive Patients Using Wavelet–Fourier Analysis (WFA) of GDx Scanning Laser Polarimetry
Author Affiliations & Notes
  • P. Gunvant
    Dept of Psychological and Brain Sciences,
    University of Louisville, Louisville, KY
  • Y. Zheng
    Dept of Psychological and Brain Sciences,
    University of Louisville, Louisville, KY
  • A. Kotecha
    Glaucoma Research Unit, Moorfields Eye Hospital, London, United Kingdom
  • D.F. Garway–Heath
    Glaucoma Research Unit, Moorfields Eye Hospital, London, United Kingdom
  • E.A. Essock
    Dept of Psychological and Brain Sciences,
    Dept of Ophthalmology &Visual Science,
    University of Louisville, Louisville, KY
  • Footnotes
    Commercial Relationships  P. Gunvant, None; Y. Zheng, UofL P; A. Kotecha, None; D.F. Garway–Heath, Laser Diagnostic Technologies R; E.A. Essock, UofL P.
  • Footnotes
    Support  Glaucoma Research Foundation
Investigative Ophthalmology & Visual Science May 2004, Vol.45, 5504. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      P. Gunvant, Y. Zheng, A. Kotecha, D.F. Garway–Heath, E.A. Essock; Predicting Visual Field Loss in Ocular Hypertensive Patients Using Wavelet–Fourier Analysis (WFA) of GDx Scanning Laser Polarimetry . Invest. Ophthalmol. Vis. Sci. 2004;45(13):5504.

      Download citation file:


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

      ×
  • Supplements
Abstract

Abstract: : Purpose: Our prior efforts have sought to characterize the "double–hump" pattern of retinal nerve fiber layer (RNFL) thickness in terms of shape parameters derived from a Fourier analysis and more recently from an improved method termed Wavelet–Fourier analysis (WFA – Zheng et al., JCIS, 2003). In the present work we use the approach to predict subsequent conversion of ocular hypertension (OHT) to glaucoma. Methods:A total of 96 subjects were followed–up during the period of April 1996 to July 2003 at Moorfields Eye Hospital of which 34 were OHT non–converters (OHTnc), 19 OHT converters (OHTc), 19 glaucoma, and 24 ocular normals, classified on the basis of visual field. RNFL scans using a GDx Nerve Fiber Analyzer (Laser Diagnostic Technologies, Inc.), Humphrey visual fields, and ophthalmoscopic exams were obtained in follow–up. Subjects were classified as converters on the basis of reproducible field defects using AGIS criteria. RNFL thickness estimates were obtained for 28 sectors around the disc at 1.75 disc diameters. The WFA classifier was trained and derived using a support vector machine from the normal and glaucoma subjects and then was applied to the 53 OHT patients to predict which would convert based on a single scan obtained prior to the visual field conversion (AGIS) (mean 7.4 months, range 1–24). The WFA method involves application of a two–level Discrete Wavelet Transform (DWT), a Fourier transform (FFT) of the detailed difference information, and application of Principal Component Analysis (PCA) to the DWT and FFT coefficients. A support vector classifier (SVC) was used to classify eyes based on these resultant features. Sensitivity, specificity and ROC area was calculated for the detection of the patients who subsequently converted. Results: Sensitivity/Specificity/ROC area was 0.71/0.77/0.83 when the WFA was applied to classify OHTc and OHTnc subjects and 0.67/0.82/0.75 when the converting and fellow non–converting eyes of the OHTc patients were used. Conclusion:WFA analysis, trained on a separate population of normals and glaucoma patients, can predict impending visual field progression in OHT subjects with moderate ability.

Keywords: imaging/image analysis: clinical • visual fields • clinical (human) or epidemiologic studies: systems/equipment/techniques 
×
×

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.

×