BMDeep

The morphology of the bone marrow forms the basis for the assessment of hematopoiesis as well as many organ functions and allows conclusions to be drawn about systemic diseases as well as pathological processes in neoplasms. The changes are reflected in quantitative and qualitative effects, which can be partially inadequately or poorly objectively recorded with the current, largely analogous approach. In the BMDeep project, we combine expertise from clinical, bioinformatics, and AI. Our goal is to automate and improve the evaluation of bone marrow smears and find characteristic (pathological) patterns in neoplasms of the blood. To this end, we will merge morphological with clinical data for high-dimensional pattern recognition and allow a neural network to emerge. In this way, the respective contribution of the different data sources can be investigated and the most important features can be merged into a final model. This, in turn, can improve the identification of new biomarkers from integrated data sources and thus improve disease understanding.