How can we use (bio)medical images to measure and predict whether therapies are effective?

Medical imaging contains a multitude of infrmation that can increase the accuracy of treatments and help further perzonalize them. Compared to a biopsy, a scan can analyze and visualize much more of the body at once. By filtering out the correct, relevant information from these scans at a large scale, analyze it and combine it with data on treatment results, researchers can find new connections between treatments and results for patients. This will help physicians to keep a better eye on what happens in the body during treatment, and will provide greater insight on the efficacy of treatment, or the results we can expect.

Example project

Predicting survival using Artificial Intelligence

Radiologists use scans to determine whether treatment is effective. This is a very time-consuming endeavor, however, and only allows them to inspect certain parts of the body. That's why we are investigating ways in which artificial intelligence could help provide in-depth analyses of these scans to predict survival rates. One thing we are working on is an algorithm that can keep track of all the changes in the body scans during immunotherapy. Physicians can use these changes to adapt the treatment if treatment isn't effective (anymore).

Medical images to measure and predict therapy efficacy

We don't want to waste time on treatments that are not effective, and we want to avoid side effects if the treatment isn't going to be effective at all. Stefano Trebeschi explains his research into AI algorithms that can measure and predict immunotherapy efficacy.