April 2010
Volume 51, Issue 13
Free
ARVO Annual Meeting Abstract  |   April 2010
Comparability of Digital vs. Film Grading of Diabetic Retinopathy Severity in DCCT/EDIC
Author Affiliations & Notes
  • L. D. Hubbard
    Ophth & Vis Sci, University of Wisconsin-Madison, Madison, Wisconsin
  • W. Sun
    Biostatistics Center, George Washington University, Rockville, Maryland
  • P. A. Cleary
    Biostatistics Center, George Washington University, Rockville, Maryland
  • D. P. Hainsworth
    Ophthalmology, University of Missouri, Columbia, Missouri
  • Q. Peng
    Ophth & Vis Sci, University of Wisconsin-Madison, Madison, Wisconsin
  • R. A. Susman
    Ophth & Vis Sci, University of Wisconsin-Madison, Madison, Wisconsin
  • R. P. Danis
    Ophth & Vis Sci, University of Wisconsin-Madison, Madison, Wisconsin
  • Diabetes Control & Complications Trial/Epidemiology of Diabetes Interventions & Complications Study
    Ophth & Vis Sci, University of Wisconsin-Madison, Madison, Wisconsin
  • Footnotes
    Commercial Relationships  L.D. Hubbard, None; W. Sun, None; P.A. Cleary, None; D.P. Hainsworth, None; Q. Peng, None; R.A. Susman, None; R.P. Danis, None.
  • Footnotes
    Support  NIH Grant NO1-DK-2-2206 (96-S03)
Investigative Ophthalmology & Visual Science April 2010, Vol.51, 4674. doi:
  • Views
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      L. D. Hubbard, W. Sun, P. A. Cleary, D. P. Hainsworth, Q. Peng, R. A. Susman, R. P. Danis, Diabetes Control & Complications Trial/Epidemiology of Diabetes Interventions & Complications Study; Comparability of Digital vs. Film Grading of Diabetic Retinopathy Severity in DCCT/EDIC. Invest. Ophthalmol. Vis. Sci. 2010;51(13):4674.

      Download citation file:


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

      ×
  • Supplements
Abstract

Purpose: : To compare diabetic retinopathy (DR) severity levels when switching from film (F) to digital (D) retinal images in a long-term multi-center study.Introduction: Long-term studies require consistent measurements across clinics and over time for valid conclusions. During the past 25 years, the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Study (DCCT/EDIC) have used 7-field stereo color film photos, per ETDRS protocol. Obsolescence of film photography has now forced EDIC to transition to digital imaging. Thus, EDIC conducted an ancillary study of the potential impact of this transition.

Methods: : At 20 clinics during 2007-2009, 310 type-1 diabetic participants, spanning the ETDRS scale, were imaged with both F and D cameras. Central grading independently assessed DR severity levels from F and D (tonally standardized at Reading Center). For retinopathy outcomes with >10% prevalence, we had ≥85% power to detect an agreement kappa (K) of 0.7 to 0.9.

Results: : Comparing DR severity gradings, D vs. F yielded weighted K of 0.74 for eye level and 0.73 for patient level. Digital grading did not systematically under- or overestimate levels (McNemar overall bias test, P=0.14). Regarding major DR outcomes (e.g., 3-step progression, mild/moderate/severe NPDR/PDR or worse), D vs.F kappas were 0.7 to 0.96 ("substantial" to "perfect"). Agreement proportion was 86% - 99%, sensitivity 75% - 98%, and specificity 72% - 99%. Major study conclusions were similar with D vs. F gradings (e.g., adjusted odds reductions with intensive DM therapy for PDR at EDIC year 10: 65.5% [D] vs. 64.3% [F]). Finally, Kappas for major DR outcomes were comparable across all clinics (Cochran’s homogeneity test, all P>0.2).

Conclusions: : This is the first multi-center, long-term study comparing DR severity with D vs. F. We show that ETDRS scores from D and F are similar, and that the high reliability of DR evaluations seen historically in DCCT/EDIC will continue with digital photography.

Clinical Trial: : www.clinicaltrials.gov NCT00360893

Keywords: diabetic retinopathy • clinical research methodology • imaging/image analysis: clinical 
×
×

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.

×