Bioengineering最新文献

Accurate interpretation of missense variants in cancer-associated genes remains a critical challenge in precision oncology, as most sequence-based predictors lack mechanistic explanations. Receptor tyrosine kinases like FGFR2 exemplify this problem: their function depends on conformational dynamics, yet most variants remain classified as variants of uncertain significance (VUS). In this paper we present DyVarMap, an interpretable structural-learning framework that integrates AlphaFold2-based ensemble generation with physics-driven refinement, manifold learning, and supervised classification using five biophysically motivated geometric features. Applied to FGFR2, the framework generates diverse conformational ensembles, identifies metastable states through nonlinear dimensionality reduction, and classifies pathogenicity while providing mechanistic attributions via SHAP analysis. External validation on ten kinase-domain variants yields an AUROC of 0.77 with superior calibration (Brier score = 0.108) compared to PolyPhen-2 (0.125) and AlphaMissense (0.132). Feature importance analysis consistently identifies K659-E565 salt-bridge distance and DFG motif dihedral angles as top predictors, directly linking predictions to known activation mechanisms. Case studies of borderline variants (A628T, E608K, L618F) demonstrate the framework's ability to provide structurally coherent mechanistic explanations. DyVarMap bridges the gap between static structure prediction and dynamics-aware functional assessment, generating testable hypotheses for experimental validation and demonstrating the value of incorporating conformational dynamics into variant effect prediction for precision oncology.

Objectives: This study presents the development of a bilingual, expert-evaluated question-answer (Q&A) dataset, named PMR-Q&A, designed for training large language models (LLMs) in the field of Physical Medicine and Rehabilitation (PMR). Methods: The dataset was created through a systematic and semi-automated framework that converts unstructured scientific texts into structured Q&A pairs. Source materials included eight core reference books, 2310 academic publications, and 323 theses covering 15 disease categories commonly encountered in PMR clinical practice. Texts were digitized using layout-aware optical character recognition (OCR), semantically segmented, and distilled through a two-pass LLM strategy employing GPT-4.1 and GPT-4.1-mini models. Results: The resulting dataset consists of 143,712 bilingual Q&A pairs, each annotated with metadata including disease category, reference source, and keywords. A representative subset of 3000 Q&A pairs was extracted for expert validation to evaluate the dataset's reliability and representativeness. Statistical analyses showed that the validation sample accurately reflected the thematic and linguistic structure of the full dataset, with an average score of 1.90. Conclusions: The PMR-Q&A dataset is a structured and expert-evaluated resource for developing and fine-tuning domain-specific large language models, supporting research and educational efforts in the field of physical medicine and rehabilitation.

Collateral status is an important therapeutic indicator for acute ischemic stroke (AIS), yet visual collateral grading remains subjective and suffers from inter-observer variability. To address this limitation, this study automatically extracted binarized vascular morphological features from CTA images and developed a convolutional neural network (CNN) for automated collateral classification. Performance trends were systematically analyzed across diverse hyperparameter combinations to meet different clinical decision needs. A total of 157 AIS patients (median age 65 [57-74] years; 61.8% were male) were retrospectively enrolled and stratified by Menon score into good (3-5, n = 117) and poor (0-2, n = 40) collateral groups. A total of 192 architectures were established, and three representative model tendencies emerged: a sensitivity-oriented model (AUC = 0.773; sensitivity = 87.18%; specificity = 65.00%), a balanced model (AUC = 0.768; sensitivity = 72.65%; specificity = 77.50%), and a specificity-oriented model (AUC = 0.753; sensitivity = 63.25%; specificity = 85.00%). These results demonstrate that kernel size, the number of filters in the first layer, and the number of convolutional layers are key determinants of performance directionality, allowing tailored model selection depending on clinical requirements. This work highlights the feasibility of CTA-based automated collateral classification and provides a systematic framework for developing models optimized for sensitivity, specificity, or balanced decision-making. The findings may serve as a reference for clinical model deployment and have potential for integration into multi-objective AI systems for endovascular thrombectomy patient triage.

Background and objectives: The increasing volume of total hip and knee arthroplasty created a significant postoperative surveillance burden. While plain radiographs are standard, the detection of aseptic loosening is subjective. This review evaluates the state of the art regarding AI in radiographic analysis for identifying aseptic loosening and mechanical failure in primary hip and knee prostheses. Methods: A systematic search in PubMed, Scopus, Web of Science, and Cochrane was conducted up to November 2025, following PRISMA guidelines. Peer-reviewed studies describing AI tools applied to radiographs for detecting aseptic loosening or implant failure were included. Studies focusing on infection or acute complications were excluded. Results: Ten studies published between 2020 and 2025 met the inclusion criteria. In internal testing, AI models demonstrated high diagnostic capability, with accuracies ranging from 83.9% to 97.5% and AUC values between 0.86 and 0.99. A performance drop was observed during external validation. Emerging trends include the integration of clinical variables and the use of sequential imaging. Conclusions: AI models show robust potential to match or outperform standard radiographic interpretation for detecting failure. Clinical deployment is limited by variable performance on external datasets. Future research must prioritize robust multi-institutional validation, explainability, and integration of longitudinal data.

Purpose: Orthognathic surgery is a cornerstone therapeutic approach for correcting dentofacial deformities; however, its Impact on neuromuscular adaptation remains incompletely understood, particularly regarding different surgical strategies. The aim of this study was to evaluate and compare neuromuscular changes in patients undergoing monomaxillary or bimaxillary orthognathic surgery.

Methods: Eighty adult patients treated with combined orthodontic-surgical therapy were included (37 monomaxillary; 43 bimaxillary). A control group of 20 healthy adult subjects with physiological occlusion and no history of orthodontic or orthognathic treatment was included. Surface electromyography (sEMG) of the masseter and anterior temporalis muscles and mandibular kinesiography were performed using standardized protocols at five treatment phases. Electromyographic symmetry indices (Percent Overlapping Coefficient-POC), muscle activity (µV), IMPACT values, and mandibular movement parameters were analyzed.

Results: During the presurgical orthodontic phase, both groups showed comparable reductions in neuromuscular activity. Postoperatively, monomaxillary patients exhibited earlier stabilization of sEMG symmetry and a faster increase in IMPACT values, approaching physiological reference ranges at the final follow-up. In contrast, bimaxillary patients showed greater variability and slower functional recovery. Mandibular opening and lateral movements improved in all patients, with more stable kinesiographic patterns observed in the monomaxillary group.

Conclusions: Within the limitations of this study, neuromuscular adaptation following orthodontic-surgical treatment appears to be associated with the surgical approach adopted, rather than representing a direct effect of surgical extent. These findings support the role of functional assessment as a complementary component in the management of orthognathic patients.

This article presents a systematic review on methods for quantifying three-dimensional, time-resolved (3D+t) deformation and motion of human arteries from Computed Tomography (CT) and Magnetic Resonance Imaging (MRI). Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we searched Scopus, Web of Science, IEEE Xplore, Google Scholar, and PubMed on 19 December 2025 for in vivo, patient-specific CT or MRI studies reporting motion or deformation of large human arteries. We included studies that quantified arterial deformation or motion tracking and excluded non-vascular tissues, in vitro or purely computational work. Thirty-five studies were included in the qualitative synthesis; most were small, single-centre observational cohorts. Articles were analysed qualitatively, and results were synthesised narratively. Across the 35 studies, the most common segmentation approaches are active contours and threshold, while temporal motion is tracked using either voxel registration or surface methods. These kinematic data are used to compute metrics such as circumferential and longitudinal strain, distensibility, and curvature. Several studies also employ inverse methods to estimate wall stiffness. The findings consistently show that arterial strain decreases with age (on the order of 20% per decade in some cases) and in the presence of disease, that stiffness correlates with geometric remodelling, and that deformation is spatially heterogeneous. However, insufficient data prevents meaningful comparison across methods.

Background: Multimodal AI integration across genomics, radiomics, and pathomics is rapidly evolving in oncology, but evidence remains heterogeneous and unevenly distributed across modalities.

Objective: To map empirical studies integrating two or more -omic modalities, summarize integration and validation approaches, and identify gaps informing future directions toward surgomics.

Methods: We conducted a scoping review in accordance with PRISMA-ScR, searching PubMed, Ovid, Wiley Online Library, and Google Scholar for English-language studies published from January 2020 to 5 March 2025. We charted study characteristics, modalities combined, fusion strategies, AI model categories, validation approaches, and reported performance metrics as presented by the original studies.

Results: From 184 records, 11 studies met inclusion criteria (n = 1078 total participants across reported studies), most focusing on radiomics-pathomics integration; fewer incorporated genomics, and tri-modal fusion was uncommon. Studies varied widely in clinical tasks, endpoints, preprocessing, and validation, limiting direct comparability.

Conclusions: The mapped evidence indicates growing methodological activity in radiopathomics and cross-scale association modeling, while tri-modal pipelines and clinically deployable multimodal workflows remain underdeveloped. Surgomics is presented as a conceptual, staged roadmap informed by these gaps rather than a current clinical capability.

Developmental Language Disorder (DLD) is associated with abnormalities in both intrinsic resting-state brain networks and task-evoked neural responses, yet direct electrophysiological evidence linking these levels remains limited. This study examined multi-level EEG markers in 21 typically developing children and 15 children with DLD across resting-state, a semantic matching task, and an auditory oddball task. Resting-state analyses revealed frequency-specific connectivity imbalances, reduced stability of intrinsic microstate dynamics, and atypical transitions between microstates in the DLD group. During the semantic matching task, DLD children showed weaker occipital P1 and N2 responses (100-300 ms) and lacked the right fronto-central difference wave (500-700 ms) observed in TD children. In the auditory oddball task, DLD children exhibited high-theta/low-alpha event-related desynchronization at left frontal electrodes (400-500 ms), in contrast to TD children. A machine learning framework integrating resting-state and task-based features discriminated DLD from TD children (test-set F1 = 70.3-80.0%) but showed limited generalizability, highlighting the constraints of small clinical samples. These findings support a translational neurophysiological signature for DLD, in which atypical intrinsic network organization constrains emergent neural computations, providing a foundation for future biomarker development and targeted intervention strategies.

Wearable sensors generate continuous physiological and behavioral data at a population scale, yet wellness prediction remains limited by noisy measurements, irregular sampling, and subjective outcomes. We introduce HybridSense, a unified framework that integrates raw wearable signals and their statistical descriptors with large language model-based reasoning to produce accurate and interpretable estimates of stress, fatigue, readiness, and sleep quality. Using the PMData dataset, minute-level heart rate and activity logs are transformed into daily statistical features, whose relevance is ranked using a Random Forest model. These features, together with short waveform segments, are embedded into structured prompts and evaluated across seven prompting strategies using three large language model families: OpenAI 4o-mini, Gemini 2.0 Flash, and DeepSeek Chat. Bootstrap analyses demonstrate robust, task-dependent performance. Zero-shot prompting performs best for fatigue and stress, while few-shot prompting improves sleep-quality estimation. HybridSense further enhances readiness prediction by combining high-level descriptors with waveform context, and self-consistency and tree-of-thought prompting stabilize predictions for highly variable targets. All evaluated models exhibit low inference cost and practical latency. These results suggest that prompt-driven large language model reasoning, when paired with interpretable signal features, offers a scalable and transparent approach to wellness prediction from consumer wearable data.

This study establishes a murine model of corticosteroid-associated osteonecrosis of the femoral head (ONFH) using a sustained-release prednisolone pellet and evaluates mitochondrial stress using ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (PET/CT) and changes in key histologic markers of bone over a 6-week period. Sixteen 12-week-old Balb/C mice were divided into two groups: a prednisolone group (PRED) and a control group (SHAM). The PRED group received a subcutaneous 60-day sustained-release pellet containing 2.5 mg of prednisolone, while the SHAM group received placebo pellets. PET/CT imaging was performed at 1, 3, and 6 weeks. Bone mineral density (BMD) measurements, and histomorphological analyses for the number of empty lacunae, osteoblasts, osteoclasts, and NADPH oxidase (NOX) 2, a marker for oxidative stress, were conducted at 4 or 6 weeks. PET/CT imaging demonstrated increased uptake in the femoral head at 3 weeks in the PRED group. This was accompanied by increased numbers of empty lacunae and osteoclasts, increased oxidative stress, and decreased alkaline phosphatase staining at 4 weeks in the PRED group. We have successfully established and validated a small murine model of ONFH. The findings of this preclinical study suggest a critical timeline for potential interventions to mitigate the early adverse effects of continuous corticosteroid exposure on bone.