Introduction: The Impact of AI in Ophthalmology
In 2020, the value of the global artificial intelligence (AI) in the healthcare market was around $6.7 billion. Estimates show in a period between 2021 and 2028; this value is likely to expand at a CAGR (compound annual growth rate) of 41.8%. Indeed, artificial intelligence is present in different aspects of healthcare, including ophthalmology. The main objective of this post is to provide more info regarding the impact of AI in ophthalmology.
Impact of AI in ophthalmology
The impact of artificial intelligence in ophthalmology is truly substantial. Evidence confirms AI can improve patient access to clinical screening, diagnosis and also reduce costs, particularly in high-risk groups.
A paper from the International Journal of Ophthalmology found that, compared to clinical grading, the sensitivity and accuracy of AI to detect eye diseases was the following:
- Age-related macular degeneration – 75% to 100%
- Cataract – over 70%
- Glaucoma – 63.7% to 93.1%
- Non-proliferative diabetic retinopathy – 75% to 94.7%
- Proliferative diabetic retinopathy – 75% to 91%
- Retinal vein occlusion – over 97%
- Retinopathy of prematurity – over 95%
The full impact of AI in ophthalmology requires further research, but one of the most recent papers on this topic revealed its implementation could be fully and semi-automated. The fully automated AI model would work without any human involvement. Artificial intelligence would be able to initiate referrals to ophthalmologists appointments when necessary. It could also determine which patients are suitable for community-based monitoring. On the other hand, the semi-automated AI model works with human involvement in different ways, which only serves to enhance classification made by deep learning. The AI technology would be able to fill all the gaps in the screening process and thereby prove to be a cost-effective solution in ophthalmology.
Source: YouTube | Intro to AI in Ophthalmology
Building blocks of AI in ophthalmology
The field of ophthalmology is a fertile ground for artificial intelligence and studies associated with it. In ophthalmology, artificial intelligence entails many techniques, devices, and approaches, including color fundus photography, computerized visual field, and optical coherence tomography.
A wide range of approaches to building artificial intelligence systems to detect and automatically measure pathologic features within an eye image is available today. These include:
- Simple automated detectors – the simplest form of artificial intelligence
- Basic machine learning – the algorithm receives the basic rules regarding disease features as well as images of the affected and non-affected eye. The machine then “learns” the differences.
- Advanced machine learning – contains one or two layers of neurons (small computing units) that are interconnected and mimic the multilayered visual cortex structure. For example, it can associate specific outputs of detectors of disease features with diagnostic outputs.
- Deep learning with CNNs (convolutional neural networks) – multiple interconnected neuron layers, works to make diagnoses that are similar to what human graders would make.
- Disease feature-based vs. image-based learning – many researchers that develop AI in ophthalmology prefer designing machine learning algorithms according to clinically known disease characteristics.
Also Read: The Role of Artificial Intelligence in Healthcare Documentation
Current challenges in implementation
Although AI can revolutionize ophthalmology, it also comes with important challenges that should be addressed, improved, and overcome. These challenges are of practical, technical, and socio-cultural in nature.
Practical challenges
The Impact of AI in Ophthalmology comes with its own set of challenges. The greatest challenge in the implementation of AI in ophthalmology is to make it as practical as possible. This mission may require combining several systems with performance that is clinically acceptable. But, these systems should also be able to receive images from frequently used devices, even if the quality of those images varies. Other practical challenges include patient selection, addressing misclassified patients such as false positives and negatives, and the restrictive nature of deep learning systems, which are only validated for independent classification of one eye-related disease at a time. What happens in the misclassification scenario is also a practical challenge that should be addressed to determine the liability, i.e., whether the AI provider is responsible or the clinician.
Technical challenges
The most significant technical challenge in the implementation of AI is the need for adequate training data and external validation. Introducing artificial intelligence also requires labeling of input data for the training process, the task that requires expert practitioners. As a result, the risk of human error increases. The whole process of labeling datasets and calibrating the system can become time-consume and thereby delays the implementation of AI. Both current and emerging AI techniques in ophthalmology require expert consensus and the development of standards and guidelines to evaluate the performance and accuracy of these systems.
Socio-cultural challenges
Implementation of AI in clinical practice is accompanied by various socio-cultural challenges. A lot of these challenges are associated with the usual differences in access to healthcare. We can take Asia as an example here. This vast continent has major differences in the access to healthcare but also spending and consumption within the healthcare industry. Many areas have limited resources, inadequate infrastructure, and other factors that would affect the implementation and effectiveness of artificial intelligence.
While some parts of this continent are some of the wealthiest and most developed places in the world, others have unreliable electricity and the internet. In these settings adopting artificial intelligence is challenging. To overcome these problems, it would be necessary to employ portable solutions with a rechargeable power supply.
Implementing AI is not just a matter of installing software and devices. Although we live in the 21st century, we still witness significant differences within the healthcare industry. Many hospitals across the globe, not just in Asia, have no finances to adopt AI. Lack of infrastructure and electricity problems are also major challenges.
Also Read: Artificial Intelligence in Healthcare Business Process Improvement
Conclusion: The Impact of AI in Ophthalmology
Artificial intelligence has become omnipresent; it is utilized in many industries, including healthcare. The use of AI technology in ophthalmology is associated with great sensitivity and accuracy, cost-effectiveness, and improved processes with minimum human involvement. However, AI implementation has several challenges that should be addressed for easier access to the screening and diagnostic processes. As technology evolves, so will AI in ophthalmology. One thing is for sure; it will be interesting to follow the latest developments in AI and see how they can improve the screening and diagnosis of patients.