This system supports highly accurate bacterial species estimation using AI image recognition technology for Gram stain images, and the selection of appropriate antibiotics connecting with antibiograms.
Currently, we need to prepare several kinds of staining solutions and rely on manual and experience-dependent techniques, or ask a central laboratory or testing center of a large hospital to do it. However, clinics or doctor offices have limited time to conduct Gram staining, and it takes at least several days to complete all the procedure if it is done at a testing center, making it impossible to immediately prescribe antibiotics to the patient at hand. The same problem exists in hospital emergency rooms and intensive care units.
Due to inappropriate use of antibiotics, reports of bacterial outbreaks of “Antimicrobial Resistance (AMR),” in which antibiotics are ineffective, are increasing worldwide.
The use of broad-spectrum antibiotics due to empirical treatment has contributed to the increase of drug-resistant bacteria.
Precision testing equipment is expensive and difficult to install in small and medium-sized medical facilities. When it is necessary to prescribe antibiotics without waiting for a culture test that takes several days, the prescription often depends on the experience of doctors and technicians.
As a result, incorrect selection of initial antibiotics can happen.
In addition, even when the causative bacteria are identified, appropriate antibiotics are not prescribed due to lack of sharing of medical knowledge.
To solve these problems, BiTTE supports highly accurate bacterial species estimation using AI image analysis technology and the selection of appropriate antibiotics in conjunction with the Antibiogram (a table of antibiotic susceptibility).
For the estimation of bacterial species, we have developed an image
recognition AI model that learns from the Gram stain images of urine samples and the results of bacterial species determination by culture tests, and use the inference results.
Gram stained images are captured with a smartphone camera attached to the eyepiece lens via an attachment to an optical microscope, and the results of drug susceptibility tests are used as statistical data in an antibiogram to support the proper use of antibiotics based on evidence data.
Aiming to reduce the occurrence of drug resistance, the system also displays spectrum scores and WHO’s AWaRe classification to enable narrower antibiotic selection.
The scope of samples will be expanded from urine to blood, sputum and other samples.