Detection of Progressive Glaucomatous Optic Nerve Damage on Fundus Photographs with Deep Learning

  • Felipe A. Medeiros
    Correspondence: Felipe A. Medeiros, MD, PhD, Duke Eye Center, Department of Ophthalmology, Duke University, 2351 Erwin Road, Durham, NC 27705.
    Vision, Imaging and Performance Laboratory (VIP), Duke Eye Center and Department of Ophthalmology, Duke University, Durham, North Carolina

    Department of Electrical and Computer Engineering, Pratt School of Engineering, Duke University, Durham, North Carolina
    Search for articles by this author
  • Alessandro A. Jammal
    Vision, Imaging and Performance Laboratory (VIP), Duke Eye Center and Department of Ophthalmology, Duke University, Durham, North Carolina
    Search for articles by this author
  • Eduardo B. Mariottoni
    Vision, Imaging and Performance Laboratory (VIP), Duke Eye Center and Department of Ophthalmology, Duke University, Durham, North Carolina
    Search for articles by this author


      To investigate whether predictions of retinal nerve fiber layer (RNFL) thickness obtained from a deep learning model applied to fundus photographs can detect progressive glaucomatous changes over time.


      Retrospective cohort study.


      Eighty-six thousand one hundred twenty-three pairs of color fundus photographs and spectral-domain (SD) OCT images collected during 21 232 visits from 8831 eyes of 5529 patients with glaucoma or glaucoma suspects.


      A deep learning convolutional neural network was trained to assess fundus photographs and to predict SD OCT global RNFL thickness measurements. The model then was tested on an independent sample of eyes that had longitudinal follow-up with both fundus photography and SD OCT. The ability to detect eyes that had statistically significant slopes of SD OCT change was assessed by receiver operating characteristic (ROC) curves. The repeatability of RNFL thickness predictions was investigated by measurements obtained from multiple photographs that had been acquired during the same day.

      Main Outcome Measures

      The relationship between change in predicted RNFL thickness from photographs and change in SD OCT RNFL thickness over time.


      The test sample consisted of 33 466 pairs of fundus photographs and SD OCT images collected during 7125 visits from 1147 eyes of 717 patients. Eyes in the test sample were followed up for an average of 5.3 ± 3.3 years, with an average of 6.2 ± 3.8 visits. A significant correlation was found between change over time in predicted and observed RNFL thickness (r = 0.76; 95% confidence interval [CI], 0.70–0.80; P < 0.001). Retinal nerve fiber layer predictions showed an ROC curve area of 0.86 (95% CI, 0.83–0.88) to discriminate progressors from nonprogressors. For detecting fast progressors (slope faster than 2 μm/year), the ROC curve area was 0.96 (95% CI, 0.94–0.98), with a sensitivity of 97% for 80% specificity and 85% for 90% specificity. For photographs obtained at the same visit, the intraclass correlation coefficient was 0.946 (95% CI, 0.940–0.952), with a coefficient of variation of 3.2% (95% CI, 3.1%–3.3%).


      A deep learning model was able to obtain objective and quantitative estimates of RNFL thickness that correlated well with SD OCT measurements and potentially could be used to monitor for glaucomatous changes over time.


      Abbreviations and Acronyms:

      AUC (area under the receiver operating characteristic curve), CI (confidence intervals), CoV (coefficient of variation), ICC (intraclass correlation coefficient), MD (mean deviation), M2M (machine-to-machine), PSD (pattern standard deviation), ResNet (residual deep neural network), ROC (receiver operating characteristic), RNFL (retinal nerve fiber layer), SD (spectral-domain)
      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic and Personal
      One-time access price info
      • For academic or personal research use, select 'Academic and Personal'
      • For corporate R&D use, select 'Corporate R&D Professionals'


      Subscribe to Ophthalmology
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect


      1. European Glaucoma Society Terminology and Guidelines for Glaucoma, 4th Edition—Chapter 3: Treatment principles and options supported by the EGS Foundation. Part 1: foreword; introduction; glossary; chapter 3 Treatment principles and options.
        Br J Ophthalmol. 2017; 101: 130-195
        • Prum Jr., B.E.
        • Rosenberg L.F.
        • Gedde S.J.
        • et al.
        Primary open-angle glaucoma Preferred Practice Pattern guidelines.
        Ophthalmology. 2016; 123: P41-P111
        • The AGIS Investigators
        The Advanced Glaucoma Intervention Study (AGIS): 7. The relationship between control of intraocular pressure and visual field deterioration.
        Am J Ophthalmol. 2000; 130: 429-440
        • Leske M.C.
        • Heijl A.
        • Hussein M.
        • et al.
        Factors for glaucoma progression and the effect of treatment: the Early Manifest Glaucoma Trial.
        Arch Ophthalmol. 2003; 121: 48-56
        • Heijl A.
        • Leske M.C.
        • Bengtsson B.
        • et al.
        Reduction of intraocular pressure and glaucoma progression: results from the Early Manifest Glaucoma Trial.
        Arch Ophthalmol. 2002; 120: 1268-1279
        • Abe R.Y.
        • Diniz-Filho A.
        • Zangwill L.M.
        • et al.
        The relative odds of progressing by structural and functional tests in glaucoma.
        Invest Ophthalmol Vis Sci. 2016; 57: OCT421-OCT428
        • Garway-Heath D.F.
        • Caprioli J.
        • Fitzke F.W.
        • Hitchings R.A.
        Scaling the hill of vision: the physiological relationship between light sensitivity and ganglion cell numbers.
        Invest Ophthalmol Vis Sci. 2000; 41: 1774-1782
        • Medeiros F.A.
        • Lisboa R.
        • Weinreb R.N.
        • et al.
        A combined index of structure and function for staging glaucomatous damage.
        Arch Ophthalmol. 2012; 130: 1107-1116
        • Silverman A.L.
        • Hammel N.
        • Khachatryan N.
        • et al.
        Diagnostic accuracy of the Spectralis and Cirrus reference databases in differentiating between healthy and early glaucoma eyes.
        Ophthalmology. 2016; 123: 408-414
        • Tatham A.J.
        • Medeiros F.A.
        Detecting structural progression in glaucoma with optical coherence tomography.
        Ophthalmology. 2017; 124: S57-S65
        • Li Z.
        • He Y.
        • Keel S.
        • et al.
        Efficacy of a deep learning system for detecting glaucomatous optic neuropathy based on color fundus photographs.
        Ophthalmology. 2018; 125: 1199-1206
        • Medeiros F.A.
        • Jammal A.A.
        • Thompson A.C.
        From machine to machine: an OCT-trained deep learning algorithm for objective quantification of glaucomatous damage in fundus photographs.
        Ophthalmology. 2019; 126: 513-521
        • Thompson A.C.
        • Jammal A.A.
        • Medeiros F.A.
        A deep learning algorithm to quantify neuroretinal rim loss from optic disc photographs.
        Am J Ophthalmol. 2019; 201: 9-18
        • Christopher M.
        • Belghith A.
        • Bowd C.
        • et al.
        Performance of deep learning architectures and transfer learning for detecting glaucomatous optic neuropathy in fundus photographs.
        Sci Rep. 2018; 8: 16685
        • Tielsch J.M.
        • Katz J.
        • Quigley H.A.
        • et al.
        Intraobserver and interobserver agreement in measurement of optic disc characteristics.
        Ophthalmology. 1988; 95: 350-356
        • Varma R.
        • Steinmann W.C.
        • Scott I.U.
        Expert agreement in evaluating the optic disc for glaucoma.
        Ophthalmology. 1992; 99: 215-221
        • Jampel H.D.
        • Friedman D.
        • Quigley H.
        • et al.
        Agreement among glaucoma specialists in assessing progressive disc changes from photographs in open-angle glaucoma patients.
        Am J Ophthalmol. 2009; 147: 39-44 e1
        • Chan H.H.
        • Ong D.N.
        • Kong Y.X.
        • et al.
        Glaucomatous Optic Neuropathy Evaluation (GONE) project: the effect of monoscopic versus stereoscopic viewing conditions on optic nerve evaluation.
        Am J Ophthalmol. 2014; 157: 936-944
        • Jammal A.A.
        • Thompson A.C.
        • Mariottoni E.B.
        • et al.
        Human versus machine: comparing a deep learning algorithm to human gradings for detecting glaucoma on fundus photographs.
        Am J Ophthalmol. 2020; 211: 123-131
        • Jammal A.A.
        • Thompson A.C.
        • Mariottoni E.B.
        • et al.
        Rates of glaucomatous structural and functional change from a large clinical population: the Duke Glaucoma Registry Study.
        Am J Ophthalmol. 2020 May 22; (30249-X) (Online ahead of print): S0002-9394
        • Varma R.
        • Bazzaz S.
        • Lai M.
        Optical tomography-measured retinal nerve fiber layer thickness in normal Latinos.
        Invest Ophthalmol Vis Sci. 2003; 44: 3369-3373
        • Patel N.B.
        • Lim M.
        • Gajjar A.
        • et al.
        Age-associated changes in the retinal nerve fiber layer and optic nerve head.
        Invest Ophthalmol Vis Sci. 2014; 55: 5134-5143
        • Bowd C.
        • Zangwill L.M.
        • Weinreb R.N.
        • et al.
        Estimating optical coherence tomography structural measurement floors to improve detection of progression in advanced glaucoma.
        Am J Ophthalmol. 2017; 175: 37-44
        • Asrani S.
        • Essaid L.
        • Alder B.D.
        • Santiago-Turla C.
        Artifacts in spectral-domain optical coherence tomography measurements in glaucoma.
        JAMA Ophthalmol. 2014; 132: 396-402
        • He K.
        • Zhang X.
        • Ren S.
        • Sun J.
        Deep residual learning for image recognition. ArXiv e-prints.
        (03385) (Accessed 12.01.18)
        • Deng J.
        • Dong W.
        • Socher R.
        • et al.
        ImageNet: a large-scale hierarchical image database.
        in: 2009 IEEE Conference on Computer Vision and Pattern Recognition. IEEE, Miami2009: 248-255
        • Kingma D.P.
        • Ba J.
        Adam: a method for stochastic optimization. ArXiv e-prints.
        (Accessed 11.01.18)
        • Ruder S.
        An overview of gradient descent optimization algorithms. ArXiv e-prints.
        (04747) (Accessed 11.01.18)
        • Smith L.N.
        Cyclical learning rates for training neural networks. ArXiv e-prints.
        (01186) (Accessed 05.01.18)
        • Medeiros F.A.
        • Leite M.T.
        • Zangwill L.M.
        • Weinreb R.N.
        Combining structural and functional measurements to improve detection of glaucoma progression using Bayesian hierarchical models.
        Invest Ophthalmol Vis Sci. 2011; 52: 5794-5803
        • Beckett L.A.
        • Tancredi D.J.
        • Wilson R.S.
        Multivariate longitudinal models for complex change processes.
        Stat Med. 2004; 23: 231-239
        • Wu Z.
        • Saunders L.J.
        • Zangwill L.M.
        • et al.
        Impact of normal aging and progression definitions on the specificity of detecting retinal nerve fiber layer thinning.
        Am J Ophthalmol. 2017; 181: 106-113
        • Leung C.K.S.
        • Ye C.
        • Weinreb R.N.
        • et al.
        Impact of age-related change of retinal nerve fiber layer and macular thicknesses on evaluation of glaucoma progression.
        Ophthalmology. 2013; 120: 2485-2492
        • Medeiros F.A.
        • Sample P.A.
        • Zangwill L.M.
        • et al.
        A statistical approach to the evaluation of covariate effects on the receiver operating characteristic curves of diagnostic tests in glaucoma.
        Invest Ophthalmol Vis Sci. 2006; 47: 2520-2527
        • Bland J.M.
        • Altman D.G.
        Measurement error.
        BMJ. 1996; 312: 1654
        • Vianna J.R.
        • Danthurebandara V.M.
        • Sharpe G.P.
        • et al.
        Importance of normal aging in estimating the rate of glaucomatous neuroretinal rim and retinal nerve fiber layer loss.
        Ophthalmology. 2015; 122: 2392-2398
        • Saunders L.J.
        • Medeiros F.A.
        • Weinreb R.N.
        • Zangwill L.M.
        What rates of glaucoma progression are clinically significant?.
        Expert Rev Ophthalmol. 2016; 11: 227-234
        • Kuang T.M.
        • Zhang C.
        • Zangwill L.M.
        • et al.
        Estimating lead time gained by optical coherence tomography in detecting glaucoma before development of visual field defects.
        Ophthalmology. 2015; 122: 2002-2009
        • Medeiros F.A.
        • Gracitelli C.P.
        • Boer E.R.
        • et al.
        Longitudinal changes in quality of life and rates of progressive visual field loss in glaucoma patients.
        Ophthalmology. 2015; 122: 293-301
        • Centers for Disease Control and Prevention (CDC)
        Using telehealth to expand access to essential health services during the COVID-19 pandemic.
        (Accessed 25.06.20)
        • Fortune B.
        • Reynaud J.
        • Cull G.
        • et al.
        The effect of age on optic nerve axon counts, SDOCT scan quality, and peripapillary retinal nerve fiber layer thickness measurements in Rhesus monkeys.
        Transl Vis Sci Technol. 2014; 3: 2