Machine Learning Assisted Electrochemical Profiling to Provide Early Identification of Bloodstream Infections Pathogens

Towards a Smart Blood Culture Bottle: Machine Learning Assisted Electrochemical Profiling to Provide Early In-situ Identification of Bloodstream Infections Pathogens

In the context of a bacteremia, although significant progress has been made in speeding up pathogen identification once a blood culture bottle turns positive, few cost-effective solutions have been proposed to improve the earlier stages of the process-specifically, from blood collection to bottle positivity. The investigators propose that transport time could be leveraged to grow and identify bacteria, enabling faster access to actionable results through innovative technologies. This project aims to develop a bacterial identification database by analyzing the electrochemical profile of bacteria growing within the blood culture bottle, using machine learning.

Inclusion Criteria:

• patient requiring a blood culture sample as standard of care procedure
• body weight > 50 Kg
• Patient for whom the collection of 2 to 4 additional blood culture bottles is feasible, depending on venous access
• patient who has not objected to participation in the project

Exclusion Criteria:

• Patient protected under the French Public Health Code (pregnant or breastfeeding women, patients under guardianship or curatorship, hospitalized under constraint, or deprived of liberty)
• patients with ongoing antibiotic treatment at the time of sampling