Comparable Accuracy Research Articles

Comparison of fully angiography-derived versus a hybrid of angiography and pressure-wire-derived approach to assess coronary microvascular resistance: the OxAMI-Hybrid study

Abstract Background Despite growing evidence of its clinical implications, assessment of coronary microvascular dysfunction (CMD) remains limited in routine clinical practice. Hence, there is an increasing interest in angiography-derived indices to encourage assessment of CMD in the cardiac catheterisation laboratory. We investigate here a novel approach, IMRhybrid, using a combination of pressure-wire-based assessment of distal coronary pressure (Pd) and an angiography-derived surrogate of coronary flow. Methods Patients were enrolled to the OxAMI-Hybrid study. The accuracy of IMRhybrid was assessed in diagnosing CMD using fully pressure wire and bolus thermodilution Index of Microcirculatory Resistance (IMR) as a reference. The accuracy of IMRhybrid was then compared against a fully angiography-derived index of microvascular resistance (IMRangio). Based on previous work, angiography-derived IMR, IMRangio, was calculated as IMRangio = Quantitative Flow Ratio (QFR) ado x Pahyperaemia x TIMI Frame Count (TFC) hyperaemia / Frame Rate. Conversely, IMRhybrid was expressed as IMRhybrid = Pdhyperaemia x TFChyperaemia / Frame Rate, where frame rate was fixed at 15 frame per second for both indices. In both cases, TFC/frame rate represented angiography-derived mean transit time, mTt in hyperaemia. Results 60 patients were enrolled prospectively, of which 28 with acute coronary syndrome and 32 with stable coronary artery disease. A total of 66 vessel analyses were performed. Both IMRhybrid and IMRangio correlated with IMR (rho=0.51, p <0.001 and rho=0.51, p <0.001 respectively) and showed good and comparable overall diagnostic accuracy in predicting IMR ≥ 25 (83.3% and 75.4% respectively) with a receiver operator curve (ROC) analyses showing an AUC of 0.82 and 0.81 respectively (p for comparison= 0.97). Conclusion Both approaches, IMRhybrid and IMRangio provide viable and comparable methods of assessing CMD which correlate with bolus thermodilution-based IMR, retaining good diagnostic accuracy. IMRhybrid is a simple alternative to conventional bolus thermodilution-based IMR, offering the advantage of being suitable with any pressure wire system.A: IMR, B: QFR, C: TFCA+C: IMRhybrid, B+D: IMRangio

Towers of Kuwait Arabic Neurocognitive Assessment: A novel executive and visuospatial functions assessment tool added to the CERAD neuropsychological battery-Arabic version (CERAD-ArNB)

Assessment of executive and visuospatial neurocognitive domains is lacking in the Omani population, especially for elderly individuals with low educational levels. Therefore, the Towers of Kuwait-Arabic Neurocognitive Assessment (ToK-ArNA) was developed using similar psychometric features of Tower of London (ToL) test, and the unique architecture of Towers of Kuwait, with the potential to overcome the limitation of ToL color perception difficulties that might be encountered in subjects with hereditary or acquired color vision disorders. We enrolled 120 older Arabic-speaking Omanis from January 2022 to November 2022 and all participants underwent screening to ensure normal cognitive function before performing the ToL and ToK-ArNA tests. Validity, reliability, and non-parametric statistical tests were used for data analysis. A total of 85 participants, 51 men (60%) and 34 women (40%) met the inclusion criteria and completed the testing. Statistical analyses confirmed the validity and reliability of ToK-ArNA compared to ToL, with comparable total Time and Accuracy scores and more preference towered the ToK-ArNA among the participants. Despite the study limitations, these results indicate that the ToK-ArNA is a reliable and applicable executive and visuospatial function assessment tool and further studies are warranted to establish its validity in patients with various neurocognitive disorders

Nationwide trends in blood pressure control among those with and without diabetes in England over last 2 decades: Health Survey of England 2003-2019

Abstract Diabetes and hypertension are both risk factors for cardiovascular disease (CVD), and often co-exist. Over recent decades practice guidance and health policy efforts have been made to improve population wide blood pressure (BP) control, particularly amongst those with diabetes, to reduce CVD. However, the nationwide impact of these ongoing efforts is yet to be documented comprehensively in England. We used data from 11 serial nationally representative health surveys of England between 2003 and 2019 and compared BP control and presence of untreated hypertension amongst those with and without diabetes. These cross-sectional, randomly sampled surveys included 94,100 individuals cumulatively. We applied sample weights, accounting for oversampling and survey nonresponse, and transformations mitigating for survey equipment changes during this period. We used ‘period definitions’ to define those with diabetes and quantify BP control at the time of each coinciding survey, facilitating accurate comparison of trends over the sixteen-year period, while accounting for changes in understanding, variations in definitions, and shifts in medical practices. Python and STATA were used to conduct analyses and Join Point regression software was used to develop and compare trends. In England, from 2003-2019, the prevalence of hypertension amongst those with diabetes and without diabetes decreased from 74.4% to 66.5% and from 32.6% to 27.1% (7.9% vs. 5.5%), respectively. This corresponded to reductions in population-wide mean systolic (S) and diastolic (D) BPs: compared to those without diabetes, those with diabetes had greater reduction in both SBP (7.9mmHg vs. 4.8mmHg) and DBP (2.7mmHg vs. 1.9mmHg) (Figure1). Both groups showed similar temporal improvement in BP-control amongst those with diagnosed hypertension, with those with diabetes showing better improvement in BP control compared to those without diabetes (14% vs. 12%). There was a significant difference in proportions of those with undiagnosed (unaware) hypertension amongst those with and without diabetes (11.4% vs. 30.0% respectively) (Figure2). Notably, there was no change in proportion of those with hypertension unaware about the condition (i.e. detection rate for hypertension in either of the group did not show any improvement over time). These results suggest that we have been able to improve BP control over this period amongst those with and without diabetes, however, those without diabetes (more than 90% of population) have shown significantly poorer improvement. In addition, around one third of hypertensive patients amongst the no-diabetes group are unaware of the condition, and overall, the detection rates for hypertension in both groups have not improved in the last two decades. Lessons learnt from improving BP control amongst those with diabetes should be applied across the entire population, and newer strategies need to develop to improve detection of those with hypertension and are unaware.

A static quantum embedding scheme based on coupled cluster theory.

We develop a static quantum embedding scheme that utilizes different levels of approximations to coupled cluster (CC) theory for an active fragment region and its environment. To reduce the computational cost, we solve the local fragment problem using a high-level CC method and address the environment problem with a lower-level Møller-Plesset (MP) perturbative method. This embedding approach inherits many conceptual developments from the hybrid second-order Møller-Plesset (MP2) and CC works by Nooijen [J. Chem. Phys. 111, 10815 (1999)] and Bochevarov and Sherrill [J. Chem. Phys. 122, 234110 (2005)]. We go beyond those works here by primarily targeting a specific localized fragment of a molecule and also introducing an alternative mechanism to relax the environment within this framework. We will call this approach MP-CC. We demonstrate the effectiveness of MP-CC on several potential energy curves and a set of thermochemical reaction energies, using CC with singles and doubles as the fragment solver, and MP2-like treatments of the environment. The results are substantially improved by the inclusion of orbital relaxation in the environment. Using localized bonds as the active fragment, we also report results for N=N bond breaking in azomethane and for the central C-C bond torsion in butadiene. We find that when the fragment Hilbert space size remains fixed (e.g., when determined by an intrinsic atomic orbital approach), the method achieves comparable accuracy with both a small and a large basis set. Additionally, our results indicate that increasing the fragment Hilbert space size systematically enhances the accuracy of observables, approaching the precision of the full CC solver.

3D color-rendered MR neurography heatmaps in visualizing normal lumbosacral (LS) plexus and increasing conspicuity of LS plexopathy.

To determine whether color-rendered 3D MR neurography (MRN) images (heatmaps) improve diagnostic accuracy, reader confidence levels, and time savings to assess LS plexus lesions compared to the conventional grayscale images. A cross-sectional study included adults of all genders with randomly chosen MRNs of LS plexus and known reference standards of normal or neuropathy (plexopathy and radiculopathy). Heatmaps were constructed using 3D MRN STIR images and color rendered with higher intensity to yellow and lower intensity to darker-red colors in 1-2 min on average and were available on PACS for the readers. 2D plus 3D grayscale MIP images and 2D plus 3D MIP heatmaps were analyzed by four musculoskeletal radiologists (two faculty and two fellows) in two separate rounds blinded to the final diagnosis. Readers evaluated: neuropathy and number of nerves affected (neuropathy score: 0-normal; 1-one nerve affected; and 2-two or more nerves affected); final diagnosis; confidence levels; and time taken to evaluate the studies. Conger's kappa and paired t-test were used for analysis. Among 70 MRNs from 70 patients, there were 32 males and 38 females with average age ± SD of 54.8 ± 20.1 years and 49.9 ± 16.6 years, respectively. There were 30 normal and 40 LS plexus lesion scans. Interreader agreements for neuropathy scores were substantial to moderate on conventional imaging and heat maps (Conger's kappa: 0.65; 95% CI: 0.55, 0.73, and 0.59; 95% CI: 0.47, 0.69), respectively. The mean neuropathy score and final diagnosis accuracies were similar in both rounds 85.7% ± 0.1% vs 83.2% ± 0.1% (p = 0.13), and 83.6% ± 0.1% vs 80.0% ± 0.1%; p = 0.16), respectively. Time savings were significant when using heatmaps for all readers (p < 0.001). Time savings using heatmaps ranged from 57.7% to 74.6% and 56.3% to 75% of the original time for the fellows and faculty, respectively. Average confidence levels for neuropathy score significantly increased using heatmaps for one fellow and one faculty (p < 0.05), while average confidence levels for final diagnosis improved for both fellows and one faculty (p < 0.05). 3D color-rendered MRN heatmaps show comparable diagnostic accuracy to conventional MRN imaging but with significant time savings to identify LS plexus lesions. Question Do color-rendered 3D MRN images (heatmaps) improve accuracy, and confidence, and save time when assessing lumbosacral (LS) plexus lesions compared to conventional grayscale images? Findings 3D-rendered heatmaps showed comparable diagnostic accuracy with time savings ranging from 56.3% to 75%. Clinical relevance 3D color-rendered heatmaps increase time efficiency in evaluating MRNs of LS plexus, allowing for improved radiologist productivity and diagnostic confidence.

A Task-driven Adversarial Channel Selection Method for Binary Classification Based on Magnetocardiography.

As the number of sensors in magnetocardiography (MCG) arrays increases to capture detailed cardiac activity, some channels contribute minimally to task performance, resulting in data redundancy and resource consumption. Although existing methods can reduce the number of channels required to meet task demands, they often struggle to balance computational time and the accuracy of the selected channels and overlook the scalability of the selected channels. This limitation means that when environmental conditions change, or when sensors malfunction, redesigning channel configurations becomes necessary, which increases experimental uncertainties. This study introduces a task-driven adversarial channel selection method tailored for binary classification of MCG signals. The optimal channel combination is determined through a group-wise search using a heuristic algorithm, whose objective function is designed to maximize the difference between the classification accuracy and cosine similarity of the selected channel. In evaluations using an MCG dataset from Qilu Hospital of Shandong University, the proposed method successfully reduced the number of channels from 36 to 5 without compromising classification performance. Furthermore, it outperforms existing hybrid sequential forward search method by achieving comparable accuracy with fewer channels, while also demonstrating superior scalability compared to both hybrid sequential forward search and pearson-rank methods. This approach strikes a balance between computational consumption and accuracy, while improving the scalability of the selected channel combinations, enhancing the efficiency and practicality of the MCG system.

Hyperdimensional computing: a framework for stochastic computation and symbolic AI

Hyperdimensional Computing (HDC), also known as Vector Symbolic Architectures (VSA), is a neuro-inspired computing framework that exploits high-dimensional random vector spaces. HDC uses extremely parallelizable arithmetic to provide computational solutions that balance accuracy, efficiency and robustness. The majority of current HDC research focuses on the learning capabilities of these high-dimensional spaces. However, a tangential research direction investigates the properties of these high-dimensional spaces more generally as a probabilistic model for computation. In this manuscript, we provide an approachable, yet thorough, survey of the components of HDC. To highlight the dual use of HDC, we provide an in-depth analysis of two vastly different applications. The first uses HDC in a learning setting to classify graphs. Graphs are among the most important forms of information representation, and graph learning in IoT and sensor networks introduces challenges because of the limited compute capabilities. Compared to the state-of-the-art Graph Neural Networks, our proposed method achieves comparable accuracy, while training and inference times are on average 14.6× and 2.0× faster, respectively. Secondly, we analyse a dynamic hash table that uses a novel hypervector type called circular-hypervectors to map requests to a dynamic set of resources. The proposed hyperdimensional hashing method has the efficiency to be deployed in large systems. Moreover, our approach remains unaffected by a realistic level of memory errors which causes significant mismatches for existing methods.

Mitigating local minima in extracting optimal parameters for photovoltaic models: An optimizer leveraging multiple initial populations (OLMIP)

The accurate identification of parameters in photovoltaic models is of paramount importance to accurately predict the electrical behavior of photovoltaic systems and improve their performance. Nowadays, this task poses a multimodal optimization problem, aiming to find the best combination of parameters that fit the model to real data. This article presents a proposal for an Optimizer Leveraging Multiple Initial Populations (OLMIP) that aims to achieve optimal solutions while effectively avoiding undesirable local optima. By utilizing a separate evolution strategy involving four distinct initial populations, followed by the construction of an elite population, the algorithm can explore multiple regions of the search space and escape local minima. Experiments were conducted with four models: the single diode, double diode, triple diode from RTC France cell, and Photowatt-PWP201 module. The mean squared errors obtained are 9.860219E-04, 9.824849E-04, 9.824849E-04, and 2.425075 E−03, respectively. These results indicate that the algorithm achieves superior or comparable accuracy to that of six competitors. Furthermore, the statistical analysis of the results, including the Wilcoxon and Friedman tests, confirms the robustness and effectiveness of the approach used for parameter estimation in photovoltaic systems. These findings open up new prospects for the improvement and optimization of photovoltaic systems.

Development of an Artificial Intelligence System for Distinguishing Malignant from Benign Soft Tissue Tumors Using Contrast-Enhanced MR Images.

The integration of artificial intelligence (AI) into orthopedics has enhanced the diagnosis of various conditions; however, its use in diagnosing soft-tissue tumors remains limited owing to its complexity. This study aimed to develop and assess an AI-driven diagnostic support system for magnetic resonance imaging (MRI)-based soft-tissue tumor diagnosis, potentially improving accuracy and aiding radiologists and orthopedic surgeons. Experienced orthopedic oncologists and radiologists annotated 720 images from 77 cases (41 benign and 36 malignant soft-tissue tumors). Eleven tumor subtypes were identified and classified into benign and malignant groups based on histological diagnosis. Utilizing the standard machine learning classifier pipeline, we examined and down-selected imaging protocols and their predominant radiomic features within the tumor's three-dimensional region to differentiate between benign and malignant tumors. Among the scan protocols, contrast-enhanced T1 weighted fat-suppressed images showed the most accurate classification based on radiomics features. We focused on the two-dimensional features from the largest tumor boundary surface and its neighboring slices, leveraging texture-based radiomic and deep convolutional neural network features from a pretrained VGG19 model. The test data comprised 44 contrast-enhanced images (22 benign and 22 malignant soft-tissue tumors) containing six malignant and five benign subtypes distinct from the training data. We compared expert and non-expert human performances against AI by assessing malignancy detection and the time required for classification. The AI model showed comparable accuracy (AUC 0.91) to that of radiologists (AUC 0.83) and orthopedic surgeons (AUC 0.73). Notably, the AI model processed data approximately 400 times faster than its human counterparts, showcasing its capacity to significantly boost diagnostic efficiency. We developed an AI-driven diagnostic support system for MRI-based soft-tissue tumor diagnosis. While additional refinement is necessary for clinical applications, our system has exhibited promising potential in differentiating between benign and malignant soft-tissue tumors based on MRI.

Unified uniaxial compressive constitutive model for C20–C120 concrete based on stochastic damage theory

The concrete uniaxial compressive stress-strain model is essential for analyzing reinforced concrete members and structures. However, models applicable to C20–C120 concrete are still relatively rare. Moreover, the previous empirical model construction methods may not reflect the physical mechanisms and randomness of concrete failure well. To this end, the database was constructed by collecting and analyzing uniaxial compressive stress-strain curve test results from different scholars. The cubic compressive strength in the database ranges from 9.9 MPa to 137.3 MPa. The adaptability of the stochastic damage model was validated using the database of test curves. Subsequently, stochastic damage models for C20–C120 concrete were calibrated based on concrete compressive strength grouping curves, and practical analysis models were suggested. The results indicate that the calculated curves closely match the test curves when the plastic strain evolution parameters η1,c and η2,c in the stochastic models are taken as 0.001–0.30 and 0.19–0.25, respectively. The adaptability of the stochastic damage model is satisfactory, but the plastic strain evolution laws vary for different concrete compressive strengths, influenced by mix proportions and the probability of failure of each component on the meso-scale. When η1,c is calculated by the empirical formula proposed in this paper and η2,c is approximated as 0.22, the calculated curves for C20–C120 concrete are consistent with the mean values of the test curves. The proposed practical analysis models provide comparable accuracy to stochastic damage models, eliminating the need for iterative calculations, making them convenient for engineering applications.

Robotic spine systems: overcoming surgeon experience in pedicle screw accuracy: a prospective study.

Prospective single-center study. To compare the accuracy of pedicle screws placed by freehand and under fluoroscopy and robotic assistance with intraoperative image acquisition. Pedicle screws are the most commonly used spinal anchors owing to their ability to stabilize all three spinal columns. Various techniques such as freehand, fluoroscopy-assisted, and navigation-assisted pedicle screw placements have been used with varying degrees of accuracy. Most studies on robotic-assisted pedicle screw placement have utilized preoperatively acquired computed tomography scans. To our knowledge, this is the only study in the literature that compared freehand with fluoroscopy-guided and robotic-assisted pedicle screw insertion with freehand and fluoroscopy. In this prospective study, a total of 1,120 pedicle screws were placed in the freehand group (n=175), 1,250 in the fluoroscopyassisted group (n=172), and 1,225 in the robotic-assisted group (n=180). Surgical parameters and screw accuracy were analyzed between the three groups. The preoperative plan overlapped with the postoperative O-arm scan to determine if the screws were executed as planned. The frequency of clinically acceptable screw placement (Gertzbein-Robbins grades A and B) in the freehand, fluoroscopy-assisted, and robotic-assisted groups were 97.7%, 98.6%, and 99.34%, respectively. With robotic assistance, an experience-neutralizing effect implied that surgeons with varying levels of experience achieved comparable pedicle screw accuracy, blood loss, O-arm time, robot time, and time per screw. No significant difference in these parameters was found between surgeries commencing before and after 2 PM. No significant differences were noted between the planned and executed screw trajectories in the robotic-assisted group irrespective of surgical experience. The third-generation robotic-assisted pedicle screw placement system used in conjunction with intraoperative threedimensional O-arm imaging consistently demonstrates safe and accurate screw placement with an experience-neutralizing effect.

Variational Bayesian EM Algorithm for Quantile Regression in Linear Mixed Effects Models

This paper extends the normal-beta prime (NBP) prior to Bayesian quantile regression in linear mixed effects models and conducts Bayesian variable selection for the fixed effects of the model. The choice of hyperparameters in the NBP prior is crucial, and we employed the Variational Bayesian Expectation–Maximization (VBEM) for model estimation and variable selection. The Gibbs sampling algorithm is a commonly used Bayesian method, and it can also be combined with the EM algorithm, denoted as GBEM. The results from our simulation and real data analysis demonstrate that both the VBEM and GBEM algorithms provide robust estimates for the hyperparameters in the NBP prior, reflecting the sparsity level of the true model. The VBEM and GBEM algorithms exhibit comparable accuracy and can effectively select important explanatory variables. The VBEM algorithm stands out in terms of computational efficiency, significantly reducing the time and resource consumption in the Bayesian analysis of high-dimensional, longitudinal data.

Impact of PM2.5 filter extraction solvent on oxidative potential and chemical analysis

ABSTRACT Fine particulate matter (PM2.5) is hypothesized to induce oxidative stress, and has been linked to acute and chronic adverse health effects. To better understand the risks and underlying mechanisms following exposure, PM2.5 is collected onto filters but prior to toxicological analysis, particles must be removed from filters. There is no standard method for filter extraction, which creates the possibility that the methods of extraction selected can alter the chemical composition and ultimately the biological implications. In this study, comparisons were made between extraction solvents (methanol (MeOH), dichloromethane (DCM), 0.9% saline, and Milli-Q water) and the results of oxidative potential and elemental concentration analysis of PM2.5 collected across sites in Arkansas, USA. Significant differences were observed between solvents, with DCM having significantly different results compared to all other extraction solvents (p ≤ 0.001). Significant correlations between element, black carbon, and PM2.5 concentrations and oxidative potential were observed. The observed correlations were extraction solvent dependent. For example, in saline extracted samples, oxidative potential had significant negative correlations with: Ba, Cd, Ce, Co, Ga, Mn and significant positive correlations with: Cr, Ni, Th, U. While in MeOH extracted samples, significant positive correlations were only between oxidative potential and Ga, U and significant negative correlations with V. This indicates that PM2.5 samples extracted with different solvents will yield different conclusions about the causal components. This study highlights the importance of filter extraction methods in interpretation of oxidative potential results and comparisons between studies. Implications While there is no standard method for PM2.5 filter extraction, variation of extraction methods impact analytical results. This project identifies that extraction method variation, particularly extraction solvent selection, leads to discrepancies in chemical and toxicological analysis for PM2.5 collected on the same filter. This work highlights the need for methods standardization to support accurate comparisons between PM2.5 research studies, thus providing better understanding of PM2.5 across the globe.

The Role of Collaboration in Prenatal Congenital Heart Disease Diagnosis: A Comparison of Maternal-Fetal Medicine Specialist and Pediatric Cardiologist Performance.

This study compared the accuracy of prenatal congenital heart disease (CHD) diagnosed by maternal-fetal medicine specialists (MFMs) and pediatric cardiologists (PCs), using postnatal cardiac findings as the reference standard. This retrospective analysis at Siriraj Hospital, Bangkok, Thailand, involved 125 pregnancies with fetal CHD diagnosed by MFMs and evaluated by PCs later. Prenatal CHD diagnoses by either MFM or PC were compared with postnatal diagnoses obtained through echocardiography, cardiac surgery/catheterization, or autopsy. Diagnostic accuracy was classified as (A) correct diagnosis, (B) minor differences not impacting clinical management or outcomes, or (C) major differences affecting prognosis or treatment. Cardiac sonography by MFM achieved diagnostic accuracies of 73.6% (A), 16% (B), and 10.4% (C), while fetal echocardiography by PC resulted in accuracies of 72% (A), 20% (B), and 8% (C). No statistically significant differences were found between MFM and PC in each category (P = .375-.832). The MFMs' accuracy was highest for tetralogy of Fallot (94.4%; 95% CI, 72.7-99.9%) and lowest for right atrial isomerism (71.4%; 95% CI, 29-96.3%) and pulmonary atresia with ventricular septal defect (57.1%; 95% CI, 18.4-90.1%). MFMs and PCs demonstrated high and comparable accuracy in prenatal CHD diagnosis. Although PCs tended to outperform MFMs in cases where misdiagnosis could significantly impact neonatal care and outcomes, MFMs can effectively perform primary screening for fetal CHD in all pregnancies. Collaboration with PCs remains essential when fetal CHD is suspected, particularly in complex cases.

Introducing edge intelligence to smart meters via federated split learning

The ubiquitous smart meters are expected to be a central feature of future smart grids because they enable the collection of massive amounts of fine-grained consumption data to support demand-side flexibility. However, current smart meters are not smart enough. They can only perform basic data collection and communication functions and cannot carry out on-device intelligent data analytics due to hardware constraints in terms of memory, computation, and communication capacity. Moreover, privacy concerns have hindered the utilization of data from distributed smart meters. Here, we present an end-edge-cloud federated split learning framework to enable collaborative model training on resource-constrained smart meters with the assistance of edge and cloud servers in a resource-efficient and privacy-enhancing manner. The proposed method is validated on a hardware platform to conduct building and household load forecasting on smart meters that only have 192 KB of static random-access memory (SRAM). We show that the proposed method can reduce the memory footprint by 95.5%, the training time by 94.8%, and the communication burden by 50% under the distributed learning framework and can achieve comparable or superior forecasting accuracy to that of conventional methods trained on high-capacity servers.

Accuracy evaluation and comparison of GSMaP series for retrieving precipitation on the eastern edge of the Qinghai-Tibet Plateau

Study RegionMin River Basin on the eastern Qinghai-Tibet Plateau. Study FocusPrecipitation is critical for hydrological processes, making accurate data essential for water management and flood forecasting. Satellite precipitation products offer valuable high-resolution spatiotemporal information, with the GSMaP series being widely used. However, comprehensive evaluations of different versions are limited. This study assesses the accuracy of Gauge and MVK products (versions 06, 07, and 08) across spatial and temporal scales and evaluates their performance in detecting precipitation events of varying intensities. New Hydrological Insights for the Study Region(1) GSMaP versions 06 and 07 exhibit higher detection rates for precipitation events, with POD values exceeding 0.8, while version 08 has a lower false alarm rate, with FAR values below 0.15. (2) GSMaP products are more successful in capturing precipitation events during the rainy season than the dry season. (3) With increasing elevation, the Gauge product consistently maintains a high hit rate and reduced false alarm rate, whereas the MVK product's hit rates improve. (4) For different rainfall intensities, GSMaP products more accurately detect moderate and heavy rain events, with the Gauge product outperforming the MVK product in terms of accuracy. These insights enhance the understanding of GSMaP product performance on eastern edge of the Qinghai-Tibet Plateau, aiding in improved water management practices.

Modular YOLOv8 optimization for real-time UAV maritime rescue object detection.

The task of UAV-based maritime rescue object detection faces two significant challenges: accuracy and real-time performance. The YOLO series models, known for their streamlined and fast performance, offer promising solutions for this task. However, existing YOLO-based UAV maritime rescue object detection methods tend to prioritize high accuracy, often at the expense of real-time performance and ease of implementation and expansion. This study proposes a modular plug-and-play optimization approach based on the YOLOv8 framework, aiming to enhance real-time performance while maintaining high accuracy for UAV maritime rescue object detection. The proposed optimization modules are flexible, easy to implement, and extendable. In experiments on the large-scale publicly available SeaDronesSee dataset, our method achieved a 13.53% improvement in accuracy over YOLOv8x while reducing computational cost by 85.63%. Additionally, it surpassed the detection speed of the SeaDronesSee official code's two-stage detector by over 20 times, while maintaining comparable accuracy. Furthermore, our analysis of the experimental results highlights differences in detection difficulty among various objects and potential biases within the dataset.

Evaluation of machine learning models for the accelerated prediction of density functional theory calculated 19F chemical shifts based on local atomic environments

The introduction of fluorine in compounds plays a crucial role in drug development as it greatly influences their final pharmacokinetic and dynamic properties. Due to the prevalence of fluorine in FDA-approved drugs in recent years, identifying the mechanisms driving their chemical transformations has become crucial in the drug discovery landscape. 19F NMR spectroscopy is a powerful analytical technique that allows for the examination of fluorine-containing compounds, offering valuable information about their structure, dynamics, and reactivity. NMR spectra can be interpreted through the leveraging of Density Functional Theory (DFT). However, the screening of compounds and discovery of feasible drug candidates is limited due to its computational cost. Here, we present a machine learning approach to accelerate the prediction of DFT-calculated 19F NMR chemical shifts. The fluorine atoms’ features in the models were derived from their local three-dimensional environments, representing their neighboring atoms within a radius of n Å away from the given fluorine atom in the compound. A comparative analysis of thirteen regression models was conducted using features extracted from 501 fluorinated compounds in our laboratory’s chemical inventory. Among the models, Gradient Boosting Regression (GBR) exhibited the highest performance, achieving a mean absolute error of 3.31 ppm with a local environment radius of 3 Å. This demonstrates a comparable accuracy to DFT calculations while reducing computational time from several hundred seconds to milliseconds. 3 Å was also found to be the most optimal radius across all models when encoding features for local atomic environments.

Rapid detection of human adenovirus by multiple cross displacement amplification combined with nanoparticle-based biosensor platform

Human adenoviruses (HAdV), particularly serotypes 3 and 7 (HAdV-3 and HAdV-7), are significant respiratory pathogens that contribute to high morbidity rates and severe pneumonia in infants and children. The lack of distinct clinical presentations and effective treatments highlights the urgent need for rapid and reliable diagnostic methods for HAdV-3 and HAdV-7. This study devises a novel detection assay, termed HAdV-MCDA-LFB, which combines isothermal multiple cross displacement amplification (MCDA) with a nanoparticle-based lateral flow biosensor (LFB). Targeting the conserved hexon gene, HAdV-MCDA-LFB demonstrated high sensitivity, detecting low to 10 fg of hexon-containing plasmid per reaction without cross-reaction under the optimized conditions. Moreover, HAdV-MCDA-LFB exhibited comparable diagnostic accuracy to real-time PCR in clinical sample analysis, indicating its practical applicability. The whole procedure, including rapid template preparation (15 min), MCDA reaction (40 min at 67 °C) and result interpretation (<5 min), can be completed within one hour. Together, this rapid turnaround time, coupled with its simplicity and accuracy, makes HAdV-MCDA-LFB a promising point-of-care diagnostic tool for HAdV-3 and HAdV-7, particularly in resource-limited settings.

18F]FDG PET/CT versus [18F]FDG PET/MRI in the diagnosis of lymph node metastasis in nasopharyngeal carcinoma: a systematic review and meta-analysis.

This meta-analysis aimed to evaluate the comparative diagnostic accuracy of [18F]FDG PET/CT versus [18F]FDG PET/MRI in identifying lymph node metastases in individuals with nasopharyngeal carcinoma. A comprehensive search was executed across PubMed, Embase, and Web of Science through September 2023 to identify studies evaluating the diagnostic precision of [18F]FDG PET/CT and [18F]FDG PET/MRI in detecting lymph node metastasis in nasopharyngeal carcinoma. Sensitivity and specificity were assessed through the DerSimonian-Laird method, incorporating the Freeman-Tukey transformation. The meta-analysis encompassed nine articles, involving a total of 916 patients. The overall sensitivity and specificity of [18F]FDG PET were 0.95 (95%CI: 0.88-1.00) and 0.95 (95%CI: 0.84-1.00). The overall sensitivity of [18F]FDG PET/CT was 0.94 (95%CI, 0.85-0.99), whereas [18F]FDG PET/MRI achieved a sensitivity of 1.00 (95%CI, 0.94-1.00). The findings reveal that [18F]FDG PET/CT demonstrates comparable sensitivity to [18F]FDG PET/MRI (p = 0.20). The overall specificity of [18F]FDG PET/CT was 0.94 (95%CI, 0.82-1.00), whereas [18F]FDG PET/MRI exhibited a specificity of 0.98 (95%CI, 0.93-1.00). Additionally, the results suggest that [18F]FDG PET/CT offers similar specificity to [18F]FDG PET/MRI (p = 0.11). [18F]FDG PET demonstrates high sensitivity and specificity in identifying lymph node metastasis in nasopharyngeal carcinoma. Furthermore, [18F]FDG PET/CT exhibits comparable sensitivity and specificity to [18F]FDG PET/MRI. https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=496006, PROSPERO (CRD42024496006).