The Radiology Society of North America will introduce machine learning labs at this year’s meeting as an effort to increase exposure of the technology throughout the field.
“No question — machine learning will change the way radiologists practice in the years ahead, sometimes dramatically. But there is much work to be done before ML becomes common place,” said Curtis Langlotz, MD, PhD, the RSNA board liaison for information technology and annual meeting.
Machine learning could help streamline data processing for large image datasets. Institutions like Stanford University are working toward implementing an automated image labeling system. However, Dr. Langlotz is encouraging more expansive training of this technology across the specialty.
“Truly mastering the problem solving requires mentorship from those who have done it before at a high level. “We have some great resources in Silicon Valley, but that’s not yet true across the board,” said Dr. Langlotz.
At the meeting, attendees will have the opportunity to explore machine learning practices. The National Cancer Institute is sponsoring “Crowds Cure Cancer: Help Annotate Data from The Cancer Imaging Archive,” an event that showcases the application of annotation tools in labeling NCI cancer imaging data sets. Carestream Health, Google Cloud, and Zebra Medical Vision will be sponsoring The Machine Learning Showcase, an exhibit that will showcase the “latest ML technology and network with companies on the forefront of ML advancements.”
The RSNA will also present for the first time a Machine Learning Challenge, in which participants submit algorithms to “automate the assessment of pediatric bone age based on hand radiographs.” So far, 29 teams have submitted algorithms. The winning team will be announced on Monday, November 27.