PhD - Machine Learning

"Knowledge becomes health" - we are filling this motto with life every day, developing new ideas and improv-ing old-established knowledge. The University Hospital Frankfurt has existed since 1914. Our around 8,500 employees contribute with their skills and knowledge to the 33 specialist clinics, theoretical clinical institutes, and administrative departments. The close connection of health care with research and teaching, as well as a climate of collegiality, internationality, and cross-professional cooperation characterize the university hospital. The Edinger Institute is a German research institution and neuropathological diagnostic section at the University Hospital Frankfurt am Main. We are part of an excellent scientific community in Frankfurt including the Mildred Scheel Career Center (MSNZ), the Univeristy Cancer Center (UCT Frankfurt-Marburg), the German Consortium for Translational Cancer Research (DKTK) and the Frankfurt Cancer Institute (FCI). The Biobank and Immunomonitoring-Platform of the FCI is located at the Edinger Institute and offer optimal conditions for tissue-based research. The AG Molecular and Computational Neuropathology uses machine and deep learning as well as genome-wide molecular analysis methods in tumor specimens and liquid biopsies to develop new diagnostic tools and to establish novel prognostic and predictive biomarkers for neurooncological patients. We are recruting a PhD student for a DFG-funded research project focused on the automated detection and classification of tumor cells in cerebrospinal fluid samples by deep learning models. The aim is to develop a fully digital workflow and clinical-grade deep learning tool to tranform the current practice in neuropathological institutes of manual cell differentiation and counting of cerebrospinal fluid samples using a microscope. The thesis will build on a previously published algorithm (PMID: 36519297; https://github.com/pseegerer/csf_cell_classification ) and use a new, multicentric, well-annotated cancer cell dataset to improve automated cell segmentation, cancer cell detection and tumor origin prediction.