On top of AIMed Cardiology, AIMed will also be hosting AIMed Radiology again this November, in association with the American College of Radiology.
Last week, AIMed invited two of the virtual conference speakers – Dr. Ty Vachon, Assistant Professor of Radiology and Radiologically Sciences, Uniformed Services University of the Health Sciences and Dr. Orest Boyko, Associate Professor of Radiology, University of Southern California to discuss some of the key trends in Radiology with AIMed Founder and Chief Artificial Intelligence Officer at Children’s Hospital of Orange County (CHOC) Dr. Anthony Chang over a 30-minute webinar.
Dr. Chang also announced the launch of the new American Board of Artificial Intelligence in Medicine (ABAIM) and welcome active participations from both medical and non-medical communities and get certified for what they have learnt.
Empowering clinicians with basic understanding of data science
Both speakers agreed the ongoing COVID-19 pandemic has generated a very interesting dynamic. For example, people are more open to collaborations. As Dr. Vachon noted, it takes a substantial amount of time to clean a CT scanner after it has been used by a suspected or confirmed COVID-19 patient. There are also many features that look like COVID-19 on CT scans and it has been tricky since the beginning to ensure a reasonable level of sensitivity.
More importantly, are there ways to better leverage clinical and imaging data to determine how long a patient is going to be on the ventilator or stay in the ICU? All these challenges have become accelerators and urge individuals with various expertise to work together. Dr. Boyko agreed. Since it’s a novel coronavirus, the medical community is learning in very strict real-time.
Given his immuno-chemistry background, Dr. Boyko has been educating some of the data scientists who have access to clinical data to try to study the immunology part of these information. He hopes by pulling all the data and scientific evidence from different sources, the system will have answer to some of these more rapid questions.
On the other hand, Dr. Chang felt it is not just having clinicians work closer with data scientists to bring about clinical relevance but also trying to empower the clinicians a little more in terms of basic understanding of data science and different ways of deploying AI, not just focusing on methods that are image-related, which are already receiving a lot of publicity.
Going back to the starting point
Dr. Vachon added it’s all about learning up the vocabularies in the respective domain. He recalled about 10 years ago, when he was still a resident, he owned a new phone which automatically allowed him to locate all his friends and allocate their faces next to their contact numbers. Yet, when he was at work, he had to manoeuvre between four processors and manually clicking and dragging radiological measurements. He wondered why can’t this be automated like what his new phone did.
However, Dr. Vachno later realized that it was still too expensive and there were not enough computing power to support such automation in radiology at that time. Nonetheless, he continued to keep up with the advancement over the years until everything took off recently. Dr. Boyko said he became interested in the field when he came across a large tech firm recruiting for radiologists five years ago. He thought the only way to find out why a tech company is going to have relationship in radiology is to respond to their recruitment. That became an opportunity for him to come on board.
For Dr. Chang, he remembered watching Star Trek when he was young and was impressed by Captain Kirk’s computer, which can answer the many questions posed to it. When he became a resident, he could not help but question why such computers don’t exist in medicine to save lives and how awesome it will be if some supercomputers can utilize all the expert knowledge; everything that has been published and patient information to make important decisions. He believed it took him 25 years before his imagination turned into a reality.
The new American Board of AI Medicine
Speakers feel it remains expensive and challenging if an individual radiologist will like to develop and adopt AI models in their work stations. Yet, it is feasible at the department level standpoint as there are ready-made, customizable applications that can be embedded into the PACS (Picture Archiving and Communication System) and allowed clinicians to explore the technology’s capabilities and limitations.
In fact, as Dr. Chang, Dr. Boyko and a large group of multi-disciplinary faculty has just launched the American Board of Artificial Intelligence in Medicine (ABAIM) last week, with a review course designed for everyone, in particular clinicians. They hope the Board will provide assistance to those who are interested in the area and fuel confidence into those who are already involved.
Dr. Chang thought the increase automation will be make radiologists impotent. The job of a radiologist is going to be more interesting in 10 years’ time and this will attract younger generation into the practice. People will also feel more empower by having data science behind medicine. Dr. Boyko added radiology is likely to march towards an interdisciplinary effort as it merges with other subspecialties like genomics. This will provide radiologists opportunities that they have never seen before.