搜索结果：expensive

Integration of LiDAR and thermal sensing has become increasingly important in robotics, infrastructure diagnostics, environmental monitoring, and autonomous perception systems. LiDAR sensors provide accurate three-dimensional geometric information but do not directly capture thermal properties of observed objects, whereas thermal cameras provide temperature distributions without explicit spatial structure. Fusion of both sensing modalities enables thermally augmented 3D scene reconstruction and spatial localization of temperature anomalies. This paper presents a practical LiDAR-thermal fusion framework for three-dimensional localization of heat sources using an Ouster OS1 LiDAR sensor and a FLIR A70 thermal camera. The proposed framework includes intrinsic thermal-camera calibration, extrinsic LiDAR-thermal calibration, multimodal data synchronization, projection of LiDAR points onto the thermal image plane, and assignment of temperature values to spatial points. Additionally, a dedicated thermally distinguishable calibration target is proposed to enable reliable multimodal feature extraction under low-contrast LWIR imaging conditions. The developed framework was experimentally validated using real radiometric thermal data and LiDAR point clouds acquired under laboratory conditions. Quantitative evaluation demonstrated reprojection errors below 1 pixel and a mean hottest-point localisation error of approximately 4.1 cm at a distance of 12.3 m. The results confirm that accurate spatial localisation of thermal anomalies can be achieved using a geometry-based multimodal fusion approach without relying on computationally expensive learning-based methods. The proposed framework emphasises practical deployment, deterministic calibration, and applicability in scenarios where limited training data or constrained computational resources make learning-based approaches difficult to apply. The proposed system may be applied to building energy diagnostics, industrial inspection, technical infrastructure monitoring, and robotic perception systems that require reliable spatial localisation of heat sources under real measurement conditions.

A virtual training environment offers clear advantages for agricultural robotics. It provides a safe setting in which perception, navigation, and control algorithms can be evaluated without risking damage to either the robot or the crop. It also supports efficient data generation: large volumes of training data can be collected under diverse environmental conditions that would be costly, slow, and often season-dependent in real-world deployments. This broader variability improves model adaptability, reduces the risk of overfitting, and leads to more robust operation. In this paper, we argue that digital twin technology should therefore be understood not merely as a passive mirror of a physical robot, but as an active training environment in which multiple sensor-related subprocesses can be developed, tested, validated, and refined jointly. This paper is based on our experiences with digital twin technology used in the development of a vineyard robot, including a self-driving rover, sensor simulation, procedural map generation, and agriculture-specific movement models. Our contribution is threefold: we reinterpret the digital twin as a training space, propose a layered framework for training agricultural robots in virtual environments, and explain why agriculture is a particularly strong use case, given variable field conditions, expensive real-world experimentation, and persistent labor scarcity. To validate this framework, we present the simulation-based evaluation of an autonomous reinforcement learning agent. The agent has been trained entirely in this virtual environment, which successfully navigated to 155 out of 161 target points in a simulated vineyard demonstration environment.

Image sensors produce high-dimensional visual data for classification algorithms. Deep Neural Networks (DNNs) achieve high accuracy but require large labeled datasets and computational and energy resources, limiting their use in embedded systems. Active Learning (ALrn) can reduce labeling effort by selecting samples based on informativeness scores, but it remains computationally expensive, especially for high-dimensional images. This work presents a hardware-accelerated approach for the instance selection stage based on a query strategy in uncertainty-based ALrn for image classification using a novel in-line top-k selection algorithm that avoids conventional sorting and reduces memory and computational requirements. The algorithm is implemented on an Xilinx ZYNQ-7000 System on Chip (SoC) using a Field Programmable Gate Array (FPGA)-based accelerator operating at 110 MHz, interfacing with an embedded Advanced RISC Machine (ARM) processor for data acquisition and communication via the Python Productivity for Zynq (PYNQ) framework. Experiments on diverse multiclass datasets demonstrate correctness within an ALrn setting, showing negligible performance deviation in the learning curves compared to software baselines. The accelerator achieves speedup of 231.7× and 22.9× over software baseline and optimized software implementation of the proposed algorithm, respectively, in query-strategy computation while consuming only 0.473 W, substantially lower than conventional Central Processing Unit (CPU)- and Graphics Processing Unit (GPU)-based platforms. These results demonstrate the efficiency and extensibility of the proposed accelerator across alternative ALrn designs and hardware platforms, where the computational cost of instance selection scales with the size of the unlabeled pool.

The process of drug discovery is one of the most expensive, time-consuming, and high-risk endeavors in modern science. Translating initial scientific insights into safe and effective therapies, supported by genomics, structural biology, and computational chemistry, typically requires more than a decade and substantial financial investment. Machine learning (ML) has emerged as a powerful tool for improving efficiency across the drug discovery pipeline. By enabling the analysis of large and complex datasets, ML supports target identification, lead discovery, optimization, and prediction of preclinical and clinical outcomes. Its integration with experimental validation and automation is illustrated by recent advances such as protein structure prediction, AI-driven antifibrotic compound discovery, and antibiotic identification. Despite these advances, significant challenges remain. Model generalizability is limited by data scarcity, heterogeneity, and hidden biases. In addition, the translation of in silico predictions into clinically validated outcomes remains a major bottleneck, and regulatory acceptance is constrained by limited model interpretability. Ethical considerations, including data privacy, equitable representation, and the potential misuse of generative models, further complicate adoption. This review examines the applications of ML across the drug discovery pipeline, with a focus on translational and regulatory considerations. It also discusses emerging directions, including hybrid physics-AI approaches, multimodal foundation models, federated learning, and explainable AI. The effective integration of ML will depend on rigorous validation, interdisciplinary collaboration, responsible data governance, and alignment with regulatory frameworks.

Background/Objectives: Pharmaceutical formulation involves designing a drug product by combining the properties of an active pharmaceutical ingredient (API) with suitable excipients and processing strategies to produce a safe, effective, and manufacturable dosage form. However, data in formulation science are often limited, expensive to generate, and frequently restricted by proprietary and confidentiality constraints. Interactive digital tools can support formulators during early drug product development by improving the structure, transparency, and efficiency of formulation decision-making. While the current system focuses on structured decision support, future extensions may incorporate machine-learning methods for recommendation and knowledge extraction. Methods: In this work, we developed the Formulation tool, an interactive decision-support and visualization system for formulation development based on a hierarchical formulation-strategy framework commonly used in pharmaceutical practice. The tool is designed to prioritize suitable formulation principles and associated processing routes, with oral solid formulation as the initial application domain. The evaluated scenarios also include pathway regions relevant to oral liquid formulations. Its modular architecture also makes it adaptable to other formulation scenarios. To assess practical applicability, the tool was evaluated in a structured expert study involving five expert users across six predefined formulation scenarios (n = 30 runs), covering three drugs under adult and pediatric conditions. Results: The tool showed agreement with the expected dosage-form families and overall formulation properties, with adult scenarios converging to oral solid regions and pediatric scenarios converging to oral liquid regions. At the downstream formulation-profile level, users converged either to the dominant expected pathway or to alternative feasible pathways within the same formulation region. Variability in full pathway completion was observed and was primarily associated with differences in user interaction behavior and exploratory usage patterns. The median task completion time was 113.5 s. Conclusions: In addition to organizing formulation knowledge, the Formulation tool records user interactions in a structured manner, which may support future learning from usage patterns. Because detailed downstream formulation constraints are often institution-specific and are typically not available in the public domain, the present evaluation focused on agreement at the dosage-form-family level and on overall formulation properties rather than on highly specialized constraint logic. The system is based on a constraint satisfaction problem (CSP) framework, which is well suited for modeling complex decision processes under explicit constraints. CSP has also been widely applied in intelligent scheduling systems, supporting its suitability for structured, constraint-rich decision-making tasks such as pharmaceutical formulation strategy development.

To improve clinical decision-making about Carbapenem-resistant Gram-negative bacteria (CR-GNB) infections and halt the spread of resistant microbes, quicker and less expensive diagnostic techniques are required. Thus, the purpose of this study was to thoroughly evaluate the diagnostic efficiency (sensitivity, specificity, and concordance) of direct-from-specimen multiplex lateral flow immunoassay (LFIA) across diverse raw clinical specimens and pathogen types from critically sick patients. A total of 300 non-duplicate samples were tested to detect CR-GNB. Five major Carbapenemase genes were detected directly from the specimen using carbapenem-resistant K.N.I.V.O. detection K-Set and from culture using culture-enhanced multiplex PCR. Turnaround time (TAT) of each method was calculated. The direct LFIA revealed 100% specificity for NDM, KPC, and IMP enzymes in all tested clinical matrices (blood, urine, and respiratory samples). The study demonstrated 100% sensitivity and specificity with perfect categorical agreement (&#x3ba; = 1.000) for the blaKPC in the Klebsiella pneumoniae and for blaOXA-48 and blaIMP in the Acinetobacter baumannii; however, sensitivity of blaVIM was significantly diminished across all isolates and samples. TAT decreased significantly (p < 0.001) from 30 to 70 h to about 50 min. The tested direct LFIA facilitates the prompt enhancement of lifesaving tailored antibiotic treatment for severe illnesses.

Gliomas are the most aggressive malignant brain tumours, occurring mostly in adults and accounting for approximately 80% of central nervous system malignant tumours. Traditional diagnostic methods are both invasive and expensive, thus accurate, minimally invasive, and cost-effective early detection is vital to guide personalised treatment plans. MicroRNAs (miRNAs) are stable non-coding RNAs detectable in various body fluids (e.g., serum, plasma, and cerebrospinal Fluid (CSF)) that regulate gene expression and influence cellular processes; their dysregulation is a significant factor in cancer development. This makes them a promising biomarker for glioma classification. In this article, we present a glioma identification methodology from miRNA data using machine learning (ML) followed by data analysis for miRNA biomarker investigation. A machine learning pipeline is applied to classify glioma from controls as well as from meningioma samples using miRNA expression data obtained by four Gene Expression Omnibus (GEO) datasets (GSE112264, GSE113486, GSE113740, GSE139031). After preprocessing, five feature selection techniques (LASSO, mRMR, ReliefF, RFE, and RF importance) were employed. Six machine learning algorithms (LR, KNN, DT, RF, SVM, XGB) were used for classification with and without SMOTE oversampling. Performance was assessed after 5-fold cross-validation, in terms of accuracy, F1-score, precision, recall, and area under the curve (AUC). The results showed that in binary classification (glioma vs controls) all models achieving up to 100% accuracy, and in multi-class classification (glioma vs meningioma vs controls) up to 100% F1-score was achieved with both KNN and XGB classifiers. The top-ranked miRNAs were also analysed and compared with biomarkers previously known from the literature. Seven miRNAs were identified as potential biomarkers, namely the miR-125a-3p, miR-4276, miR-4648, miR-4763-3p, miR-663a, miR-6784-5p and miR-873-3p, and were independently validated on the GSE211692 dataset.

Early and reliable identification of cardiac disorders from Phonocardiogram (PCG) signals acquired from wearable biosensors is critical to support clinical decision making and reduce subjectivity in auscultation-based assessments. This study proposes a multi-stage hybrid feature selection-classification approach to increase diagnostic accuracy without requiring computationally expensive deep learning (DL) architectures. First, the most statistically discriminative features were identified using mRMR, ReliefF, and Kruskal-Wallis filtering methods. Particle swarm optimization (PSO) and ant colony optimization (ACO) were then applied to optimize the solution space. Finally, the selected feature subsets were tested with k-nearest neighbor (k-NN), support vector machines (SVMs), and Bagged Tree (BT) classifiers. Experimental results show that the proposed method significantly increases the model robustness and generalizability. In particular, the Kruskal-Wallis+k-NN and ReliefF+k-NN combinations achieved competitive performance compared to many DL-based approaches in the literature, with 99.80% accuracy and 99.50% F1-score. Furthermore, hybrid models augmented with PSO and ACO also achieved 99.60% accuracy. The findings demonstrate that well-designed feature selection strategies offer high accuracy and enhanced clinical applicability while using only a small set of handcrafted features and conventional classifiers. Therefore, the proposed framework is a strong candidate for smart stethoscope-based early screening solutions.

The contamination by foodborne pathogens posed a significant health threat and huge economic burden. Traditional detection methods were limited by cumbersome and time-consuming procedures, low automation, and reliance on expensive instrumentation, making them inadequate for on-site detection. This paper presented a centrifugal microfluidic chip that integrated sample pretreatment, nucleic acid extraction, amplification reaction, and signal detection. The chip featured an innovative design that combined a bursting valve with the siphon channel and employed a dual-channel configuration for splitting and directing the flow of different reagents, thereby overcoming the instability issue of unintended pre-activation or interruption that often occurred in the cascade design of multilevel siphon channels. Moreover, by synergistically combining with immunomagnetic separation as well as multi-enzyme isothermal rapid amplification, a portable, easy, rapid, high-sensitivity, and low-cost point-of-care testing (POCT) system for foodborne pathogens was developed. Under optimized conditions, the system enabled detection of Salmonella in spiked milk samples at 10 CFU/mL in 1 h. The recoveries ranged from 83.22% to 127.60%, with relative standard deviations of &#x2264;13.7%, indicating that this system had great potential for rapid and high-sensitivity detection of foodborne pathogens in resource-limited settings.

Visual impairment affects approximately 2.2 billion people worldwide, yet existing assistive technologies remain fragmented and prohibitively expensive. This paper presents Munir, an integrated multimodal assistive system designed to enhance human-computer interaction through a combination of a mobile application and Bluetooth-enabled smart glasses. Munir leverages a hybrid machine learning architecture to provide inclusive, real-time support for daily living activities. The system integrates ten core capabilities-including face recognition, optical character recognition, and scene description-all accessible through a unified bilingual (Arabic/English) voice interface. By employing on-device processing for biometric tasks, Munir ensures user privacy and trust while maintaining high responsiveness. End-to-end system evaluation on the SCface dataset achieves a 96.69% recognition rate with 0% False Accept Rate. At an estimated first-year total cost of $806, Munir demonstrates a 4-5&#xd7; cost advantage over commercial alternatives, providing a scalable and affordable multimodal solution for global digital inclusion.

Binding affinity prediction is about estimating the degree to which a drug binds to a protein. Predicting the binding affinity between a drug and a protein in a computational process helps researchers filter huge libraries of compounds before performing expensive biochemical lab experiments. Currently, there is interest in predicting binding affinity through computational pattern recognition or machine learning methods instead of the classical physics-inspired methods, which are computationally intractable except for tiny chemical compounds. In the last five years, several machine learning-based methods have been presented, whose experimental validations have achieved increasing Pearson coefficients while trained and tested in the PDBBind 2016 and CASF 2016 databases, respectively. These methods have an important diversity of architectures that provide different properties. The aim of this paper is to discern which binary properties (existence or absence) of these methods make them return higher Pearson coefficients. Basically, the properties introduced are related to the level of structural knowledge, the presence of 3D information, and the introduction of the relationship between the drug and the protein in the input of the model. The t-test confirms that the important binary properties for having a high Pearson coefficient are the protein (or part of the protein) being represented and introduced into the computational model as a graph, the pocket and the drug-protein interaction being part of the input, and incorporating the distance between atoms and the type of chemical bonds into the model.

Low back pain (LBP) is a major global health problem and can result in a variety of movement impairments. Advances in smart technology have enabled the collection of novel streams of movement data, and machine learning (ML) methods have been increasingly used for data analysis. However, many existing technologies remain expensive and unsuitable for widespread clinical use, and ML approaches have largely focused on distinguishing people with LBP from healthy controls rather than identifying meaningful subgroups within the LBP population. Motion Tape (MT) is a recently developed wearable strain sensor that translates skin deformation from underlying movement and muscle engagement into electrical signals. In this exploratory study involving 10 participants with LBP, we demonstrate that MT data from six sensors applied on the lower back capture rich movement information capable of characterizing movement patterns among participants with LBP. We propose a feature engineering approach based on biomechanical features as well as time-series causal discovery applied to multivariate sensor time-series data to extract directed inter-segment coordination patterns. We further develop an exploratory subgroup discovery pipeline by aggregating clustering coassociation information across diverse movement tasks. Our causal coordination features show promising discriminative information across several movement types, capturing aspects of motor control not reflected in amplitude-based or embedding-based features alone, such as asymmetries and movement restrictions. Preliminary ensemble clustering analysis indicates three potential LBP subgroups distinguished by biomechanical and inter-segment coordination patterns, which may reflect varied strategies under different movement demands. We investigate the differences in clinical characteristics among these LBP subgroups. We show that time-series foundation models are not well suited for LBP subgrouping due to their uninterpretability, which is improved in our feature engineering pipeline. This framework could reveal additional subgroups with larger cohorts and may generalize to other sensor modalities.

Texture modified foods are recommended for people with dysphagia but the visual appeal and nutritional quality of the foods may be reduced through modification. Moulding pureed foods back into their original shape rather than serving as a scoop in a ramekin may improve the visual appeal of texture modified foods, and their intake. The current study describes the cost, nutritional composition, intakes and satisfaction of two methods for preparing and serving texture modified meals in hospital, one cooked-fresh and scooped in ramekins and one commercially prepared and moulded. Data on patient nutritional intakes and adequacy of energy and protein were collected retrospectively from two cohorts of patients receiving texture modified foods in hospital using data recorded in the electronic menu management system and compared to theoretical nutrition requirements calculated based on height and weight. The first cohort (n&#x2009;=&#x2009;120) were served texture modified meals made on-site, served in ramekins. The second cohort (n&#x2009;=&#x2009;152) were served externally sourced commercially prepared, moulded food items. Overall, we found the externally sourced commercially-prepared, moulded meals to be more nutrient dense and approximately 10% more expensive than cooked-fresh, scooped meals, but there were no differences in nutritional intakes between the two cohorts (with patients meeting 43%-47% of energy requirements and 55%-59% of protein requirements). Patient satisfaction reports appeared to favour the commercially-prepared meals. Our findings suggest that for a 10% increase in cost, an externally sourced commercially prepared, moulded texture modified meal can deliver a more energy dense product though significant improvements in intakes of energy and protein were not achieved.

Rhamnolipid biosurfactants have garnered significant attention as renewable, biodegradable alternatives to conventional, petroleum-derived surfactants, aligning closely with global sustainable development goals. Nevertheless, their commercial viability remains hampered by intrinsic production challenges mainly suboptimal production titers, expensive downstream recovery, and biosafety constraints of native producers. This review synthesizes current advances within an integrated framework that couples metabolic pathway optimization with strategies to enhance titer, rate, and yield (TRY), high-cell-density cultivation, and cost-effective downstream processing. Key innovations in pathway engineering such as elimination of competing routes and cofactor rebalancing are examined alongside dynamic regulatory systems that decouple growth from product formation. Process intensification techniques, including in situ foam control and membrane-based separations, are evaluated for their impact on purification efficiency and overall process economics. Critical research gaps and limitations are identified, with recommendations for future efforts to bridge metabolic landscape and scale-up implementation. Finally, potential market-ready rhamnolipid formulations are discussed, illustrating their application spectrum from agriculture and personal care to environmental remediation and healthcare. By contextualizing synthetic biology breakthroughs within rigorous techno-economic and life-cycle assessments, this review maps a clear trajectory toward the industrial deployment of truly sustainable glycolipid biosurfactants.

Lipid-lowering therapies, particularly statins, are central to the prevention of atherosclerotic cardiovascular disease. However, their effectiveness is often compromised by statin-associated muscle symptoms, leading to non-adherence, discontinuation and/or switching to alternative, more expensive therapies such as Proprotein Convertase Subtilisin/Kexin type 9 (PCSK9) inhibitors or bempedoic acid. Recent studies suggest that N=1-interventions may help distinguish true side effects from nocebo-driven symptoms and support reinitiation of statin therapy. The 'N=1-studies In Statin-intolerance; Objectifying Nocebo Effects' (NISONE) trial is a study with 249 patients with atherosclerotic cardiovascular disease or familial hypercholesterolaemia who stopped using two or more statins due to perceived symptoms. Participants are randomised (2:1) to an N=1-intervention or usual care. The intervention consists of four double-blind 6-week periods of statin (rosuvastatin 10&#x2009;mg 1-2&#x2009;tablets/day or atorvastatin 20&#x2009;mg 1-2&#x2009;tablets/day) or placebo treatment during which patients are required to record their symptoms through questionnaires in an application developed for this study. During the subsequent fifth treatment period, feedback on symptoms during the intervention periods is provided in a personalised report, which will be discussed by a healthcare professional. Statin continuation is encouraged if symptoms are similar for statin and placebo periods, but remains voluntary. Statin-intolerant patients in the usual care group will be treated according to the cardiovascular risk management guidelines. The primary outcome, the percentage of patients continuing their statin after 1 year, will be analysed using an odds ratio and its 95% Confidence Interval. Secondary outcomes will be analysed similarly, and cost-effectiveness will be assessed using seemingly unrelated regression equations, adjusted for baseline scores, costs and quality of life. The study protocol was approved by a Medical Ethical Research Committee in The Netherlands (EU CT-number 2023-507489-20-00). The study results will be disseminated via peer-reviewed medical journals, conference presentations, advisory boards and, if possible, by using various media channels. This trial is registered in the EU Clinical Trials Information System (CTIS) (https://euclinicaltrials.eu/search-for-clinical-trials/?lang=en): 2023-507489-20-00.

Lignin is the most abundant renewable source of aromatic carbon, and yet it remains a mostly underutilized byproduct of the biorefinery and paper industries. Factors such as complexity and a heterogeneous structure make lignin recalcitrant to conventional valorization, the utility of which often requires harsh conditions and expensive catalysts. Electrochemical conversion has emerged as a highly promising, sustainable alternative due to the use of electricity produced by renewable sources to drive depolymerization under mild, ambient conditions. This review summarizes recent progress in this field and provides a comprehensive overview of the primary electrochemical pathways used to promote the valorization of lignin. Herein, we critically examine oxidative strategies that include both direct electrooxidation at the anode surface and indirect oxidation using redox mediators, and provide details of the key challenges of electrode deactivation and product overoxidation. We then discuss reductive strategies with a focus on electrocatalytic hydrogenolysis for C-O bond cleavage. Furthermore, we explore advanced integrated systems that combine electrochemistry with microbial, enzymatic, and photochemical processes to enhance selectivity and efficiency. Finally, this review addresses persistent challenges and offers future perspectives and suggests opportunities with an emphasis on the critical need for innovations in electrocatalyst design, green electrolytes, and integrated reactor engineering to unlock the full potential of lignin as a renewable feedstock for a circular carbon economy.

Prognostic health management of rolling element bearings requires feature representations that reliably track degradation while remaining tractable for real-time deployment. This paper investigates whether uniform time-delay embedding can serve as a near-optimal substitute for computationally expensive non-uniform embedding in recurrence-based vibration analysis. We show empirically that optimally chosen uniform delay vectors yield phase-space reconstructions of bearing vibration signals not significantly inferior to those produced by globally optimized non-uniform delay vectors, compressing the parameter search from a combinatorial optimization to a single scalar selection. Building on this near-optimality result, we construct color recurrence plots from uniformly embedded phase spaces and apply them to remaining useful life (RUL) prediction on the Intelligent Maintenance Systems (IMS) bearing dataset. We further demonstrate that standard binary recurrence plots are poorly suited for RUL estimation: their dense and erratically varying local patterns obscure the degradation trends required for reliable prognostics. Color recurrence plots, by contrast, suppress these local instabilities by averaging recurrence structures across multiple phase-space projections, exposing a globally evolving intensity that tracks bearing health throughout its degradation trajectory. This work establishes uniform delay embedding combined with color recurrence representation as an efficient, principled, and practically deployable approach to recurrence-based condition monitoring in industrial predictive maintenance.

Background/Objectives: High-grade B-cell lymphomas (HGBCLs) with MYC and BCL2 or BCL6 translocations, colloquially referred to as double-hit lymphomas (and abbreviated here to DHL), are aggressive malignancies. Differentiating DHL from non-DHL HGBCLs is important, as DHL patients may benefit from more intensive treatment regimes. We aimed to identify predictive clinicopathological, morphological, and immunophenotypic features that could guide selection of HGBCLs for fluorescence in situ hybridization (FISH), which is expensive and less accessible in some centers. Methods: We conducted a PRISMA systematic review and meta-analysis on 29 studies identified from four databases (PubMed (MEDLINE), Ovid (Embase), Web of Science, and Scopus). We calculated risk ratios (RRs) to compare features between DHL and non-DHL HGBCL and between MYC/BCL2 and MYC/BCL6 DHL patients. Results: DHL patients were associated with higher Ann Arbor stage (RR 1.15, p = 0.028, I2 = 38.7%), International Prognostic Index (IPI) score (RR 1.27, p = 0.047, I2 = 37.9%), elevated lactate dehydrogenase (RR 1.26, p = 0.012, I2 = 34.0%), and germinal center B-cell-like (GCB) immunophenotype (RR 1.21, p = 0.043, I2 = 35.8%) compared to non-DHL HGBCL patients. c-Myc immunopositivity, extranodal disease, and bone marrow involvement were more likely in DHL, albeit not reaching statistical significance. Extranodal disease (p = 0.015, I2 = 0.0%), central nervous system involvement (p = 0.044, I2 = 0.0%), and non-GCB immunophenotype (p = 0.016, I2 = 71.1%) were more likely in MYC/BCL6 compared to MYC/BCL2 DHL patients. BCL2 immunopositivity, CD10 immunopositivity, and MUM1 immunonegativity were more likely in MYC/BCL2 DHL, although the differences were not statistically significant. Conclusions: Our results have associated DHL with features of aggressive disease and found GCB immunophenotype as a histopathological feature with statistically significant predictive value for MYC/BCL2 DHL. Heterogeneity within the non-DHL HGBCL group and variation in immunohistochemical cut-off values between studies limited identification of other predictive features. Larger, consistently designed, prospective cohort studies could provide further evidence for a screening strategy for DHL.

Thermoresponsive composites made from phase change material particles embedded in polymer matrices show great promise for regulating sunlight transmittance in smart windows. However, their current performance is hindered by critical limitations. Composites obtained from water-soluble matrices present excellent optical performance, but they exhibit intrinsic poor weathering resistance, as they swell and dissolve upon water exposure (e.g., during rainfall or cleaning with aqueous solutions). In contrast, those prepared from water-resistant curable polymers typically require expensive or custom-made phase change materials, and very often show undesired optical properties (e.g., opaque states at room temperature or below). To address this challenge, in this work we developed novel thermoresponsive composite films comprising readily available paraffin particles dispersed within a cross-linked, water-insoluble acrylate matrix. The resulting materials exhibit excellent resistance to water while preserving the characteristic smart window behavior of paraffin-polymer composites; i.e., they remain transparent at low temperatures due to refractive index matching between the matrix and the solid paraffin particles, and become opaque once this condition is lost upon thermally induced paraffin melting, enabling efficient modulation of solar heating gain. In addition, because paraffin-acrylate films are produced via photopolymerization, they are obtained in a fast, straightforward and scalable manner. These features, combined with their low cost, mechanical flexibility, and multistimuli-responsive behavior, make paraffin-acrylate composites a robust and scalable platform for next-generation smart window technologies.