Five research abstracts accepted at the Radiological Society of North America (RSNA), the world’s largest radiology society, focusing on solving clinical problems with medical AI models
Demonstrating key solutions, including DEEP:NEURO and M4CXR, at an event that drew about 700 companies and 38,000 attendees last year
“RSNA is where global medical AI trends are set; we will actively pursue international research collaboration opportunities”
DEEPNOID, one of Korea’s first-generation medical AI companies (CEO Woo Sik Choi), announced that five of its research abstracts have been accepted at RSNA 2025, the annual meeting of the Radiological Society of North America. The meeting will be held for five days, from November 30 to December 4, 2025 (local time), at McCormick Place in Chicago, USA.
Founded in 1915, the Radiological Society of North America is the world’s largest radiology society, with more than 51,000 members in 157 countries. In 2024, about 700 companies and 38,000 registrants from around the world attended the meeting.
The five accepted abstracts cover different imaging modalities, but they share a common goal: improving the reliability, interpretability, and clinical reproducibility of medical AI models. Their academic significance was recognized because they go beyond improving AI model performance to demonstrate how AI can solve real clinical problems and be used in practice.
The five abstracts cover:
A dual-task learning architecture that addresses the performance gap between posteroanterior (PA) and anteroposterior (AP) chest X-ray views
AI-assisted double reading to detect chest X-ray interpretation errors using a vision-language model (VLM)
Evaluating the visualization performance and clinical usefulness of chest X-ray reports through heatmaps generated by a vision-language model
Developing a model for automated cerebral aneurysm detection and segmentation by fine-tuning the Segment Anything Model (SAM), a general-purpose segmentation AI model, and comparing its performance
A multicenter clinical trial evaluating the diagnostic performance of DEEP:LUNG, a lung nodule computer-aided diagnosis (CADx) system
On December 1, 2025 (local time), DEEPNOID will also present at the RSNA 2025 Seminar, hosted by the Korea Health Industry Development Institute (KHIDI) as part of the meeting program. Kyoyun Choi, Senior Researcher in AI Advanced Technology at DEEPNOID’s AI Research Lab, will give a talk on next-generation modalities.
DEEPNOID will also exhibit at a booth at RSNA from November 30 to December 3, 2025. At the booth, the company will introduce and demonstrate two medical AI solutions to clinicians, hospital representatives, and AI researchers from around the world: DEEP:NEURO, an AI solution for cerebrovascular disease diagnostic assistance, and M4CXR, a generative AI-based draft report generation solution for chest X-ray. Through the booth, DEEPNOID aims to share its research achievements and lay the groundwork for entering global markets.
“The acceptance of multiple abstracts means DEEPNOID’s technology has earned international recognition at RSNA, the global stage that leads medical AI trends,” said Woo Sik Choi, CEO of DEEPNOID. “At this meeting, we will communicate directly with potential global customers and gather their feedback, while actively pursuing collaboration opportunities with international research institutions.”
