Clinical Data Analytics

Researchers Ghim Siong Ow and Vladimir A. Kuznetsov have optimized their Prognostic Signature Vector Matching (PSVM) model for patient classification. They found that using negative log-rank p-values improves the model's robustness. They also combined PSVM with other machine learning techniques in a multi-test system for more precise patient stratification.

Researchers Jonas Scherer, Marco Nolden, Jens Kleesiek, Jasmin Metzger, Klaus Kades, Verena Schneider, Hanno Gao, Peter Neher, Ralf Floca, Heinz-Peter Schlemmer, and Klaus Maier-Hein discuss the potential of AI in healthcare, particularly in image analysis for disease prediction, diagnosis, and treatment. They address the challenges of data sharing between institutions and propose the Joint Imaging Platform (JIP) of the German Cancer Consortium (DKTK) as a solution for secure and compliant federated data analysis.