OCT imaging with AI screening improves retinal diagnosis

22 May 2019

System from Optovue and Ping An Technology has now completed a clinical trial.

A screening system said to be the first integration of OCT retinal imaging and artificial intelligence-based image processing for detection of retinal lesions has completed a multi-center clinical trial.

The system has been developed jointly by US OCT vendors Optovue and Ping An Technology, a subsidiary of China’s Ping An finance conglomerate founded in 2008 to focus on new technologies in health care, finance and smart cities.

According to Ping An Technology, the system's clinical trial commenced in December 2018 and was conducted at three centers in Shanghai: Fudan University; Shanghai Jiao Tong University; and Tongji University. A total of 784 study subjects were included.

Announcing the results, Ping An stated that the workflow from the start of an OCT examination to the generation of an intelligent screening report using the system could be completed in three minutes. The accuracies of the tests in terms of image quality assessment, lesion detection, and referral urgency assessment were 99.2 percent, 98.6 percent, and 96.7 percent respectively, according to the company.

"This is the first time that multi-center prospective research on AI-assisted ophthalmic medical products has been conducted," commented Xie Guotong of Ping An. "With the strong healthcare ecosystem of Ping An Group, primary medical institutions will be empowered with high-quality ophthalmic AI models to assist in carrying out diagnosis and treatment at different levels."

The potential use of AI and deep learning (DL) in ophthalmology has been a focus of interest for some time, given the parallel emergence of the technology in wider image recognition and data processing scenarios.

A 2018 survey published by the British Journal of Ophthalmology concluded that AI-based models were potentially capable of high levels of accuracy in the diagnosis of many ophthalmic diseases, but challenges relating to the diversity of patient data used to train the system and the natural resistance of clinicians to automated "black-box" systems were likely to persist.

"Although there are challenges ahead, DL will likely impact on the practice of medicine and ophthalmology in the coming decades," the review said.

Analysis and diagnosis

Ping An Technology first demonstrated its AI healthcare technologies in August 2018 at the China Smart City International Expo in Shenzhen, as part of the company's push towards publicising the medical dimension of its smart city systems. That vision ultimately includes other topics such as the diagnosis and prevention of lung disease and the prediction of risk for cardiovascular conditions, as well as ophthalmology.

At the Shenzhen presentation Ping An described an eye screening platform designed to study and analyze thousands of data points collected during an examination of the fundus, to screen for diabetic retinopathy and age-related macular degeneration (AMD) by identifying the physiological signs of the conditions.

"Based on the AI algorithms, the platform identifies the lesions automatically and smartly assists in the diagnosis," it said at the product announcement.

"The system is able to control the quality of the imaging diagnosis using AI technology with the aim of reducing the number of missed imaging-based diagnoses and improving the usefulness of imaging for clinical diagnosis as well as productivity at hospitals. The platform provides physicians with real-time visibility of data on quality control results and the comparisons between quality and quantity of the images examined in terms of the device, location of the lesions, profile of the patient and timestamp."

Discussing the new clinical trial results, Sun Xinghuai from Fudan University commented that the research "presented a highly accurate evaluation of the intelligent OCT fundus screening system developed by Ping An, which covers the quality of image acquisition, lesion detection and urgency, suggesting that AI will be extremely helpful to primary medical institutions in the future."