M4CXR Study Wins Best Oral Presentation Award at ESTI

Deepnoid, a leading first-generation medical AI company in Korea, announced on the 18th that a study evaluating the performance of the generative AI-based chest X-ray (CXR) report draft generation model 'M4CXR' by the research team led by Professor Jung Yeonju at Pusan National University Hospital received the Best Oral Presentation Award at 'ESTI 2025'.

ESTI 2025 was held on June 14 in Bergen, Norway. The award is a prestigious recognition given by the European Society of Thoracic Imaging to abstracts with significant clinical and scientific contributions, and this year, it represents the highest honor among all abstracts presented at ESTI.

The award demonstrates, in an objective manner, that the M4CXR model can be effectively used as a diagnostic support tool at international academic conferences. Its significance lies in confirming the practical potential of using medical AI software.

The study evaluated the performance of the generative AI-based report draft generation model using Deepnoid's 'M4CXR' and explored its clinical applicability as a diagnostic support tool.

The research team conducted a comprehensive analysis of M4CXR’s diagnostic accuracy, report quality, and efficiency. The model achieved an accuracy of 85% for clinically significant findings. It recorded diagnostic accuracies of 89.2% in screening environments and 87.6% in emergency room tests. The average time to generate a draft report was 3.4 seconds (ranging from 2.0 to 9.6 seconds), confirming its potential as a diagnostic support tool.

Professor Jung Yeonju, who received the Best Oral Presentation Award, said, "I am delighted that our research achievements have been recognized at such a prestigious international venue," and added, "Recently, the use of AI as an auxiliary tool in diagnostics has become a major topic." She further stated, "We will continue to conduct research on the performance and clinical effectiveness of medical AI solutions."

A Deepnoid representative commented, "This award is an international recognition of M4CXR’s performance and clinical applicability in the field of radiology," and added, "We plan to accelerate technological advancement in preparation for clinical trial approval from the Ministry of Food and Drug Safety."

This year, Deepnoid developed the generative AI-based chest X-ray interpretation model 'M4CXR' and is currently undergoing procedures related to clinical trials with the Ministry of Food and Drug Safety. The company is accelerating the development of next-generation solutions, aiming to launch the first and highest-performing generative AI-based medical device in Korea, surpassing the limitations of existing imaging data-based diagnostic support technologies.