Microscopic techniques examining for the presence of acid fast bacilli using the Ziehl-Neelson (ZN) method directly from patient samples has long been the benchmark for presumptive diagnosis of mycobacterial infections, including tuberculosis.
A special breed of scientist is required for this work, methodically examining at least 100 fields over 5 minutes under oil immersion to find what might be as little as 1 organism in an attempt to guide appropriate treatment, given that cultures could take up to weeks to grow.
Indeed, the difficulties of using ZN stains in larger laboratories are so well known that official guidelines are in place to limit the number of ZN stains a scientist can read in a shift, and a different fluorescent stain can be introduced in an attempt to deal with larger volumes and eye fatigue.
These considerations led us to develop a method of scanning ZN stains which are then assessed via an in house developed artificial intelligence algorithm, allowing scientists to accurately and comfortably assess for the likelihood of acid fast bacilli being present in the range and volume of samples encountered in a large clinical mycobacteriology lab.
Join the session for a discussion of the development, the positives, and the challenges faced in the verification and implementation of this new approach.