shutterstock_740559436 [Converted] 01_800

AI and the Eye: What we need to know about artificial intelligence and big data analytics in ophthalmology

The first symposium on “Artificial Intelligence and Big Data Analytics in Ophthalmology for Global Eye Health: Opportunities and Challenges” kicked off with a panel of seven speakers sharing their findings on this timely topic. 

Dr. Marcus Ang, head of Cornea and External Eye Diseases Service, Singapore National Eye Center (SNEC); Dr. Michael Chiang, director of National Eyes Institute, United States; R.V. Paul Chan, chair, Department of Ophthalmology and Visual Sciences at University of Illinois College of Medicine; Dr. Linda Lam, professor of ophthalmology from USC Keck School of Medicine, Los Angeles; Dr. Mohamad Aziz Salowi, ophthalmologist at Selayang Hospital, Selangor, Malaysia; Dr. Jude Stern, head of Knowledge Management at International Agency for the Prevention of Blindness (IAPB); and Wei-Chi Wu, professor and chairman of ophthalmology at Chang Gung Memorial Hospital, Taoyuan, Taiwan, discussed the impacts and challenges of artificial intelligence (AI) and big data analytics in ophthalmology. 

Challenges in eye care: AI and Big Data to the rescue?

Dr. Jude Stern began by presenting five bullet points on the challenges in global eye care that can be addressed with AI and Big Data:

  • inadequate access to eye care services in many parts of the world;
  • shortage of trained eye care professionals in certain regions; 
  • limited availability of accurate and timely patient data for diagnoses and treatment;
  • difficulty in identifying and prioritizing patients for treatment due to a large patient population and limited resources; 
  • and the need for continuous monitoring and follow-up care for patients with chronic eye conditions.

“There are many opportunities to increase access to eye health services using AI and Big Data,” she said. “AI-powered screening tools can aid early detection of eye diseases, allowing for timely interventions to prevent vision impairment. Big data analytics can enable healthcare providers to identify eye health trends, the development of targeted interventions, and resource location.

Teleophthalmology, enabled by AI, can extend eye health services to remote areas. Not to mention, AI algorithms can assist in personalized treatment plans by analyzing patient data leading to reduced healthcare costs, thus enabling providers to see more patients and overall enhance the efficiency of eye health services.

“We need to make sure that all populations benefit from these developments, and that’s a huge opportunity for anyone in technology or if we have an opportunity in the global health agenda. This means that they need services [to be] accessible, affordable, and inclusive,” she added.

The role of AR and VR in ophthalmology

Prof. Dr. Linda Lam said that augmented reality (AR) and virtual reality (VR) in eye care can create immersive and interactive experiences in multiple arenas in ophthalmology, including those affecting patient care, surgery, education, and research. 

Opportunities for patient care with this technology include visual rehabilitation, patient education, treatment option for immersive simulations, and improved access and in-home disease monitoring. 

“With refinements and innovation, wearables in VR and AR for vision rehabilitation now allows for high magnification, text-to-speech or optical character recognition, facial recognition, code readers, and assistive technologies from live people,” she said. 

VR and AR, she added, also offer the opportunity in research, by enabling data visualization and clinical trials. Patient conditions can be monitored remotely, reducing the burden of eye care visits when data collection can be done in the comfort of the patient’s home.

Improving access to eye care with AI and telemedicine

In his presentation entitled “The Opportunities of AI ROP (Retinopathy of Prematurity) Screening in Global Eye Settings,” R.V. Paul Chan highlighted some of the research findings on AI’s usage for the prevention and screening of ROP in countries such as India, Mongolia, and Nepal. 

He said that on the health equity and disparities in ROP care, there is a need for systematic evaluation to remove the barrier to care. 

“Telemedicine and AI can be part of the solution in ophthalmology and technological innovation for improving access to health care in different patient populations,” he shared. “Prevention of ROP is the ultimate goal, thus there should be advocacy and strong partnerships to train the workforce.”

Dr. Marcus Ang, on the other hand, shared some of the AI opportunities beyond the pandemic, and how SNEC is implementing some of these technologies and the challenges associated with them. At the same time, the institute is looking at the accuracy of assessing refractive error and visual acuity in children based on their telemedicine and home monitoring program.

He shared a ‘virtual clinic’ model for myopia, which features three stages: primary eye care, specialist review, and clinical management. This includes myopia evaluation on early identification of complications, potential biomarkers for progression, and monitoring treatment.

“Myopia screening now includes new models for detection and triage whereby telemedicine, virtual clinic, and AI work together,” he said. 

The need for data sharing and harmonization

While there are now new opportunities, such as imaging plus digital infrastructure, OCT, Widefield imaging, and biomechanics, among others, there are also challenges, including implementation, infrastructure, and integration. 

Dr. Michael Chiang said, as we have unprecedented access to large-scale data in many fields, data sharing and establishing standard data representations are important to avoid biases. 

He reckoned that some key unanswered research questions in AI include a variety of findings in individual patients, generalizability and bias, and unclear “ground truth.” 

“Another problem is one can’t easily exchange data among imaging systems as there are enormous barriers to clinical care and research involving images,” he shared. “Therefore, we need data sharing and harmonization. We also need a common data model for ophthalmology.”

Editor’s Note: A version of this article was first published in APAO 2023 Show Daily, Issue 2.

Notify of
Inline Feedbacks
View all comments