The efficacy of “disease/no disease” grading for diabetic retinopathy in a systematic screening programme

S Philip, A D Fleming, K A Goatman, S Fonseca, P McNamee, G S Scotland, G J Prescott, P F Sharp, J. A. Olson

Research output: Contribution to journalArticle

129 Citations (Scopus)

Abstract

Aim: To assess the efficacy of automated “disease/no disease” grading for diabetic retinopathy within a systematic screening programme.
Methods: Anonymised images were obtained from consecutive patients attending a regional primary care based diabetic retinopathy screening programme. A training set of 1067 images was used to develop automated grading algorithms. The final software was tested using a separate set of 14406 images from 6722 patients. The sensitivity and specificity of manual and automated systems operating as “disease/no disease” graders (detecting poor quality images and any diabetic retinopathy) were determined relative to a clinical reference standard.
Results: The reference standard classified 8.2% of the patients as having ungradable images (technical failures) and 62.5% as having no retinopathy. Detection of technical failures or any retinopathy was achieved by manual grading with 86.5% sensitivity (95% confidence interval 85.1 to 87.8) and 95.3% specificity (94.6 to 95.9) and by automated grading with 90.5% sensitivity (89.3 to 91.6) and 67.4% specificity (66.0 to 68.8). Manual and automated grading detected 99.1% and 97.9%, respectively, of patients with referable or observable retinopathy/maculopathy. Manual and automated grading detected 95.7% and 99.8%, respectively, of technical failures.
Conclusion: Automated “disease/no disease” grading of diabetic retinopathy could safely reduce the burden of grading in diabetic retinopathy screening programmes.
Original languageEnglish
Pages (from-to)1512-1517
Number of pages6
JournalBritish Journal of Ophthalmology
Volume91
Issue number11
Early online date15 May 2007
DOIs
Publication statusPublished - Nov 2007

Fingerprint

Diabetic Retinopathy
Primary Health Care
Software
Confidence Intervals
Sensitivity and Specificity

Keywords

  • Aged
  • Aged, 80 and over
  • Algorithms
  • Diabetic Retinopathy
  • Female
  • Humans
  • Male
  • Mass Screening
  • Middle Aged
  • Program Evaluation
  • Sensitivity and Specificity
  • Severity of Illness Index

Cite this

The efficacy of “disease/no disease” grading for diabetic retinopathy in a systematic screening programme. / Philip, S; Fleming, A D; Goatman, K A; Fonseca, S; McNamee, P; Scotland, G S; Prescott, G J; Sharp, P F; Olson, J. A.

In: British Journal of Ophthalmology, Vol. 91, No. 11, 11.2007, p. 1512-1517.

Research output: Contribution to journalArticle

@article{6f3ee76104ea47b2aa7bb5f4dea0558b,
title = "The efficacy of “disease/no disease” grading for diabetic retinopathy in a systematic screening programme",
abstract = "Aim: To assess the efficacy of automated “disease/no disease” grading for diabetic retinopathy within a systematic screening programme. Methods: Anonymised images were obtained from consecutive patients attending a regional primary care based diabetic retinopathy screening programme. A training set of 1067 images was used to develop automated grading algorithms. The final software was tested using a separate set of 14406 images from 6722 patients. The sensitivity and specificity of manual and automated systems operating as “disease/no disease” graders (detecting poor quality images and any diabetic retinopathy) were determined relative to a clinical reference standard. Results: The reference standard classified 8.2{\%} of the patients as having ungradable images (technical failures) and 62.5{\%} as having no retinopathy. Detection of technical failures or any retinopathy was achieved by manual grading with 86.5{\%} sensitivity (95{\%} confidence interval 85.1 to 87.8) and 95.3{\%} specificity (94.6 to 95.9) and by automated grading with 90.5{\%} sensitivity (89.3 to 91.6) and 67.4{\%} specificity (66.0 to 68.8). Manual and automated grading detected 99.1{\%} and 97.9{\%}, respectively, of patients with referable or observable retinopathy/maculopathy. Manual and automated grading detected 95.7{\%} and 99.8{\%}, respectively, of technical failures. Conclusion: Automated “disease/no disease” grading of diabetic retinopathy could safely reduce the burden of grading in diabetic retinopathy screening programmes.",
keywords = "Aged, Aged, 80 and over, Algorithms, Diabetic Retinopathy, Female, Humans, Male, Mass Screening, Middle Aged, Program Evaluation, Sensitivity and Specificity, Severity of Illness Index",
author = "S Philip and Fleming, {A D} and Goatman, {K A} and S Fonseca and P McNamee and Scotland, {G S} and Prescott, {G J} and Sharp, {P F} and Olson, {J. A.}",
year = "2007",
month = "11",
doi = "10.1136/BJO.2007.119453",
language = "English",
volume = "91",
pages = "1512--1517",
journal = "British Journal of Ophthalmology",
issn = "0007-1161",
publisher = "BMJ Publishing Group",
number = "11",

}

TY - JOUR

T1 - The efficacy of “disease/no disease” grading for diabetic retinopathy in a systematic screening programme

AU - Philip, S

AU - Fleming, A D

AU - Goatman, K A

AU - Fonseca, S

AU - McNamee, P

AU - Scotland, G S

AU - Prescott, G J

AU - Sharp, P F

AU - Olson, J. A.

PY - 2007/11

Y1 - 2007/11

N2 - Aim: To assess the efficacy of automated “disease/no disease” grading for diabetic retinopathy within a systematic screening programme. Methods: Anonymised images were obtained from consecutive patients attending a regional primary care based diabetic retinopathy screening programme. A training set of 1067 images was used to develop automated grading algorithms. The final software was tested using a separate set of 14406 images from 6722 patients. The sensitivity and specificity of manual and automated systems operating as “disease/no disease” graders (detecting poor quality images and any diabetic retinopathy) were determined relative to a clinical reference standard. Results: The reference standard classified 8.2% of the patients as having ungradable images (technical failures) and 62.5% as having no retinopathy. Detection of technical failures or any retinopathy was achieved by manual grading with 86.5% sensitivity (95% confidence interval 85.1 to 87.8) and 95.3% specificity (94.6 to 95.9) and by automated grading with 90.5% sensitivity (89.3 to 91.6) and 67.4% specificity (66.0 to 68.8). Manual and automated grading detected 99.1% and 97.9%, respectively, of patients with referable or observable retinopathy/maculopathy. Manual and automated grading detected 95.7% and 99.8%, respectively, of technical failures. Conclusion: Automated “disease/no disease” grading of diabetic retinopathy could safely reduce the burden of grading in diabetic retinopathy screening programmes.

AB - Aim: To assess the efficacy of automated “disease/no disease” grading for diabetic retinopathy within a systematic screening programme. Methods: Anonymised images were obtained from consecutive patients attending a regional primary care based diabetic retinopathy screening programme. A training set of 1067 images was used to develop automated grading algorithms. The final software was tested using a separate set of 14406 images from 6722 patients. The sensitivity and specificity of manual and automated systems operating as “disease/no disease” graders (detecting poor quality images and any diabetic retinopathy) were determined relative to a clinical reference standard. Results: The reference standard classified 8.2% of the patients as having ungradable images (technical failures) and 62.5% as having no retinopathy. Detection of technical failures or any retinopathy was achieved by manual grading with 86.5% sensitivity (95% confidence interval 85.1 to 87.8) and 95.3% specificity (94.6 to 95.9) and by automated grading with 90.5% sensitivity (89.3 to 91.6) and 67.4% specificity (66.0 to 68.8). Manual and automated grading detected 99.1% and 97.9%, respectively, of patients with referable or observable retinopathy/maculopathy. Manual and automated grading detected 95.7% and 99.8%, respectively, of technical failures. Conclusion: Automated “disease/no disease” grading of diabetic retinopathy could safely reduce the burden of grading in diabetic retinopathy screening programmes.

KW - Aged

KW - Aged, 80 and over

KW - Algorithms

KW - Diabetic Retinopathy

KW - Female

KW - Humans

KW - Male

KW - Mass Screening

KW - Middle Aged

KW - Program Evaluation

KW - Sensitivity and Specificity

KW - Severity of Illness Index

U2 - 10.1136/BJO.2007.119453

DO - 10.1136/BJO.2007.119453

M3 - Article

VL - 91

SP - 1512

EP - 1517

JO - British Journal of Ophthalmology

JF - British Journal of Ophthalmology

SN - 0007-1161

IS - 11

ER -