Researchers at University Medicine Oldenburg have developed an AI tool that delivers fewer false-positive results than conventional screening methods when testing bacteria for resistance to reserve antibiotics. The research group calls the new AI model CarbaDetector. It is better at identifying bacteria that are resistant to reserve antibiotics (also known as "last-resort" antibiotics) compared to other methods currently in use.

Researchers at University Medicine Oldenburg have developed an AI tool that delivers fewer false-positive results than conventional screening methods when testing bacteria for resistance to reserve antibiotics.

Forschende der Universitätsmedizin Oldenburg haben ein KI-Modell entwickelt, das Bakterien, die gegen Reserveantibiotika (Carbapeneme) resistent sind, deutlich präziser identifiziert als herkömmliche Screening-Algorithmen. Das Modell „CarbaDetector“ produziert deutlich weniger falsch-positive Ergebnisse und könnte unnötige, teure Folgeuntersuchungen vermeiden. Die Ergebnisse wurden in „Nature Communications“ veröffentlicht. Treffen Antibiotikaplättchen auf eine Bakterienkultur, entstehen Hemmzonen, in denen das Bakterienwa ...Quelle: Axel Hamprecht

Nature Communications, Published online: 14 November 2025; doi:10.1038/s41467-025-66183-zCarbapenems are last-resort antibiotics, but resistance is rising due to hydrolyzing enzymes called carbapenemases. The authors present a machine learning algorithm and web-app to rapidly predict carbapenemase-production in Enterobacterales.

We investigated and tested current screening methods for CPE prediction and aimed to develop a machine learning tool to accurately predict the presence of carbapenemases in Enterobacterales using simple inhibition zone diameters. EUCAST has established cut-offs that trigger further carbapenemase...

Nature Communications, Published online: 14 November 2025; doi:10.1038/s41467-025-66183-zCarbapenems are last-resort antibiotics, but resistance is rising due to hydrolyzing enzymes called carbapenemases. The authors present a machine learning algorithm and web-app to rapidly predict carbapenemase-production in Enterobacterales.

Carbapenemase-producing Enterobacterales (CPE) are considered as the highest threat to global health in the WHO’s 2024 Bacterial Priority Pathogens List. Their detection is difficult, time-consu...