Jeonbuk National University (JBNU) student Se-hyun Lee (College of Medicine, Premedical Course, 2nd year) won the Best Poster Presentation Award at the 8th International Ho Medical Symposium hosted by Korea University College of Medicine, receiving international recognition for the excellence of his artificial intelligence (AI)-based medical imaging research outcomes.

The International Ho Medical Symposium is an international academic event attended by medical students from domestic and overseas medical schools. All presentations and Q&A sessions are conducted in English, and the event comprehensively evaluates the academic completeness, creativity, and clinical significance of research. Students from foreign medical schools also participated, creating fierce competition, and Se-hyun Lee was selected as an outstanding researcher in the poster category.

Lee's award-winning study is an AI-based image synthesis project that automatically generates T1-weighted images from a single T2-weighted acquisition in pediatric brain magnetic resonance imaging (MRI). It proposed a novel approach to address limitations in the pediatric patient examination environment. The study was conducted under the supervision of Professor Kim Hyun-ho of the Department of Pediatrics at Jeonbuk National University Hospital.

Notably, the study applied an age-conditioned diffusion model, a diffusion-based deep learning architecture that explicitly incorporates age information into the model. It was designed to precisely learn structural differences according to stages of pediatric brain development. As a result, it effectively modeled age-related brain structural differences that previous image synthesis studies did not fully capture. It demonstrated superior performance over prior studies in both quantitative image quality metrics (PSNR, SSIM, etc.) and visual assessments, achieving state-of-the-art (SOTA) level results.

The study suggested the potential to mitigate clinical problems commonly encountered during pediatric MRI examinations, such as motion artifacts, increased scan time, and the burden of sedation. If applied in real clinical settings, it could provide diagnostically necessary information even when T1 images are not acquired, thereby contributing to improved safety and examination efficiency for pediatric patients.

The judging panel commented, 'The research design is impressive not only for its technical completeness but also for its deep understanding of the unique clinical environment of pediatric patients, and it has high potential for practical application in clinical settings.'

Lee said, 'This study is the result of considering medical AI technology that will provide tangible benefits to pediatric patients beyond mere performance improvements,' and added, 'We plan to extend the research to elderly patients for whom long MRI scans are difficult.' He also remarked, 'I am honored that the significance and potential of the research were recognized on the international academic stage alongside medical students from abroad, and I hope to further develop the work through publication in international journals and follow-up studies.'

Meanwhile, follow-up research is currently underway with the goal of publishing in an international academic journal, and studies are being considered to extend the work beyond brain MRI to various medical imaging modalities.