Microscopic examination of clinical specimens is a critical initial screening tool for the laboratory detection of mycobacterial infections worldwide.
Microscopy using an auramine-rhodamine stained smear which may be followed by a Ziehl-Neelsen stain to confirm positives is a standard algorithm to screen for acid fast bacilli in specimens from patients suspected of infection with mycobacterium species.
This process can be a laborious task. Our aim was to automate this and increase the sensitivity of microscopy by developing a Rapid Screening System (RSS) which digitises a stained smear on a glass slide and performs automatic ranking on the smear regions.
Based on the ranking, the regions are sorted and displayed to the microscopist in descending order from the most likely regions with auramine-rhodamine stained bacilli present, to the least likely.
The RSS consists of three components: the scanning system; the analytic system that provides automatic region ranking and the smart viewer displaying the images captured by the scanning system that are ranked based on information from the analytic system.
In a preliminary study of 107 samples, this RSS had good negative predictive value (NPV) and significantly increased sensitivity, which are critically important for a primary screening test particularly when testing a low prevalence patient population. The developed RSS also reduces screening time for positives to only seconds. The RSS is designed to be an open system as it can be used to scan and process other stained slides such as Gram stained specimen preparations. The RSS system increases screening sensitivity and reduces labour thereby allowing better utilisation of laboratory staff.