Clinical Performance Evaluation of AI-Enabled Automated Gram Staining Device

Gram staining,whitch provides crucial information for selecting antibiotics in the diagnosis and treatment of infectious diseases, is often manually performed in many laboratories, and the microscopic findings are subjectively interpreted and reported based on the experience of laboratory technicians. Automating the entire process from staining to microscopic examination, as achieved by the AI-equipped automatic Gram staining device, can reduce the labor required for Gram staining. This automation enables even less experienced technicians to quickly report Gram staining results.

The aim of this study is to evaluate the accuracy of the Auto microscopic examination of Gram-stained samples using anonymized processed slides created from clinical specimens. This evaluation will be compared with the accuracy of microscopic examinations performed by hospital laboratory technicians.

If the accuracy of the machine is demonstrated, it can be registered as a medical device and proceed to the implementation stage in actual clinical settings. The implementation of this device in clinical laboratories is expected to reduce workload, expedite processes, and contribute to the optimization of antibiotic use based on Gram staining results.