Representatives of the Artificial Intelligence for Health Imaging (AI4HI) cluster and the EUCAIM project organized a mini-symposium at the Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2023), celebrated in Sydney, Australia, from the 24^th to the 27^th of July. Susanna Aussó, Head of Artificial Intelligence at TIC Salut Social Foundation, participated on behalf of the INCISIVE project.

The AI4HI cluster is made up of five collaborative projects funded under the AI for Health Imaging action of the European Union’s Horizon 2020 research and innovation programme: Chaimeleon, EuCanImage, ProCancer-I, Primage and INCISIVE. The cluster gathers over 100 institutions, from universities to companies, from all European regions and several non-EU partners.

The mini-symposium, entitled ‘The European Cancer Imaging Initiative – Status, Challenges and Opportunities’, provided details on the achievements of the cluster in creating large interoperable cancer imaging repositories, and their experiences in developing novel AI-based models. It also provided hands-on experience on the use of the FUTURE-AI guidelines and reported on the challenges addressed by the EUCAIM project.

INCISIVE experience in Explainable AI

Aussó gave a lecture on Explainable AI (XAI) in cancer imaging, focusing on the work carried out in this field in the INCISIVE project. She pointed out that “XAI plays a critical role in making AI models more transparent, trustworthy, and safe for all stakeholders involved in the healthcare ecosystem.” Before giving examples of XAI in breast and lung cancer models developed for the INCISIVE project, she emphasized that XAI’s implementation is an iterative process that should involve clinicians from the very beginning in the design phase, and that the definition of requirements must take into account multiple factors in clinical and technical areas.

She also admitted that “although theoretical XAI requirements are well defined, its implementation is not easy because each model has its peculiarities.” Finally, Aussó listed metrics for objective and subjective XAI evaluation.

About the conference

As it is described on their webpage, the Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) dedicates itself to helping the biomedical engineering community achieve its goals. Since its first technical conference in 1960 in London, England, EMBC has been known for convening academics, clinicians, industry, students and researchers in the biomedical engineering and the Medtech community. This year’s theme was ‘Engineering Better and More Resilient Healthcare for All’.