Work

Contextualized biomedical language processing enhances ICU survival prediction

Year: 2025

Type: article

Abstract: = 58,615) during external validation. Excluding text features or replacing free-text ICD entries with coded formats reduced performance (AUROC from 0.983 to 0.946-0.947), highlighting the importance of contextual embeddings. As a secondary task, cerebrospinal fluid culture prediction achieved AUROC = 0.853. Overall, integrating contextualized biomedical language representations significantly improves multimodal learning and ICU survival prediction.

Source: iScience

DOI: 10.1016/j.isci.2025.114442

Authors: Rui Chen, Yu Cai, Sitong Zhang, Zirong Huo, Mingming Song, Wenqing Li, Dongyan Yang, Seungyong Hwang, Ling Bai, Fuxin Han, Xi Zhang