MRI-based deep learning and radiomics for preoperative prediction of P53abn endometrial cancer: A multicenter study.

Acute febrile illnesses are frequently attributed to malaria in endemic countries, despite the circulation of arboviruses whose vectors share the same breeding environments. This study assessed the molecular prevalence of arboviral and arboviral-malaria co-infections and associated epidemiological and hematological characteristics among febrile children in Libreville, Gabon. Plasma samples from 151 febrile children attending a malaria sentinel surveillance site between September and October 2022 were retrospectively analyzed. Samples were screened by pan-DENV, CHIKV and ZIKV RT-qPCR for arboviruses diagnosis. DENV-positive samples underwent additional molecular characterization, including serotype-specific RT-PCR and pan-flavivirus PCR targeting the NS5 region. Malaria was diagnosed by microscopy, and demographic and hematological data were recorded. DENV RNA was detected in 18.5% of patients, with serotype identification confirming DENV-1 and DENV-3 circulation and suggesting possible DENV-4. No CHIKV or ZIKV infections were detected. Malaria prevalence was 56.9%, and 9.3% had dengue-malaria co-infection. Dengue predominated in children <5 years, whereas those aged 5-10 years had higher odds of malaria (OR: 2.61 IC95%: [1.07-7.02]; p=0.03) and co-infection (OR:16.28 IC95% [1.85-79.40]; p=0.002). DENV infections occurred mainly in urban/peri‑urban areas , while malaria was associated with rural residence (p=0.03). Severe anemia was most common in malaria (p=0.02), and thrombocytopenia was most marked in co-infected patients (p < 0.01). These findings confirm active DENV circulation and malaria-dengue co-endemicity in Libreville, highlighting the need to strengthen differential diagnostic approaches and surveillance strategies for febrile illnesses in urban Gabon.

Abstract Reconstructing the trajectories of charged particles in high-energy collisions requires high precision to ensure reliable event reconstruction and accurate downstream physics analyses. In particular, both precise hit selection and transverse momentum estimation are essential to improve the overall resolution of reconstructed physics observables. Enhanced momentum resolution also enables more efficient trigger threshold settings, leading to more effective data selection within the given data acquisition constraints. In this paper, we introduce a novel endto-end tracking approach that employs the differentiable programming paradigm to incorporate physics priors directly into a machine learning model. This results in an optimized pipeline capable of simultaneously reconstructing tracks and accurately determining their transverse momenta. The model combines a graph attention network with differentiable clustering and fitting routines, and is trained using a composite loss that, due to its differentiable design, allows physical constraints to be back-propagated effectively through both the neural network and the fitting procedures. This proof of concept shows that introducing differentiable connections within the reconstruction process improves overall performance compared to an equivalent factorized and more standard-like approach, highlighting the potential of integrating physics information through differentiable programming.

Alzheimer’s disease (AD) develops as a progressive dementia condition through the step-by-step breakdown of nerve cells. Neurodegeneration in this context primarily results from metal ions, including copper, iron, zinc, and aluminum, building up in the system. The aggregation of amyloid-beta (Aβ) peptides and oxidative stress generation stem from metal ion involvement acting as defining characteristics of Alzheimer’s disease pathology. We developed a comprehensive mathematical model based on 24 coupled ordinary differential equations (ODEs) to represent the interactions between metal ions, Aβ peptides, reactive oxygen species (ROS), antioxidant defenses, and tau protein phosphorylation. The mathematical model monitors how metal ion concentrations change over time and examines their competitive binding effects, which trigger a series of reactions, resulting in oxidative stress and subsequent tau protein damage. The model uses analytical and numerical mathematical methods to expose nonlinear behaviors and threshold effects while offering mechanistic insights into the course of disease development. This model functions as a quantitative framework for assessing how therapeutic interventions that target metal dyshomeostasis and oxidative stress can potentially affect outcomes.

Life tables are essential tools to describe insect biology and to understand population responses to environmental factors. Two main approaches are used to represent life tables data: the differential approach provides the distribution of the stage-development times, from which statistics can be computed, and supports the development of biodemographic models; the integral approach provides the stage-frequency matrices and offers a global view of the biological life cycle. However, the two representations are not equivalent, and it is not possible to switch from the one to the other if the original raw data are no longer available. As stage-frequency matrices are available for many species, this study introduces a novel method to estimate the statistics coming from the differential life tables representation, such as the mean development time and its standard deviation, using frequency matrices when raw data are not available. The approach combines constrained least-squares optimization with stage-impulse response models to infer the distribution of the development time from stage-frequency data. The method was validated using datasets of Corcyra cephalonica and Drosophila suzukii, for which individual life history data was available. The results show that the estimated distributions of the development times and their statistics are in accordance with validation data, especially for early developmental stages. This method provides a valuable methodology for estimating the distribution of the stage-development times using stage-frequency matrices when raw data is inaccessible or when individual rearing is impractical.

A vital question in neuroscience is whether and how efficiently cellular models may be differentiated into functional neuronal cells in culture. Despite the frequent use of the human neuroblastoma cell line SH-SY5Y, differentiation protocols vary extensively, with the most common being differentiation via the addition of retinoic acid and brain-derived neurotrophic factor. However, due to the lack of a reliable evaluation method, their adequacy as synaptic models remains unclear. Here, we investigate whether SH-SY5Y cells constitute a functional model for synaptic studies by phenotypically and ultrastructurally analyzing synaptogenesis in SH-SY5Y cells subjected to different differentiation protocols. Electron microscopy (EM) techniques, including conventional EM, cryo-EM, and cryo-electron tomography, were systematically applied to characterize synaptogenesis in SH-SY5Y cells. Further characterization was performed using immunostaining and functional assays, such as live exocytosis assays and whole-cell patch-clamp electrophysiology. Despite exhibiting some presynaptic-like features, differentiated SH-SY5Y cells do not form morphologically or functionally complete synapses under the conditions tested. Immunostaining results were consistent with previous findings, showing synaptic markers. However, functional investigations did not detect synaptic activity. High-throughput EM analyses revealed an absence of synaptic structures in these cells. Additionally, an alternative differentiation approach incorporating additional neurotrophic factors promoted the formation of presynaptic-like compartments containing synaptic vesicle-like vesicles (SVLVs). In contrast to typical synaptic vesicles, these SVLVs exhibited a pleomorphic size distribution and lacked connectors. These findings underscore the need for cautious interpretation of results derived from SH-SY5Y cells when investigating molecular synaptic architecture or function, as well as neurodegenerative diseases.

Prospective memory anchors upcoming activities in future time and space either cued externally (event-based prospective memory) or internally (time-based prospective memory). It is key to remain autonomous and can be enhanced using incentives. To examine the neuronal underpinnings of prospective memory and any performance improvement, we randomly assigned 58 healthy older participants (60-75 years old, 36 women) to time-based and focal event-based prospective memory tasks using 7T-fMRI, with or without incentives. Event-based and time-based prospective memory differ in their functional neuroanatomy. The focal event-based task required activity important for the perception of relevant changes in the environment and for adapting the focus of attention to then respond swiftly. The time-based task, in contrast, involved regions supporting cognitive control or serving as a multimodal information buffer. Incentives had no effect in time-based prospective memory but recruited brain regions important for attention thus improving cue recognition in event-based prospective memory. This highlights the differences between these two prospective memory types and the need for differential approaches depending on the type of memory demand.

Interval differential equation approach for inventory systems with green and price-dependent nonlinear demand

Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) have potential applications in treating cardiovascular disease but are currently limited in their clinical translation. This can be attributed in large part to the complex molecular and cellular interactions that underly cardiac differentiation, with current differentiation approaches yielding heterogeneous outcomes due to inadequate understanding and control of these interactions. We hypothesize that clonal lineage-dependent responses to differentiation contribute to these heterogeneous outcomes, and as such cardiac differentiations can be improved by tracking and controlling for hiPSC clonal heterogeneity, a variable often overlooked in current differentiation approaches. "Fate priming", wherein clonal lineage identity determines differentiation fate, has been demonstrated in other stem cell differentiation pathways. We investigated fate priming in hiPSC cardiac differentiation using the ClonMapper cell barcoding platform to label, track, and isolate distinct hiPSC lineages from the same cell line. We show that certain hiPSC lineages preferentially differentiate into hiPSC-CMs or non-CMs. After isolating lineages with apparent fate priming, we found significant differences in cardiac differentiation outcomes between these single-clone populations and heterogeneous, multi-clone hiPSC populations. These findings indicate that lineage identity influences hiPSC cardiac differentiation outcomes.

The differentiation between pathological subtypes of non-small cell lung cancer (NSCLC) is an essential step in guiding treatment options and prognosis. However, current clinical practice relies on multi-step staining and labelling processes that are time-intensive and costly, requiring highly specialised expertise. In this study, we propose a label-free methodology that facilitates autofluorescence imaging of unstained NSCLC samples and deep learning (DL) techniques to distinguish between non-cancerous tissue, adenocarcinoma (AC), squamous cell carcinoma (SqCC), and other subtypes (OS). We conducted DL-based classification and generated virtual immunohistochemical (IHC) stains, including thyroid transcription factor-1 (TTF-1) for AC and p40 for SqCC. We evaluated these methods using two types of autofluorescence imaging: intensity imaging and lifetime imaging. The results demonstrate the exceptional ability of this approach for NSCLC subtype differentiation, achieving an area under the curve above 0.981 and 0.996 for binary- and multi-class classification. Furthermore, this approach produces clinical-grade virtual IHC staining, which was blind-evaluated by three experienced thoracic pathologists. Our label-free NSCLC subtyping approach enables rapid and accurate diagnosis without the need for conventional tissue processing and staining. Both strategies can significantly accelerate diagnostic workflows and support efficient lung cancer diagnosis, without compromising clinical decision-making.

Special coupled 1D-2D procedure for predicting heat transfer regimes at supercritical parameters using differential approach

Pareto improving taxes with non-separable externalities: A differentiable approach

Introduction: Liver cancer globally remains the fourth-leading cause of cancer-related mortality. Metastatic lesions are more common in the liver tissue due to its rich portal and systemic venous blood supply. Even though an accurate histopathological diagnosis of tumours is possible in some cases. Distinguishing the hepatocellular carcinoma from metastatic lesion with similar morphology, poor differentiation and sub classification of carcinomas with unknown primary tumours always remains a problem. A multidisciplinary approach of histomorphological evaluation combined with systematic Immunohistochemistry (IHC) study is essential in differentiating neoplastic lesions. Aim: To evaluate the role of IHC in differentiating the neoplastic lesions of liver biopsy to hepatic origin and metastatic lesion. Materials and Methods: The present cross-sectional study of 50 neoplastic liver biopsy reported in the Departments of Pathology in oncquest laboratory for 3 years from march 2021 and February 2023, Bangalore, Karnataka, India. Clinical details were retrieved by request forms. All the tissue blocks were processed for Haematoxylin and Eosin (H&amp;E) sections and for IHC. Primary panel includes HepPar1, CK7,CK20,CK19 and KI67. The second panel includes arginase, glypican 3, CEA, CDX2 and SATB2. The third panel was decided further based on results of second panel, clinical details and radiological findings. A known positive and negative controls were used for each batch of slides. The slides were reviewed by two Pathologist and results were analysed. Results: Of the total 50 cases 37 were male (74%). The primary panel classified and second panel confirmed the liver biopsy into hepatic origin and metastatic lesion. The lesions classified were Hepatocellular Carcinoma (HCC) 7 (14%) cases, cholangiocarcinoma 2 (4%) cases, hepatic adenoma 6 (12%) cases and metastatic deposits 35 (70%) cases by the end of secondary panel. The third panel determines the origin of metastatic deposits. Metastatic lesions were the most common lesions in liver biopsy 35 (70%) cases. Colorectal adenocarcinoma was the most common lesion in metastatic deposits (74%). Conclusion: Diagnosing liver tumours on core biopsy specimen can be challenging. Accurate diagnosis is essential to establish the optimal treatment strategy. A systematic approach of initial differentiation by primary panel, confirmation by secondary panel and narrowing down the primary site of metastasis using third panel is highly recommended.

Measuring diffusion in strongly coupled systems is challenging due to collective interactions and dynamic heterogeneity, which limit the effectiveness of traditional methods. We propose a novel differential dynamic microscopy approach based on short-time dynamics, which directly extracts diffusion coefficients by focusing on short-time behavior, avoiding the effects of dynamic heterogeneity in strongly coupled systems without requiring particle tracking. Using Langevin simulations of dusty plasmas, we validate the method by comparing it with theoretical results. This approach provides a new pathway for precise diffusion measurements in strongly interacting systems and offers a reliable framework for studying transport in complex, correlated environments.

Safety-certified pursuit-evasion game of underactuated autonomous surface vehicles with experiments: A differential game approach with rate-tunable CBF optimization

Towards unified frameworks for fair and privacy-preserving graph neural networks.

Bilevel optimization is encountered in many relevant real-world applications. The main feature of this type of problem is that an upper-level optimization problem is constrained by a nested lower-level optimization problem. Because of this nested structure, bilevel problems are usually computationally expensive to solve. Differential Evolution has demonstrated promising results in solving bilevel problems of relatively small scales. As the problem scale increases, the decision space becomes intrinsically larger, requiring a growing number of function evaluations for the method to work properly. In this context, heavy parallelization and high-performance computing techniques are indispensable to enable the resolution of more complex and challenging optimization problems. Hence, we propose a hierarchical many-threaded parallel Differential Evolution approach for bilevel problems, where both levels are parallelized. The computational experiments demonstrate that the parallel implementation achieved runtime speeds ranging from 44 to 2559 times faster than the sequential version on a well-known scalable SMD benchmark test problem when executed on an NVIDIA A100 GPU. The findings indicate that the algorithm’s convergence is strongly influenced by the number of both upper-and lower-level generations. Moreover, the success of experiments with large-scale problems is closely linked to the choice of small population sizes.

Dissecting microglial contributions to neurodegenerative disease pathophysiology using human pluripotent stem cells.

Human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) hold tremendous promise for in vitro modeling to assess native myocardial function and disease mechanisms, as well as testing drug safety and efficacy. However, current hiPSC-CMs are functionally immature, resembling in vivo CMs of fetal or neonatal developmental states. The use of targeted culture media and organoid formats have been identified as potential high-yield contributors to improve CM maturation. This study presents an hiPSC-CM maturation medium formulation, designed using a differential evolutionary approach targeting metabolic functionality for iterative optimization. Relative to existing high-performing reference formulations, our medium significantly matured morphology, Ca2+ handling, electrophysiology, and metabolism, which was further validated by multi-omic screening, for cells in either pure or co-cultured microtissue formats. Together, these findings not only provide a reliable workflow for highly functional hiPSC-CMs for downstream use, but also demonstrate the power of high-dimensional optimization processes in evoking advanced biological function in vitro.

Location-Based Services (LBS) is an integral component of the modern digital environment, enabling the delivery of applications like smart city management and personalized mobile services. However, the service is challenged by privacy concerns due to the perpetual requirement for accurate spatial-temporal data. Although conventional privacy protection mechanisms like Differential Privacy (DP) offer strong theoretical guarantees of privacy, the approach is often based on a static privacy budget (ε) that does not offer satisfactory results when considering the data utility-privacy trade-off across various applications. In the present study, an adaptive Differential Privacy approach is proposed, enabled by the power of Artificial Intelligence (AI) technology. The approach allows the system to adapt the static privacy budget by adjusting the value of ε based on the contextual sensitivity of the query. Using the Random Forest algorithm, the system is able to evaluate the risks of the query by considering various parameters like the nature of the location, time factor, and user behavior. The experimental results based on a synthetic dataset for 5,000 location queries show that the proposed method significantly enhances the privacy utility trade-off. In specific terms, the proposed method reduces the success rate of inference attack by 49.2%, while utility is preserved with an increase in MAE by 63.7% for a controlled scenario. This demonstrates the applicability of the proposed method for adaptive and scalable privacy protection using ML and DP for LBS environments.