DeepNoid's generative AI-based chest X-ray preliminary report generation solution is being introduced for the first time to a medical imaging specialty institutional investor. The company has expanded its supply channels to include large-scale remote reading centers, following introductions at university hospitals and private general hospitals.
DeepNoid announced on the 16th that it signed a supply contract for 'M4CXR' with Human Medical Imaging Center, a medical imaging specialty institutional investor, on the 1st of this month.
Human Medical Imaging Center is a specialized institutional investor where approximately 150 radiologists work across eight subspecialty reading centers. It collaborates with 20 university hospitals nationwide and over 800 hospitals to provide remote reading services.
M4CXR is a digital medical device that analyzes chest X-ray images using generative AI and automatically generates preliminary reports in the form of written findings. It analyzes normal findings and 41 types of abnormal chest findings to present draft reports for review by medical staff. Final interpretation and diagnosis remain the responsibility of radiologists.
Unlike existing medical imaging AI that marks the presence or suspected areas of specific diseases in images, M4CXR generates preliminary reports in sentence form used by medical professionals. It reduces repetitive report-writing tasks and supports radiologists to focus on image review and final judgment.
According to DeepNoid, M4CXR was designated as an innovative medical device by the Ministry of Food and Drug Safety (MFDS) in November last year—the first generative AI-based medical device to receive such designation—and obtained product approval in June this year.
M4CXR was previously introduced to three university hospitals, including Pusan National University Hospital, through the regional responsibility medical institution AI-based clinical support system project. In August, it also signed a supply contract with Pohang Se-myung Christian Hospital, a private general hospital.
Through this contract, DeepNoid has expanded M4CXR's supply channels to include university hospitals and private general hospitals participating in government-supported projects, as well as medical imaging specialty centers. In particular, Human Medical Imaging Center receives images from multiple medical institutions for large-scale reading, making it an ideal site to validate the solution's efficiency in actual reading operations.
Kim Sung-hyun, CEO of Human Medical Imaging Center, stated, "Chest X-rays are among the most frequently performed examinations in clinical practice, but limited by a shortage of interpreting physicians, their speed has been constrained. We expect M4CXR to assist radiologists in interpretation and improve workflow efficiency."
DeepNoid plans to evolve M4CXR from a simple preliminary report generation tool into a service that supports the entire reading process. The company envisions applying AI across the workflow—from image assignment to report writing and result delivery—to reduce repetitive tasks for medical staff.
To achieve this, DeepNoid completed the MFDS licensing procedure for its remote reading operation platform 'AxonFlow' in August. The goal is to integrate AxonFlow with M4CXR to unify the workflow encompassing medical image reception and classification, report assignment, preliminary report generation, and result delivery.
Choi Woo-sik, CEO of DeepNoid, said, "Human Medical Imaging Center serves as a reading hub where all processes—from image assignment to report writing and result delivery—converge. Through this contract, we will expand our product into an agent AI-based medical service that supports the entire reading workflow."
in TECH
