Skip to main content

Advertisement

SpringerLink
Log in

Deep learning prediction of steep and flat corneal curvature using fundus photography in post-COVID telemedicine era

  • Original Article
  • Published:
Medical & Biological Engineering & Computing Aims and scope Submit manuscript

Abstract

Recently, fundus photography (FP) is being increasingly used. Corneal curvature is an essential factor in refractive errors and is associated with several pathological corneal conditions. As FP-based examination systems have already been widely distributed, it would be helpful for telemedicine to extract information such as corneal curvature using FP. This study aims to develop a deep learning model based on FP for corneal curvature prediction by categorizing corneas into steep, regular, and flat groups. The EfficientNetB0 architecture with transfer learning was used to learn FP patterns to predict flat, regular, and steep corneas. In validation, the model achieved a multiclass accuracy of 0.727, a Matthews correlation coefficient of 0.519, and an unweighted Cohen’s κ of 0.590. The areas under the receiver operating characteristic curves for binary prediction of flat and steep corneas were 0.863 and 0.848, respectively. The optic nerve and its peripheral areas were the main focus of the model. The developed algorithm shows that FP can potentially be used as an imaging modality to estimate corneal curvature in the post-COVID-19 era, whereby patients may benefit from the detection of abnormal corneal curvatures using FP in the telemedicine setting.

Graphical Abstract

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11

Similar content being viewed by others

Data availability

Sample FP images of corneal curvature groups were deposited into a public repository (https://data.mendeley.com/datasets/bc2jfr7dv9) to validate the study. Data will be made available by the corresponding author upon reasonable request.

References

  1. Varadarajan AV, Poplin R, Blumer K et al (2018) Deep learning for predicting refractive error from retinal fundus images. Invest Ophthalmol Vis Sci 59:2861–2868. https://doi.org/10.1167/iovs.18-23887

    Article  PubMed  Google Scholar 

  2. Xu L, Wang YX, Guo Y et al (2012) Prevalence and associations of steep cornea/keratoconus in Greater Beijing. The Beijing Eye Study. PLoS One 7:e39313. https://doi.org/10.1371/journal.pone.0039313

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Elliott JH, Feman SS, O’Day DM, Garber M (1985) Hereditary sclerocornea. Arch Ophthalmol 103:676–679. https://doi.org/10.1001/archopht.1985.01050050068020

    Article  CAS  PubMed  Google Scholar 

  4. Vanhonsebrouck E, Consejo A, Coucke PJ et al (2021) The corneoscleral shape in Marfan syndrome. Acta Ophthalmol 99:405–410. https://doi.org/10.1111/aos.14636

    Article  PubMed  Google Scholar 

  5. Mohamed Mostafa E (2015) Effect of flat cornea on visual outcome after LASIK. J Ophthalmol 2015:e794854. https://doi.org/10.1155/2015/794854

    Article  Google Scholar 

  6. Qin Y, Liu L, Mao Y et al (2023) Accuracy of intraocular lens power calculation based on total keratometry in patients with flat and steep corneas. Am J Ophthalmol 247:103–110. https://doi.org/10.1016/j.ajo.2022.11.011

    Article  PubMed  Google Scholar 

  7. Yoo TK, Ryu IH, Kim JK et al (2022) A deep learning approach for detection of shallow anterior chamber depth based on the hidden features of fundus photographs. Comput Methods Programs Biomed 219:106735. https://doi.org/10.1016/j.cmpb.2022.106735

    Article  PubMed  Google Scholar 

  8. Gong W, Cheng T, Wang J et al (2022) Role of corneal radius of curvature in early identification of fundus tessellation in children with low myopia. Br J Ophthalmol. https://doi.org/10.1136/bjo-2022-321295

    Article  PubMed  Google Scholar 

  9. Sommer AC, Blumenthal EZ (2020) Telemedicine in ophthalmology in view of the emerging COVID-19 outbreak. Graefes Arch Clin Exp Ophthalmol 258:2341–2352. https://doi.org/10.1007/s00417-020-04879-2

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Parikh D, Armstrong G, Liou V, Husain D (2020) Advances in telemedicine in ophthalmology. Semin Ophthalmol 35:210–215. https://doi.org/10.1080/08820538.2020.1789675

    Article  PubMed  Google Scholar 

  11. Sunnetci KM, Alkan A (2023) Biphasic majority voting-based comparative COVID-19 diagnosis using chest X-ray images. Expert Syst Appl 216:119430. https://doi.org/10.1016/j.eswa.2022.119430

    Article  PubMed  Google Scholar 

  12. Sunnetci KM, Kaba E, Beyazal Çeliker F, Alkan A (2023) Comparative parotid gland segmentation by using ResNet-18 and MobileNetV2 based DeepLab v3+ architectures from magnetic resonance images. Concurr Comput: Pract Experience 35:e7405. https://doi.org/10.1002/cpe.7405

    Article  Google Scholar 

  13. Babenko B, Mitani A, Traynis I et al (2022) Detection of signs of disease in external photographs of the eyes via deep learning. Nat Biomed Eng 6:1370–1383. https://doi.org/10.1038/s41551-022-00867-5

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Huang D, Li R, Qian Y et al (2023) Fundus tessellated density assessed by deep learning in primary school children. Transl Vision Sci Technol 12:11. https://doi.org/10.1167/tvst.12.6.11

    Article  CAS  Google Scholar 

  15. Korot E, Pontikos N, Liu X et al (2021) Predicting sex from retinal fundus photographs using automated deep learning. Sci Rep 11:10286. https://doi.org/10.1038/s41598-021-89743-x

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Kim BR, Yoo TK, Kim HK et al (2022) Oculomics for sarcopenia prediction: a machine learning approach toward predictive, preventive, and personalized medicine. EPMA J. https://doi.org/10.1007/s13167-022-00292-3

    Article  PubMed  PubMed Central  Google Scholar 

  17. Masiwa LE, Moodley V (2020) A review of corneal imaging methods for the early diagnosis of pre-clinical keratoconus. J Optom 13:269–275. https://doi.org/10.1016/j.optom.2019.11.001

    Article  PubMed  PubMed Central  Google Scholar 

  18. Luft N, Siedlecki J, Reinking F et al (2021) Impact of extreme (flat and steep) keratometry on the safety and efficacy of small incision lenticule extraction (SMILE). Sci Rep 11:17854. https://doi.org/10.1038/s41598-021-97375-4

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  19. Schornack MM, Patel SV (2010) Relationship between corneal topographic indices and scleral lens base curve. Eye Contact Lens 36:330. https://doi.org/10.1097/ICL.0b013e3181eb8418

    Article  PubMed  Google Scholar 

  20. Rowsey JJ, Waring GO, Monlux RD et al (1991) Corneal topography as a predictor of refractive change in the prospective evaluation of radial keratotomy (PERK) study. Ophthalmic Surg Lasers Imaging Retina 22:370–380. https://doi.org/10.3928/1542-8877-19910701-04

    Article  CAS  Google Scholar 

  21. Reitblat O, Levy A, Kleinmann G et al (2017) Intraocular lens power calculation for eyes with high and low average keratometry readings: comparison between various formulas. J Cataract Refract Surg 43:1149–1156. https://doi.org/10.1016/j.jcrs.2017.06.036

    Article  PubMed  Google Scholar 

  22. Yoo TK, Choi JY, Kim HK et al (2021) Adopting low-shot deep learning for the detection of conjunctival melanoma using ocular surface images. Comput Methods Programs Biomed 205:106086. https://doi.org/10.1016/j.cmpb.2021.106086

    Article  PubMed  Google Scholar 

  23. Gupta A, Anjum GS, Katarya R (2021) InstaCovNet-19: a deep learning classification model for the detection of COVID-19 patients using Chest X-ray. Appl Soft Comput 99:106859. https://doi.org/10.1016/j.asoc.2020.106859

    Article  PubMed  Google Scholar 

  24. Eminaga O, Abbas M, Shen J et al (2023) PlexusNet: a neural network architectural concept for medical image classification. Comput Biol Med 154:106594. https://doi.org/10.1016/j.compbiomed.2023.106594

  25. Alkan A, Abdullah M, Abdullah H et al (2021) A smart agricultural application: automated detection of diseases in vine leaves using hybrid deep learning. Turk J Agric For 45:717–729. https://doi.org/10.3906/tar-2007-105

    Article  Google Scholar 

  26. Yoo TK, Kim SH, Kim M et al (2022) DeepPDT-Net: predicting the outcome of photodynamic therapy for chronic central serous chorioretinopathy using two-stage multimodal transfer learning. Sci Rep 12:18689. https://doi.org/10.1038/s41598-022-22984-6

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  27. Altun S, Alkan A, Altun H (2022) Application of deep learning and classical machine learning methods in the diagnosis of attention deficit hyperactivity disorder according to temperament features. Concurr Comput: Pract Experience 34:e6908. https://doi.org/10.1002/cpe.6908

    Article  Google Scholar 

  28. Choi JY, Yoo TK, Seo JG et al (2017) Multi-categorical deep learning neural network to classify retinal images: a pilot study employing small database. PLoS ONE 12:e0187336. https://doi.org/10.1371/journal.pone.0187336

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  29. Akpokiro V, Martin T, Oluwadare O (2022) EnsembleSplice: ensemble deep learning model for splice site prediction. BMC Bioinforma 23:413. https://doi.org/10.1186/s12859-022-04971-w

    Article  CAS  Google Scholar 

  30. Madani A, Arnaout R, Mofrad M, Arnaout R (2018) Fast and accurate view classification of echocardiograms using deep learning. npj Digit Med 1:1–8. https://doi.org/10.1038/s41746-017-0013-1

    Article  Google Scholar 

  31. Yoo TK, Choi JY, Kim HK (2021) Feasibility study to improve deep learning in OCT diagnosis of rare retinal diseases with few-shot classification. Med Biol Eng Comput 59:401–415. https://doi.org/10.1007/s11517-021-02321-1

    Article  PubMed  PubMed Central  Google Scholar 

  32. Yeniad B, Yiğit B, Işsever H, Közer Bilgin L (2003) Effects of contact lenses on corneal thickness and corneal curvature during usage. Eye Contact Lens 29:223–229. https://doi.org/10.1097/01.icl.0000086494.50288.70

    Article  PubMed  Google Scholar 

  33. Rim TH, Lee CJ, Tham Y-C et al (2021) Deep-learning-based cardiovascular risk stratification using coronary artery calcium scores predicted from retinal photographs. Lancet Digit Health 3:e306–e316. https://doi.org/10.1016/S2589-7500(21)00043-1

    Article  CAS  PubMed  Google Scholar 

  34. Ishii K, Asaoka R, Omoto T et al (2021) Predicting intraocular pressure using systemic variables or fundus photography with deep learning in a health examination cohort. Sci Rep 11:3687. https://doi.org/10.1038/s41598-020-80839-4

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Bernardes R, Serranho P, Lobo C (2011) Digital ocular fundus imaging: a review. Ophthalmologica 226:161–181. https://doi.org/10.1159/000329597

    Article  PubMed  Google Scholar 

  36. Zhang J, Zou H (2023) Artificial intelligence technology for myopia challenges: a review. Front Cell Dev Biol 11:1124005

    Article  PubMed  PubMed Central  Google Scholar 

  37. Yang D, Li M, Li W et al (2022) Prediction of refractive error based on ultrawide field images with deep learning models in myopia patients. Front Med 9:834281

  38. Kim J, Ryu IH, Kim JK et al (2022) Machine learning predicting myopic regression after corneal refractive surgery using preoperative data and fundus photography. Graefes Arch Clin Exp Ophthalmol 260:3701–3710. https://doi.org/10.1007/s00417-022-05738-y

    Article  PubMed  Google Scholar 

  39. Tham Y-C, Goh JHL, Anees A et al (2022) Detecting visually significant cataract using retinal photograph-based deep learning. Nat Aging 2:264–271. https://doi.org/10.1038/s43587-022-00171-6

    Article  PubMed  PubMed Central  Google Scholar 

  40. Li Dong HuXY, Yan YN et al (2021) Deep learning-based estimation of axial length and subfoveal choroidal thickness from color fundus photographs. Front Cell Dev Biol 9:653692. https://doi.org/10.3389/fcell.2021.653692

    Article  PubMed  Google Scholar 

  41. Yoo TK, Ryu IH, Kim JK, Lee IS (2022) Deep learning for predicting uncorrected refractive error using posterior segment optical coherence tomography images. Eye 36:1959–1965. https://doi.org/10.1038/s41433-021-01795-5

    Article  PubMed  Google Scholar 

  42. Langenbucher A, Szentmáry N, Cayless A et al (2022) Prediction of corneal back surface power – deep learning algorithm versus multivariate regression. Ophthalmic Physiol Opt 42:185–194. https://doi.org/10.1111/opo.12909

    Article  PubMed  Google Scholar 

  43. Fırat M, Çankaya C, Çınar A, Tuncer T (2022) Automatic detection of keratoconus on Pentacam images using feature selection based on deep learning. Int J Imaging Syst Technol 32:1548–1560. https://doi.org/10.1002/ima.22717

    Article  Google Scholar 

  44. Kamiya K, Ayatsuka Y, Kato Y et al (2021) Prediction of keratoconus progression using deep learning of anterior segment optical coherence tomography maps. Ann Transl Med 9:1287. https://doi.org/10.21037/atm-21-1772

    Article  PubMed  PubMed Central  Google Scholar 

  45. Jonas JB, Kling F, Gründler AE (1997) Optic disc shape, corneal astigmatism, and amblyopia. Ophthalmology 104:1934–1937. https://doi.org/10.1016/S0161-6420(97)30004-9

    Article  CAS  PubMed  Google Scholar 

  46. Chen M-J, Liu Y-T, Tsai C-C et al (2009) Relationship between central corneal thickness, refractive error, corneal curvature, anterior chamber depth and axial length. J Chin Med Assoc 72:133–137. https://doi.org/10.1016/S1726-4901(09)70038-3

    Article  PubMed  Google Scholar 

  47. Fard AM, Patel SP, Sorkhabi RD et al (2020) Posterior pole retinal thickness distribution pattern in keratoconus. Int Ophthalmol 40:2807–2816. https://doi.org/10.1007/s10792-020-01464-8

    Article  PubMed  Google Scholar 

  48. Hashemi H, Heirani M, Ambrósio R et al (2022) The link between keratoconus and posterior segment parameters: an updated, comprehensive review. Ocul Surf 23:116–122. https://doi.org/10.1016/j.jtos.2021.12.004

    Article  PubMed  Google Scholar 

  49. Somani S, Tuan KA, Corneal Chernyak D (2004) Asphericity and retinal image quality: a case study and simulations. J Refract Surg 20:581–585. https://doi.org/10.3928/1081-597X-20040901-32

    Article  Google Scholar 

  50. Maceo Heilman B, Mohamed A, Ruggeri M et al (2021) Age-dependence of the peripheral defocus of the isolated human crystalline lens. Invest Ophthalmol Vis Sci 62:15. https://doi.org/10.1167/iovs.62.3.15

    Article  PubMed  PubMed Central  Google Scholar 

  51. Abramovich O, Pizem H, Van Eijgen J et al (2023) FundusQ-Net: a regression quality assessment deep learning algorithm for fundus images quality grading. Comput Methods Programs Biomed 239:107522. https://doi.org/10.1016/j.cmpb.2023.107522

    Article  PubMed  Google Scholar 

  52. Yu AC, Mohajer B, Eng J (2022) External validation of deep learning algorithms for radiologic diagnosis: a systematic review. Radiol: Artif Intell 4:e210064. https://doi.org/10.1148/ryai.210064

    Article  PubMed  Google Scholar 

  53. Wagner SK, Fu DJ, Faes L et al (2020) Insights into systemic disease through retinal imaging-based oculomics. Trans Vis Sci Tech 9:6–6. https://doi.org/10.1167/tvst.9.2.6

    Article  Google Scholar 

  54. Xiao W, Huang X, Wang JH et al (2021) Screening and identifying hepatobiliary diseases through deep learning using ocular images: a prospective, multicentre study. Lancet Digit Health 3:e88–e97. https://doi.org/10.1016/S2589-7500(20)30288-0

    Article  CAS  PubMed  Google Scholar 

  55. Tariq YM, Samarawickrama C, Pai A et al (2010) Impact of ethnicity on the correlation of retinal parameters with axial length. Invest Ophthalmol Vis Sci 51:4977–4982. https://doi.org/10.1167/iovs.10-5226

    Article  PubMed  Google Scholar 

  56. Steyerberg EW, Harrell FE (2016) Prediction models need appropriate internal, internal-external, and external validation. J Clin Epidemiol 69:245–247. https://doi.org/10.1016/j.jclinepi.2015.04.005

    Article  PubMed  Google Scholar 

Download references

Acknowledgements

Bo Young Lee and Hee Jin Kim played significant roles in preprocessing the data.

Funding

This research was supported by the “Regional Innovation Strategy (RIS)” through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (MOE) in 2023 (2022RIS-005) and by a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (RS-2023–00251484).

Author information

Authors and Affiliations

  1. Department of Biomedical Engineering, Yonsei University, Wonju, South Korea

    Joon Yul Choi

  2. Yonsei Eye Clinic, Seoul, South Korea

    Hyungsu Kim

  3. Department of Refractive Surgery, B&VIIT Eye Center, B2 GT Tower, 1317-23 Seocho-Dong, Seocho-Gu, Seoul, South Korea

    Jin Kuk Kim, In Sik Lee, Ik Hee Ryu & Tae Keun Yoo

  4. Research and Development Department, VISUWORKS, Seoul, South Korea

    Ik Hee Ryu, Jung Soo Kim & Tae Keun Yoo

Authors
  1. Joon Yul Choi
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Hyungsu Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Jin Kuk Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. In Sik Lee
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Ik Hee Ryu
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Jung Soo Kim
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Tae Keun Yoo
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

TKY and JYC had full access to all the data in the study and take responsibility for the integrity of the data and accuracy of the data analysis. Tae Keun Yoo and Joon Yul Choi contributed equally to the work presented here and should be regarded as equivalent authors. JYC, IHR, and TKY developed the algorithms. JKK, ISL, and TKY consolidated the data and performed data analyses. JYC and TKY drafted the manuscript. IHR and JKK conceived and designed the study. All authors contributed to the revisions and finalization of the submitted manuscript. All authors met the following criteria: (1) substantial contribution to the conception or design of the work or the acquisition, analysis, or interpretation of the data; (2) drafting the work or critically revising it for important intellectual content; (3) final approval of the completed version; and (4) accountability for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work were appropriately investigated and resolved.

Corresponding author

Correspondence to Tae Keun Yoo.

Ethics declarations

Ethical approval

All procedures of studies involving human participants were performed in accordance with the ethical standards of the Institutional Review Board of the Korean National Institute for Bioethics Policy (KNIBP No. 2023–0088-001) and the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Informed consent

The Institutional Review Board waived the requirement for informed consent because the data were fully de-identified to protect patient confidentiality.

Competing interests

Ik Hee Ryu and Jin Kuk Kim are directors of VISUWORKS and own company stock. Ik Hee Ryu serves on the Advisory Board of Carl Zeiss Meditec AG and Avellino Lab USA/MAB for Avellino Lab Korea. Jin Kuk Kim is an executive of the Korea Intelligent Medical Industry Association (KIMIA). Tae Keun Yoo is an employee of VISUWORKS and receives salary and stock as part of the standard compensation package. The authors declare no conflicts of interest. VISUWORKS has received research grants for SMILE surgery from Carl Zeiss Meditec AG, Germany. The research grants had no effect on this study.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Choi, J.Y., Kim, H., Kim, J.K. et al. Deep learning prediction of steep and flat corneal curvature using fundus photography in post-COVID telemedicine era. Med Biol Eng Comput 62, 449–463 (2024). https://doi.org/10.1007/s11517-023-02952-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11517-023-02952-6

Keywords

Search

Navigation