November 28, 2022 – The Radiological Society of North America (RSNA) announced the launch of the RSNA Screening Mammography Breast Cancer Detection AI Challenge. The latest in a series of such research competitions that RSNA has been running since 2015, this competition will evaluate competitors in their ability to develop artificial intelligence (AI) to aid in the detection of breast cancer when screening mammography images.
According to the World Health Organization, breast cancer is the most common type of cancer worldwide. In 2020 alone, there were 2.3 million new breast cancer diagnoses and 685,000 deaths.
Breast cancer screening of the most vulnerable population has been shown to reduce cancer deaths. AI tools have the potential to make screening more efficient and effective.
“Although there is a global shortage of radiologists to interpret screening mammograms, radiologists remain concerned about how well AI systems will perform in their patient population,” said Linda Moy, MD, professor of radiology at NYU Grossman School of Medicine and editor designation of the journal radiology. “This diverse, well-curated dataset can be used to assess generalizability to different patient populations. The RSNA Screening Mammography Breast Cancer Detection AI Challenge will catalyze collaboration to improve the diagnostic accuracy of screening mammography and save patient lives.”
The dataset to be used for the challenge was contributed by mammography screening programs in Australia and the United States. It includes detailed labels with radiologist assessments and pathological follow-up results for suspected malignancies.
For the challenge competition, the accuracy of machine learning models developed by participants for cancer detection will be assessed using this ground truth dataset. After completing the challenge, the dataset remains available for further research.
This challenge is part of a broader research project examining how competitively generated models perform compared to previously unseen data, and comparing their performance to that of experienced human observers. These questions are critical in determining the performance of AI tools in clinical settings.
“Large, curated data sets that the RSNA compiles for AI challenges are a key resource driving improvements in radiology AI,” said John Mongan, MD, Ph.D., professor of radiology at the University of California, San Francisco and Chair of the RSNA Machine Learning Steering Committee. “We expect that after the release of this dataset, we will see an acceleration in mammography AI activity, as we have seen in other areas with the release of previous datasets.”
The RSNA Screening Mammography Breast Cancer Detection AI Challenge will be conducted on a platform provided by Kaggle, Inc. The competition will run until March 2023. Once the results have been validated, the winners will be announced at the end of April. The top-performing participants will receive a total of $30,000 and will also be recognized in a presentation at the 109th RSNA Scientific Assembly and Annual Meeting, November 26-30, 2023 in Chicago.
More information about RSNA22 can be found here