https://scholar.dgist.ac.kr/handle/20.500.11750/1194 Wed, 24 Jun 2026 16:46:46 GMT 2026-06-24T16:46:46Z https://scholar.dgist.ac.kr/handle/20.500.11750/60415 Title: A robust vision language model for molecular status prediction and radiology report generation in adult-type diffuse gliomas Author(s): Park, Yae Won; Kang, Myeongkyun; Ryu, Huiseung; Han, Kyunghwa; Sim, Yongsik; Park, Ji Eun; Chang, Jong Hee; Kim, Se Hoon; Lee, Seung-Koo; Park, Sang Hyun; Ahn, Sung Soo Abstract: We aimed to establish a robust vision-language model ("Glio-LLaMA-Vision") for molecular status prediction and radiology report generation (RRG) in adult-type diffuse gliomas. Multiparametric MRI data and paired radiology reports from 1001 patients with adult-type diffuse gliomas were included in the institutional training set. A vision-language model, Glio-LLaMA-Vision, was developed from LLaMA 3.1 pre-trained on 2.79 million biomedical image-text pairs from PubMed Central and further fine-tuned from the institutional training set. The performance was validated in 100 patients and 75 patients with paired MRI-radiology reports from an institutional validation set and another tertiary institution (AMC), and in 170 and 477 patients with MRI from TCGA and UCSF datasets, respectively. In terms of IDH mutation status prediction, Glio-LLaMA-Vision showed AUCs ranging from 0.85-0.95 in the internal validation and external datasets. In terms of RRG, the BLEU-1 and ROUGE-L scores were 0.50 and 0.49 in the internal validation, respectively, and 0.32 and 0.36 on the AMC dataset, respectively. Overall, 37.8% of generated reports were considered superior or equal to the original reports, while 91.0% of generated reports were considered clinically acceptable by neuroradiologists. In conclusion, Glio-LLaMA-Vision demonstrates promising performance in molecular status prediction and RRG in adult-type diffuse gliomas, showing potential for clinical assistance. Tue, 31 Mar 2026 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60415 2026-03-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60410 Title: Sustainable Bamboo-Based Magnetic Activated Carbon for Adsorption of Cationic and Anionic Dyes from Wastewater: Kinetics, Isotherms, and Thermodynamics Author(s): Ali, Asif; Matsumoto, Michiaki; Tahara, Yoshiro; Khan, Shahzad; Ali, Abbas; Rahman, Atta Ur Abstract: This study presents the synthesis and use of a novel bamboo-derived magnetic activated carbon (BMAC) for the effective removal of cationic and anionic dyes, specifically methylene blue (MB), methyl orange (MO), and sunset yellow (SY), from aqueous solutions. The adsorbent was synthesized using thermal carbonization and subsequent inclusion of magnetic oxide, yielding a porous structure with improved adsorption and magnetic separation properties. Thorough characterization utilizing SEM, EDX, BET, FTIR, XRD, and TGA/DTA validated the creation of a highly porous material including uniformly dispersed magnetic particles and several surface functional groups. Batch adsorption tests were performed to examine the influences of contact time, adsorbent dosage, initial dye concentration, pH, and temperature. The findings indicated rapid adsorption kinetics, with equilibrium reached in around 60-70 min, and adsorption capacity ranked as MB > MO > SY. Augmenting adsorbent dosage enhanced removal efficiency but diminished adsorption capacity per unit mass due to site unsaturation. The maximum adsorption capacities (q(m)) of BMAC were 58.9, 56.3, and 32.7 mg/g for MB, MO, and SY, respectively, as determined from the Langmuir isotherm model, indicating superior performance compared with other reported magnetic activated carbon. The adsorption process was determined to be exothermic and spontaneous, as evidenced by thermodynamic characteristics. The equilibrium data were optimally characterized by the Langmuir isotherm model, indicating monolayer adsorption, whereas the kinetic studies conformed to the pseudo-second-order model, signifying that chemisorption is predominant. The adsorption mechanism encompasses electrostatic interactions, pi-pi stacking, hydrogen bonding, van der Waals forces, pore filling, and surface complexation with magnetic oxides. The findings indicate that BMAC is an efficient, sustainable, and magnetically recoverable adsorbent for the elimination of both cationic and anionic dyes from wastewater. Thu, 30 Apr 2026 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60410 2026-04-30T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60408 Title: DEA 이중 부트스트랩(Double Bootstrap)을 이용한 국내 대학 창업지원 요소의 효율성 영향요인 분석 Author(s): 김흥희; 김대건 Abstract: 본 연구는 국내 대학 창업지원 시스템의 효율성에 영향을 미치는 핵심 요인을 규명하고자 하였다. 이를 위해 2020년부터 2024년까지 전국 17개 광역시·도의 대학 창업지원 관련 패널 데이터를 구축하고, Simar & Wilson(2007)이 제안한 2단계 DEA 이중 부트스트랩 방법론을 적용하였다. 1단계에서는DEA를 통해 지역별 창업지원 시스템의 상대적 비효율성 점수를 측정하고, 통계적 편향을 보정하였다. 2단계에서는 편향이 보정된 비효율성 점수를 종속변수로 하여, 어떤 창업지원 활동이 효율성에유의미한 영향을 미치는지 부트스트랩 절단 회귀분석을 통해 분석하였다. 분석 결과, 대학의 수많은창업지원 활동 중 ‘창업강좌 개설 수’가 비효율성을 유의하게 낮추는(즉, 효율성을 높이는) 핵심적인요인으로 나타났다. 반면, ‘전담 교직원 수’는 오히려 비효율성을 높이는(즉, 효율성을 저해하는) 경향을 보였다. 창업경진대회나 캠프와 같은 단기 이벤트성 프로그램들은 통계적으로 유의미한 영향을 미치지 못했다. 본 연구의 결과는 한정된 대학 창업지원 자원을 배분함에 있어, 단발성 행사나단순한 인력 확충보다는 장기적인 관점에서 학생들의 역량을 강화하는 교육 기반을 확충하는 방향으로 정책적 우선순위를 설정해야 함을 시사한다. Sun, 31 Aug 2025 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60408 2025-08-31T15:00:00Z https://scholar.dgist.ac.kr/handle/20.500.11750/60406 Title: Comprehensive Insights into Prion Diseases: Classification, Mechanisms of Action, Detection Methods, and Preventive Strategies Author(s): Bora, Jutishna; Kumari, Pallavi; Tripathi, Anjali; Mondal, Sagar; Pandey, Vinay Kumar; Malik, Sumira M.; Rustagi, Sarvesh; Ramniwas, Seema; Talukder, Nayan; Preetam, Subham Abstract: Health concerns have become increasingly prominent in society, with even minor injuries demonstrating an exaggerated potential for complications. This phenomenon parallels an unrelated yet historically significant observation from the 18th century, wherein Merino sheep exhibited abnormal behavioural manifestations indicative of a mysterious ailment. Upon further scientific investigation, this condition was identified as the first recorded instance of a novel class of neurodegenerative disorders now classified as prion diseases, or transmissible spongiform encephalopathies (TSEs), which are brought on by aberrant folding of the brain's cellular prion protein (PrPc). Rapidly progressive neurodegenerative disorders due to prion disease can be seen that can be tough to analyse and are transmissible under certain circumstances. Transmissible spongiform encephalopathies, sometimes known as prion diseases, are a fatal class of inherited, spontaneous, and epidemic neurodegenerative illnesses that affect both humans and animals. Prion disease arises in three ways: acquired or sporadic, inherited, or causal. Many pharmacological techniques and therapies aimed at various stages of the illness progression have been developed and assessed over an extended period, with a select number making progress in clinical trials. As yet, there are no helpful treatment medications authorised for prion illnesses. This paper contains a comprehensive study of the prion protein, prion disease, and different types of prion disease found in animals and humans. The mechanism of this disease, as well as its detection, diagnosis, treatment, and prevention, is discussed in this article. © 2025 Elsevier B.V., All rights reserved. Mon, 30 Jun 2025 15:00:00 GMT https://scholar.dgist.ac.kr/handle/20.500.11750/60406 2025-06-30T15:00:00Z