Why have you decided to join medDARE?
I was inspired by the conference «Intelligence. Innovation. Imaging». In April 2019, I firmly decided on what I was looking for in my future career and started looking for a company I could join.
medDare is a unique medical data storage company. By now, it is head and shoulders above the rest of the competition, and I have a chance to be a part of it, helping to create something that, in the nearest future, will relieve doctors, scientists and technicians of certain unnecessary work.
How do you see the future of AI in Healthcare? What are the main trends now, especially taking into account COVID-19 and global pandemic?
Artificial intelligence is a magnificent tool for innovation and improving doctors’ jobs and also the lives of patients, by realizing tasks that are typically done by humans, but in less time, probably with fewer mistakes and at a lesser cost.
Based on the past two years we can already assume that COVID-19 has been an unprecedented catalyzer to all facets of the healthcare industry in a very short amount of time. While it also showed us certain soft spots, for the most part we can observe how rapidly the healthcare technology industry has matured with the necessary solutions and innovations to deal with the pandemic. For example, COVID-19 advanced the use of telemedicine. Detection and recognition features were improved as well – thermal screening, CT pneumonia detection and I am sure this list will keep growing. On the other hand, by no means do we have the variety of health care applications needed to monitor chronic illnesses.
How can AI ease the everyday job of doctors of different specialties?
Deep learning techniques will help doctors of different specialties in a variety of diagnostic tasks. Awareness with the tendency, strengths and limitations of computer-based techniques is critical to support optimal patient care.
Why is it so important to have huge amounts of datasets for validating AI algorithms?
To establish the ground truth for Deep learning – powerful and universal artificial intelligence techniques that can resolve image detection, recognition, and classification tasks that previously required human intelligence, you need to have large amounts of data, but quantity and quality should be of equal importance
You are a seasoned radiologist with many years of experience in medical data annotation. Can you share some examples when data annotation was impossible due to poor quality of datasets?
This is not a matter of producing examples, but more as an axiom, a basic rule for companies that produce datasets should follow, which is strict quality control for the medical data they are offering. Visual inspection of images is a necessary component of understanding large image datasets. Just like most radiologists, I personally believe that the quality of the annotation significantly depends on the quality of the images. In order to create a good end-product, you need to have good components at the beginning.
Which pain points of their future clients can medDARE solve?
medDare is an exclusive storage of a variety of anonymized medical data. Data scientists, radiologists and doctors from different specialties don’t need to look for proper data since medDare can offer any medical data upon request, which will be of high quality without artifacts, anonymized, with or without metadata.