Artificial intelligence (AI) can be helpful in analyzing medical images and support radiotherapy planning workflows. However, huge amounts of sensitive patient data are needed to train robust AI models. Researchers at the Karlsruhe Institute of Technology (KIT) have investigated how hospitals can jointly develop AI models for radiation therapy applications without exchanging such data. In their project, they use federated learning, a process in which a single model is trained at a decentralized location while the data remains in place. Their results demonstrate not only the technical feasibility of the approach but also its safe, more sustainable, and accessible use.

Dead branches of trees are by no means static but clearly move in the course of the day. Geoinformatics researchers at Heidelberg University tracked these movements in a study conducted in cooperation with the Finnish Geospatial Research Institute (FGI). With the help of an autonomously operating laser scanning system, the scientists mapped a specific forest plot at regular intervals. Using a specially developed tracking algorithm, they were able to quantify precisely even small branch movements, which are driven by changes in humidity. Their approach opens up new opportunities for automated monitoring, which could make it possible to assess tree health based on characteristic movement patterns.

In a prospective cohort of 91 patients undergoing laparoscopic colorectal cancer surgery, electrical impedance tomography (EIT) quantified the ratio of dorsal ventilation (RDV). A ≥14.5% reduction in RDV from baseline to surgery end (RDV_T0–T3) was associated with higher PPC rates and longer hospital stay. RDV_T0–T3 ≥14.5% was independently associated with PPCs (adjusted OR 11.28; 95% CI 2.58–49.25). These results establish a clinically applicable threshold, supporting EIT-guided individualized lung-protective ventilation strategies.

Creatinine Rises on a Breast Cancer Drug, but the Kidneys May Be Fine: Why Cystatin C Should Be Checked Before Treatment