Skip to main content
SpringerLink
Log in

Explaining Eye Diseases Detected by Machine Learning Using SHAP: A Case Study of Diabetic Retinopathy and Choroidal Nevus

  • Original Research
  • Published:
SN Computer Science Aims and scope Submit manuscript

Abstract

Most visual impairment and eye cancers are preventable if detected in their early stages. Diabetic retinopathy (DR) is a significant cause of blindness worldwide and a serious public health concern in a population aged 20–65. With the increasing number of diabetes globally and its effects on patients’ vision, the automatic detection of DR has received wide attention from the machine learning field. Uveal melanoma (UM) is one of the most severe intraocular cancers in adults aged 50–80. A choroidal nevus (CN) is one of the most common intraocular tumours that can transform into UM, which can cause eyesight loss and spiteful melanoma with a high risk of melanoma-relevant metastasis and even death. Early prediction of UM can mitigate the risk of death caused by skeptical diagnosis decisions. In this paper, we use a transfer learning technique with a convolutional neural network (CNN)-based algorithm to detect UM and improve the interpretation of the diagnosis results. However, due to the black-box nature of deep learning and machine learning models, the interpretation and reliability of the predictions are still an issue that needs to be addressed before deploying these predictive models successfully. In this paper, we use the SHapley Additive exPlanations (SHAP) analysis approach to detect areas of an eye image that contribute the most to the DR and CN prediction using transfer learning. Our predictive model achieves an accuracy of 97% and 81% for binary and multi-class classification of DR and 82.5% accuracy for binary classification of CN. The SHAP analysis of the proposed method shows that regardless of the performance of the predictive models, this approach can be used as a tool to interpret the prediction results with more context-sensitive information about each sample and better understand the reasons for the classification results.

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

Similar content being viewed by others

Data availability

Obtaining fundus images from the two datasets used in this study is challenging due to privacy concerns and strict regulations, resulting in limited availability of these images.

References

  1. Saydah SH, Gerzoff RB, Saaddine JB, Zhang X, Cotch MF. Eye care among us adults at high risk for vision loss in the united states in 2002 and 2017. JAMA Ophthalmol. 2020;138:479–89.

    Article  Google Scholar 

  2. Cotter SA, et al. Causes of low vision and blindness in adult latinos: the los angeles latino eye study. Ophthalmology. 2006;113:1574–82.

    Article  Google Scholar 

  3. Ting DSW, Cheung GCM, Wong TY. Diabetic retinopathy: global prevalence, major risk factors, screening practices and public health challenges: a review. Clin Exp Ophthalmol. 2016;44:260–77.

    Article  Google Scholar 

  4. Wong TY, Cheung CMG, Larsen M, Sharma S, Simó R. Erratum: Diabetic retinopathy. Nat Rev Dis Primers. 2016;2:1–1.

    Google Scholar 

  5. Consejo A, Melcer T, Rozema JJ. Introduction to machine learning for ophthalmologists. Semin Ophthalmol. 2019;34:19–41.

    Article  Google Scholar 

  6. Alam M, Le D, Lim JI, Chan RV, Yao X. Supervised machine learning based multi-task artificial intelligence classification of retinopathies. J Clin Med. 2019;8:872.

    Article  Google Scholar 

  7. Statista. Percentage of diabetics in the global adult population in 2019 and 2045. DATAREPORTAL. (2019). https://www.statista.com/statistics/271464/percentage-of-diabetics-worldwide/. Accessed 3 Dec 2021.

  8. Klein R, Knudtson MD, Lee KE, Gangnon R, Klein BE. The wisconsin epidemiologic study of diabetic retinopathy xxii: the twenty-five-year progression of retinopathy in persons with type 1 diabetes. Ophthalmology. 2008;115:1859–68.

    Article  Google Scholar 

  9. Lim LS, Liew G, Cheung N, Mitchell P, Wong TY. Mixed messages on systemic therapies for diabetic retinopathy. Lancet. 2010;376:1461.

    Article  Google Scholar 

  10. Stefánsson E, et al. Screening and prevention of diabetic blindness. Acta Ophthalmologica Scandinavica. 2000;78:374–85.

    Article  Google Scholar 

  11. Care D. Medical care in diabetes 2018. Diabet Care. 2018;41:S105–18.

    Article  Google Scholar 

  12. Marous CL, et al. Malignant transformation of choroidal nevus according to race in 3334 consecutive patients. Indian J Ophthalmol. 2019;67:2035.

    Article  Google Scholar 

  13. Chien JL, Sioufi K, Surakiatchanukul T, Shields JA, Shields CL. Choroidal nevus: a review of prevalence, features, genetics, risks, and outcomes. Curr Opin Ophthalmol. 2017;28:228–37.

    Article  Google Scholar 

  14. Qiu M, Shields CL. Choroidal nevus in the united states adult population: racial disparities and associated factors in the national health and nutrition examination survey. Ophthalmology. 2015;122:2071–83.

    Article  Google Scholar 

  15. Sumich P, Mitchell P, Wang JJ. Choroidal nevi in a white population: the blue mountains eye study. Arch Ophthalmol. 1998;116:645–50.

    Article  Google Scholar 

  16. Kaliki S, Shields C. Uveal melanoma: relatively rare but deadly cancer. Eye. 2017;31:241–57.

    Article  Google Scholar 

  17. Shields CL, et al. Choroidal nevus transformation into melanoma: analysis of 2514 consecutive cases. Arch Ophthalmol. 2009;127:981–7.

    Article  Google Scholar 

  18. Shields CL, et al. White paper on ophthalmic imaging for choroidal nevus identification and transformation into melanoma. Transl Vis Sci Technol. 2021;10:24–24.

    Article  Google Scholar 

  19. Al Rasheed R, Al Adel F. Diabetic retinopathy: knowledge, awareness and practices of physicians in primary-care centers in Riyadh, Saudi Arabia. Saudi J Ophthalmol. 2017;31:2–6.

    Article  Google Scholar 

  20. Delorme C, Boisjoly H, Baillargeon L, Turcotte P, Bernard P. Screening for diabetic retinopathy. Do family physicians know the Canadian guidelines? Can Fam Phys. 1998;44:1473.

    Google Scholar 

  21. Sengupta S, Singh A, Leopold HA, Gulati T, Lakshminarayanan V. Ophthalmic diagnosis using deep learning with fundus images—a critical review. Artif Intell Med. 2020;102: 101758.

    Article  Google Scholar 

  22. Ganin Y, Lempitsky V. \(\hat{n}4\)-fields: neural network nearest neighbor fields for image transforms. In: Asian conference on computer vision. Springer; 2014. p. 536–51.

  23. Yadav S, et al. Performance analysis of deep neural networks through transfer learning in retinal detachment diagnosis using fundus images. Sādhanā. 2022;47:1–13.

    Article  Google Scholar 

  24. Qummar S, et al. A deep learning ensemble approach for diabetic retinopathy detection. IEEE Access. 2019;7:150530–9.

    Article  Google Scholar 

  25. Khalifa NEM, Loey M, Taha MHN, Mohamed HNET. Deep transfer learning models for medical diabetic retinopathy detection. Acta Informatica Medica. 2019;27:327.

    Article  Google Scholar 

  26. Pak A, Ziyaden A, Tukeshev K, Jaxylykova A, Abdullina D. Comparative analysis of deep learning methods of detection of diabetic retinopathy. Cogent Eng. 2020;7:1805144.

    Article  Google Scholar 

  27. Gangwar AK, Ravi V. Diabetic retinopathy detection using transfer learning and deep learning. In: Evolution in computational intelligence. Springer; 2021. p. 679–89.

  28. Tymchenko B, Marchenko P, Spodarets D. Deep learning approach to diabetic retinopathy detection. 2020. arXiv preprint arXiv:2003.02261.

  29. Dutta S, Manideep B, Basha SM, Caytiles RD, Iyengar N. Classification of diabetic retinopathy images by using deep learning models. Int J Grid Distrib Comput. 2018;11:89–106.

    Article  Google Scholar 

  30. Sinha A, RP, ANS N. Eye tumour detection using deep learning. In: 2021 Seventh international conference on bio signals, images, and instrumentation (ICBSII). 2021. p. 1–5.

  31. LeCun, et al. Lenet-5, convolutional neural networks. 2015;20:14. http://yann.lecun.com/exdb/lenet.

  32. Lundberg SM, Lee S-I. A unified approach to interpreting model predictions. In: Proceedings of the 31st international conference on neural information processing systems. 2017. p. 4768–77.

  33. He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016. p. 770–8.

  34. Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the inception architecture for computer vision. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2016. p. 2818–26.

  35. Chollet F. Xception: Deep learning with depthwise separable convolutions. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2017. p. 1251–8.

  36. Huang G, Liu Z, Van Der Maaten L, Weinberger KQ. Densely connected convolutional networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition. 2017. p. 4700–8.

  37. Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Commun ACM. 2017;60:84–90.

    Article  Google Scholar 

  38. Iandola FN, et al. Squeezenet: Alexnet-level accuracy with 50x fewer parameters and \(<\) 0.5 mb model size. 2016. arXiv preprint arXiv:1602.07360.

  39. Shah R, Yang Y. Googlenet. popul. Health Manag. 2015.

  40. Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition. 2014. arXiv preprint arXiv:1409.1556.

  41. Oyedotun O, Khashman A. Iris nevus diagnosis: convolutional neural network and deep belief network. Turk J Electr Eng Comput Sci. 2017;25:1106–15.

    Article  Google Scholar 

  42. Ahmed IO, Ibraheem BA, Mustafa ZA. Detection of eye melanoma using artificial neural network. J Clin Eng. 2018;43:22–8.

    Article  Google Scholar 

  43. Ganguly B, Biswas S, Ghosh S, Maiti S, Bodhak S. A deep learning framework for eye melanoma detection employing convolutional neural network. In: 2019 international conference on computer, electrical communication engineering (ICCECE). IEEE; 2019. p. 1–4.

  44. Zabor EC, Raval V, Luo S, Pelayes DE, Singh AD. A prediction model to discriminate small choroidal melanoma from choroidal nevus. Ocular Oncol Pathol. 2022;8:71–8.

    Article  Google Scholar 

  45. Oyedotun OK, Olaniyi EO, Helwan A, Khashman A. Decision support models for iris nevus diagnosis considering potential malignancy. Int J Sci Eng Res. 2014;5:421.

    Google Scholar 

  46. Dai H, MacBeth C. Application of back-propagation neural networks to identification of seismic arrival types. Phys Earth Planet Inter. 1997;101:177–88.

    Article  Google Scholar 

  47. Zurada J. Introduction to artificial neural systems. Eagan: West Publishing Co.; 1992.

    Google Scholar 

  48. Simon H. Neural networks: a comprehensive foundation. Hoboken: Prentice Hall; 1999.

    MATH  Google Scholar 

  49. Shorfuzzaman M, Hossain MS, El Saddik A. An explainable deep learning ensemble model for robust diagnosis of diabetic retinopathy grading. In: ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), vol. 17. 2021. p. 1–24.

  50. Society, A. P. T.-O. Aptos 2019 blindness detection dataset. 2019.

  51. Jeelani H, Martin J, Vasquez F, Salerno M, Weller DS. Image quality affects deep learning reconstruction of mri. In: 2018 IEEE 15th international symposium on biomedical imaging (ISBI 2018). 2018. p. 357–60.

  52. Kugelman J, et al. Effect of altered oct image quality on deep learning boundary segmentation. IEEE Access. 2020;8:43537–53.

    Article  Google Scholar 

  53. Onken M, Eichelberg M, Riesmeier J, Jensch P. Digital imaging and communications in medicine. In: Biomedical image processing. Springer; 2010. p. 427–54.

  54. Rafi TH, Shubair RM. A scaled-2d cnn for skin cancer diagnosis. In: 2021 IEEE conference on computational intelligence in bioinformatics and computational biology (CIBCB). IEEE; 2021. p. 1–6.

  55. Liang Y, He L, Fan C, Wang F, Li W. Preprocessing study of retinal image based on component extraction. In: 2008 IEEE international symposium on it in medicine and education. IEEE; 2008. p. 670–2.

  56. Wu Z, Shen C. Van, Den Hengel A. Wider or deeper: revisiting the resnet model for visual recognition. Pattern Recogn. 2019;90:119–33.

    Article  Google Scholar 

  57. Guo H, Viktor HL. Learning from imbalanced data sets with boosting and data generation: the databoost-im approach. ACM Sigkdd Explor Newsl. 2004;6:30–9.

    Article  Google Scholar 

  58. Santos CFGD, Papa JP. Avoiding overfitting: a survey on regularization methods for convolutional neural networks. ACM Comput Surv (CSUR). 2022;54:1–25.

    Article  Google Scholar 

  59. Abadi M, et al. Tensorflow: a system for large-scale machine learning. In: 12th USENIX symposium on operating systems design and implementation (OSDI 16). 2016. p. 265–83.

  60. Chollet F, et al. Keras: the python deep learning library. Astrophysics Source Code Library. 2018.

  61. Pedregosa F, et al. Scikit-learn: machine learning in python. J Mach Learn Res. 2011;12:2825–30.

    MathSciNet  MATH  Google Scholar 

  62. Shakeri E, Mohammed EA, HA ZS, Far B. Exploring features contributing to the early prediction of sepsis using machine learning. In: 2021 43rd annual international conference of the IEEE engineering in medicine biology society (EMBC). IEEE; 2021. p. 2472–5.

  63. Shakeri E, Crump T, Weis E, Souza R, Far B. Using shap analysis to detect areas contributing to diabetic retinopathy detection. In: 2022 IEEE 23rd international conference on information reuse and integration for data science (IRI). IEEE; 2022. p. 166–71.

  64. Covert I, Lundberg S, Lee S-I. Understanding global feature contributions with additive importance measures. 2020. arXiv preprint arXiv:2004.00668.

  65. Arcadu F, et al. Deep learning algorithm predicts diabetic retinopathy progression in individual patients. NPJ Dig Med. 2019;2:1–9.

    Google Scholar 

  66. Parsa AB, Movahedi A, Taghipour H, Derrible S, Mohammadian AK. Toward safer highways, application of xgboost and shap for real-time accident detection and feature analysis. Accid Anal Prev. 2020;136: 105405.

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. Department of Electrical and Software Engineering, University of Calgary, Calgary, Canada

    Esmaeil Shakeri, Roberto Souza & Behrouz Far

  2. Cumming School of Medicine, University of Calgary, Calgary, Canada

    Trafford Crump & Ezekiel Weis

  3. Department of Engineering, Faculty of Science, Thompson Rivers University, Kamloops, Canada

    Emad Mohammed

Authors
  1. Esmaeil Shakeri
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Trafford Crump
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Ezekiel Weis
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Emad Mohammed
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Roberto Souza
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Behrouz Far
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Esmaeil Shakeri.

Ethics declarations

Conflict of interest

The authors declare that they have no conflict of interest.

Additional information

Publisher's Note

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

This article is part of the topical collection “Recent Trends on AI for HealthCare” guest edited by Lydia Bouzar-Benlabiod.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Shakeri, E., Crump, T., Weis, E. et al. Explaining Eye Diseases Detected by Machine Learning Using SHAP: A Case Study of Diabetic Retinopathy and Choroidal Nevus. SN COMPUT. SCI. 4, 433 (2023). https://doi.org/10.1007/s42979-023-01859-1

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1007/s42979-023-01859-1

Keywords

Search

Navigation