P-AEB performance and limiting factors for superior-rated P-AEB systems based on simulations of real-world pedestrian crashes: A simulation study on the VIPA database.

Success and failure of human-AI collaboration in clinical reasoning: An experimental study on challenging real-world cases.

Enhancing object detection accuracy in visually challenging scenarios using super-resolution

Lung ultrasound is essential for rapid, radiation-free bedside pneumothorax diagnosis but limited by variability in human interpretation. Key gaps include insufficiently large and diverse human datasets, inconsistent image acquisition, lack of rigorous expert benchmarking, and inadequate clinical interpretability of existing artificial intelligence models. We aimed to develop and validate a robust, explainable artificial intelligence (AI) ensemble model addressing these critical gaps. With our multidisciplinary team, we developed an explainable soft-voting ensemble model trained on 1,856 diverse ultrasound clips from critically ill patients, healthy volunteers, and tailored cadaver models. Model interpretability was ensured using visualization, with heatmaps validated by expert clinicians. The model's diagnostic performance was rigorously benchmarked against 11 experienced clinicians using an independent, balanced test set. Statistical analyses included sensitivity, specificity and inter-rater reliability. The ensemble model achieved 100% sensitivity (95% CI: 85·8%-100·0%) and 100% specificity (95% CI: 85·8%-100·0%), surpassing expert sensitivity and specificity. Diagnostic performance of experts significantly differed by ultrasound mode, with notably lower specificity in M-mode imaging (p < 0·001). The AI consistently maintained perfect sensitivity and significantly reduced false positives compared to clinicians across all conditions, including challenging diagnostic scenarios (e.g., subtle pleural motions), and showed excellent generalizability to both cadaveric and clinical cases. Our explainable AI ensemble robustly matches the consensus-level performance of an expert "committee," significantly reducing diagnostic variability and false-positive diagnoses. This AI tool can serve as a critical second reader, standardize clinical decisions at the bedside, and substantially improve patient safety.

A Stepwise Multimodality Imaging Approach to Out-of-Hospital Cardiac Arrest in Cystic Fibrosis.

Underground coal mine images often suffer from severe blurring and low-resolution degradation due to harsh lighting, dust, and machinery motion, which hinder accurate visual inspection and automated analysis. This study proposes a transformer-based super-resolution (SR) network that integrates local convolution with adaptive interaction mechanisms for effective local-global feature modeling. The network employs a hierarchical architecture consisting of shallow feature extraction, cascaded spatial and channel transformer blocks, and a reconstruction module. Each transformer block incorporates a bidirectional adaptive interaction module (BAIM) to fuse convolutional local features with transformer-based global representations through adaptive reweighting in both spatial and channel dimensions. A dual-group feedforward network (DGFN) decouples channel feature preservation from spatial information enhancement, while cross-group interactions ensure balanced channel modeling and spatial perception without information loss. Additionally, a local convolution block (LCB) with SE-based channel weighting is used to restore fine-grained details. Extensive experiments on both a dedicated coal mine dataset and public benchmarks demonstrate that the proposed method consistently outperforms existing state-of-the-art (SOTA) SR approaches. Specifically, for ×2 super-resolution, it achieves a PSNR/SSIM of 32.07/0.9688 on the coal mine dataset, improving over the previous best by 0.59 dB and 0.0036, respectively. For ×4 super-resolution, it attains 28.10/0.8836, surpassing the previous best by 0.24 dB and 0.0013. Similar improvements are observed on public datasets, confirming the method's effectiveness in both general and challenging industrial scenarios.

Newly detected splenic lesions represent a challenging diagnostic scenario in internal medicine, particularly when lymphoproliferative disease is suspected. Evidence on the role of contrast-enhanced ultrasound (CEUS) in this setting is limited, and its integration into structured clinical pathways has been poorly explored. This study evaluated the impact of a structured clinical-ultrasound diagnostic pathway incorporating CEUS on diagnostic latency and use of advanced imaging, as well as its exploratory ability to discriminate hematologic/lymphoproliferative from non-hematologic splenic lesions. We conducted a multicenter, procedure-based quality Improvement study including 100 patients with newly detected splenic lesions. Fifty patients underwent a structured pathway based on an integrated clinical, laboratory, and ultrasound tool (intervention group), while 50 received standard management (control group). Primary outcomes were diagnostic latency and number of advanced imaging examinations (CT, MRI, PET). Secondary outcomes included the diagnostic performance of CEUS and of the integrated tool for discriminating hematologic/lymphoproliferative versus non-hematologic splenic lesions. Mean diagnostic latency was significantly shorter in the intervention group (13.8 ± 21.2 vs 32.4 ± 35.7 days; p = 0.004). Median latency was 5 [IQR 3-14] versus 14 [IQR 3-30] days (p = 0.002). Early diagnoses (≤ 7days) were more frequent in the intervention group (58% vs 22%), while delayed diagnoses (> 30days) were more common in controls (38% vs 10%; p < 0.001). Use of advanced imaging was reduced in the intervention group (CT: 46% vs 100%, p < 0.001; PET: 4% vs 20%, p = 0.018). For the discrimination of hematologic/lymphoproliferative versus non-hematologic lesions, CEUS showed high sensitivity (90.9%) and good specificity (79.5%), whereas the integrated tool improved specificity (94.9%) while maintaining good sensitivity (81.8%). A structured CEUS-centered diagnostic pathway appears to shorten diagnostic latency and reduce advanced imaging use while maintaining adequate diagnostic accuracy. The diagnostic performance findings should be interpreted as exploratory and limited to binary stratification of hematologic/lymphoproliferative versus non-hematologic lesions, rather than lesion-specific diagnosis.

Adaptive USV trajectory tracking: Preview-optimized rudder controller with self-tuning dynamics model

Incidental peritoneal metastases in the setting of emergency surgery: Review of evidence and algorithm for oncologic decision-making.

Efficacy and safety of electrochemotherapy in the treatment of cutaneous and sub-cutaneous recurrence from breast cancer: A single-center cohort study.

Conventional volumetric modulated arc therapy (VMAT) is limited by its longitudinal field size for large targets, often requiring multiple isocenters, while Helical Tomotherapy (HT) offers superior longitudinal conformity but suffers from prolonged treatment times due to its narrow fan beam. This simulation study proposes a novel spiral volumetric modulated arc therapy (SVMAT) technique designed to bridge this gap by synergizing continuous couch movement with dynamic MLC modulation. The SVMAT technique was implemented on a model of ring-gantry linac with dual-layer staggered MLC. Its core is a direct aperture optimization algorithm that discretizes the delivery path into finite projections, concurrently optimizing MLC aperture, monitor unit weight, gantry angle, and couch position. A comprehensive dosimetric and efficiency comparison was conducted against state-of-the-art VMAT and HT plans for three clinically challenging scenarios: hippocampal-sparing whole-brain radiotherapy (HS-WBRT), bilateral breast radiotherapy (BBRT), and craniospinal irradiation (CSI), including a pediatric subgroup. SVMAT demonstrated comparable or superior target coverage and conformity index to VMAT and HT across all cases. Its most significant advantage was in organ-at-risk (OAR) sparing. For HS-WBRT, SVMAT significantly reduced the maximum (Dmax) and mean (Dmean) doses to the hippocampus compared to both VMAT and HT (p<0.05). For BBRT, SVMAT notably reduced the heart Dmean (4.78±0.86Gy vs. 9.47±3.44Gy) for VMAT. In CSI, SVMAT reduced the lens Dmax by over 33% and the heart Dmean by 36.1%. Also, the pediatric CSI analysis confirmed these benefits, with SVMAT significantly reducing doses to developing organs. Regarding efficiency, SVMAT's beam on time was significantly shorter than HT's across all plans (reductions of 33.1% to 55.3%) and was comparable to the multi-isocenter VMAT in CSI. The SVMAT technique successfully integrates the dynamic delivery of VMAT with the longitudinal integration and single-isocenter capability of HT. By offering enhanced OAR sparing and reduced treatment times, SVMAT represents a significant advancement in radiotherapy, showing immense potential for improving outcomes, especially in vulnerable populations such as pediatric patients.

Combined surgery and proton radiotherapy in the management of craniopharyngiomas: an update with paradigmatic and challenging case scenarios.

Mycoplasma genitalium is an increasingly recognized sexually transmitted infection (STI) associated with urethritis, cervicitis, and pelvic inflammatory disease. Its management, though, is complicated by rapid expansion of antimicrobial resistance. Because M. genitalium lacks a cell wall, treatment options are limited to macrolides, tetracyclines, and fluoroquinolones. Widespread macrolide resistance, along with rising fluoroquinolone and dual-class resistance, has made refractory infection a common and challenging clinical scenario. Current guidelines recommend staged, resistance guided-therapy using a doxycycline lead-in followed by azithromycin for macrolide-susceptible infection or moxifloxacin for macrolide-resistant infection. However, access to macrolide resistance testing remains inconsistent; fluoroquinolone resistance testing is largely unavailable in routine practice, and correlations between resistance markers and clinical failure are incomplete. This review summarizes contemporary standard-of-care treatment strategies and emerging approaches for resistant or refractory M. genitalium infection. It also highlights critical gaps in diagnostic capacity, clinical trial data, and the antimicrobial development pipeline for STIs.

Oriented object detection (OOD) has rapidly advanced in recent years. However, the performance of existing methods is unsatisfactory when dealing with challenging scenarios, especially in scenes involving small-scale objects or objects with extreme aspect ratio. Inspired by recent advances in vision-language pre-training, we propose a novel Text-Guided Dual-Awareness Network (TG-DANet), which addresses these challenges from two complementary perspectives: robust feature interaction for multi-scale and long-range context modeling, and semantic-aware feature learning through textual guidance. Specifically, we design a Bi-Directional Feature Interaction Module (BDFIM) to capture horizontal and vertical contextual features via spatial interactions, which improves the representation of small and elongated objects. Additionally, a Text-Semantic Guidance Framework (TSGF) is supposed to align and fuse textual embeddings with visual features at multiple levels, which enhances model interpretability and discriminability for objects with ambiguous appearances or complex layouts. Extensive experiments on three benchmark datasets (DOTA, DIOR-R, and HRSC2016) show that TG-DANet achieves improvements of 3.05%, 3.49%, and 2.32% in mAP over baseline methods, respectively. These results demonstrate the effectiveness of our dual-perspective strategy in handling complex scenes with cluttered backgrounds and multi-scale objects, which highlights the promising potential of vision-language fusion in OOD.

This study presents an annotated multi-sensor, multimodal, and hyperspectral dataset designed to support deep learning-based classification and segmentation of bulky waste. The dataset comprises four distinct sensor modalities: high-resolution visible RGB images (VIS), hyperspectral near-infrared (NIR), temporally resolved thermal infrared (IR), and terahertz (THz) imaging with depth information, providing complementary multimodal information. An image registration process aligns all modalities to a common reference frame, enabling near pixel-precise fusion across sensors. WoodVIT contains 56 registered multi-sensor scenes, partitioned into 22,659 annotated patches with two main classes (wood and non-wood) and 16 subclass labels. It includes pixel-masks and patch-wise annotations to facilitate both segmentation and classification tasks. The primary benchmark task is binary discrimination of wood versus non-wood. The dataset also includes challenging scenarios involving occlusion and concealed contaminants (e.g., embedded metals) to motivate robust multimodal fusion approaches. We provide predefined train/validation/test splits and report baseline results using convolutional neural networks and fusion architectures to establish reference performance. WoodVIT is publicly available to support research on multi-sensor learning for waste sorting.

This paper introduces a novel channel estimation method for Orthogonal Time Frequency Space (OTFS) systems affected by nonlinear distortion from High-Power Amplifiers (HPA). The method integrates a Bidirectional Gated Recurrent Unit (Bi-GRU) with a dynamic gating mechanism driven by the Input Back-Off (IBO) parameter of the HPA, combined with a multi-head attention network. The dynamic gating mechanism adaptively adjusts the update gate of the Gated Recurrent Unit (GRU) based on real-time IBO values, optimizing the trade-off between historical memory and current input during training. The multi-head attention module further captures long-range dependencies in the channel response. Theoretical analysis indicates that the proposed IBO-driven dynamically gated Bi-GRU achieves a computational complexity reduction of 20–46.7% compared to a Bi-GRU architecture. Simulation results demonstrate the superior performance of the proposed method across both bit error rate (BER) and normalized mean square error (NMSE) metrics under high mobility and nonlinear distortion. It achieves up to 22.6 quantified in decibels (dB) lower NMSE and, at a signal-to-noise ratio (SNR) of 30 dB, a 15.2 dB reduction in logarithmic BER compared to conventional methods, along with a 3–4 dB improvement over deep learning baselines at the same SNR. It also provides over 7 dB peak-to-average power ratio (PAPR) reduction over traditional methods, confirming strong robustness and accuracy in challenging communication scenarios.

Robust plant disease detection in real-world agricultural environments remains challenging due to dynamic environmental conditions. Accurate and reliable disease identification is essential for precision agriculture and effective crop management. Although computer vision and Artificial Intelligence (AI) have shown promising results in controlled settings, their performance often drops under lesion scale variability, inter- and intra-class similarity among diseases, class imbalance, and illumination fluctuations. To overcome these challenges, we propose a Heterogeneous Feature Aggregation Network (HFA-Net) that brings together architectural improvements, illumination-aware preprocessing, and training-level enhancements into a single cohesive framework. To extract richer and more discriminative features from the early layers of the network, HFA-Net introduces a multi-scale, multi-level feature aggregation stem. The Reduction-Expansion (RE) mechanism helps preserve important lesion details while adapting to variations in scale. Considering real agricultural environments, an Illumination-Adaptive Contrast Enhancement (IACE) preprocessing pipeline is designed to address illumination variability in real agricultural environments. Experimental results show that HFA-Net achieves 96.03% accuracy under normal conditions and maintains strong performance under challenging lighting scenarios, achieving 92.95% and 93.07% accuracy in extremely dark and bright environments, respectively. Furthermore, quantitative explainability analysis using perturbation-based metrics demonstrates that the model's predictions are not only accurate but also faithful to disease-relevant regions. Finally, Grad-CAM-based visual explanations confirm that the model's predictions are driven by disease-specific regions, enhancing interpretability and practical reliability.

Fever or inflammation of unknown origin (FUO/IUO), bacteremia, and sepsis represent challenging diagnostic scenarios in clinical practice. Over the past decade, multiple systematic reviews and meta-analyses have reported quantitative data regarding the diagnostic accuracy, clinical impact, and cost-effectiveness of fluorine-18 fluorodeoxyglucose positron emission tomography combined with computed tomography ([18F]FDG PET/CT) in these complex infectious and inflammatory conditions. We performed an umbrella review of published systematic reviews and meta-analyses to provide a comprehensive evidence-based summary of the diagnostic performance, clinical utility, and cost-effectiveness of [18F]FDG PET/CT across different patient populations and clinical scenarios. A comprehensive literature search of PubMed/MEDLINE and Cochrane Library databases was conducted to identify published evidence-based articles evaluating [18F]FDG PET/CT in adult FUO/IUO, pediatric FUO, bacteremia, sepsis, febrile neutropenia, and chronic Q fever. Quality assessment of included reviews was performed using the critical appraisal framework of the Oxford Centre for Evidence-Based Medicine (OCEBM). A total of 14 published systematic reviews met the inclusion criteria. [18F]FDG PET/CT demonstrated high diagnostic sensitivity (81-94%) in adult FUO/IUO with diagnostic yield ranging from 56-75%. In pediatric FUO, sensitivity was 83% with specificity of 77.6% and area under the curve of 0.83. In bacteremia and sepsis, particularly in critically ill patients, [18F]FDG PET/CT achieved sensitivity of 94%. The imaging modality showed diagnostic utility in specialized populations including chronic Q fever (focus detection rate 79.5%), febrile neutropenia (83% added clinical value), and ICU patients (mean sensitivity 94.6%). Management modifications were observed in 20-48% of patients based on [18F]FDG PET/CT findings. Cost-effectiveness analysis supports selective use in high-risk patient populations, with numbers-needed-to-scan ranging from 3-6 in complex bacteremia. Evidence from systematic reviews demonstrates that [18F]FDG PET/CT is a valuable diagnostic tool in FUO/IUO, bacteremia, sepsis, and related infectious/inflammatory conditions, with clinically significant impact on patient management. Strategic patient selection based on clinical risk factors optimizes cost-effectiveness. Future standardization of imaging protocols and prospective comparative studies are needed to define optimal clinical indications across different populations.

Abstract Student injury incidents serve as institutional nexuses where health emergencies intersect with peer conflicts, requiring teachers to negotiate the competing demands of medical immediacy and moral culpability. Drawing on conversation analysis of audio-recorded student incident calls between teachers and parents, this study investigates how teachers navigate these challenging scenarios through the systematic management of agency and urgency . Analysis reveals two distinct yet interrelated patterns in teachers’ reporting practices. First, teachers treat agency (injurer) and urgency (injured student) as discrete components linked to responsibility attribution; they systematically background both elements when reporting to injured students’ parents while foregrounding them in communications with injurers’ parents. Second, the initial attenuation of urgency, while serving to mitigate conflict, necessitates subsequent upgrades to secure immediate involvement from injured students’ parents. These findings illuminate how teachers’ institutional practices systematically prioritize conflict mediation over medical urgency.

This study explores the use of a supervised Retrieval-Augmented Generation (RAG) pipeline for automatic misinformation classification in Brazilian Portuguese. The method combines semantic retrieval based on dense embeddings with traditional machine learning classification using TF–IDF and Support Vector Machine. Experiments were conducted on the largest publicly available Brazilian Portuguese fake news dataset, previously consolidated by the authors from multiple sources, and performance was compared with results from the literature using classical methods such as SVM. The results indicate that while supervised RAG provides competitive performance, its gains over traditional approaches may be limited when the dataset is balanced and linguistically homogeneous. A detailed error analysis is presented, and the potential of supervised RAG in more challenging and low-resource scenarios is discussed.