Abstract
A urinary tract infection (UTI) can be diagnosed via urinalysis, consisting of a dipstick test and manual microscopic examination. Point-of-care (POC) image-based systems have been designed to automate the microscopic examination for low-volume laboratories or low-resource clinics. In this pilot study, acridine orange (AO) was evaluated as a fluorescence-based contrast agent to aid in detecting and enumerating urine sediment specific for diagnosing a UTI. Acridine orange staining of epithelial cells, leukocytes, and bacteria provided sufficient contrast to successfully implement image segmentation techniques, which enabled the extraction of classifiable morphologic features. Surface area bounded by each cell border was used to differentiate the sediment; epithelial cells were larger than 500μm2, bacteria were less than 30μm2, and leukocytes in between. This image-based semi-automated technique using AO resulted in similar cell counts to the clinical results, which demonstrates the feasibility of AO as an aid for POC urinalysis systems.