The first joint goal within the game is to strengthen immunotherapy research for patients with solid breast cancer patients, being the first step to determine which patient is eligible for it. Our previous studies found that patients with more than 5% immune cells had a better response after treatment with immunotherapy. But we look not only at the number but now also at the location of the immune cells. However, for a pathologist, it is difficult, if not impossible, to determine with the naked eye the number of immune cells a tumor cell touches. Can artificial intelligence offer a solution? Yes, but such a model needs many annotations and millions of data points that are currently lacking.

By playing the game we can collectively feed this algorithm. The idea is that everyone, from young to old, can playfully learn to recognize different cell types (such as immune cells, cancer cells and fibroblasts) and become better and better at annotation with automated feedback.

More details here (dutch):