Particle.news

Download on the App Store

Microfluidic-AI Method Detects Bloodstream Bacteria in About Two Hours

Next steps focus on clinical validation plus a solution for Staphylococcus aureus to move beyond days-long cultures.

Overview

  • KTH Royal Institute of Technology and Uppsala University reported the culture-free approach in npj Digital Medicine, led by doctoral students Henar Marino Miguélez and Mohammad Osaid.
  • A density-agent “smart centrifugation” step separates bacteria from blood cells before a microfluidic chip traps them for automated, time-lapse microscopy.
  • Machine-learning image analysis flags bacterial growth directly from the captured cells without conventional culture.
  • In spiked-blood tests, the system detected E. coli, K. pneumoniae, and E. faecalis at 7–32 CFU per milliliter.
  • The method did not detect Staphylococcus aureus, which hides in clots, and the team is working on a fix while aiming for 4–6 hour clinical results versus the days required for standard blood cultures.