https://pubs.kist.re.kr/handle/201004/153344 2026-04-13T05:34:30Z https://pubs.kist.re.kr/handle/201004/154556 Title: DeformMLP: Effective Deformation Prediction for Breast Cancer Using Graph Topology-Assisted MLPs Authors: Shin, Yong-Min; Lee, Kyunghyun; Lim, Sunghwan; Yoon, Kyungho; Shin, Won-Yong Abstract: Early diagnosis of breast cancer is crucial, enabling the establishment of appropriate treatment plans and markedly enhancing patient prognosis. While direct magnetic resonance imaging (MRI)-guided biopsy demonstrates promising performance in detecting cancer lesions, its practical application is limited by prolonged procedure times and high costs. To overcome these issues, an indirect MRI-guided biopsy that allows the procedure to be performed outside of the MRI room has been proposed, but it still faces challenges in creating an accurate realtime deformable breast model. In our study, we propose DeformMLP, a deformation prediction method that uses graph topology-assisted multilayer perceptrons (MLPs) as the main backbone architecture. DeformMLP is able to effectively predict the deformation of nodal surfaces given a point force with significantly faster training and low memory requirements. As DeformMLP is designed to take force vectors and graph features as input, along with nontrivial graph structure encoding, which performs feature propagation based on the underlying graph constructed from the element information. Our experimental results demonstrate that DeformMLP outperforms graph neural network (GNN)based alternatives with respect to both test root mean squared error (RMSE) and efficiency in time and memory costs. The source code is publicly available at https://github.com/jordan7186/DeformMLP. 2025-09-01T00:00:00Z https://pubs.kist.re.kr/handle/201004/154555 Title: Efforts to Improve Nonclinical Guidelines for Enhancing the Safety of Prophylactic mRNA Vaccines in South Korea Authors: Che, J. -H.; Keum, G.; Kang, B. -C.; Yun, J. W.; Youn, H.; Lee, S. -M.; Nam, J. -H. Abstract: The emergence of safety concerns regarding novel COVID-19 vaccines four years ago underscored the urgent need for nonclinical safety evaluation methods tailored to the unique characteristics of mRNA vaccines. In response, the World Health Organization (WHO) recommended in 2021 that five key aspects be specifically considered in in vivo safety assessments of mRNA vaccines. However, the current South Korean vaccine guidelines do not fully reflect these considerations, and WHO's recommendations lack specific protocols for their assessment. Therefore, there is a critical need to establish concrete protocols and guidelines for the nonclinical evaluation of prophylactic mRNA vaccines to enhance their safety while aligning with WHO's framework. To address this issue, this study consolidates findings from the “Research on the Development of Toxicity Evaluation Methods for mRNA Vaccines” project, which has been supported by the National Institute of Food and Drug Safety (NIFDS) in South Korea since 2022. The findings aim to contribute to the revision of South Korea's nonclinical evaluation guidelines for mRNA vaccines. For the nonclinical safety assessment, various mRNA vaccine platform components were produced and analyzed. First, characterization, immunogenicity, and efficacy evaluations were conducted for each produced substance to determine its viability as a vaccine candidate. Subsequently, each of the five WHO-recommended factors was assessed. Regarding biodistribution and persistence, we evaluated distribution patterns in normal animals as well as in models representing vulnerable populations and pre-existing conditions. Analytical methods were refined, and target organs were identified to propose a preliminary guideline for mRNA vaccines. For inflammation risk evaluation, studies were conducted on chronic inflammation, mitochondrial toxicity, and toxicity assessments using interleukin-1 receptor antagonist knockout (IL1ra KO) mice, aiming to propose appropriate nonclinical inflammation evaluation methodologies. For general toxicity evaluation, a comprehensive assessment was performed across multiple species, from mice to non-human primates, examining different mRNA platform types, lipid nanoparticle (LNP) components, administration routes, durations, and dosage levels. This enabled an in-depth analysis of target organ toxicities and provided recommendations for improving Good Laboratory Practice (GLP) toxicity studies of mRNA vaccines. Additionally, comparative analyses between normal animals and those with pre-existing conditions were conducted to assess potential differences in toxicity profiles, contributing to the feasibility of utilizing specialized animal models in future assessments. Based on these findings, we will propose a nonclinical safety evaluation protocol for prophylactic mRNA vaccines in South Korea and present a draft guideline as a foundation for assessing future mRNA platform-based vaccines. 2025-09-01T00:00:00Z https://pubs.kist.re.kr/handle/201004/154553 Title: Preliminary Study on the Connective Tissue Sheath Removal Device to Facilitate Insertion of Peripheral Nerve Interfaces Authors: Jang, Namseon; Ji, Suhyun; Yu, Soomin; Park, Shinsuk; Hwang, Donghyun Abstract: In inserting neural interfaces to obtain neural signals from peripheral nerves, the connective tissue sheath of the peripheral nerve, which protects the peripheral nerves, makes it challenging to insert interfaces. Therefore, to penetrate the connective tissue sheath and insert the interface into the nerve, the mechanical properties of the interface should be higher than those of the nerve fibers, or a high-stiffness insertion device should be additionally used, which can cause nerve damage during the insertion process. In this study, we propose a connective tissue sheath device to partially remove the connective tissue sheath as a method for inserting low-stiffness neural interfaces. The effect of collagenase on peripheral nerves to remove the connective tissue sheath was investigated by exvivo experiments, and the amount of the removed connective tissue sheath was analyzed. An in-vivo animal experiment was conducted using the developed connective tissue sheath removal device, and the amount of connective tissue sheath was determined at 15-minute intervals using an optical coherence tomography (OCT) device. Melamine foam is used as a carrier to react 5 mg/mL of collagenase solution with peripheral nerves, and the melamine foam is designed to move back and forth according to the scanning situation. As a result, the connective tissue sheath reacts with the collagenase, and the residual amount is reduced from approximately 96 mu m to approximately 29 mu m after 1 hour. 2025-01-01T00:00:00Z https://pubs.kist.re.kr/handle/201004/154552 Title: Deep Learning-Based Sarcopenia Classification through Gait Video Analysis with a Single Mobile Camera Authors: Jamsrandorj, Ankhzaya; Jung, Heeeun; Lee, Daehyun; Kim, Jinwook; Mun, Kyung-Ryoul Abstract: As the global population ages and life expectancy increases, early detection and continuous monitoring of sarcopenia-an age-related decline in muscle mass and strength-are critical for promoting healthy aging. Traditional assessment methods rely on expensive, specialized medical equipment and expert intervention, limiting their practicality for everyday use. To address these challenges, this study proposes a novel vision-based approach for identifying sarcopenia using gait. A total of 92 elderly individuals participated, including 60 patients with sarcopenia and 32 healthy controls. Digital cameras captured each participant's walking motion, from which 2D skeleton sequences were extracted. Our deep learning model, trained on these 2D skeleton sequences along with additional gait-related features, classified sarcopenia and healthy controls with 82.88% sample-wise accuracy and 94.44% subject-wise accuracy on the test dataset. 2025-07-01T00:00:00Z