Launching an Artificial Intelligence Platform for Diagnosis and Treatment of Eye Disease in Zhongshan Ophthalmic Center, Sun Yat-sen University
Source: Zhongshan Ophthalmic Center of Sun Yat-sen University
Edited by: Liu Nian
On May 4th, a press conference entitled “Establishment of an Artificial Intelligence (AI) Platform for Diagnosis and Treatment of Eye Disease” was held in the Conference Hall at Zhongshan Ophthalmic Center (ZOC), Sun Yat-sen University. Reporters from the National Xinhua news agency together with others, most notably from the Guangming Daily, the Science and Technology Daily and Nanfang Daily were among the total of more than 20 media representatives present. As the ZOC press spokesman and the Deputy Director of ZOC, Professor Xiaofeng Lin hosted the press conference.
Professor Haotian Lin, as the Principle Investigator of the “Artificial Intelligence team”, gave a presentation about the implementation of artificial intelligence in the diagnosis and treatment of eye diseases.
Professor Haotian Lin at the Press Conference of AI application in ophthalmic clinic
The uneven distribution of high-quality medical resources and the huge conflict between supply and demand is the challenging our health care reform. In an editorial in the April Editions of the People Daily newspaper，it was pointed out that the barrier to health care reform lies in the lack of “manpower” rather than money. Over recent years, AI has made tremendous progress and promoted the clinical application of AI robotics, which provides innovative ideas to solving the problem of the short supply of medical resources. The physicians and scientists at ZOC have developed the first AI platform for the diagnosis of eye disease, and took the lead in exploring an entirely new mode of delivering medical care.
Computerizing ocular images provided a vital chance for a breakthrough of AI application
The diagnosis of eye diseases are mainly based on direct imaging. Portable instruments equipped with a high definition camera are becoming increasingly common, and the images they produce are one of the most available tools with direct clinical merit. Deep convolution neural networks (CNN), are the core technology of artificial intelligence and based on the profound learning of tons of images. In conclusion, image- based diagnosis of eye diseases provides a vital chance for a breakthrough of AI application.
In February 2017, the team from ZOC led by Professor Yizhi Liu joined a team from Xidian University. They established the “CC-Cruiser Congenital Cataracts Artificial Intelligence Platform” by applying deep learning algorithm. This platform simulates the human brain, learns and analyses vast congenital cataract images, and improves itself by continuous feedback. After embedding of the program into cloud-based platform, diagnosis, risk evaluation, and treatment strategy of congenital cataract could be obtained by simply uploading ocular images of patients. Original article was published in Nature Biomedical Engineering 2017 February edition as the cover story (“Auspicious machine learning”) and the elected news (“A cybernetic eye for rare disease”).
Based on this artificial intelligence platform, the project entitled “Artificial Intelligence Diagnosis of Eye Diseases” was initiated on April 2017 at Zhongshan Ophthalmic Center. Besides regular diagnosis and recommending treatments, the ocular images of patients in the AI Clinics are spontaneously uploaded to the CC-Cruiser Platform, which were also diagnosed by AI Robots.
An ophthalmologist in ZOC working with the CC-Cruiser Intelligence Platform
AI assisted health care radiates high-quality medical resources by providing “professional diagnosis and treatment” for more patients
Rare and complex eye diseases, such as congenital cataracts, require specialists for accurate diagnosis and treatment. Currently, most of our specialists are located in big cities. Qualified medical resources are very difficult to access in some remote rural areas. The development of Web-communication and the aide of artificial intelligence make it possible for our qualified medical resources to reach-out to these remote rural areas.
The performance of CC-Cruiser is compatible to a senior physician. Establishing artificial intelligence clinics, as well as joining the AI-aide system with primary care hospitals gives hope that patients can see senior physicians without travelling outside their hometown. By now, CC-Cruiser has incorporated with 3 hospitals to for pilot clinical application.
In addition, Zhongshan Ophthalmic Center has connected with the local hospitals in Tibet, Xinjiang, Yunnan, Qinghai and other remote areas. Large-volume ocular clinical data have been uploaded to the Cloud Platform for diagnosis assistance. With the application of artificial intelligence, cases and images could be screened by an artificial intelligence program and then reviewed by specialists. By doing this, the efficiency could be improved by as much as 70%.
Development plan of AI application in ZOC
Promoting health care reform by exploring new modes of intervention
Long training-period and uneven distribution of qualified specialists have greatly hindered our health care reform. The over burdening of physicians in the tertiary hospitals compared with the light work-loads of the ones in the primary level hospitals is the dilemma of current medical confusion in China. Artificial intelligence can integrate regional medical data and aide the diagnosis for the physicians in primary care hospitals. It provides new strategies to rapidly improve the diagnosis level in Ophthalmology for the primary care hospital.
There are approximately 30,000 physicians based in primary care hospitals. By promoting and implementing the artificial intelligent aide diagnosis system, the accuracy and efficiency of diagnosis can be greatly and rapidly improved. Meanwhile, “Guangdong Ophthalmic Diagnosis and Treatment Innovation Technology and Engineering Center” will be established and long-term specialized ophthalmologic AI physicians will be assigned to the areas where physicians are scarce to secure and improve the ophthalmologic care, offered in those areas.