A5012088444- Radiomics and Machine Learning in Medical Imaging
- Nuclear reactor physics and engineering
- Fault Detection and Control Systems
- Brain Tumor Detection and Classification
- Cancer Genomics and Diagnostics
- Multimodal Machine Learning Applications
- Advanced Image and Video Retrieval Techniques
- AI in cancer detection
- Image Retrieval and Classification Techniques
- Nuclear Physics and Applications
- Medical Imaging Techniques and Applications
- Nuclear Materials and Properties
- Nuclear physics research studies
- RNA modifications and cancer
- Machine Learning in Materials Science
- Artificial Intelligence in Healthcare and Education
- COVID-19 diagnosis using AI
- Nuclear Engineering Thermal-Hydraulics
- Advanced Radiotherapy Techniques
- Astronomical and nuclear sciences
- Domain Adaptation and Few-Shot Learning
- Radiation Therapy and Dosimetry
- Epigenetics and DNA Methylation
- Graphite, nuclear technology, radiation studies
- Boron Compounds in Chemistry
RIKEN Center for Advanced Intelligence Project
2019-2024
University of Illinois Urbana-Champaign
2022-2024
Tokyo Metropolitan University
2023
National Cancer Research Institute
2021-2022
Tokyo Medical and Dental University
2019-2021
National Cancer Center Hospital East
2018
National Cancer Center
2017
Rikkyo University
2013-2017
Kogakuin University
1995
Mortality attributed to lung cancer accounts for a large fraction of deaths worldwide. With increasing mortality figures, the accurate prediction prognosis has become essential. In recent years, multi-omics analysis emerged as useful survival tool. However, methodology relevant not yet been fully established and further improvements are required clinical applications. this study, we developed novel method accurately predict patients with using data. unsupervised learning techniques,...
Lung cancer is one of the leading causes death worldwide. Therefore, understanding factors linked to patient survival essential. Recently, multi-omics analysis has emerged, allowing for groups be classified according prognosis and at a more individual level, support use precision medicine. Here, we combined RNA expression miRNA with clinical information, conduct analysis, using publicly available datasets (the genome atlas (TCGA) focusing on lung adenocarcinoma (LUAD)). We were able...
Abstract Deep learning is a promising method for medical image analysis because it can automatically acquire meaningful representations from raw data. However, technical challenge lies in the difficulty of determining which types internal representation are associated with specific task, feature vectors vary dynamically according to individual inputs. Here, based on magnetic resonance imaging (MRI) gliomas, we propose novel extract shareable set that encode various parts tumor phenotypes. By...
Machine learning models for automated magnetic resonance image segmentation may be useful in aiding glioma detection. However, the differences among facilities cause performance degradation and impede This study proposes a method to solve this issue. We used data from Multimodal Brain Tumor Image Segmentation Benchmark (BraTS) Japanese cohort (JC) datasets. Three tumor are developed. In our methodology, BraTS JC trained on datasets, respectively, whereas fine-tuning developed model...
In medical imaging, the characteristics purely derived from a disease should reflect extent to which abnormal findings deviate normal features. Indeed, physicians often need corresponding images without of interest or, conversely, that contain similar regardless anatomical context. This is called comparative diagnostic reading images, essential for correct diagnosis. To support reading, content-based image retrieval (CBIR) can selectively utilize and features in as two separable semantic...
Several challenges appear in the application of deep learning to genomic data. First, dimensionality input can be orders magnitude greater than number samples, forcing model prone overfitting training dataset. Second, each variable’s contribution prediction is usually difficult interpret, owing multiple nonlinear operations. Third, genetic data features sometimes have no innate structure. To alleviate these problems, we propose a modification Diet Networks by adding element-wise scaling. The...
To contribute to automating the medical vision-language model, we propose a novel Chest-Xray Different Visual Question Answering (VQA) task. Given pair of main and reference images, this task attempts answer several questions on both diseases and, more importantly, differences between them. This is consistent with radiologist's diagnosis practice that compares current image before concluding report. We collect new dataset, namely MIMIC-Diff-VQA, including 700,703 QA pairs from 164,324...
Abstract DNA methylation is an epigenetic modification that results in dynamic changes during ontogenesis and cell differentiation. patterns regulate gene expression have been widely researched. While tools for analysis developed, most of them focused on intergroup comparative within a dataset; therefore, it difficult to conduct cross-dataset studies, such as rare disease studies or cross-institutional studies. This study describes novel method analysis, namely, methPLIER, which enables...
Abstract Background In an extensive genomic analysis of lung adenocarcinomas (LUADs), driver mutations have been recognized as potential targets for molecular therapy. However, there remain cases where target genes are not identified. Super-enhancers and structural variants frequently identified in several hundred loci per case. Despite this, most cancer research has approached the these data sets separately, without merging comparing data, no examples integrated LUAD. Methods We performed...
This study aimed to evaluate the residual radioactivity in mice induced by neutron irradiation with an accelerator-based boron capture therapy (BNCT) system using a solid Li target. The radionuclides and their activities were evaluated high-purity germanium (HP-Ge) detector. saturated of irradiated mouse was estimated assess radiation protection needs for BNCT system. 24Na, 38Cl, 80mBr, 82Br, 56Mn, 42K identified, radioactivities (1.4 ± 0.1) × 102, (2.2 101, (3.4 0.4) 2.8 0.1, 8.0 (3.8 101...
The volume of medical images stored in hospitals is rapidly increasing; however, the utilization these accumulated remains limited. Existing content-based image retrieval (CBMIR) systems typically require example images, leading to practical limitations, such as lack customizable, fine-grained retrieval, inability search without and difficulty retrieving rare cases. In this paper, we introduce a sketch-based (SBMIR) system that enables users find interest need for images. key concept feature...
The neutron-rich, even-even 122,124,126Pd isotopes were studied via in-beam γ-ray spectroscopy at the RIKEN Radioactive Isotope Beam Factory. Excited states 499(9), 590(11), and 686(17) keV found in three isotopes, which we assign to respective 21+→0gs+ decays. In addition, candidates for 41+ 1164(20) 1300(22) observed 122Pd 124Pd, respectively. resulting Ex(21+) systematics are similar Xe (Z=54) isotopic chain theoretical predictions by proton-neutron interacting boson model (IBM-2),...
The amount of medical images stored in hospitals is increasing faster than ever; however, utilizing the accumulated has been limited. This because existing content-based image retrieval (CBMIR) systems usually require example to construct query vectors; nevertheless, cannot always be prepared. Besides, there can with rare characteristics that make it difficult find similar images, which we call isolated samples. Here, introduce a novel sketch-based (SBMIR) system enables users interest...
Explainability is key to enhancing the trustworthiness of artificial intelligence in medicine. However, there exists a significant gap between physicians' expectations for model explainability and actual behavior these models. This arises from absence consensus on physician-centered evaluation framework, which needed quantitatively assess practical benefits that effective should offer practitioners. Here, we hypothesize superior attention maps, as mechanism explanation, align with...
Abstract This study introduces a physics-regularized neural network (PRNN) as computational approach to predict silicon carbide’s (SiC) swelling under irradiation, particularly at high temperatures. The PRNN model combines physics-based regularization with methodologies generalize the behavior of SiC, even in conditions beyond traditional empirical model’s valid range. ensures continuity and accuracy SiC predictions extreme environments. A key aspect this research is using nested...