Digital holographic microscopy: use of artificial intelligence to enable rapid blood culture Gram stain digitization and interpretation.

Gram staining is one of the most commonly performed tests in the clinical microbiology laboratory. Results are used to both guide empiric antimicrobial therapy as well as downstream laboratory procedures, including culture workup. Despite its ubiquity, Gram staining is time-intensive, operator-dependent, and demands a rapid turnaround time. As such, it is an ideal target for automation. However, currently available automated microscopes struggle to rapidly and accurately collect image data due to physical challenges posed by focusing on micron-sized bacteria on slides with variable stain thickness and intensity. Here, we present the development of digital holographic microscopy, an optical system that scans slides using a multispectral coherent light instead of traditional lenses. This allows for the collection of all image data from a large portion (50%) of a slide. These data are then reconstructed and displayed as full color images. We additionally demonstrate that images collected by this method are qualitatively equivalent to those collected with optical microscopy and can be interpreted by individuals with no previous exposure to automated microscopy.IMPORTANCEGram staining is a commonly performed test that demands significant time and experience. Automation of this task is therefore attractive but has been hindered by physical limitations of traditional optics as currently implemented in automated microscopes. Here, we present the development of digital holographic microscopy, which can collect and display image data without the need for optical focusing. This technology may therefore provide benefit to clinical microbiology laboratories by increasing efficiency of the Gram staining process.