Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers

Lyne Racette, Christine Y. Chiou, Jiucang Hao, Christopher Bowd, Michael H. Goldbaum, Linda M. Zangwill, Te Won Lee, Robert N. Weinreb, Pamela A. Sample

Research output: Contribution to journalArticle

19 Citations (Scopus)

Abstract

Purpose To investigate whether combining optic disc topography and short-wavelength automated perimetry (SWAP) data improves the diagnostic accuracy of relevance vector machine (RVM) classifiers for detecting glaucomatous eyes compared with using each test alone. Methods One eye of 144 glaucoma patients and 68 healthy controls from the Diagnostic Innovations in Glaucoma Study were included. RVM were trained and tested with cross-validation on optimized (backward elimination) SWAP features (thresholds plus age; pattern deviation; and total deviation) and on Heidelberg retina tomograph II (HRT) optic disc topography features, independently and in combination. RVM performance was also compared with 2 HRT linear discriminant functions and to SWAP mean deviation and pattern standard deviation. Classifier performance was measured by the area under the receiver operating characteristic curves (AUROCs) generated for each feature set and by the sensitivities at set specificities of 75%, 90%, and 96%. Results RVM trained on combined HRT and SWAP thresholds plus age had significantly higher AUROC (0.93) than RVM trained on HRT (0.88) and SWAP (0.76) alone. AUROCs for the SWAP global indices (mean deviation: 0.68; pattern standard deviation: 0.72) offered no advantage over SWAP thresholds plus age, whereas the linear discriminant functions AUROCs were significantly lower than RVM trained on the combined SWAP and HRT feature set and on HRT alone feature set. Conclusions Training RVM on combined optimized HRT and SWAP data improved diagnostic accuracy compared with training on SWAP and HRT parameters alone. Future research may identify other combinations of tests and classifiers that can also improve diagnostic accuracy.

Original languageEnglish (US)
Pages (from-to)167-175
Number of pages9
JournalJournal of Glaucoma
Volume19
Issue number3
StatePublished - Mar 2010
Externally publishedYes

Fingerprint

Visual Field Tests
Routine Diagnostic Tests
Retina
ROC Curve
Optic Disk
Glaucoma

Keywords

  • Glaucoma
  • Machine learning classifier
  • Neural networks
  • Optic disc
  • Structure-function relationship
  • Visual function

ASJC Scopus subject areas

  • Ophthalmology

Cite this

Racette, L., Chiou, C. Y., Hao, J., Bowd, C., Goldbaum, M. H., Zangwill, L. M., ... Sample, P. A. (2010). Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers. Journal of Glaucoma, 19(3), 167-175. https://doi.org/10.1097/IJG.0b013e3181a98b85

Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers. / Racette, Lyne; Chiou, Christine Y.; Hao, Jiucang; Bowd, Christopher; Goldbaum, Michael H.; Zangwill, Linda M.; Lee, Te Won; Weinreb, Robert N.; Sample, Pamela A.

In: Journal of Glaucoma, Vol. 19, No. 3, 01.01.2010, p. 167-175.

Research output: Contribution to journalArticle

Racette, L, Chiou, CY, Hao, J, Bowd, C, Goldbaum, MH, Zangwill, LM, Lee, TW, Weinreb, RN & Sample, PA 2010, 'Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers', Journal of Glaucoma, vol. 19, no. 3, pp. 167-175. https://doi.org/10.1097/IJG.0b013e3181a98b85
Racette, Lyne ; Chiou, Christine Y. ; Hao, Jiucang ; Bowd, Christopher ; Goldbaum, Michael H. ; Zangwill, Linda M. ; Lee, Te Won ; Weinreb, Robert N. ; Sample, Pamela A. / Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers. In: Journal of Glaucoma. 2010 ; Vol. 19, No. 3. pp. 167-175.
@article{2ed133827a724beea751efdf345bed79,
title = "Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers",
abstract = "Purpose To investigate whether combining optic disc topography and short-wavelength automated perimetry (SWAP) data improves the diagnostic accuracy of relevance vector machine (RVM) classifiers for detecting glaucomatous eyes compared with using each test alone. Methods One eye of 144 glaucoma patients and 68 healthy controls from the Diagnostic Innovations in Glaucoma Study were included. RVM were trained and tested with cross-validation on optimized (backward elimination) SWAP features (thresholds plus age; pattern deviation; and total deviation) and on Heidelberg retina tomograph II (HRT) optic disc topography features, independently and in combination. RVM performance was also compared with 2 HRT linear discriminant functions and to SWAP mean deviation and pattern standard deviation. Classifier performance was measured by the area under the receiver operating characteristic curves (AUROCs) generated for each feature set and by the sensitivities at set specificities of 75{\%}, 90{\%}, and 96{\%}. Results RVM trained on combined HRT and SWAP thresholds plus age had significantly higher AUROC (0.93) than RVM trained on HRT (0.88) and SWAP (0.76) alone. AUROCs for the SWAP global indices (mean deviation: 0.68; pattern standard deviation: 0.72) offered no advantage over SWAP thresholds plus age, whereas the linear discriminant functions AUROCs were significantly lower than RVM trained on the combined SWAP and HRT feature set and on HRT alone feature set. Conclusions Training RVM on combined optimized HRT and SWAP data improved diagnostic accuracy compared with training on SWAP and HRT parameters alone. Future research may identify other combinations of tests and classifiers that can also improve diagnostic accuracy.",
keywords = "Glaucoma, Machine learning classifier, Neural networks, Optic disc, Structure-function relationship, Visual function",
author = "Lyne Racette and Chiou, {Christine Y.} and Jiucang Hao and Christopher Bowd and Goldbaum, {Michael H.} and Zangwill, {Linda M.} and Lee, {Te Won} and Weinreb, {Robert N.} and Sample, {Pamela A.}",
year = "2010",
month = "1",
day = "1",
doi = "10.1097/IJG.0b013e3181a98b85",
language = "English (US)",
volume = "19",
pages = "167--175",
journal = "Journal of Glaucoma",
issn = "1057-0829",
publisher = "Lippincott Williams and Wilkins",
number = "3",

}

TY - JOUR

T1 - Combining functional and structural tests improves the diagnostic accuracy of relevance vector machine classifiers

AU - Racette, Lyne

AU - Chiou, Christine Y.

AU - Hao, Jiucang

AU - Bowd, Christopher

AU - Goldbaum, Michael H.

AU - Zangwill, Linda M.

AU - Lee, Te Won

AU - Weinreb, Robert N.

AU - Sample, Pamela A.

PY - 2010/1/1

Y1 - 2010/1/1

N2 - Purpose To investigate whether combining optic disc topography and short-wavelength automated perimetry (SWAP) data improves the diagnostic accuracy of relevance vector machine (RVM) classifiers for detecting glaucomatous eyes compared with using each test alone. Methods One eye of 144 glaucoma patients and 68 healthy controls from the Diagnostic Innovations in Glaucoma Study were included. RVM were trained and tested with cross-validation on optimized (backward elimination) SWAP features (thresholds plus age; pattern deviation; and total deviation) and on Heidelberg retina tomograph II (HRT) optic disc topography features, independently and in combination. RVM performance was also compared with 2 HRT linear discriminant functions and to SWAP mean deviation and pattern standard deviation. Classifier performance was measured by the area under the receiver operating characteristic curves (AUROCs) generated for each feature set and by the sensitivities at set specificities of 75%, 90%, and 96%. Results RVM trained on combined HRT and SWAP thresholds plus age had significantly higher AUROC (0.93) than RVM trained on HRT (0.88) and SWAP (0.76) alone. AUROCs for the SWAP global indices (mean deviation: 0.68; pattern standard deviation: 0.72) offered no advantage over SWAP thresholds plus age, whereas the linear discriminant functions AUROCs were significantly lower than RVM trained on the combined SWAP and HRT feature set and on HRT alone feature set. Conclusions Training RVM on combined optimized HRT and SWAP data improved diagnostic accuracy compared with training on SWAP and HRT parameters alone. Future research may identify other combinations of tests and classifiers that can also improve diagnostic accuracy.

AB - Purpose To investigate whether combining optic disc topography and short-wavelength automated perimetry (SWAP) data improves the diagnostic accuracy of relevance vector machine (RVM) classifiers for detecting glaucomatous eyes compared with using each test alone. Methods One eye of 144 glaucoma patients and 68 healthy controls from the Diagnostic Innovations in Glaucoma Study were included. RVM were trained and tested with cross-validation on optimized (backward elimination) SWAP features (thresholds plus age; pattern deviation; and total deviation) and on Heidelberg retina tomograph II (HRT) optic disc topography features, independently and in combination. RVM performance was also compared with 2 HRT linear discriminant functions and to SWAP mean deviation and pattern standard deviation. Classifier performance was measured by the area under the receiver operating characteristic curves (AUROCs) generated for each feature set and by the sensitivities at set specificities of 75%, 90%, and 96%. Results RVM trained on combined HRT and SWAP thresholds plus age had significantly higher AUROC (0.93) than RVM trained on HRT (0.88) and SWAP (0.76) alone. AUROCs for the SWAP global indices (mean deviation: 0.68; pattern standard deviation: 0.72) offered no advantage over SWAP thresholds plus age, whereas the linear discriminant functions AUROCs were significantly lower than RVM trained on the combined SWAP and HRT feature set and on HRT alone feature set. Conclusions Training RVM on combined optimized HRT and SWAP data improved diagnostic accuracy compared with training on SWAP and HRT parameters alone. Future research may identify other combinations of tests and classifiers that can also improve diagnostic accuracy.

KW - Glaucoma

KW - Machine learning classifier

KW - Neural networks

KW - Optic disc

KW - Structure-function relationship

KW - Visual function

UR - http://www.scopus.com/inward/record.url?scp=79952048463&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79952048463&partnerID=8YFLogxK

U2 - 10.1097/IJG.0b013e3181a98b85

DO - 10.1097/IJG.0b013e3181a98b85

M3 - Article

C2 - 19528827

AN - SCOPUS:79952048463

VL - 19

SP - 167

EP - 175

JO - Journal of Glaucoma

JF - Journal of Glaucoma

SN - 1057-0829

IS - 3

ER -