Artificial intelligence and ophthalmology: where does the future lead?

Artificial intelligence (AI) is a branch of computer science focusing on designing methods and frameworks that can simulate humans’ way of thinking. Research area of AI includes robotics, image recognition, language recognition, computer vision, etc. Deep learning method is a class of machine learning algorithms that use a cascade of multiple layers of nonlinear processing units for feature extraction and transformation, which grant computers the ability to learn from and make predictions on data. Deep learning networks have been utilized in diagnosis of certain ophthalmic diseases. A recent study by Liu et al. showed us a very promising algorithm (CC-Cruiser) based on deep convolutional network, achieving excellent diagnostic accuracy of congenital cataract (1). Besides high accuracy of diagnosis, CC-Cruiser also provided doctors with a cloud platform to upload their pictures and get evaluation results. CC-Cruiser is not the first time for AI to be combined with ocular diseases. In 2015, Google released its first product on screening of diabetic retinopathy (DR) (2). Using deep convolutional network, machines are able to find traces of DR in fundus photos.

It seems that the era of AI has arrived. With invention of deep neural networks, machines become able to make diagnosis after learning through big data. However, before we can march forward, we must conquer several obstacles.

First, most of studies chose diseases that can be diagnosed by imaging tests. In ophthalmology, we have various diseases having no anatomical or structural anomalies, but functional abnormalities, which cannot be simply diagnosed by photo. For example, it won’t be easy for machines to learn how to diagnose glaucoma, because it often has no structural abnormality but functional impairment. Second, machines learn from the diagnostic principles by humans, which means without humans’ experience as the guide, machines won’t achieve high learning efficiency and efficacy. Thus, it’s possible for machines to catch up with humans, but nearly impossible to overtake humans. Third, current deep learning methods must be accompanied by large leaning sample size, which limits learning of rare diseases. It’s convenient to capture a photo, but for other tests, such as OCT scans or visual fields, it’s difficult to accumulate clinical data.

In conclusion, AI is a promising tool which can help humans in disease diagnosis and screening. But only after an evolution and optimization in basic theories of deep learning, can we obtain further progress in AI-assisted medical care.

Acknowledgments

Funding: None.

Footnote

Provenance and Peer Review: This article was commissioned and reviewed by Section Editor Dr. Kangjun Li, PhD (State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangzhou, China).

Conflicts of Interest: Both authors have completed the ICMJE uniform disclosure form (available at http://dx.doi.org/10.21037/aes.2018.03.03). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Long E, Lin H, Liu Z, et al. An artificial intelligence platform for the multihospital collaborative management of congenital cataracts. Nature Biomedical Engineering 2017;1:0024.
2. Gulshan V, Peng L, Coram M, et al. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs. JAMA 2016;316:2402-10. [Crossref] [PubMed]