Mammography image dataset useful for researchers around the world

The CSAW-CC dataset contains thousands of images used in mammography screening showing subjects with and without a breast cancer diagnosis. In 2022, Fredrik Strand contacted KIs Data Access Unit for assistance publishing the dataset via DORIS, Swedish National Data Service's system for sharing data. Since being made available in 2022, the dataset has been requested over 40 times by researchers in 23 different countries. Researchers have used the dataset as a way of benchmarking their own AI models or in training their models.   

AI in combination with human radiologists has the potential to increase the quality of screening at the same time as it makes screening more available. A recent systematic review published by an international group of researchers including Fredrik Strand found that AI performed as well as or better than radiologists. 

Given this promising outlook, it is easy to see the importance of AI models as a tool in mammography screening.  Making the CSAW-CC dataset openly available to researchers has helped various projects to refine their models. The continual development of better models will eventually contribute to higher quality and more readily available screening that will save lives. While the role of AI is obviously critical here and will gain the most attention, the importance of open research data in the process is also crucial. In the years to come, we hope to see the combination of open research data and AI lead to advancements in many areas of medical research. 

KIs Data Access Unit assists researchers in publishing their research data via DORIS.  Don't hesitate to contact us with questions or if you are interested in attending a workshop on publishing research data.