Browse Articles
Discover research articles across all indexed journals
Dosimetric comparison of 3D-conformal radiotherapy, volumetric modulated arc therapy, and h-VMAT in left breast radiotherapy under free-breathing and breath-hold conditions
Abstract This study investigates the role of the h-VMAT (a hybrid of 3DCRT and VMAT) technique in the Deep Inspirational Breath-Hold (DIBH) method for left-breast irradiation by comparing dosimetric parameters of 3DCRT, VMAT, and h-VMAT plans in free breathing and breath-hold conditions. The study enrolled fourteen left breast cancer patients with nodal involvement. Five sets of plans, namely, 3DCRT, VMAT, and three different dose ratios of h-VMAT, were created in both the DIBH and Free-Breathing (FB) data sets. All plans had similar planning criteria and were compared using Dose-Volume parameters of the target and critical organs. Both h-VMAT and VMAT achieved acceptable PTV parameters compared to 3DCRT. A significant reduction in heart mean dose was observed in the h-VMAT80/20 (p < 0.001) under both DIBH and FB methods and was found to be minimal compared to 3DCRT and VMAT. Similarly, Left Anterior Descending Artery (LAD) dose was also reduced in hybrid. Additionally, the left lung volume receiving 20 Gy and 30 Gy were less in hybrid. We conclude that h-VMAT, along with DIBH, is successful in achieving the best trade-off between heart exposed to high and low doses of radiation compared to 3DCRT and VMAT, respectively, without compromising target coverage for left breast radiotherapy.
Mechanistic analysis of frost heave in airport runway subgrades in seasonal frost regions and pavement structural tolerance thresholds
Secure federated learning with metaheuristic optimized dimensionality reduction and multi-head attention for DDoS attack mitigation
Axial strength prediction of FRP reinforced concrete columns under concentric and eccentric loading using machine learning models
Identification of cytotoxic constituents from Siegesbeckiae Herba and network pharmacology prediction of their anti-pancreatic cancer mechanisms
Abstract Pancreatic cancer remains a challenging malignancy with limited treatment options. This study aimed to isolate bioactive compounds from Siegesbeckiae Herba and explore their potential mechanisms against pancreatic cancer through network pharmacology and molecular docking. Twenty compounds, including one new compound, were isolated via chromatographic methods and structurally identified by NMR and mass spectrometry. Cytotoxicity of these compounds was evaluated against the human pancreatic cancer cell line PANC-1 using the CCK-8 assay. Two flavonoid constituents, compound 16 (8, 3’-dihydroxy-3, 7, 4’-trimethoxy-6-O-β-D-glucopyranosyl) and compound 17 (quercetin-3-methyl ether), exhibited cytotoxicity, with compound 16 showing the most potent activity (IC 50 = 4.48 ± 0.74 µg/mL). Network pharmacology analysis identified 182 potential targets for these compounds, including key targets such as AKT1, PIK3CA, SRC, and HRAS etc. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses indicated enrichment of critical pathways such as PI3K-Akt, Ras, and HIF-1 signaling, which regulate cell proliferation, migration, inflammation, and apoptosis. Molecular docking confirmed strong interactions between compounds 16 and 17 and key pancreatic cancer targets, yet the two flavonoids occupied distinct active pockets. Notably, compound 16 displayed stronger binding affinities (<-7.8 kcal/mol) toward top-ranked targets including HRAS, PIK3CB, PIK3CA, and HSP90AA1. Our results suggest flavonoids from Siegesbeckiae Herba, especially compound 16 , as promising candidates for pancreatic cancer treatment, warranting further pharmacological investigation.
Scene image visual layout based on deep encoder–decoder network and visual image attention model
Machine learning model for early prediction of acute kidney injury in heatstroke patients based on the first 24 h hospitalization data
A unified probabilistic energy and carbon footprint appraisal for intermittent water supply systems in arid regions
Establishing the relationship between heavy oil viscosity and molecular markers using an enhanced neural network model
Correction: Association between bipolar disorder and diabetic ketoacidosis/hyperosmolar hyperglycemic state
Decoding the relationship between oxidative stress and antiseizure medications using network pharmacology and molecular docking
Fault diagnosis of rolling bearing failures using a multi-stage e-CNN-GRU-SAM network
Abstract This study presents a forensic diagnostic framework aimed at enhancing the early detection, fault classification and remaining useful life (RUL) prediction of rolling bearing failures. The proposed network integrates a novel three-stage machine learning formulation – (1) identification of health state using voting ensemble, (2) prognostic analysis via a hybrid convolutional neural network and gated recurrent unit (CNN-GRU), and (3) fault type identification through the segment anything model (SAM) based on time-frequency representations. The ensemble and CNN-GRU models are trained on both time- and frequency-domain features from vibration signals, while SAM leverages this data in visual sense through iterative masking for zero-shot spatial-temporal fault segmentation. Pre-processing techniques, including piecewise aggregate approximation and singular spectrum analysis, are used to denoise and compress the vibration response without impacting key statistical traits. The proposed e-CNN-GRU-SAM network demonstrates better accuracy in diagnosing fault types, predicting RUL and identifying root causes under different operational conditions. This is established using diverse operating benchmark datasets that simulate induced and real-world degradation scenarios for generalization. Thus, the proposed framework offers a comprehensive forensic analysis toolkit for diagnosis and prognosis of bearings.
Field study of compound microbial agents for soil improvement and microbial community dynamics on rocky slopes in Southwest China
Abstract Engineering construction and mining activities have resulted in numerous exposed rocky slopes, posing significant geological and ecological challenges. To address these issues, this study developed compound microbial agents composed of functional microorganisms and applied them in a field-based ecological restoration project on typical high and steep rocky slopes in Southwest China. The aim was to restore the micro-ecological environment of the slopes by reconstructing the soil conditions. After 240 days of restoration, soil available phosphorus increased from < 2.0 to 6.10 mg/kg, available potassium from 62.80 to 75.00 mg/kg, organic matter from 8.90 to 12.86 g/kg, and organic carbon from 0.70 to 0.73%. Total nitrogen and total phosphorus slightly increased, indicating improved soil fertility. The addition of compound microbial agents enhanced microecological stability while maintaining the overall structure of the indigenous microbial communities. The progressive development of biological soil crusts and rock fissures facilitated the colonization of algae, lichens, mosses, and higher plants. By the 8th month, vegetation coverage exceeded 30% in some areas. This study presents an effective field-based model for the microbial ecological restoration of rocky slopes and offers insights into ecosystem-recovery mechanisms supporting sustainable land management.
Symplectic physics-embedded learning via Lie groups Hamiltonian formulation for serial manipulator dynamics prediction
Abstract Accurate dynamic modeling is critical for advanced robotic control, yet conventional methods struggle with manipulator nonlinear complexity. While Hamiltonian neural networks leveraging Lie group symmetries improve physical consistency of the network, existing methods overlook key limitations: unconstrained sparsity in mass, dissipation, and control matrices; redundancy in mass network outputs; and lack of validation on multi-rigid-body systems. This paper proposes a symplectic physics-embedded learning approach (SPEL) based on Lie group Hamiltonian formulations for enhanced dynamics modeling of serial manipulators. By systematically encoding physical priors such as Lie group symmetries and Hamiltonian dynamics into neural network design, SPEL enforces sparsity in mass, dissipation, and control input matrices via physics-driven constraints and replaces input-independent matrix elements with trainable parameters. These mechanisms structurally optimize the network topology, significantly reducing output dimensionality while preserving physical consistency of the network. Experimental validation on simulated two-link and revolute-prismatic-revolute (RPR) manipulators, as well as a real 6-DOF manipulator, demonstrates that SPEL reduces over 52% of the parameters, enhances computational efficiency by more than 75%, and achieves higher prediction accuracy. Additionally, Symplectic Physics-Embedded Learning Kolmogorov-Arnold Networks (SPEL-KAN) reduce over 63% of the parameters and improve computational efficiency by more than 39%. This approach embeds geometric-mechanical principles into architectures, balancing efficiency with interpretable predictions.