Contact
Privacy Policy

Copyright © DEEPNOID Inc. All rights reserved.

Back to list
Research

DEEPNOID’s M4CXR Study Wins Best Oral Presentation at ESTI 2025, Demonstrating Clinical Potential at an International Congress

June 18, 2025
DEEPNOID’s M4CXR Study Wins Best Oral Presentation at ESTI 2025, Demonstrating Clinical Potential at an International Congress

- Professor Yeon Joo Jeong of Pusan National University Hospital receives the Best Oral Presentation award in recognition of the study’s clinical and scientific contribution
- M4CXR, a generative AI-based draft report generation model, demonstrates its utility as a diagnostic support tool in clinical practice
DEEPNOID, one of Korea’s first-generation medical AI companies (CEO Woo Sik Choi), announced that a research team led by Professor Yeon Joo Jeong of Pusan National University Hospital received the Best Oral Presentation – Magna Cum Laude award at ESTI 2025, the 31st Annual Scientific Meeting of the European Society of Thoracic Imaging, held in Bergen, Norway, on June 14, 2025. The award recognized the team’s study evaluating the performance of M4CXR, DEEPNOID’s generative AI-based draft report generation model for chest X-ray (CXR).
The European Society of Thoracic Imaging (ESTI) presents this prestigious award to abstracts that make a significant clinical and scientific contribution. It is the highest honor among the abstracts presented at this year’s ESTI meeting.
The award objectively demonstrates, at an international congress, that M4CXR can serve as an efficient diagnostic support tool, and it confirms the real-world potential of medical AI software.
The study used DEEPNOID’s M4CXR to evaluate the performance of a generative AI-based draft report generation model and examined its clinical applicability as a diagnostic support tool. The research team analyzed M4CXR from multiple angles, including diagnostic accuracy, report quality, and efficiency. M4CXR achieved 85% accuracy for clinically significant findings, with diagnostic accuracy of 89.2% in health screening settings and 87.6% in emergency department cases. It generated AI-generated draft reports in an average of 3.4 seconds (range: 2.0–9.6 seconds), confirming its potential as a diagnostic support tool.
“I am very pleased that our research has been recognized on such a prestigious international stage,” said Professor Jeong. “The supportive use of AI in diagnosis has recently become a major topic, and we will continue our research on the performance and clinical validity of medical AI solutions.”
“This award is international recognition of M4CXR’s performance and clinical utility in radiology,” said a DEEPNOID representative. “We will accelerate the advancement of our technology in preparation for MFDS clinical trial approval.”
As of June 2025, DEEPNOID had developed M4CXR, a generative AI-based chest X-ray interpretation model, and was undergoing clinical trial procedures with Korea’s Ministry of Food and Drug Safety (MFDS). Moving beyond the limitations of conventional image-based diagnostic support technologies, DEEPNOID is accelerating the development of next-generation solutions, with the goal of launching Korea’s first and highest-performing generative AI-based medical device.

###