Use of Digital Image Analysis to Interpret Urine Cultures on Blood and MacConkey Agar

Introduction
Urine culture interpretation can be complicated by several variables, including the presence of small numbers of colonies and the growth of more than one bacterial type. In general, voided-urine cultures containing ≥10,000 CFU/mL should be reported as potential pathogens as long as there are not more than 3 pathogens or these organisms are not normal skin flora. This complexity of interpretation requires significant hands-on time and technical expertise, which imposes a significant work burden on the clinical laboratory. In this multi-center study, we evaluated the accuracy of the WASPLab (Copan, Brescia, IT) software to differentiate negative and non-negative urine cultures from sheep blood, MacConkey, and Columbia CNA agar compared to manual analysis.

Method
Urine specimens submitted for bacterial culture from 3 different sites were plated on sheep blood, MacConkey, and CNA agar (CNA only site 1). All specimens were processed by the WASPLab using a 1-μL loop, and images were captured after 0 and 18 h incubation. The software quantified each plate and reported the specimen as non-negative if any plate contained more than 10 colonies (10,000 CFU/mL). Results were then compared to manual interpretation as either positive or negative for pathogens based on each laboratory’s urine culture policy. These data were also analyzed by separating laboratory-negative specimens depending on site-specific rules when growth was detected, but no further work up was needed due to either > 3 pathogens present or no significant growth (skin or fecal contamination). Manual-positive, automation-negative cultures were reviewed by a second technologist.

Conclusions

• High Positive percent agreement 99.9% and Negative percent agreement 96.4% for colony couting
• 88.6% of Manual Negative, Automation Positive specimen were due to complex interpretation of urine cultures that use parameters other than basic colony count including: >3 pathogens or only normal genital flora identified that do not require work-up.
• After closer look at the MP/AN specimens rules such as any growth for catheters reduced the total specimens to 3 FN results.
• Expert ruling could be created to change counts based on source or location that can be specific for each site.
• Current software could improve laboratory workflow by removing 42.2% of urine cultures based on growth less than the laboratories threshold.

Visit Product Pages to Learn More!
PhenoMATRIX™