SOE 2023: AI in Eye Care in Japan: Facing the Emerging Challenges


Artificial intelligence (AI) is going to be highly instrumental in patient care in all medical specialties and will be highly relevant in eye care.

©Lee /

Artificial intelligence (AI) is going to be highly instrumental in patient care in all medical specialties and will be highly relevant in eye care. (Image Credit: Adobe Stock/Lee)

Reviewed by Tetsuro Oshika, MD, PhD

Tetsuro Oshika, MD, PhD, shed light on how AI will be useful in caring for patients in Japan at the European Society of Ophthalmology in Prague. He is President of the Japanese Ophthalmological Society, President-Elect of Asia Pacific Academy of Ophthalmology, and Professor/Chairman, Department of Ophthalmology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan.

Ophthalmologists are faced with a number of challenges, specifically, an increased number of aging patients and those who rely more and more on excellent vision at all distances and fewer ophthalmologists to evaluate patients as well as the clinicians having the ability to balance their work with their private lives. The last is the areain which AI will be of the greatest assistance.

He demonstrated that in a great number of cases, the retinal examinations may appear to be normal when in reality the accuracy of the normal retinal examinations may be totally lacking. When AI is applied to those same normal-appearing retinas, early disease states become apparent in 91.1% of the cases.

The same is true for patients who present with corneal diseases. During the standard corneal examination, patients can present and appear to be normal, when in reality a number of conditions can be present: infection, infiltration, scarring, deposits, bullous keratopathy, cataract, tumors, and acute glaucoma.

The numbers do not lie, as he demonstrated when he compared the respective AL findings (percentages of cases diagnosed correctly) to those in which the ophthalmologists were correct: normal eyes, 94% vs. 84%; infection, 88% vs.75%; infiltration, 8 74% vs. 58%; scarring, 83% vs. 60%; tumors, 98% vs. 92%; acute glaucoma 100% vs. 24%; cataract, 75% vs. 74%; and bullous keratopathy, 85% vs. 70%. AI was also superior when the physician accuracy rates were determined based on experience, with residents reaching correct diagnoses 73.6% of the time, specialists 82.2%, and AI 89.8%.

The percentages involving the presence of acute glaucoma underscore the high clinical relevance of the use of AI in detecting the disease.

Dr. Oshika also discussed the important of the smartphone AI project in which patients will be able to monitor themselves and transfer the information to clinicians. This ability will ensure early treatment as it is needed and lighten the in-office treatment burden for both patients and clinicians. He showed that when the smartphone images were assessed by AI clinicians were able to determine easily which patients need urgent assessment, consultation, non-urgent consultations, and no consultations.

AI also will be useful in ophthalmic surgery. In his recent study,1 Dr. Oshika and colleagues used an AI-based system for preoperative safety management in cataract surgery, including facial recognition, ocular laterality confirmation, and intraocular lens (IOL) parameter verification. In 171 patients undergoing phacoemulsification and IOL implantation an iPad mini (Apple Inc.) camera captured patients’ faces, location of surgical drape aperture, and IOL parameter descriptions on the packages, which were then checked with the information stored in the referral database. They reported that “the authentication rates on the first attempt and after repeated attempts were 92.0% and 96.3% for facial recognition, 82.5% and 98.2% for laterality confirmation, and 67.4% and 88.9% for IOL parameter verification, respectively. After authentication, both the false rejection rate and the false acceptance rate were 0% for all three parameters. An artificial intelligence-based system for preoperative safety management was implemented in real cataract surgery with a passable authentication rate and very high accuracy.”

AI can also analyze hand motions intraoperatively, ie, arm vertical and horizontal movements movements and those of the wrist and fingers, which, he said, will scientifically decipher the surgeon’s expertise during surgery.

Kiuchi G, Tanabe M, Nagata K, et al. Deep learning-based system for preoperative safety management in cataract surgery. J Clin Med. 2022;11:5397;

Related Videos
Retinal Inner Layer Disorganization and OCT in Uveitic Macular Edema: Insights from Dr. Amitha Domalpally
ARVO 2024: Study Reveals Faricimab's Potential for Extended Dosing in nAMD
© 2024 MJH Life Sciences

All rights reserved.