Intravascular imaging modalities have significantly advanced cardiovascular research by providing highly detailed structural insights that traditional x-ray imaging cannot achieve. Over the last decade, there has been a growing focus on multimodal intravascular imaging to access functional and compositional information, extending the capabilities of conventional intravascular methods. Beyond the successful development and preclinical validation of such technologies, the translation into clinical practice stands as the next essential milestone. From an engineering perspective, the clinical applicabiity of novel multimodal imaging techniques depends on several practical considerations: ensuring patient safety regarding laser exposure, maintaining catheter compatibility with commercially available profiles, and minimizing procedural burdens for clinicians in real-world environments.
This presentation will briefly review recent developments in multimodal intravascular imaging, highlighting fluorescence lifetime imaging (FLIm) as a label-free modality with significant potential for assessing atherosclerotic plaque characteristics. We will discuss the development of multispectral FLIm for high-speed intravascular imaging and its integration with OCT into a single catheter system. Lastly, we will outline the essential technical considerations for clinical implementation, which are key to establishing this technology as a routine tool in clinical settings.
•Using a large-scale Korean medical corpus from diverse resources such as textbooks, news articles, and academic journals, we have developed a Korean medical term-pair reference standard, a specialized pre-trained BERT model for medical applications, and an algorithm for recommending medical specialties.
•Natural language processing models are inevitably influenced by the surrounding context in which they are applied (such as culture and institutional factors). Recent large language models, like ChatGPT and GPT-4, also require additional fine-tuning to achieve precise and specialized applications.
•Developing Korean NLP technology in the medical field is essential. In addition to building large, high-quality corpora, as demonstrated in this study, creating standard ized datasets is a meaningful step toward validating future medical NLP technology.
