Validating the Usefulness of the "Random Forests" Classifier to Diagnose Early Glaucoma With Optical Coherence Tomography.

PURPOSE To validate the usefulness of the "Random Forests" classifier to diagnose early glaucoma with spectral-domain optical coherence tomography (SDOCT). METHODS design: Comparison of diagnostic algorithms. SETTING Multiple institutional practices. STUDY PARTICIPANTS Training dataset included 94 eyes of 94 open-angle glaucoma (OAG) patients and 84 eyes of 84 normal subjects and testing dataset included 114 eyes of 114 OAG patients and 82 eyes of 82 normal subjects. In both groups, OAG eyes with mean deviation (MD) values better than -5.0 dB were included. OBSERVATION PROCEDURE Using the training dataset, classifiers were built to discriminate between glaucoma and normal eyes using 84 OCT measurements using the Random Forests method, multiple logistic regression models based on backward or bidirectional stepwise model selection, a least absolute shrinkage and selection operator regression (LASSO) model, and a Ridge regression model. MAIN OUTCOME MEASURES Diagnostic accuracy. RESULTS With the testing data, the area under the receiver operating characteristic curve (AROC) with the Random Forests method (93.0%) was significantly (P < .05) larger than those with other models of the stepwise model selections (71.9%), LASSO model (89.6%), and Ridge model (89.2%). CONCLUSION It is useful to analyze multiple SDOCT parameters concurrently using the Random Forests method to diagnose glaucoma in early stages.

[1]  M. T. Leite,et al.  Spectral-domain optical coherence tomography for early glaucoma assessment: analysis of macular ganglion cell complex versus peripapillary retinal nerve fiber layer. , 2011, Canadian journal of ophthalmology. Journal canadien d'ophtalmologie.

[2]  J Katz,et al.  Evaluation of nerve fiber layer assessment. , 1984, Archives of ophthalmology.

[3]  A. Iwase,et al.  Risk factors for open-angle glaucoma in a Japanese population: the Tajimi Study. , 2006, Ophthalmology.

[4]  C. Glymour,et al.  Optical coherence tomography machine learning classifiers for glaucoma detection: a preliminary study. , 2005, Investigative ophthalmology & visual science.

[5]  H. Quigley,et al.  The number of people with glaucoma worldwide in 2010 and 2020 , 2006, British Journal of Ophthalmology.

[6]  R S Harwerth,et al.  Ganglion cell losses underlying visual field defects from experimental glaucoma. , 1999, Investigative ophthalmology & visual science.

[7]  E. DeLong,et al.  Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. , 1988, Biometrics.

[8]  H. Rao,et al.  Effect of spectrum bias on the diagnostic accuracy of spectral-domain optical coherence tomography in glaucoma. , 2012, Investigative ophthalmology & visual science.

[9]  H. Rao,et al.  Retinal nerve fiber layer and macular inner retina measurements by spectral domain optical coherence tomograph in Indian eyes with early glaucoma , 2012, Eye.

[10]  Achim Zeileis,et al.  BMC Bioinformatics BioMed Central Methodology article Conditional variable importance for random forests , 2008 .

[11]  M. Bruce Shields,et al.  Textbook of glaucoma , 1987 .

[12]  Eun Suk Lee,et al.  Structure-function relationship and diagnostic value of macular ganglion cell complex measurement using Fourier-domain OCT in glaucoma. , 2010, Investigative ophthalmology & visual science.

[13]  Tin Aung,et al.  Classification algorithms enhance the discrimination of glaucoma from normal eyes using high-definition optical coherence tomography. , 2012, Investigative ophthalmology & visual science.

[14]  H. Quigley,et al.  Number of ganglion cells in glaucoma eyes compared with threshold visual field tests in the same persons. , 2000, Investigative ophthalmology & visual science.

[15]  Robert N Weinreb,et al.  Comparison of different spectral domain OCT scanning protocols for diagnosing preperimetric glaucoma. , 2013, Investigative ophthalmology & visual science.

[16]  Makoto Aihara,et al.  Cross-sectional study: Does combining optical coherence tomography measurements using the ‘Random Forest’ decision tree classifier improve the prediction of the presence of perimetric deterioration in glaucoma suspects? , 2013, BMJ Open.

[17]  W. Youden,et al.  Index for rating diagnostic tests , 1950, Cancer.

[18]  Sumio Watanabe,et al.  The Relationship between H. pylori Infection and Osteoporosis in Japan , 2014, Gastroenterology research and practice.

[19]  A. Sommer,et al.  An evaluation of optic disc and nerve fiber layer examinations in monitoring progression of early glaucoma damage. , 1992, Ophthalmology.

[20]  Takuhei Shoji,et al.  Impact of high myopia on the performance of SD-OCT parameters to detect glaucoma , 2012, Graefe's Archive for Clinical and Experimental Ophthalmology.

[21]  D. R. Anderson,et al.  The mode of progressive disc cupping in ocular hypertension and glaucoma. , 1980, Archives of ophthalmology.

[22]  R. Tibshirani.  Regression Shrinkage and Selection via the Lasso , 1996 .

[23]  M. Segal,et al.  Relating Amino Acid Sequence to Phenotype: Analysis of Peptide‐Binding Data , 2000, Biometrics.

[24]  Ryo Asaoka,et al.  Applying "Lasso" Regression to Predict Future Visual Field Progression in Glaucoma Patients. , 2015, Investigative ophthalmology & visual science.

[25]  A E Maumenee,et al.  Optic disc parameters and onset of glaucomatous field loss. I. Methods and progressive changes in disc morphology. , 1979, Archives of ophthalmology.

[26]  David R. Anderson,et al.  Multimodel Inference , 2004 .

[27]  T. Wong,et al.  Refractive error, axial dimensions, and primary open-angle glaucoma: the Singapore Malay Eye Study. , 2010, Archives of ophthalmology.

[28]  Janez Demsar,et al.  Statistical Comparisons of Classifiers over Multiple Data Sets , 2006, J. Mach. Learn. Res..

[29]  G. Wollstein,et al.  Detection of macular ganglion cell loss in glaucoma by Fourier-domain optical coherence tomography. , 2009, Ophthalmology.

[30]  Earl L. Smith,et al.  Neural losses correlated with visual losses in clinical perimetry. , 2004, Investigative ophthalmology & visual science.

[31]  Robert N Weinreb,et al.  Comparison of different spectral domain optical coherence tomography scanning areas for glaucoma diagnosis. , 2010, Ophthalmology.

[32]  João Maroco,et al.  Data mining methods in the prediction of Dementia: A real-data comparison of the accuracy, sensitivity and specificity of linear discriminant analysis, logistic regression, neural networks, support vector machines, classification trees and random forests , 2011, BMC Research Notes.

[33]  T. Zimmerman,et al.  Clinical Pathways in Glaucoma , 2001 .

[34]  S. Holm.  A Simple Sequentially Rejective Multiple Test Procedure , 1979 .

[35]  J. Jonas,et al.  High myopia and glaucoma susceptibility the Beijing Eye Study. , 2007, Ophthalmology.

[36]  Nikunj C. Oza,et al.  Online Ensemble Learning , 2000, AAAI/IAAI.

[37]  J. J. Wang,et al.  The relationship between glaucoma and myopia: the Blue Mountains Eye Study. , 1999, Ophthalmology.

[38]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[39]  I. Schmidtmann,et al.  Diagnostic ability of retinal ganglion cell complex, retinal nerve fiber layer, and optic nerve head measurements by Fourier-domain optical coherence tomography , 2011, Graefe's Archive for Clinical and Experimental Ophthalmology.

[40]  Roberto Cipolla,et al.  Semantic texton forests for image categorization and segmentation , 2008, 2008 IEEE Conference on Computer Vision and Pattern Recognition.

[41]  Eun Suk Lee,et al.  Comparing the ganglion cell complex and retinal nerve fibre layer measurements by Fourier domain OCT to detect glaucoma in high myopia , 2010, British Journal of Ophthalmology.

[42]  A. Sommer,et al.  The nerve fiber layer in the diagnosis of glaucoma. , 1977, Archives of ophthalmology.

[43]  G. Holló,et al.  Diagnostic accuracy of nerve fibre layer, macular thickness and optic disc measurements made with the RTVue-100 optical coherence tomograph to detect glaucoma , 2011, Eye.

[44]  Ryo Asaoka,et al.  Discriminating between Glaucoma and Normal Eyes Using Optical Coherence Tomography and the ‘Random Forests’ Classifier , 2014, PloS one.

[45]  Jean-Claude Mwanza,et al.  Combining spectral domain optical coherence tomography structural parameters for the diagnosis of glaucoma with early visual field loss. , 2013, Investigative ophthalmology & visual science.

[46]  G. Dunkelberger,et al.  Retinal ganglion cell atrophy correlated with automated perimetry in human eyes with glaucoma. , 1989, American journal of ophthalmology.

[47]  M Araie,et al.  Influence of myopic disc shape on the diagnostic precision of the Heidelberg Retina Tomograph. , 1999, Japanese journal of ophthalmology.

[48]  Douglas R. Anderson.  Automated Static Perimetry , 1992 .

[49]  K. Lunetta,et al.  Screening large-scale association study data: exploiting interactions using random forests , 2004, BMC Genetics.

[50]  M. Aickin,et al.  Adjusting for multiple testing when reporting research results: the Bonferroni vs Holm methods. , 1996, American journal of public health.

[51]  Trevor Hastie,et al.  Regularization Paths for Generalized Linear Models via Coordinate Descent. , 2010, Journal of statistical software.

[52]  Kazunori Miyata,et al.  Glaucoma-Induced Optic Disc Morphometric Changes and Glaucoma Diagnostic Ability of Heidelberg Retina Tomograph II in Highly Myopic Eyes , 2014, PloS one.

[53]  Debin Zhao,et al.  Building Emerging Pattern (EP) Random forest for recognition , 2010, 2010 IEEE International Conference on Image Processing.

[54]  B. Bengtsson,et al.  False-negative responses in glaucoma perimetry: indicators of patient performance or test reliability? , 2000, American journal of ophthalmology.

[55]  Mohak Shah,et al.  Evaluating Learning Algorithms: A Classification Perspective , 2011 .

[56]  A. Sommer,et al.  Clinically detectable nerve fiber atrophy precedes the onset of glaucomatous field loss. , 1991, Archives of ophthalmology.

[57]  Ryo Asaoka,et al.  Measuring Visual Field Progression in the Central 10 Degrees Using Additional Information from Central 24 Degrees Visual Fields and ‘Lasso Regression’ , 2013, PloS one.

[58]  Alan L. Yuille,et al.  Performance comparison of machine learning algorithms and number of independent components used in fMRI decoding of belief vs. disbelief , 2011, NeuroImage.

[59]  D. Apple,et al.  Congenital anomalies of the optic disc. , 1982, Survey of ophthalmology.

[60]  Ryo Asaoka,et al.  The Relationship between Central Visual Field Damage and Motor Vehicle Collisions in Primary Open-Angle Glaucoma Patients , 2014, PloS one.

[61]  Sung Yong Kang,et al.  Relationship between visual field sensitivity and macular ganglion cell complex thickness as measured by spectral-domain optical coherence tomography. , 2010, Investigative ophthalmology & visual science.

[62]  Yuan Bo Liang,et al.  Prevalence of primary open angle glaucoma in a rural adult Chinese population: the Handan eye study. , 2011, Investigative ophthalmology & visual science.