Data Science in Clinics

Tuesday, 14 November • 13:30 – 15:00

This session showcases how data science drives precision medicine in hospitals. How does data science drive state-of-the-art molecular diagnostics? How digitalization enables smart clinics? How will the Swiss Electronic Patient Record revolutionize healthcare in the 21 century? How can health-relevant data be made interoperable and shareable for research in Switzerland? 

Session Chair:

Abdullah Kahraman, Ph.D. (FHNW University of Applied Sciences and Arts Northwestern Switzerland)

Speakers:

EpiDiP – Epigentic Digital Pathology
Jürgen Hench, M.D. (University Hospital Basel)
Epigenetic analyses have significantly advanced human brain tumor classification, starting out around 2014 with childhood cancers that are histologically similar but represent a highly diverse biological spectrum. Therefore, adapted therapeutic strategies depend on reliable diagnoses. The diagnostic principle was extended by the development of a brain tumor classifier released in 2017 by the German Cancer Research Centre. While the WHO classification of tumors listed epigenetic testing in the context of very few newly discovered brain tumor types in 2016, 2021 has grown to almost twice the number of entities and recommends epigenetic testing for a large portion of those tumors. While tumor epigenetics are still not well understood mechanistically, the epigenomic patterns are highly indicative of cellular lineages and can be diagnostically exploited, mostly through supervised and unsupervised machine learning approaches. In parallel to DNA methylation signatures, whole-genomic epigenetic analysis determines genome-wide copy number profiles that - for many tumor types - are of diagnostic, prognostic, or predictive value. We developed an open-source, freely accessible data analysis platform termed "Epigenetic Digital Pathology" - epidip.org. This online tool is based on UMAP dimension reduction and places every uploaded tumor methylation pattern in a larger context of currently approx. 20,000 reference cases. In addition, its companion tool, NanoDiP (Nanopore Digital Pathology) enables methylation- and copy number profiling-based diagnostics within hours through nanopore sequencing on small-scale, portable embedded computers.

Clinical AI-Adoption and Radiomics
Felice Burn (Cantonal Hospital Aarau)

One Google Health
Dominik Steiner (Google)
An overview of Google's point of view on the future of Health. It covers our efforts in different areas through the Google Ecosystem, including Generative AI, AL/ML and Data.