Artificial Intelligence (AI) tools for medical applications have been largely developed over the last years. However, their clinical use has been limited. One of the reasons for this is the lack of confidence of clinicians in AI-powered tools, which is motivated by a set of unresolved challenges.

In this project (AIMED), major medical imaging challenges observed in data (e.g. limitation, biased) and AI models (e.g. robustness, fairness, explainability) are investigated. Novel methodologies are proposed to improve trustworthiness of AI models which will be validated in clinical sites.

To demonstrate the clinical application of the proposed methods and tools, three use cases will be used: 

1.  ChemoTherapy – breast cancer treatment through response prediction of chemotherapy in breast cancer patients. 
2.  CardioVascular – patient diagnose and deployment of endovascular devices within the artery during intervention.
3.  RadioTherapy – multiorgan dose estimation during breast cancer radiotherapy treatment.