Fracture injuries often lead to complex bone fragmentations, posing significant challenges for accurate segmentation in surgical planning and trauma assessment. Manual annotation of each fragment is time-consuming and inconsistent, while existing automated methods often fail to separate individual fragments due to the wide variation in fracture types, irregular fracture surface, and close inter-fragment contact. To address these challenges, we introduce FracSegmentator, a deep learning approach for bone fragment instance segmentation. The model takes extracted bone regions in CT as input and isolates individual fragments by identifying fracture surfaces and separating closely contacting structures. Central to our approach is a Trauma-Prior-Guided Contrastive Learning module, which incorporates clinical knowledge through memory-based attention to better distinguish fractured surfaces from healthy regions. We evaluate FracSegmentator on four datasets that cover a range of anatomical sites and fracture patterns. The method achieves state-of-the-art results across all datasets and demonstrates strong generalization capabilities. By delivering accurate and efficient fragment-level segmentation, FracSegmentator supports critical downstream tasks such as automated fracture diagnosis, surgical planning, and preoperative reduction simulation.

Lunar polar resource studies necessitate probes capable of large-scale exploration on complex terrain. Aiming at creating a walking-leaping multimode probe for the lunar surface, this paper proposes a probe leg configuration synthesis method that combines the atlas and screw constraint methods. The reciprocity between wrench and twist is used to establish design constraints for the probe leg's landing cushion requirement, and freedom constraints are derived from the walking function requirement. Using the screw constraint method, three leg configurations, namely, series, parallel, and hybrid, were synthesized to obtain candidate solutions that met the design principles. The degree-of-freedom distribution of these solutions was visualized using the atlas method. Compared with existing wheeled mobile detectors, walking-leaping detectors require higher repeated buffering capabilities in the legs. This paper proposes a method for analyzing detector buffering capacity based on the Lie group distance and based on this, analyzes the buffering capacity of the legs. The proposed design and analysis methods for the multimode mobile leg configuration are universal, enabling leg configuration synthesis for diverse lunar probes and other multimode mobile mechanisms operating on complex terrains.

The segmentation of pelvic fracture fragments in CT and X-ray images is crucial for trauma diagnosis, surgical planning, and intraoperative guidance. However, accurately and efficiently delineating the bone fragments remains a significant challenge due to complex anatomy and imaging limitations. The PENGWIN challenge, organized as a MICCAI 2024 satellite event, aimed to advance automated fracture segmentation by benchmarking state-of-the-art algorithms on these complex tasks. A diverse dataset of 150 CT scans was collected from multiple clinical centers, and a large set of simulated X-ray images was generated using the DeepDRR method. Final submissions from 16 teams worldwide were evaluated under a rigorous multi-metric testing scheme. The top-performing CT algorithm achieved an average fragment-wise intersection over union (IoU) of 0.930, demonstrating satisfactory accuracy. However, in the X-ray task, the best algorithm achieved an IoU of 0.774, which is promising but not yet sufficient for intra-operative decision-making, reflecting the inherent challenges of fragment overlap in projection imaging. Beyond the quantitative evaluation, the challenge revealed methodological diversity in algorithm design. Variations in instance representation, such as primary-secondary classification versus boundary-core separation, led to differing segmentation strategies. Despite promising results, the challenge also exposed inherent uncertainties in fragment definition, particularly in cases of incomplete fractures. These findings suggest that interactive segmentation approaches, integrating human decision-making with task-relevant information, may be essential for improving model reliability and clinical applicability.

BACKGROUND:Displaced pelvic fractures present real surgical challenges because of complex three-dimensional deformity patterns and proximity to vital structures, with conventional manual reduction techniques limited by accuracy constraints and radiation exposure. Although robotic assistance shows promise in preclinical studies, its clinical effectiveness remains unproven in randomized clinical trials (RCTs). QUESTIONS/PURPOSES:(1) Does robotic closed reduction improve reduction quality compared with manual closed reduction in displaced pelvic fractures? (2) Can robotic closed reduction reduce intraoperative radiation exposure while maintaining functional outcomes? METHODS:In this multicenter RCT conducted at six tertiary trauma centers in China involving 10 senior orthopaedic traumatologists, 92 adult patients with acute closed, displaced pelvic fractures (Tile Type B or C) were randomized 1:1 to robotic closed reduction (n = 46) or manual closed reduction (n = 46) groups. At 12 weeks, loss to follow-up for patient-reported outcomes was 9% (4 of 46) in the robotic group and 4% (2 of 46) in the manual group; the remainder were handled in a prespecified per-protocol analysis. In the robot group, reduction was planned using CT-based three-dimensional reconstruction with contralateral pelvic symmetry as the target and executed by a robotic arm with adjunct elastic traction and contralateral pelvic stabilization. In the manual group, reduction was performed using traction and manual manipulation under fluoroscopic guidance. Surgeons and patients were not blinded; radiographic outcome assessors and data analysts were blinded. Primary outcome was reduction quality assessed using Matta criteria (excellent ≤ 4 mm residual displacement, good 5 to 10 mm, acceptable 10 to 20 mm, poor > 20 mm), analyzed as the proportion of excellent to good reductions. Secondary outcomes included intraoperative surgeon fluoroscopic exposure and 12-week Majeed pelvic scores (0 to 100 points across seven domains; higher scores indicate better function). The primary analysis was intention to treat. RESULTS:In the intention-to-treat analysis, a higher proportion of patients who underwent robotic closed reduction achieved an excellent or good reduction than did those who received manual closed reduction (96% [44 of 46] versus 48% [22 of 46], relative risk 2.00 [95% confidence interval (CI) 1.47 to 2.72]; p < 0.001). Median (IQR) intraoperative surgeon fluoroscopic exposure was lower in the robotic closed reduction group (0 [0 to 0] versus 38 [14 to 78] fluoroscopic exposures; p < 0.001). No differences were found in 12-week Majeed functional scores between groups (mean ± SD 69 ± 16 versus 71 ± 17, mean difference -3 [95% CI -11 to 6]; p = 0.55). One superficial infection occurred in the manual closed reduction group, and there were no serious complications in either group. CONCLUSION:Surgeons treating acute displaced pelvic ring fractures should consider robotic closed reduction, when available, to improve reduction quality and reduce intraoperative fluoroscopic exposure, although it did not result in improved patient-reported outcome scores at short term in this randomized trial. Future studies should evaluate longer term functional benefits, define the fracture patterns most likely to benefit, and evaluate implementation factors including learning curve and cost-effectiveness across varied trauma settings. LEVEL OF EVIDENCE:Level I, therapeutic study.

BACKGROUND:Osteoporotic and fragility fractures impose a significant global health burden, especially with the aging population. Despite advancements in imaging, risk assessment, and surgical techniques, underdiagnosis and undertreatment persists. Artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL), offers promise for enhancing fracture risk prediction, imaging-based diagnosis, clinical decision support, and postoperative outcome monitoring. OBJECTIVE:To systematically review AI applications in the evaluation and treatment of osteoporotic and fragility fractures, summarizing performance, limitations, evidence gaps, and future directions for clinical translation. METHODS:A PRISMA-compliant search was conducted in PubMed, IEEE Xplore, Google Scholar, and Web of Science from inception to October 2025. Inclusion criteria targeted original English-language studies in adults using AI/ML/DL for fracture risk prediction, diagnosis, treatment planning, intraoperative guidance, or postoperative management. Data extraction focused on study design, population, AI methods, performance metrics, and validation. RESULTS:Of 1286 records, 21 studies were included, clustering into three domains: fracture risk prediction (n = 10) using clinical, biochemical, and imaging data, often outperforming tools like FRAX; bone mineral density (BMD) estimation and osteoporosis screening from CT, X-ray, or opportunistic imaging (n = 6); and prognosis/postoperative outcomes (n = 5). CONCLUSIONS:AI demonstrates robust performance in BMD estimation, fracture detection, and risk prediction, frequently surpassing traditional methods. However, methodological heterogeneity, bias risks, and limited prospective/multicenter validation hinder translation. Future efforts should prioritize transparent reporting, external validation, regulatory compliance, and user-centered integration.

Introduction: Robot ‑assisted surgery is becoming increasingly popular and its application is expanding to various spinal surgical procedures, including endoscopic spinal surgery. Aim: The aim of this study was to describe a novel small parallel orthopedic surgical robot and evaluate its feasibility in assisting surgeons during percutaneous lumbar laminectomy on cadaveric specimens. Materials and methods: The authors of the study developed a new orthopedic surgical navigation system (R ‑Pharos, Rossum Robot Co., Ltd, Beijing, China), consisting of a navigation cart and a hybrid serial ‑parallel bedside robotic arm. The system is equipped with interactive software for selecting and planning the percutaneous lumbar laminectomy target and path. A cadaveric specimen was selected for a right ‑side partial laminectomy at L4. During the procedure, the surgeon used the robotic arm to guide the saw to the target lamina and perform the percutaneous resection. Postoperative cone beam computed tomography (CBCT) and endoscopic assessments were used to confirm the resection outcome. Results: After optimizing the precision of the small parallel orthopedic surgical robot to 1 mm, it was shown to meet the navigational requirements for percutaneous lumbar laminectomy. The surgeon utilized the interactive software to design the resection range and path for the right L4 lamina which was suc cessfully resected, as confirmed by endoscopic observation. A postoperative CBCT scan revealed that the resection area precisely matched the preoperative design. Conclusions: This study demonstrated that the small parallel orthopedic surgical robot was capable of preoperatively planning the lamina resection area and could assist the surgeon in performing percutane ous lumbar laminectomy with high navigational precision.

IntroductionAccurate segmentation of pelvic fractures from computed tomography (CT) is crucial for trauma diagnosis and image-guided reduction surgery. The traditional manual slice-by-slice segmentation by surgeons is time-consuming, experience-dependent, and error-prone. The complex anatomy of the pelvic bone, the diversity of fracture types, and the variability in fracture surface appearances pose significant challenges to automated solutions.MethodsWe propose an automatic pelvic fracture segmentation method based on deep learning, which effectively isolates hipbone and sacrum fragments from fractured pelvic CT. The method employs two sequential networks: an anatomical segmentation network for extracting hipbones and sacrum from CT images, followed by a fracture segmentation network that isolates the main and minor fragments within each bone region. We propose a distance-weighted loss to guide the fracture segmentation network's attention on the fracture surface. Additionally, multi-scale deep supervision and smooth transition strategies are incorporated to enhance overall performance.ResultsTested on a curated dataset of 150 CTs, which we have made publicly available, our method achieves an average Dice coefficient of 0.986 and an average symmetric surface distance of 0.234 mm.DiscussionThe method outperformed traditional max-flow and a transformer-based method, demonstrating its effectiveness in handling complex fracture.

The design of continuum robots often involves a dilemma between flexibility and stiffness, where increased flexibility may reduce stiffness and control precision. The human hand achieves both power grasp and precision grasp by leveraging different joint structures, particularly in the thumb, which plays a key role in balancing dexterity and stability. Inspired by the three distinct joints of the human thumb, we designed three types continuum manipulators featuring uniaxial, ball-and-socket, and saddle joints (SJ). A templated surface design was employed to control all other variables, ensuring that the only difference among the joint contact surfaces was their Gaussian curvature. The analysis covers aspects such as kinematic modeling, finite element simulations, workspace measurement, and stiffness experiments. Experimental results show that the workspace of the SJ manipulator is 0.73 times that of the ball-and-socket joint (BSJ) and 1.69 times that of the uniaxial joint (UJ). In terms of stability performance, the SJ achieves a maximum increase of 5.51 times in torsional stiffness and 2.68 times in bending stiffness compared to the BSJ. Compared to the UJ, the maximum improvements are 3.73 times in torsional stiffness and 2.44 times in bending stiffness. This suggests that the SJ continuum structure design can enhance stiffness while maintaining flexibility. This work provides a new approach for achieving a balanced flexibility and stability in continuum robot design.

Pelvic fractures are among the most complex challenges in orthopedic trauma, which usually involve hipbone and sacrum fractures, as well as joint dislocations. Traditional preoperative surgical planning relies on the operator’s subjective interpretation of CT images, which is both time-consuming and prone to inaccuracies. This study introduces an automated preoperative planning solution for pelvic fracture reduction, addressing the limitations of conventional methods. The proposed solution includes a novel multi-scale distance-weighted neural network for segmenting pelvic fracture fragments from CT scans, and a learning-based approach to restore pelvic structure, combining a morphable model-based method for single-bone fracture reduction and a recursive pose estimation module for joint dislocation reduction. Comprehensive experiments on a clinical dataset of 30 fracture cases demonstrated the efficacy of our methods. Our segmentation network outperformed traditional max-flow segmentation and networks without distance weighting, achieving a Dice similarity coefficient (DSC) of 0.986 ± 0.055 and a local DSC of 0.940 ± 0.056 around the fracture sites. The proposed reduction method surpassed mirroring and mean template techniques, and an optimization-based joint matching method, achieving a target reduction error of (3.265 ± 1.485) mm, rotation errors of (3.476 ± 1.995)°, and translation errors of (2.773 ± 1.390) mm. In the proof-of-concept cadaver studies, our method achieved a DSC of 0.988 in segmentation and 3.731 mm error in reduction planning, which senior experts deemed excellent. In conclusion, our automated approach significantly improves traditional preoperative planning, enhancing both efficiency and accuracy in pelvic fracture reduction.

One of the challenges in applying soft robots in real-world environments is their ability to perceive both their shape and external forces. In this paper, we propose a method that combines an embedded pressure sensor and deep learning to achieve three-dimensional perception of both self-shape and external forces acting on tendon-driven soft robots. By combining the actuator state as input, the system can accurately estimate the fingertip position during movement. Furthermore, the proposed sensing system effectively detects the magnitude of the change in force at five selected contact points outside the finger.

To evaluate the accuracy and reliability of a novel automated 3D CT-based method for measuring femoral neck anteversion (FNA) compared to three traditional manual methods. A total of 126 femurs from 63 full-length CT scans (35 men and 28 women; average age: 52.0 ± 14.7 years) were analyzed. The automated method used a deep learning network for femur segmentation, landmark identification, and anteversion calculation, with results generated based on two axes: Auto_GT (using the greater trochanter-to-intercondylar notch center axis) and Auto_P (using the piriformis fossa-to-intercondylar notch center axis). These results were validated through manual landmark annotation. The same dataset was assessed using three conventional manual methods: Murphy, Reikeras, and Lee methods. Intra- and inter-observer reliability were assessed using intraclass correlation coefficients (ICCs), and pairwise comparisons analyzed correlations and differences between methods. The automated methods produced consistent FNA measurements (Auto_GT: 17.59 ± 9.16° vs. Auto_P: 17.37 ± 9.17° on the right; 15.08 ± 9.88° vs. 14.84 ± 9.90° on the left). Intra-observer ICCs ranged from 0.864 to 0.961, and inter-observer ICCs between Auto_GT and the manual methods were high, except for the Lee method. No significant differences were observed between the two automated methods or between the automated and manual verification methods. Moreover, strong correlations (R > 0.9, p < 0.001) were found between Auto_GT and the manual methods. The novel automated 3D CT-based method demonstrates strong reproducibility and reliability for measuring femoral neck anteversion, with performance comparable to traditional manual techniques. These results indicate its potential utility for preoperative planning, postoperative evaluation, and computer-assisted orthopedic procedures. Not applicable.

Kinematic modeling of soft continuum robots in constrained environments remains challenging due to their complex nonlinear dynamics. Data-driven processing of proprioceptive signals offers a promising pathway to enhance robot perception of both its own state and the external environment. This paper proposes a soft continuum robot structure with multiple integrated proprioceptive sensing, including measurements of tendon tension and intersegmental pressure. A long short-term memory (LSTM) network is employed to jointly estimate multiple perception tasks, including tip position, orientation, and the magnitude and direction of external forces. Ablation studies demonstrate that fused proprioceptive inputs yield significantly higher accuracy in multi-task estimation than single-modality inputs. The proposed method provides a novel and effective approach for advancing perception and control in soft continuum robotics.

Accurate orthopedic fracture reduction planning is essential for ensuring successful postoperative recovery and improving patient outcomes. However, current automatic methods are challenged by the complex and irregular fracture geometries and the scarcity of annotated training data. To address these challenges, we propose a novel approach that integrates learning-based shape restoration and fracture simulation. A transformer-based model is developed, which utilizes patch-to-patch shape translation and recursive fragment registration to iteratively refine fracture reduction poses. A deformable fracture generation model (DFGM) combines statistical shape modeling with clinically representative fracture patterns to generate diverse and realistic datasets, reducing the dependence on annotated samples. Tested on extensive clinical data with hipbone, sacrum, and femoral shaft fractures, the proposed method achieved mean errors of 1.85 mm and 3.40°, outperforming both template-based and existing learning-based methods. In addition, models trained solely on DFGM-synthesized data presented strong generalizability to real clinical data. The ablation experiments demonstrate the effectiveness of the fragment-aware network pipeline and the synthesis steps. Finally, a cadaver study with ground truth derived from the pre-injury scan further validated the performance of the method.

MR merges virtual and physical worlds, with MLLM enhancing contextual understanding and multisensory perception for smarter interactions. This study explores user behaviors and task needs through contextual inquiry and retrospective interviews (N=12), identifying key daily tasks and information requirements. A participatory design workshop (N=18) further refines interface displays and interactions based on the physical and semantic attributes of objects, resulting in a dynamic design space. By leveraging MLLM, this design space adapts to user needs, paving the way for improving intelligent perception and redefining the role of everyday objects in MR.

Background and objectives Computer-assisted orthopedic surgical techniques and robotics has improved the therapeutic outcome of pelvic fracture reduction surgery. The preoperative reduction path is one of the prerequisites for robotic movement and an essential reference for manual operation. As the largest irregular bone with complicated morphology, the rotational motion of pelvic fracture fragments impacts the reduction process directly. To address this, the primary objective of this study is to develop an efficient and effective algorithm for automatically planning the reduction trajectory in robot-assisted pelvic fracture surgeries. Methods After obtaining rotational and reorientated translational degrees of freedom through the initial and target positions of the fracture fragments, the initial path is acquired through improved path planning method combined with specific designed collision detection algorithm. The final reduction path is post-processed to be shortened and smoothed. The effectiveness of the algorithm was evaluated in various pelvic fracture models with surrounding muscles and was compared with prior relevant implementations. Results Simulation results showed the ability of the planner to save time and overcome the state of art in terms of collision detection, path length and smoothness, search time, and surrounding muscle stretching conditions. Conclusions The proposed method enables a reasonable reduction path for pelvic fracture, which is demonstrated to be superior in various pelvic fracture scenarios.

Pelvic fracture surgery is a highly complex and skill-dependent procedure due to adjacent neurovascular structures. These challenges often result in prolonged operative time, elevated complication rates, and repeated reduction attempts. To support surgeons in overcoming these difficulties, our study introduces a rapid modeling approach that enables precise preoperative planning and quantitative evaluation of reduction strategies. We developed an automated patient-specific modeling method that integrates Statistical Shape Models with the personalized modeling modules of the OpenSim software’s Application Programming Interface for generating personalized musculoskeletal models. Using this approach, we rapidly reconstructed reduction models for 10 patients (age range: 49-72) with pelvic fractures and validated the results against clinical reduction force data. Here we show that the SMAG framework generates patient-specific models 78

In this paper, a disturbance observer-based command filtered backstepping control scheme is proposed to handle parametric uncertainties of hypersonic flight vehicles. The dynamics of an electromechanical actuator with parametric uncertainties are investigated effectively. Then, the lumped disturbances caused by parametric uncertainties and external disturbance torque are estimated by fixed-time observers. Fixed-time filters and compensation signals are constructed to handle the explosion of terms problem and ensure that the filter errors converge to zeros. The tracking error converges to zero in fixed time. Finally, the effectiveness of the proposed control scheme is illustrated through simulation results.

Magnetic soft robots offer significant potential for biomedical applications, as recent advances in magnetic programming and 3D printing technologies have enabled the fabrication of increasingly complex and multifunctional structures. However, current magnetic programming soft robots face limitations in injection and long-range manipulation. Here, we present a magnetic programming-assisted ultrasonic printing (MPUP) technique - a noncontact approach for the in situ fabrication of magnetically programmable soft robots within enclosed or hard-to-access spaces. We developed a magneto-responsive ultrasonic ink that can be selectively and rapidly cured by focused ultrasound. During printing, an externally applied magnetic field induces directional alignment of magnetic particles within the ink, enabling programmable magnetization profiles in the printed structures. The resulting soft robots exhibit diverse shape-morphing behaviors and programmable responses under external magnetic stimuli. In vitro through-tissue printing experiments demonstrate the feasibility of this method for clinical scenarios such as gastric perforation sealing and bleeding control. Overall, MPUP leverages the deep penetration and remote controllability of ultrasound to enable noninvasive, on-demand fabrication of untethered magnetic soft robots, offering a promising strategy for advancing their use in biomedical interventions.

Accurate orthopedic fracture reduction planning is essential for ensuring successful postoperative recovery and improving patient outcomes. However, current methods are challenged by the complex and irregular fracture geometries and the scarcity of annotated training data. To address these challenges, we propose a novel approach that integrates learning-based shape restoration and fracture simulation. A transformer-based model is developed, which utilizes patch-to-patch restoration and recursive fragment registration to iteratively refine fracture reduction poses. To generate diverse and anatomically realistic fractured datasets for model training, we develop a fracture data simulation approach that combines statistical shape modeling with clinically representative fracture patterns, reducing reliance on annotated samples. Tested on extensive clinical data with hipbone and sacrum fractures, the proposed method achieved mean translational and rotational errors of 2.34 mm and 4.54 ^∘ , respectively, outperforming both template-based and existing learning-based methods. Our approach enhances learning and generalization for automated fracture reduction by connecting synthetic and real-world fracture data.