CLABSI prediction in children with acute leukemia

"These findings highlight the potential of machine learning models as effective decision-support tools for CLABSI risk stratification and underscore the need for further validation to facilitate broader clinical application" Zhang et al (2026).

"These findings highlight the potential of machine learning models as effective decision-support tools for CLABSI risk stratification and underscore the need for further validation to facilitate broader clinical application" Zhang et al (2026).

ReTweet

CLABSI prediction in children with acute leukemia

Abstract:

Central line-associated bloodstream infection (CLABSI) is a frequent and severe complication in children undergoing treatment for acute leukemia, substantially compromising therapeutic outcomes. This study aimed to develop a machine learning-based predictive model to facilitate the early identification of patients at high risk. A retrospective analysis was performed using clinical data from 407 pediatric patients with acute leukemia, and the predictive performance of six machine learning algorithms was evaluated. Among these, the TabPFN model demonstrated the highest accuracy, achieving a rate of 91.2%. Feature importance analysis showed that corticosteroid type, body temperature, neutrophil count and white blood cell count were important factors associated with CLABSI. These findings highlight the potential of machine learning models as effective decision-support tools for CLABSI risk stratification and underscore the need for further validation to facilitate broader clinical application.

Reference:

Zhang Y, Wu Y, Gao H, Liu Q, Lv M, Wu H, Liu L, Li N. Machine Learning-Based Prediction of Central Line-Associated Bloodstream Infection in Children with Acute Leukemia. J Hosp Infect. 2026 May 20:S0195-6701(26)00189-1. doi: 10.1016/j.jhin.2026.05.016. Epub ahead of print. PMID: 42167607.