AI to Detect Tuberculosis

Tuberculosis is treatable but time-consuming... could AI burst through the bottleneck?
19 December 2019
Presented by Phil Sansom
Production by Phil Sansom.


A scientist examining samples of bacteria growing on petri dishes.


The lung disease tuberculosis is still one of the world’s top ten causes of death. And while it’s completely treatable, patients need constant monitoring to make sure the treatment is working. The monitoring is fairly labour-intensive: it requires taking a sample of phlegm and counting the bacteria inside by eye under a microscope. It's a treatment bottleneck. But now, some engineers from Cambridge Consultants have been developing a technology to get rid of that bottleneck using AI - as Phil Sansom found out from developer Matthew Murchie...


Add a comment