San Jose, CA, April 01, 2018 --(PR.com
)-- To nearly 400 million Chinese people affected by diabetes and pre-diabetes -- more than the population of the United States -- serious eye trouble including blindness are a daily risk. Early detection can prevent most of these cases, and now for the first time, to the 125 million without access to trained ophthalmologists, new hope is available thanks to Airdoc DR Detection, an NVIDIA-powered screening system developed by Airdoc, China’s leading medical artificial intelligence healthcare company.
Speaking at NVIDIA’s GTC 2018 Conference in San Jose, California, Dr. Li You from Shanghai Changzheng Hospital Department of Ophthalmology, described her medical team's support activities in Xinjiang and Tibet, remote areas of Western China that lack not only drugs and doctors, but also basic hospital support staff. Some, like the Haibei Tibetan Autonomous Region in Ningxia, are so rural that her staff can reach them only through three days on horseback -- hardly ideal conditions for treating conditions like diabetic retinopathy, where prompt treatment is essential.
Now, AI scientists at Airdoc, working with experts from the Changzheng Hospital developed an ophthalmic artificial intelligence algorithm. Hospital experts marked the disease sites one by one on a large number of fundus maps, each one of which was cross-labeled by multiple ophthalmologists. The accuracy of the algorithm is comparable to that of professional ophthalmologists in the Sanjia Specialist Hospital.
Prof. Li You said, "Airdoc's algorithm is deployed in a number of rural hospitals. After a patient has taken a photo of the fundus in a primary hospital, she will soon be able to receive artificial intelligence interpretation and doctor-confirmed screening on her mobile phone. This allows a doctor and patient to not see the results of the eye fundus disease recognition, but if necessary, it can also preserve the eye fundus photos for future comparison."
One patient who benefited is Sheng Lao Lao, from Zhejiang, a local county hospital that cooperated with the Long March Hospital. After Airdoc’s DR Detection system identified his high risk for retinopathy, he was immediately referred to the Shanghai Changzheng Hospital for active intervention. Now his vision has been restored.
In order to solve the screening problem among rural patients, Airdoc collected millions of fundus photographs from top domestic and foreign top hospitals. Several eye specialists were asked to cross-mark and choose to label together as the gold standard. Airdoc specialists then built a multi-layer convolutional neural network based on a large amount of data. Here Nvidia's powerful GPU provides essential horsepower for medical artificial intelligence.
Airdoc has been working for over three years with different governments and agencies in areas of China as diverse as Hainan, Anhui, Shanghai and Guandong, conducting large scale screenings for chronic diseases.
Given the contribution of Airdoc in the medical AI field, the company was selected as a member of China’s prestigious Standing Committee of the Artificial Intelligence Society and has submitted its AI based diagnosis application with CFDA3 in China.
According to Emma Xu, Airdoc ophthalmic product manager, "No medical product is successful unless it's proven in a clinical setting and we are proud to see Airdoc's persistence being rewarded with success in among patients whose lives are being helped. Our dream is to make artificial intelligence technology to ubiquitous thateveryone can enjoy the best possible medical services."