New Australian research has identified clinical registries like The Save Sight Registries to be a treasure trove of high-volume training data that could power artificial intelligence (AI) diagnostic models of the future.
AI is set to revolutionise the way doctors and patients interact with ophthalmic healthcare, with several published diagnostic AI models already boasting performance on par with eye specialists in the detection of diabetic retinopathy, macular degeneration, and glaucoma.
To be useful in the clinic, however, AI requires large volumes of real patient data to produce good predictions.
The results of a systematic literature review – ‘Artificial Intelligence and Ophthalmic Clinical Registries’ have been published in the latest issue of The American Journal of Ophthalmology1 as a result of research out of the Save Sight Institute, the University of Sydney, and the Sydney Eye Hospital.
The paper sets out the various ways in which AI has been applied to other clinical registries in ophthalmology, and highlights the challenges associated with creating, maintaining, and mobilising collected data.
Authors, Drs Luke Tran, Himal Kandel, Daliya Sari, Christopher Chiu, and Professor Stephanie Watson OAM now aim to create their own AI model, based on the Save Sight Registries, that may overcome these challenges.
“Through our literature search, we observed a distinctive lack of deep learning models being applied to clinical registry data and a wide variation in methods and validation measures suggesting that the potential of clinical registries to train AI models has not been fully realised,” said lead author, Dr Tran.
The majority of studies included in the review employed conventional machine learning algorithms as opposed to deep learning algorithms, which have recently garnered widespread attention in both the scientific literature and general media. Deep learning algorithms, while usually more accurate and powerful than conventional algorithms in a range of tasks, are computationally more expensive and technically complex, often requiring heavy involvement and collaboration with computer science domain experts to implement.
Prof Watson, head of the Corneal Research Group who directed the study said, “It’s important for registry custodians to understand the administrative and ethical barriers that may limit effective collaborations with computer scientists. High quality collaborations with domain experts will allow us to create safe and clinically impactful AI.”
“Having observed the great potential of registry trained AI and (the) potential pitfalls, we intend to make use of the lessons learned here to explore ways of applying AI to the Save Sight Keratoconus Registry with the goal of improving the delivery of ophthalmic healthcare,” added Dr Kandel, a senior researcher within the Corneal Research Group.
The researchers hope that these findings will help clinicians understand the current state of AI applied to ophthalmic clinical registries, the opportunities and barriers associated with mobilising registry data, and the need for early involvement of machine learning experts in the development of clinically deployable AI.
Reference
- Tran L, Himal Kandel H, Watson SL. Artificial intelligence and ophthalmic clinical registries, American Journal of Ophthalmology, 2024, doi.org/10.1016/j.ajo.2024.07.039