DR-185 A human interpreter validation on results generated by artificial intelligence (AI) algorithms developed for outcome and response predictions for CRC and melanoma patients: optimal use and validation of multi-layer research data

We want to achieve better care for patients with CRC and melanoma, by allowing a multidisciplinary team of researchers to being able to query, analyse and integrate DNA and RNA sequencing data with matching clinical, pathology and radiology images and annotations, using different approaches: (1) artificial intelligence (AI), in order to determine whether predictive disease outcome correlations/algorithms using these data types can be identified, and (2) exploration of available data (in different formats than required for AI) on a human readable platform, allowing researchers to determine whether it is possible to explain AI findings by human interpretation of the data.

Mariska Bierkens, Netherlands Cancer Institute, the Netherlands