Making its grand debut at ECR 2019 in Vienna, the ESR was proud to present a new exhibit: The Artificial Intelligence Exhibition (AIX). Located in the X1 hall, this special exhibition was devoted to the application of artificial intelligence (AI) in radiology and healthcare, focusing on areas such as machine learning, deep learning, and big data – and the positive feedback from attendees, speakers and exhibitors alike was overwhelming.
Features of AIX
AIX was comprised of several different areas: the exhibitor area, the AIX Theatre, and a self-paced training programme provided by NVIDIA’s Deep Learning Institute. The exhibitor area created a space in which twenty-five companies, mainly small and medium-sized enterprises (SMEs), who are at the forefront of AI in healthcare, could present their programs and products and discuss all things AI with attendees. The AIX Theatre provided a space where attendees could take in informative sessions on everything from the future of AI in radiology to ethical and regulatory issues. Finally, NVIDIA’s Deep Learning Institute offered hands-on training and courses to individuals, regardless of their level of knowledge on AI and accelerated computing.
This was the first time that companies, mainly SMEs, involved in AI were invited to exhibit in a larger feature exhibition at the ECR and interact with congress attendees; altogether, twenty-five companies participated as AIX exhibitors.
The makeup of the exhibiting companies was of interest, as a majority were companies that have emerged over the past five years and currently operate through investment for their income, as opposed to marketing a large product line. Although most of these companies are involved in the development of algorithms, their solutions deal with a wide range of things – many assist with image interpretation, others with case searches and report structuring. When it comes down to it, there is no shortage of possibilities when applying AI to radiology and healthcare, and these companies are there to fill these gaps.
Moderated by Dr. Hugh Harvey, Clinical Director at Kheiron Medical, and Dr. Wim Van Hecke, co-founder and CEO of icometrix, the AIX Theatre programme took place over three days, from February 28th through March 2nd, and offered a variety of programmes. The idea of the content presented in the AIX Theatre was that it should be accessible, in that interested attendees with little to no knowledge of AI could experience, enjoy, and learn about an assortment of topics relating to AI tools and technologies. Throughout this programme, congress-goers experienced “AI in 8”, which gave exhibiting companies the opportunity to present their product(s) for eight minutes, a variety of sessions covering various aspects of AI, keynote lectures followed by Q&As with their respective speakers, and three different AIX panels comprised of experts in their fields. The three keynote lectures covered the topics “Assuring Safe and Effective AI in Medical Imaging” with Professor Charles Kahn, “AI From a Patient’s Perspective” with Dr. Liz O’Riordan, and “AI Ethics and Regulation: An Open Issue” with Dr. Filippo Pesapane and Dr. Marina Codari.
In the first keynote lecture by Professor Kahn, the importance of testing and making sure that the data you train and test your model on is reflective of your population and the patients you see in your setting, the perils of machine learning, and what we need to consider going forward with AI in healthcare. The second lecture came from Dr. O’Riordan, which was the only talk given by a patient. Dr. O’Riordan is a breast surgeon who was diagnosed with Stage 3 breast cancer and survived. In this lecture, she dives into the importance of the perspective of patients, especially when it comes to AI. Finally, in the third lecture, Dr. Pesapane and Dr. Codari explore the legal and regulatory issues and challenges that have evolved alongside AI systems and tools, including data protection, cyber security, and liability.
The final sessions each day at the AIX Theatre consisted of informative panels: “From Research to the Clinical Validation of AI: One Small Step“, “Implementing Artificial Intelligence in Clinical Care“, and “Where to next? The Regulation of Radiological AI in Europe and Abroad,” with each panel followed by a question and answer session with the panelists. The first panel, hosted by Dr. Hugh Harvey, featured Professor Charles Kahn, Professor Paul M. Parizel, Elad Walach, and Professor Wieland Sommer, began with some opinions from the panel on where we are on the “hype curve” with regards to AI in radiology and covered the future of AI in radiology in relation to integration, trends, and reporting. The AIX panel on day two was hosted by Dr. Wim Van Hecke, and featured Antoine Jomier, Jonny Hancox, Professor Greg Zaharchuk, and Raymond Moh. This panel explored each panelist’s organization’s work and focus within AI, as well as navigating the GDPR and HIPAA when using massive data sets to train algorithms. In the final panel, which consisted of a mix of doctors, researchers, and entrepreneurs, Dr. Hugh Harvey returned as moderator, with panelists Dr. Bibb Allen, Abdul Hamid Halabi, Professor Greg Zaharchuk, Dr. Lawrence Tanenbaum, and Dr. Marina Codari. The panel discussed global regulations, liability, and litigation, as AI becomes more integrated into the field of healthcare and companies develop systems for the benefit of patients and doctors alike.
NVIDIA’s Deep Learning Institute
NIVIDIA’s Deep Learning Institute provided a space at AIX in which attendees could sit down with an expert and participate in courses and electives on AI and accelerated computing. The course and elective offering from NVIDIA ranged from beginner to expert, focusing on real-world problems that could be solved with the use of AI, which allowed anyone, regardless of their level of knowledge, the ability to expand their knowledge on AI. Some of the training offered included working with medical imaging, deep learning, and data augmentation and segmentation.
What does the future hold for AIX?
After exceptional feedback from participating individuals and companies, and attendees, the first bookings for AIX at ECR 2020 are already underway. Moreover, due to its popularity and success, it has been decided that AIX will see an expanision at next year’s ECR.
Impressions from the AIX and the Deep Learning Institute