ACR Data Science Institute™ (DSI) Chief Science Officer Keith Dreyer, DO, PhD, recently told a packed house that there must be strong dialogue between radiologists and data scientists developing algorithms for medical imaging and radiation oncology (RO) artificial intelligence (AI) applications. This was part of the message of his keynote address at the GPU Technology Conference in Washington, DC, known as GTC DC.
Dreyer pointed out that radiologists have years of training and often decades of experience, providing clinical expertise with which data scientists are not equipped. He stressed that radiology AI is different than developing a self-driving car. Driving is a skill which most data scientists already have and understand very well. That is generally not the case with medical imaging and RO. It is essential that radiologists serve patients by directly managing use of clinical data science to improve care.
DSI Panel ACR Chief Information Officer and Executive Vice President for Technology, Mike Tilkin and DSI Senior Scientist for Education Kathy Andriole, PhD, also served primary roles at GTC DC. The two served on a panel titled “AI in Healthcare: Beyond the Hype Cycle” that explored the impact of AI in medical imaging. Practical issues such as the need for well-qualified data sets, the importance of surveillance of AI algorithms to continually monitor accuracy and regulatory challenges were all primary themes.
The ACR DSI continues to press forward with its efforts to improve patient care by advancing medical imaging AI. Stay tuned to the new ACR DSI website for further updates.