Search

Researchers at Mass General Brigham have developed two large-scale foundation models, UNI and CONCH, in the field of computational pathology (CPath). These models have the potential to advance diagnostic accuracy, prognostic insights, and therapeutic response predictions. UNI is a model for understanding pathology images and outperformed established pathology models in 34 tasks, while CONCH is a model for understanding pathology images and language and excelled in tasks like identifying rare diseases and tumor segmentation. The research team has made the code publicly available for other academic groups to use. [Extracted from the article]

Researchers from the University of Liverpool have developed a framework for using artificial intelligence (AI) to address the global challenge of antimicrobial resistance (AMR). The researchers outline an adaptive implementation and maintenance framework for AI to improve antimicrobial use and infection care. They emphasize the need for collaboration between the fields of AMR and AI to design, maintain, and globalize AI systems that support clinician-driven strategies to minimize AMR. The full research paper is available for further reading. [Extracted from the article]

A preprint abstract from medrxiv.org discusses the development and validation of an explainable, prognostic prediction model for catheter-associated urinary tract infections (CA-UTIs) among hospitalized individuals receiving urinary catheterization. The study used data from three hospitals in Taiwan and applied statistical and computational machine learning algorithms to develop the prediction model. The best model, a random forest (RF), accurately detected most CA-UTI positive cases and correctly ruled out most predicted negative individuals. The researchers also provided a web-based application and a paper-based nomogram for using the best model. However, further studies are needed to evaluate the implementation, validity, and reliability of the prediction model and its impact on patient outcomes. [Extracted from the article]