Engineers at Duke University have combined light-based holographic imaging with deep-learning algorithms to come up with a no-pathologist-needed way to identify blood cells infected with malaria.
Their hope is that the work will lead to a fast and consistently accurate exam that could be administered by just about any health worker, anytime and anywhere the mosquito-borne disease is a threat.
“With this technique, the path is there to be able to process thousands of cells per minute,” biomedical engineer Adam Wax, PhD, told Duke’s news office. “That’s a huge improvement to the 40 minutes it currently takes a field technician to stain, prepare and read a slide to personally look for infection.”
The team has launched a startup to build a diagnostic device on the technique—the imaging portion of which is called quantitative phase spectroscopy—and Wax has received funding to look into adopting the advance for catching cancerous cells in the blood.
Read the rest: