JOIN NOW

How is the advent of Machine Learning enhancing Ocular Diagnostics?

by | Dec 1, 2024 | Manuscripts | 0 comments

Kristi Sharma, M. Optom Student

Sankara Nethralaya Academy, Chennai, India

 

What is Machine Learning?

Machine learning, a branch of artificial intelligence (AI), adjusts its parameters based on input data to improve performance. Deep learning, a type of machine learning, utilises various algorithms to analyse input data and autonomously detect patterns and structures within the dataset. Machine learning aims to utilise computer generated algorithms to make predictions and responses to data. The layers of algorithms used in deep learning form a neural network that replicates the human brain.(1)

Integration into Ophthalmic Sciences

Machine learning is integrated into tele ophthalmology to enhance screening efficacy by covering large number of images and reducing clinician review.(2) It was introduced in the year 1959 by Arthur Samuel and saw an exponential growth into Ophthalmology during the COVID-19 pandemic.(2)

The various fields of ophthalmology that incorporated Machine Learning are as follows:

  • Dry Eye: Dryness of the front surface of the eye associated with irritation and burning sensation.
  • Conjunctivitis: Inflammation of the conjunctiva due to causes like bacterial or viral infestation or any non-infectious causes like seasonal allergy.
  • Keratoconus: Progressive conical protrusion of the cornea of the eye resulting in distorted vision.
  • Cataract: Opacification of the natural crystalline lens of the eye.
  • Amblyopia: Lazy eye due to uncorrected refractive error, squint, or vision deprivation.
  • Glaucoma: Pathology of the optic nerve behind the eyeball resulting in peripheral vision loss.
  • Diabetic Retinopathy: Retinal condition caused due to prolonged and uncontrolled Diabetes Mellitus.
  • Age Related Macular Degeneration: Pathology of the central retina due to ageing process of the eye.
  • Retinopathy of Prematurity: Retinal pathology caused due to premature birth and low birth weight.

Instruments Used

Machine learning and AI is used in posterior segment diagnostic procedures such as standard fundus photography, ultra-widefield fundus photography, optical coherence tomography, optical coherence tomography angiography and smart-phone based fundus photography.(3) In anterior segment diseases, slit lamp images, lacrimal scintigraphy, meibography, and anterior segment optical coherence tomography utilises deep learning algorithms.(4-6) For glaucoma, disc photography and spectral domain optical coherence tomography have been used with deep learning.(7)

Screening Efficacy

Studies have shown that in anterior chamber, there have been 97.6% and 95.4% accuracy for eyelid and atrophy segmentations, respectively.(6) In identifying keratoconus, accuracy was reported to be 99.45% to 99.57%.(8) High sensitivity has been reported by various studies in diagnosing diabetic retinopathy using AI models. The sensitivity ranged from 91.7% to 95.7%.(9,10) For age related macular degeneration, combining deep learning modalities increased the accuracy from 91% to 96%.(11)

Conclusion

The inclusion of AI in medical sciences brings promises of enhanced healthcare and wider coverage of facilities. However, it is of utmost importance to monitor all services, keep check on regular updates, ensure patient safety and data security that may pose challenges in smooth operations. Machine learning will provide the best of care that is possible and increase the efficiency of ophthalmic diagnostics.

 

References

  1. Srivastava, O., Tennant, M., Grewal, P., Rubin, U., & Seamone, M. (2023). Artificial intelligence and machine learning in ophthalmology: A review. Indian journal of ophthalmology, 71(1), 11–17. https://doi.org/10.4103/ijo.IJO_1569_22
  2. Nikolaidou, A., & Tsaousis, K. T. (2021). Teleophthalmology and Artificial Intelligence As Game Changers in Ophthalmic Care After the COVID-19 Pandemic. Cureus, 13(7), e16392. https://doi.org/10.7759/cureus.16392
  3. Grzybowski, A., & Brona, P. (2021). Analysis and Comparison of Two Artificial Intelligence Diabetic Retinopathy Screening Algorithms in a Pilot Study: IDx-DR and Retinalyze. Journal of clinical medicine, 10(11), 2352. https://doi.org/10.3390/jcm10112352
  4. Kim MC, Okada K, Ryner AM, Amza A, Tadesse Z, Cotter SY, et al. (2019). Sensitivity and specificity of computer vision classification of eyelid photographs for programmatic trachoma assessment. PloS One.14:e0210463. https://doi.org/10.1371/journal.pone.0210463.
  5. Park Y-J, Bae JH, Shin MH, Hyun SH, Cho YS, Choe YS, et al. (2019). Development of predictive models in patients with epiphora using lacrimal scintigraphy and machine learning. Nucl Med Mol Imaging. 53:125–35.
  6. Wang J, Yeh TN, Chakraborty R, Stella XY, Lin MC.(2019). A deep learning approach for meibomian gland atrophy evaluation in meibography images. Transl Vis Sci Technol. https://doi.org/10.1167/tvst.8.6.37
  7. Thomas SM, Jeyaraman MM, Hodge WG, Hutnik C, Costella J, Malvankar-Mehta MS. (2014). The effectiveness of teleglaucoma versus in-patient examination for glaucoma screening:A systematic review and meta-analysis. PLoS One. 9:e113779.https://doi.org/10.1371/journal.pone.0113779.
  8. Dos Santos VA, Schmetterer L, Stegmann H, Pfister M, Messner A, Schmidinger G, et al. (2019). CorneaNet:Fast segmentation of cornea OCT scans of healthy and keratoconic eyes using deep learning. Biomed Optics Express; 10:622–41
  9. Heydon P, Egan C, Bolter L, Chambers R, Anderson J, Aldington S, et al. (2021) Prospective evaluation of an artificial intelligence-enabled algorithm for automated diabetic retinopathy screening of 30 000 patients. Br J Ophthalmol ;105:723–8
  10. Wang X, Ji Z, Ma X, Zhang Z, Yi Z, Zheng H, et al. (2021). Automated grading of diabetic retinopathy with ultra-widefield fluorescein angiography and deep learning. J Diabetes Res ; 2611250. https://www.doi.org/10.1155/2021/2611250
  11. Vaghefi E, Hill S, Kersten HM, Squirrell D. (2020) Multimodal retinal image analysis via deep learning for the diagnosis of intermediate dry age-related macular degeneration:A feasibility study. J Ophthalmol ;2020:7493419. https://doi.org/10.1155/2020/7493419

 

Submit a Comment