LOWER RESPIRATORY TRACT INFECTION DIAGNOSTICS

LOWER RESPIRATORY TRACT INFECTION DIAGNOSTICS

Researchers at UCSF, CZ Biohub SF, CZI and University of Colorado have developed a host/pathogen classifier to differentiate patients with lower respiratory tract infections (LRTI) from those with non-infectious acute respiratory illnesses.

Lower respiratory tract infections (LRTI) are a leading cause of hospitalizations in the pediatric population. About 12% of children under 5 years old contract a LRTI globally. Due to difficult diagnosis, LRTIs cause more deaths per year than any other type of infection. Early detection of the responsible organism reduces length of hospital stay by 24% and use of antibiotics by 13%. Yet, existing diagnostic techniques have low sensitivity and specificity, low turnaround time, and a narrow spectrum of pathogen targets. Thus, LRTI treatment is typically empirical which leads to antimicrobial overuse creating more resistant pathogens and adverse reactions. Lack of highly sensitive and specific diagnostic tools, especially in the pediatric population, is a critical barrier in timely and effective treatment of LRTIs.

Stage of Research

The inventors have developed a machine-learning model that can predict LRTIs based on analysis of biological samples (i.e., tracheal aspirates). Tracheal aspirate samples were used to profile host gene expression and respiratory microbiota. A host classifier was used to process host gene expression levels to determine whether a subject has an increased likelihood of LRTI. The host classifier was trained on a dataset including patients with a diagnosis of LRTI with microbiologic findings and on patients with respiratory failure from non-infectious causes. A rules-based model was applied to the bacterial micrbiota results to differentiate pathogens from colonizing bacteria. Preliminary studies showed that in individuals of suspected or indeterminate LRTI status, this method identified LRTI in 52% of cases and identified the likely pathogens in 98% of those cases.

Applications

• Accurate LRTI diagnosis with pathogen identification in critically ill subjects using lower airway metagenomics

Advantages

• Quick turnaround time

• >90% sensitivity in pediatric populations

Stage of Development

Research- in vivo

Related Web Links

https://profiles.ucsf.edu/charles.langelier

Keywords

Disease classifier, respiratory, pathogen

Technology Reference