Automated grading for diabetic retinopathy

a large-scale audit using arbitration by clinical experts

Alan D. Fleming, Keith A. Goatman, Sam Philip, Gordon J. Prescott, Peter F. Sharp, John A. Olson*

*Corresponding author for this work

Research output: Contribution to journalArticle

65 Citations (Scopus)

Abstract

Background/aims Automated grading software has the potential to reduce the manual grading workload within diabetic retinopathy screening programmes. This audit was undertaken at the request of Scotland's National Diabetic Retinopathy Screening Collaborative to assess whether the introduction of automated grading software into the national screening programme would be safe, robust and effective.

Methods Automated grading, performed by software for image quality assessment and for microaneurysm/dot haemorrhage detection, was carried out on 78 601 images, obtained from 33 535 consecutive patients, which had been manually graded at one of two regional diabetic retinopathy screening programmes. Cases where the automated grading software assessment indicated gradable images with no disease but the screening programme indicated ungradable images or disease more severe than mild retinopathy were arbitrated by seven senior ophthalmologists.

Results 100% (180/180) of patients with proliferative retinopathy, 100% (324/324) with referable background retinopathy, 100% (193/193) with observable background retinopathy, 97.3% (1099/1130) with referable maculopathy, 99.2% (384/387) with observable maculopathy and 99.8% (1824/1827) with ungradable images were detected by the software.

Conclusion The automated grading software operated to previously published results when applied to a large, unselected population attending two regional screening programmes. Manual grading workload reduction would be 36.3%.

Original languageEnglish
Pages (from-to)1606-1610
Number of pages5
JournalBritish Journal of Ophthalmology
Volume94
Issue number12
DOIs
Publication statusPublished - Dec 2010

Keywords

  • clinical audit
  • diabetic retinopathy
  • diagnosis, computer-assisted
  • female
  • humans
  • male
  • mass screening
  • negotiating
  • program evaluation
  • retinal hemorrhage
  • Scotland
  • sensitivity and specificity
  • severity of illness index
  • software
  • screening-program
  • cost-effectiveness
  • photography
  • population
  • disease

Cite this

Automated grading for diabetic retinopathy : a large-scale audit using arbitration by clinical experts. / Fleming, Alan D.; Goatman, Keith A.; Philip, Sam; Prescott, Gordon J.; Sharp, Peter F.; Olson, John A.

In: British Journal of Ophthalmology, Vol. 94, No. 12, 12.2010, p. 1606-1610.

Research output: Contribution to journalArticle

Fleming, Alan D. ; Goatman, Keith A. ; Philip, Sam ; Prescott, Gordon J. ; Sharp, Peter F. ; Olson, John A. / Automated grading for diabetic retinopathy : a large-scale audit using arbitration by clinical experts. In: British Journal of Ophthalmology. 2010 ; Vol. 94, No. 12. pp. 1606-1610.
@article{5270664b4d5743d6b711adbd5b7f1565,
title = "Automated grading for diabetic retinopathy: a large-scale audit using arbitration by clinical experts",
abstract = "Background/aims Automated grading software has the potential to reduce the manual grading workload within diabetic retinopathy screening programmes. This audit was undertaken at the request of Scotland's National Diabetic Retinopathy Screening Collaborative to assess whether the introduction of automated grading software into the national screening programme would be safe, robust and effective.Methods Automated grading, performed by software for image quality assessment and for microaneurysm/dot haemorrhage detection, was carried out on 78 601 images, obtained from 33 535 consecutive patients, which had been manually graded at one of two regional diabetic retinopathy screening programmes. Cases where the automated grading software assessment indicated gradable images with no disease but the screening programme indicated ungradable images or disease more severe than mild retinopathy were arbitrated by seven senior ophthalmologists.Results 100{\%} (180/180) of patients with proliferative retinopathy, 100{\%} (324/324) with referable background retinopathy, 100{\%} (193/193) with observable background retinopathy, 97.3{\%} (1099/1130) with referable maculopathy, 99.2{\%} (384/387) with observable maculopathy and 99.8{\%} (1824/1827) with ungradable images were detected by the software.Conclusion The automated grading software operated to previously published results when applied to a large, unselected population attending two regional screening programmes. Manual grading workload reduction would be 36.3{\%}.",
keywords = "clinical audit, diabetic retinopathy, diagnosis, computer-assisted, female, humans, male , mass screening, negotiating, program evaluation, retinal hemorrhage, Scotland, sensitivity and specificity, severity of illness index, software, screening-program, cost-effectiveness, photography, population, disease",
author = "Fleming, {Alan D.} and Goatman, {Keith A.} and Sam Philip and Prescott, {Gordon J.} and Sharp, {Peter F.} and Olson, {John A.}",
year = "2010",
month = "12",
doi = "10.1136/bjo.2009.176784",
language = "English",
volume = "94",
pages = "1606--1610",
journal = "British Journal of Ophthalmology",
issn = "0007-1161",
publisher = "BMJ Publishing Group",
number = "12",

}

TY - JOUR

T1 - Automated grading for diabetic retinopathy

T2 - a large-scale audit using arbitration by clinical experts

AU - Fleming, Alan D.

AU - Goatman, Keith A.

AU - Philip, Sam

AU - Prescott, Gordon J.

AU - Sharp, Peter F.

AU - Olson, John A.

PY - 2010/12

Y1 - 2010/12

N2 - Background/aims Automated grading software has the potential to reduce the manual grading workload within diabetic retinopathy screening programmes. This audit was undertaken at the request of Scotland's National Diabetic Retinopathy Screening Collaborative to assess whether the introduction of automated grading software into the national screening programme would be safe, robust and effective.Methods Automated grading, performed by software for image quality assessment and for microaneurysm/dot haemorrhage detection, was carried out on 78 601 images, obtained from 33 535 consecutive patients, which had been manually graded at one of two regional diabetic retinopathy screening programmes. Cases where the automated grading software assessment indicated gradable images with no disease but the screening programme indicated ungradable images or disease more severe than mild retinopathy were arbitrated by seven senior ophthalmologists.Results 100% (180/180) of patients with proliferative retinopathy, 100% (324/324) with referable background retinopathy, 100% (193/193) with observable background retinopathy, 97.3% (1099/1130) with referable maculopathy, 99.2% (384/387) with observable maculopathy and 99.8% (1824/1827) with ungradable images were detected by the software.Conclusion The automated grading software operated to previously published results when applied to a large, unselected population attending two regional screening programmes. Manual grading workload reduction would be 36.3%.

AB - Background/aims Automated grading software has the potential to reduce the manual grading workload within diabetic retinopathy screening programmes. This audit was undertaken at the request of Scotland's National Diabetic Retinopathy Screening Collaborative to assess whether the introduction of automated grading software into the national screening programme would be safe, robust and effective.Methods Automated grading, performed by software for image quality assessment and for microaneurysm/dot haemorrhage detection, was carried out on 78 601 images, obtained from 33 535 consecutive patients, which had been manually graded at one of two regional diabetic retinopathy screening programmes. Cases where the automated grading software assessment indicated gradable images with no disease but the screening programme indicated ungradable images or disease more severe than mild retinopathy were arbitrated by seven senior ophthalmologists.Results 100% (180/180) of patients with proliferative retinopathy, 100% (324/324) with referable background retinopathy, 100% (193/193) with observable background retinopathy, 97.3% (1099/1130) with referable maculopathy, 99.2% (384/387) with observable maculopathy and 99.8% (1824/1827) with ungradable images were detected by the software.Conclusion The automated grading software operated to previously published results when applied to a large, unselected population attending two regional screening programmes. Manual grading workload reduction would be 36.3%.

KW - clinical audit

KW - diabetic retinopathy

KW - diagnosis, computer-assisted

KW - female

KW - humans

KW - male

KW - mass screening

KW - negotiating

KW - program evaluation

KW - retinal hemorrhage

KW - Scotland

KW - sensitivity and specificity

KW - severity of illness index

KW - software

KW - screening-program

KW - cost-effectiveness

KW - photography

KW - population

KW - disease

U2 - 10.1136/bjo.2009.176784

DO - 10.1136/bjo.2009.176784

M3 - Article

VL - 94

SP - 1606

EP - 1610

JO - British Journal of Ophthalmology

JF - British Journal of Ophthalmology

SN - 0007-1161

IS - 12

ER -