RSNA, other medical imaging groups launch brain tumor artificial intelligence challenge

A handful of imaging groups have joined together to launch the 10th annual Brain Tumor Segmentation challenge.

The event, known as BraTs 2021, will have participants use artificial intelligence to detect and classify brain tumors on large, annotated datasets consisting of multi-parametric MRI (mpMRI) scans. The three groups co-sponsoring the challenge are the Radiological Society of North America, the American Society of Neuroradiology, and the Medical Image Computing and Computer Assisted Interventions Society.

“RSNA has significantly ‘upped their game’ with this year’s Brain Tumor Classification Challenge,” Adam E. Flanders, MD, who serves on the RSNA Machine Learning Subcommittee, said Wednesday. “It is our first AI Challenge to use MRI, and it is also our first to address an oncology problem—brain cancer.”

Teams will help solve two important clinical tasks: developing an automated way to measure visual components of cancer that can help track growth or treatment response, and creating a reliable method to predict genetic features within tumors using only images.

BraTs 2021 consists of two challenge tasks and participants can choose to compete in one or both. In the first, teams must build a model that segments brain tumor sub-regions that correspond to those developed by neuroradiologists.

In part two, challengers will develop a tool based on mpMRI to predict MGMT (O[6]-methylguanine-DNA methyltransferase) promoter methylation status, a prognostic factor that can influence treatment decision-making, according to the statement.

Final submissions are due October 12, with the winners revealed on November 23. Challenge champions will also be recognized at RSNA 2021 during an event scheduled for Monday, November 29, at the AI Showcase Theater.

Prize money for the top entries in each task is provided by Intel, NeoSoma and RSNA.