— Identifies causative bacteria with accuracy comparable to microbiology specialists, presenting new possibilities in diagnostic testing —
— Joint research conducted with the National Center for Global Health and Medicine (NCGM, now the Japan Institute for Global Health – JIHS), the Center for Infectious Disease Research, and the Research Center for Urban Safety and Security / University Hospital of Kobe University —
CarbGeM Inc. (Headquarters: Shibuya-ku, Tokyo; CEO: Masakazu Nakajima) is pleased to announce that research findings on BiTTE®-Urine, an AI software for microbial estimation, has been published in the international peer-reviewed journal “The Journal of Medical Microbiology”, issued by the Microbiology Society in the UK (published on April 23, 2025).
This study is the first of its kind internationally to demonstrate that AI can analyze Gram stain images and identify causative pathogens of urinary tract infections (UTIs) with accuracy equivalent to that of microbiology specialists.
Title: Accuracy of classification of urinary Gram-stain findings by a computer-aided diagnosis app compared with microbiology specialists
Journal: Journal of Medical Microbiology
Publication Date: April 23, 2025
Authors: Kei Yamamoto, Goh Ohji, Isao Miyatsuka, Kei Furui-Ebisawa, Ataru Moriya, Shogo Maeta, Hidetoshi Nomoto, Masami Kurokawa, Kenichiro Ohnuma, Mari Kusuki, Yukari Uemura, Norio Ohmagari
DOI: https://doi.org/10.1099/jmm.0.002008
Urinary tract infections (UTIs) are commonly encountered in daily clinical practice, and prompt, accurate identification of the pathogen is crucial for effective treatment.
Traditionally, interpreting Gram stain findings has required specialized knowledge, posing challenges in ensuring diagnostic consistency and accuracy.
To address this issue, CarbGeM developed BiTTE®-Urine, AI-powered diagnostic support software that analyzes Gram-stained images taken with a smartphone.
The study evaluated using 306 Gram stain images collected from the National Center for Global Health and Medicine and Kobe University Hospital.
A non-inferiority study was conducted comparing the identification accuracy of the AI (BiTTE®-Urine) with 20 microbiology specialists (10 from each institution). The AI was statistically proven to be non-inferior.
The AI identified seven types: Gram-positive cocci, Gram-positive bacilli, Gram-negative cocci, Gram-negative bacilli, yeast-like fungi, polymicrobial, and no organisms.
Accuracy rates: BiTTE®-Urine: 87.9% / Microbiology specialists: 83.0%
Particularly high accuracy was noted in identifying Gram-negative bacilli and yeast-like fungi.
Gram staining is a valuable tool for diagnosing UTIs, providing rapid results on the same day, enabling appropriate initial antibiotic selection and treatment decisions. However, interpretation requires specialized training, and facilities that perform Gram staining are limited.
BiTTE®-Urine offers high-accuracy AI diagnostic support even in settings without specialists or during off-hours, contributing to appropriate antibiotic use and improved efficiency of medical resources.
Details on BiTTE®-Urine:
🔗 https://carbgem.com/product/bitte-urine/ (Japanese only)
Details on BiTTE-iE (RUO: Research Use Only)
🔗 https://carbgem.com/en/product-en/bitte-ie-en/
Moving forward, we aim to expand the application of this technology beyond UTIs and continue enhancing accuracy toward real-world implementation.
CarbGeM is a technology company advancing diagnostic support, research efficiency, and quality management in the life sciences through cutting-edge innovations including regulatory-compliant Software as a Medical Device (SaMD). We are committed to addressing social issues such as the shortage of skilled technicians and healthcare disparities by providing fast, accurate, and accessible results that support both clinical and industrial settings.
Leveraging AI and digital platforms, CarbGeM strengthens collaboration among researchers and healthcare professionals, contributing to solutions for global challenges like antimicrobial resistance (AMR).
Our efforts have been recognized with numerous awards, including the Tokyo Social Innovation Tech Award 2024. Through open innovation with top research institutions and industry-academia-government collaboration, we are shaping the future of healthcare and scientific discovery.
• Headquarters: 1-5-13 Jinnan, Shibuya-ku, Tokyo, Japan
• CEO: Masakazu Nakajima
• Website: https://carbgem.com/en/
This information and the following content are intended to provide information about our products for healthcare professionals in Japan. Please note that it is not intended to provide information to the general public and/or healthcare professionals outside Japan.
