On last July 7th, in APTOS 2018 Singapore, Dr. Rim gave a presentation about his research, “Retinal Damage as the Window to Cardiovascular Organ Damage: Use of Big Data & Machine Learning”.
Dr. Rim, one of the co-founders of Medi-Whale Inc., is engaged in Yonsei University College of Medicine as a professor and ophthalmologist in Severance Hospital.
As a professional ophthalmologist, he suggested his idea that a retina is a window to investigate the association between microcirculation abnormalities and cardiovascular diseases.
However, among the researchers already studying the relationship between cardiovascular diseases and eye vessels, Dr. Rim focused on the question: What is the most important marker in cardiology?
He and Medi-Whale Inc. attempted to evaluate the association between retinal image and coronary artery calcium score using deep learning, as it can analyze and catch distinctive features of individual retinal fundus images and heart CT.
While processing this research, Dr. Rim used big data from Korea National Health Insurance Service (KNHIS) and stored a huge amount of data to enhance the accuracy of predicting cardiovascular risk factors.
Additionally, Dr. Rim mentioned deep convolutional network architectures, which can enhance the learning ability and reduce error rates between official results and research results.
Medi-Whale Co. sponsored the 3RD Asia Pacific Tele-Ophthalmology Symposium (APTOS) held in Singapore. In their exhibition booth the ‘Dr. Noon’, Medi-Whale’s product, was displayed on July 7-8, 2018.
APTOS was founded by tele-ophthalmologists in Asia-Pacific region in May 2016, aggregating professional researchers, technicians, and clinicians to share, collaborate, exchange individual perspectives, knowledge or ideas.
Along with Medi-Whale, Zeiss, Topcon, Novartis, and Alcon also sponsored APTOS 2018 and introduced technology which can precisely diagnose patient’s eye diseases.
At APTOS 2018, Medi-Whale opened a booth to introduce ‘Dr. Noon’, a deep learning based retinal fundus screening program, specially attracting companies from India, Philippine, and middle-east Asia.
By using red and green colored dots, this program will identify if a patient’s eye will contain glaucoma, opacity, or retina diseases that can be frequently triggered from elder or diabetes.
As Dr. Noon can assist basic eye examination for elderly and diabetic patients, by analyzing retinal fundus images, physicians are now expected to be able to decide if patients need further diagnosis from ophthalmologists or not.
Dr. Noon is expected to assist internal clinics to provide better service for patients, especially for those in area which professional ophthalmologists are not localized.
Medi-Whale is now planning to develop a program that predicts cardiovascular diseases via analyzing retinal fundus images.
In terms of heart diseases, researchers from Asan hospital, Seoul represented AI which can concisely predict ventricular tachycardia before 1 hour from when it strokes.
This increased the viability of patients from heart attack but required the human effort to set an artificial neural network which was based on specific feature.
Furthermore, researchers from Vuno institution, developed DeepEWS, which used RNN (Recurrent Neural Network) deep learning system to learn 7 types of information from individual patient.
RNN was used for this system as researchers wanted to predict heart disease before 24 hours and concentrate on data variance and reducing false warning alert.
Similarly, Cardiogram applied deep learning to the heartbeat sensor in wearable device to identify atrial fibrillation and atrial flutter as heartbeat’s distinctive rhythm was able to recognize some irregular pulse.
Irregular pulse specificity of wearable device was evaluated with two criteria (Sequence F1- figure out arrhythmia on proper time, Set F1- detect arrhythmia from whole data).
As a result, deep learning was showing much higher specificity with both sequence F1 and set F1 when it compared to cardiologist.
This research identified that AI could easily classify two types of atrioventricular block which most professional cardiologists were having difficulty to identify.
If data aggregation (electrocardiogram, heartbeat rate…etc), analyzing, and classifying can be constantly processed through wearable device or apple watch, predicting heart diseases with high accuracy will be developed.