RESUMO
In the rapidly evolving field of healthcare, the integration of artificial intelligence (AI) has become a pivotal component in the automation of clinical workflows, ushering in a new era of efficiency and accuracy. This study focuses on the transformative capabilities of the fine-tuned KoELECTRA model in comparison to the GPT-4 model, aiming to facilitate automated information extraction from thyroid operation narratives. The current research landscape is dominated by traditional methods heavily reliant on regular expressions, which often face challenges in processing free-style text formats containing critical details of operation records, including frozen biopsy reports. Addressing this, the study leverages advanced natural language processing (NLP) techniques to foster a paradigm shift towards more sophisticated data processing systems. Through this comparative study, we aspire to unveil a more streamlined, precise, and efficient approach to document processing in the healthcare domain, potentially revolutionizing the way medical data is handled and analyzed.
RESUMO
This study explores the progression of intracerebral hemorrhage (ICH) in patients with mild to moderate traumatic brain injury (TBI). It aims to predict the risk of ICH progression using initial CT scans and identify clinical factors associated with this progression. A retrospective analysis of TBI patients between January 2010 and December 2021 was performed, focusing on initial CT evaluations and demographic, comorbid, and medical history data. ICH was categorized into intraparenchymal hemorrhage (IPH), petechial hemorrhage (PH), and subarachnoid hemorrhage (SAH). Within our study cohort, we identified a 22.2% progression rate of ICH among 650 TBI patients. The Random Forest algorithm identified variables such as petechial hemorrhage (PH) and countercoup injury as significant predictors of ICH progression. The XGBoost algorithm, incorporating key variables identified through SHAP values, demonstrated robust performance, achieving an AUC of 0.9. Additionally, an individual risk assessment diagram, utilizing significant SHAP values, visually represented the impact of each variable on the risk of ICH progression, providing personalized risk profiles. This approach, highlighted by an AUC of 0.913, underscores the model's precision in predicting ICH progression, marking a significant step towards enhancing TBI patient management through early identification of ICH progression risks.
Assuntos
Lesões Encefálicas Traumáticas , Progressão da Doença , Aprendizado de Máquina , Humanos , Masculino , Feminino , Lesões Encefálicas Traumáticas/diagnóstico por imagem , Lesões Encefálicas Traumáticas/patologia , Lesões Encefálicas Traumáticas/complicações , Pessoa de Meia-Idade , Estudos Retrospectivos , Adulto , Hemorragia Cerebral/diagnóstico por imagem , Hemorragia Cerebral/patologia , Tomografia Computadorizada por Raios X , Idoso , Medição de Risco/métodosRESUMO
Matrix metalloproteinases (MMP) are enzymes that degrade the extracellular matrix and regulate essential normal cell behaviors. Inhibition of these enzymes has been a strategy for anti-cancer therapy since the 1990s, but with limited success. A new type of MMP-targeting strategy exploits the innate selective hydrolytic activity and consequent catalytic signal amplification of the proteinases, rather than inhibiting it. Using nanomaterials, the enzymatic chemical reaction can trigger the temporal and spatial activation of the anti-cancer effects, amplify the associated response, and cause mechanical damage or report on cancer cells. We analyzed nearly 60 literature studies that incorporate chemical design strategies that lead to spatial, temporal, and mechanical control of the anti-cancer effect through four modes of action: nanomaterial shrinkage, induced aggregation, formation of cytotoxic nanofibers, and activation by de-PEGylation. From the literature analysis, we derived chemical design guidelines to control and enhance MMP activation of nanomaterials of various chemical compositions (peptide, lipid, polymer, inorganic). Finally, the review includes a guide on how multiple characteristics of the nanomaterial, such as substrate modification, supramolecular structure, and electrostatic charge should be collectively considered for the targeted MMP to result in optimal kinetics of enzyme action on the nanomaterial, which allow access to amplification and additional levels of spatial, temporal, and mechanical control of the response. Although this review focuses on the design strategies of MMP-responsive nanomaterials in cancer applications, these guidelines are expected to be generalizable to systems that target MMP for treatment or detection of cancer and other diseases, as well as other enzyme-responsive nanomaterials.
Assuntos
Antineoplásicos , Nanoestruturas , Neoplasias , Humanos , Nanoestruturas/química , Neoplasias/tratamento farmacológico , Peptídeos/química , Antineoplásicos/uso terapêutico , HidróliseRESUMO
One of the artificial intelligence applications in the biomedical field is knowledge-intensive question-answering. As domain expertise is particularly crucial in this field, we propose a method for efficiently infusing biomedical knowledge into pretrained language models, ultimately targeting biomedical question-answering. Transferring all semantics of a large knowledge graph into the entire model requires too many parameters, increasing computational cost and time. We investigate an efficient approach that leverages adapters to inject Unified Medical Language System knowledge into pretrained language models, and we question the need to use all semantics in the knowledge graph. This study focuses on strategies of partitioning knowledge graph and either discarding or merging some for more efficient pretraining. According to the results of three biomedical question answering finetuning datasets, the adapters pretrained on semantically partitioned group showed more efficient performance in terms of evaluation metrics, required parameters, and time. The results also show that discarding groups with fewer concepts is a better direction for small datasets, and merging these groups is better for large dataset. Furthermore, the metric results show a slight improvement, demonstrating that the adapter methodology is rather insensitive to the group formulation.
Assuntos
Inteligência Artificial , Unified Medical Language System , Benchmarking , Conhecimento , IdiomaRESUMO
Accurate scatter estimation is important in quantitative SPECT for improving image contrast and accuracy. With a large number of photon histories, Monte-Carlo (MC) simulation can yield accurate scatter estimation, but is computationally expensive. Recent deep learning-based approaches can yield accurate scatter estimates quickly, yet full MC simulation is still required to generate scatter estimates as ground truth labels for all training data. Here we propose a physics-guided weakly supervised training framework for fast and accurate scatter estimation in quantitative SPECT by using a 100× shorter MC simulation as weak labels and enhancing them with deep neural networks. Our weakly supervised approach also allows quick fine-tuning of the trained network to any new test data for further improved performance with an additional short MC simulation (weak label) for patient-specific scatter modelling. Our method was trained with 18 XCAT phantoms with diverse anatomies / activities and then was evaluated on 6 XCAT phantoms, 4 realistic virtual patient phantoms, 1 torso phantom and 3 clinical scans from 2 patients for 177Lu SPECT with single / dual photopeaks (113, 208 keV). Our proposed weakly supervised method yielded comparable performance to the supervised counterpart in phantom experiments, but with significantly reduced computation in labeling. Our proposed method with patient-specific fine-tuning achieved more accurate scatter estimates than the supervised method in clinical scans. Our method with physics-guided weak supervision enables accurate deep scatter estimation in quantitative SPECT, while requiring much lower computation in labeling, enabling patient-specific fine-tuning capability in testing.
Assuntos
Redes Neurais de Computação , Tomografia Computadorizada de Emissão de Fóton Único , Humanos , Tomografia Computadorizada de Emissão de Fóton Único/métodos , Simulação por Computador , Tronco , Imagens de Fantasmas , Método de Monte Carlo , Espalhamento de Radiação , Processamento de Imagem Assistida por Computador/métodosRESUMO
We investigated the use of amphiphilic, protease-cleavable peptides as encapsulation moieties for hydrophobic metallodrugs, in order to enhance their bioavailability and consequent activity. Two hydrophobic, gold-containing anticancer agents varying in aromatic ligand distribution (Au(I)-N-heterocyclic carbene compounds 1 and 2) were investigated. These were encapsulated into amphiphilic decapeptides that form soluble filamentous structures with hydrophobic cores, varying supramolecular packing arrangements and surface charge. Peptide sequence strongly dictates the supramolecular packing within the aromatic core, which in turn dictates drug loading. Anionic peptide filaments can effectively load 1, and to a lesser extent 2, while their cationic counterparts could not, collectively demonstrating that loading efficiency is dictated by both aromatic and electrostatic (mis)matching between drug and peptide. Peptide nanofilaments were nontoxic to cancerous and noncancerous cells. By contrast, those loaded with 1 and 2 displayed enhanced cytotoxicity in comparison to 1 and 2 alone, when exposed to Caki-1 and MDA-MB-231 cancerous cell lines, while no cytotoxicity was observed in noncancerous lung fibroblasts, IMR-90. We propose that the enhanced in vitro activity results from the enhanced proteolytic activity in the vicinity of the cancer cells, thereby breaking the filaments into drug-bound peptide fragments that are taken up by these cells, resulting in enhanced cytotoxicity toward cancer cells.
Assuntos
Antineoplásicos , Neoplasias , Humanos , Antineoplásicos/farmacologia , Antineoplásicos/química , Linhagem Celular Tumoral , Endopeptidases , Ouro/química , Peptídeo Hidrolases , Peptídeos/farmacologia , Peptídeos/química , CápsulasRESUMO
The paper describes guidelines for the planning, organization, and successful execution of virtual, global scientific conferences for global audiences. The guidelines are based on experience and lessons learned during the organization of the 3-day 2020 Virtual Systems Chemistry Symposium hosted on Zoom webinar and Twitter, held on May 2020 with over 1000 registered participants from 46 different countries.
RESUMO
Supramolecular materials have gained substantial interest for a number biological and nonbiological applications. However, for optimum utilization of these dynamic self-assembled materials, it is important to visualize and understand their structures at the nanoscale, in solution and in real time. Previous approaches for imaging these structures have utilized super-resolution optical imaging methods such as STORM, which has provided important insights, but suffers from drawbacks of complex sample preparation and slow acquisition times, thus limiting real-time in situ imaging of dynamic processes. We demonstrate a noncovalent fluorescent labeling design for STED-based super-resolution imaging of self-assembling peptides. This is achieved by in situ, electrostatic binding of anionic sulfonates of Alexa-488 dye to the cationic sites of lysine (or arginine) residues exposed on the peptide nanostructure surface. A direct, multiscale visualization of static structures reveals hierarchical organization of supramolecular fibers with sub-60 nm resolution. In addition, the degradation of nanofibers upon enzymatic hydrolysis of peptide could be directly imaged in real time, and although resolution was compromised in this dynamic process, it provided mechanistic insights into the enzymatic degradation process. Noncovalent Alexa-488 labeling and subsequent imaging of a range of cationic self-assembling peptides and peptide-functionalized gold nanoparticles demonstrated the versatility of the methodology for the imaging of cationic supramolecular structures. Overall, our approach presents a general and simple method for the electrostatic fluorescent labeling of cationic peptide nanostructures for nanoscale imaging under physiological conditions and probe dynamic processes in real time and in situ.
Assuntos
Nanopartículas Metálicas , Nanoestruturas , Ouro , Microscopia de Fluorescência , PeptídeosRESUMO
Overexpression and activation of matrix metalloproteinase-9 (MMP-9) is associated with multiple diseases and can serve as a stimulus to activate nanomaterials for sensing and controlled release. In order to achieve autonomous therapeutics with improved space-time targeting capabilities, several features need to be considered beyond the introduction of an enzyme-cleavable linker into a nanostructure. We introduce guiding principles for a customizable platform using supramolecular peptide nanostructures with three modular components to achieve tunable kinetics and morphology changes upon MMP-9 exposure. This approach enables (1) fine-tuning of kinetics through introduction of ordered/disordered structures, (2) a 12-fold variation in hydrolysis rates achieved by electrostatic (mis)matching of particle and enzyme charge, and (3) selection of enzymatic reaction products that are either cell-killing nanofibers or disintegrate. These guiding principles, which can be rationalized and involve exchange of just a few amino acids, enable systematic customization of enzyme-responsive peptide nanostructures for general use in performance optimization of enzyme-responsive materials.
Assuntos
Metaloproteinase 9 da Matriz/metabolismo , Metaloproteinases da Matriz/metabolismo , Nanoestruturas/química , Peptídeos/química , Cinética , Metaloproteinase 9 da Matriz/química , Metaloproteinases da Matriz/químicaRESUMO
A central challenge in cancer care is to ensure that therapeutic compounds reach their targets. One approach is to use enzyme-responsive biomaterials, which reconfigure in response to endogenous enzymes that are overexpressed in diseased tissues, as potential site-specific anti-tumoral therapies. Here we report peptide micelles that upon MMP-9 catalyzed hydrolysis reconfigure to form fibrillar nanostructures. These structures slowly release a doxorubicin payload at the site of action. Using both in vitro and in vivo models, we demonstrate that the fibrillar depots are formed at the sites of MMP-9 overexpression giving rise to enhanced efficacy of doxorubicin, resulting in inhibition of tumor growth in an animal model.