Technology: For Nations That Lack the Expertise, an Automated System for Detecting TB

The NYTimes has an interesting article in their Technology section: For Nations That Lack the Expertise, an Automated System for Detecting TB:

One of the difficulties of diagnosing tuberculosis is that there is no simple blood or urine test. Instead, a laboratory technician must take a sample of sputum coughed up from the lungs, stain it and inspect it under a microscope for the telltale bacteria, which resemble long-grain rice. It takes expertise that is often rare in poor countries.

Guardian Technologies, a Virginia company that was started to help airport X-ray scanners distinguish explosives in luggage from innocuous plastics and liquids, has developed a system that automatically scans microscope slides for the bacillus. The company’s software algorithms can spot distinctive shapes, colors and densities that untrained eyes may miss. In a recent test in South Africa, which has some of the highest TB rates in the world, the technology was 93 percent accurate in detecting the bacillus on microscope slides and had a false positive rate of less than 2 percent, the company said.

Wow, how cool is that?  *This* is the promise of digital pathology, realized.  Searching for AFB on lung slides is painstaking and time-consuming, not to mention error-prone.  Automating this process with image analysis on digitized slides makes so much sense.