Back

Knowledge Connector: Decision support system for multiomics-based precision oncology

Huebschmann, D.; Kreutzfeldt, S.; Roth, B.; Glocker, K.; Schoop, J.; Oeser, L.; Hausmann, S.; Koch, C.; Uhrig, S.; Huellein, J.; Hutter, B.; Froehlich, M.; Heilig, C. E.; Teleanu, M.-V.; Lipka, D. B.; Kerle, I. A.; Baude, A.; Beck, K.; Heining, C.; Glimm, H.; Ueckert, F.; Knurr, A.; Froehling, S.; Horak, P.

2025-02-25 genetic and genomic medicine
10.1101/2025.02.23.25322403 medRxiv
Show abstract

Precision cancer medicine aims to improve patient outcomes by providing individually tailored recommendations for clinical management based on the evaluation of biological disease profiles in multidisciplinary molecular tumor boards (MTBs). The quality of MTB decisions depends on the comprehensive, reliable, and reproducible interpretation of increasingly complex molecular data. We developed and implemented, as part of a multicenter precision oncology program, the Knowledge Connector (KC), a decision support system that integrates individual patients molecular and clinical data with world knowledge to generate and document MTB recommendations. The KC supports data curation, database integration, and discussion based on multiomics data and provides an interface for creating a cross-institutional knowledge base. Furthermore, it extracts relevant biomarker-drug associations and increases the efficacy of data interpretation in a clinically relevant manner by reducing reliance on external sources and optimizing inter-curator concordance. Our results demonstrate that the KC is a versatile tool that supports medical decision-making in MTBs, thus enabling the scalability of precision cancer medicine.

Published in Nature Communications (predicted rank #1) · training set

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.