DeepNoid presents 8 AI studies

DeepNoid said on September 4 that eight of its medical AI abstracts had been accepted for the Korean Congress of Radiology 2026, a four-day meeting that opens on September 9 at KINTEX 2 in Goyang, South Korea. The Seoul-based company said the studies cover chest and brain imaging and were selected as three oral presentations and five posters. The company framed the work around clinical use in reading rooms and emergency settings, rather than model scores alone. KCR 2026 is the 82nd Korean Congress of Radiology, according to the conference website. ### What exactly is DeepNoid presenting? DeepNoid said the eight accepted abstracts break evenly across two imaging areas: four in chest imaging and four in brain imaging. The company said the common thread was evidence from clinical data on whether AI improved accuracy, efficiency or consistency when used alongside physicians in real reading workflows. The chest studies include a double-reading workflow with Seoul National University Hospital in which AI flags potentially missed cases after a physician’s first read, according to DeepNoid. (deepnoid.com) Another study with Seoul Metropolitan Government Boramae Medical Center tested AI-generated draft reports for chest X-rays in the emergency room and found higher sensitivity for major findings and shorter reading time, the company said. Two additional chest projects focused on underlying vision-language methods for report generation and lesion classification or localization in chest X-ray and CT. ### Which studies look closest to day-to-day radiology work? Two of the chest abstracts are described by DeepNoid as workflow studies rather than pure model-development work. One inserts AI after the first read to propose a second look only for cases with possible misses, and the other uses AI-generated preliminary text for emergency chest X-ray interpretation, according to the company. (deepnoid.com) DeepNoid described those studies in operational terms that are familiar to radiology departments: missed findings, reading time and sensitivity for major abnormalities. That emphasis also appeared in the company’s statement that it is expanding from supplying individual software products to supporting the broader interpretation workflow, a characterization attributed to Chief Executive Woosik Choi. (deepnoid.com) ### What is in the brain imaging group? DeepNoid said four accepted abstracts focus on neuroimaging, including two tied directly to its DEEP:NEURO product line. One study with Eunpyeong St. Mary’s Hospital examined whether AI-assisted reading of MRA scans improved detection of small cerebral aneurysms, while another with Chung-Ang University Hospital evaluated whether AI support improved agreement between readers and reduced reading time. (deepnoid.com) Two additional brain projects addressed algorithm development. DeepNoid said one combines a foundation model to segment white matter hyperintensity lesions more stably, and another with Gangnam Severance Hospital uses deep-learning style transfer to improve MRA image quality in patients with cerebral aneurysms. ### Which products is the company using to anchor the event? (deepnoid.com) DeepNoid said it will run an exhibition booth during the congress and demonstrate three products: M4CXR for generative-AI-based chest X-ray preliminary findings, DEEP:NEURO for brain aneurysm image interpretation and diagnostic support, and DEEP:NEURO BrainSeg for quantitative brain MRI analysis. The company said BrainSeg received approval last month from South Korea’s Ministry of Food and Drug Safety as software that automatically segments major anatomical brain structures and supports regional volumetric assessment. (deepnoid.com) ### Who else is involved at KCR? KCR 2026 is scheduled for September 9 through September 12 in Goyang, according to the congress website. DeepNoid said it will hold a luncheon symposium on September 11 under the theme “The Future of Neurovascular Imaging: DEEP:NEURO and Advanced MR Angiography.” DeepNoid said Lee Areum of Soonchunhyang University Bucheon Hospital will chair that session and Jung Seung-chae of Asan Medical Center will speak on research trends in neurovascular imaging and AI-enabled MR angiography. (deepnoid.com) The next public markers for the story are the conference presentations themselves and any abstracts or presentation materials released through KCR 2026 or DeepNoid after the meeting. (2026.kcr4u.org)