Design, Optimization, and Modeling of a Hydraulic Soft Robot for Chronic Total Occlusions.

Chronic total occlusion (CTO) is one of the most severe and sophisticated vascular stenosis because of complete blockage, greater operation difficulty, and lower procedural success rate. This study proposes a hydraulic-driven soft robot imitating the earthworm's locomotion to assist doctors or operators in actively opening thrombi in coronary or peripheral artery vessels. Firstly, a three-actuator bionic soft robot is developed based on earthworms' physiological structure. The soft robot's locomotion gait inspired by the earthworm's mechanism is designed. Secondly, the influence of structure parameters on actuator deformation, stress, and strain is explored, which can help us determine the soft actuators' optimal structure parameters. Thirdly, the relationship between hydraulic pressure and actuator deformation is investigated by performing finite element analysis using the bidirectional fluid-structure interaction (FSI) method. The kinematic models of the soft actuators are established to provide a valuable reference for the soft actuators' motion control.

Giant bile duct dilatation in newborn: A case report.

BACKGROUND: Giant congenital biliary dilation (CBD) is a rare condition observed in clinical practice. Infants born with this condition often experience a poor overall health status, and the disease progresses rapidly, leading to severe biliary obstruction, infections, pressure exerted by the enlarged CBD on abdominal organs, disturbances in the internal environment, and multiple organ dysfunction. The treatment of giant CBD using laparoscopy is challenging due to the high degree of variation in the shape of the bile duct and other organs, making it difficult to separate the bile duct wall from adjacent tissues or to control bleeding. CASE SUMMARY: Herein, we present the details of an 11-d-old male newborn who was diagnosed with giant CBD. The patient was admitted to the neonatal surgery department of our hospital due to a history of common bile duct cyst that was detected more than 3 mo ago, and also because the patient had been experiencing yellowish skin for the past 9 d. The abnormal echo in the fetal abdomen was first noticed by the patient's mother during a routine ultrasound examination at a local hospital, when the patient was at 24 wk + 6 d of pregnancy. This finding raised concerns about the possibility of congenital biliary dilatation (22 mm × 21 mm). Subsequent ultrasound examinations at different hospitals consistently confirmed the presence of a congenital biliary dilatation. No specific treatment was administered for biliary dilatation during this period. A computed tomography scan conducted during the hospitalization revealed a large cystic mass in the right upper quadrant and pelvis, measuring approximately 9.2 cm × 7.4 cm × 11.3 cm. Based on the scan, it was classified as a type I biliary dilatation. CONCLUSION: The analysis reveals that prenatal imaging techniques, such as ultrasound and magnetic resonance imaging, play a crucial role in the early diagnosis, fetal prognosis, and treatment plan for giant CBD. Laparoscopic surgery for giant CBD presents certain challenges, including difficulties in separating the cyst wall, anastomosis, and hemostasis, as well as severe biliary system infection and ulceration. Consequently, there is a high likelihood of converting to laparotomy. The choice between surgical methods like hepaticojejunostomy (HJ) or hepaticoduodenostomy has not been standardized yet. However, we have achieved favorable outcomes using HJ. Preoperative management of inflammation, biliary drainage, liver function protection, and supportive treatment are particularly vital in improving children's prognosis. After discharge, it is essential to conduct timely reexamination and close follow-up to identify potential complications.

High-resolution feature based central venous catheter tip detection network in X-ray images.

Hospital patients can have catheters and lines inserted during the course of their admission to give medicines for the treatment of medical issues, especially the central venous catheter (CVC). However, malposition of CVC will lead to many complications, even death. Clinicians always detect the malposition based on position detection of CVC tip via X-ray images. To reduce the workload of the clinicians and the percentage of malposition occurrence, we propose an automatic catheter tip detection framework based on a convolutional neural network (CNN). The proposed framework contains three essential components which are modified HRNet, segmentation supervision module, and deconvolution module. The modified HRNet can retain high-resolution features from start to end, ensuring the maintenance of precise information from the X-ray images. The segmentation supervision module can alleviate the presence of other line-like structures such as the skeleton as well as other tubes and catheters used for treatment. In addition, the deconvolution module can further increase the feature resolution on the top of the highest-resolution feature maps in the modified HRNet to get a higher-resolution heatmap of the catheter tip. A public CVC Dataset is utilized to evaluate the performance of the proposed framework. The results show that the proposed algorithm offering a mean Pixel Error of 4.11 outperforms three comparative methods (Ma's method, SRPE method, and LCM method). It is demonstrated to be a promising solution to precisely detect the tip position of the catheter in X-ray images.

Dibenzocyclooctadiene lignans from the family Schisandraceae: A review of phytochemistry, structure-activity relationship, and hepatoprotective effects.

Liver injury is a common pathological process characterized by massive degeneration and abnormal death of liver cells. With increase in dead cells and necrosis, liver injury eventually leads to nonalcoholic fatty liver disease (NAFLD), hepatic fibrosis, and even hepatocellular carcinoma (HCC). Consequently, it is necessary to treat liver injury and to prevent its progression. The drug Bicylol is widely employed in China to treat chronic hepatitis B virus (HBV) and has therapeutic potential for liver injury. It is the derivative of dibenzocyclooctadiene lignans extracted from Schisandra chinensis (SC). The Schisandraceae family is a rich source of dibenzocyclooctadiene lignans, which possesses potential liver protective activity. This study aimed to comprehensively summarize the phytochemistry, structure-activity relationship and molecular mechanisms underlying the liver protective activities of dibenzocyclooctadiene lignans from the Schisandraceae family. Here, we had discussed the analysis of absorption or permeation properties of 358 compounds based on Lipinski's rule of five. So far, 358 dibenzocyclooctadiene lignans have been reported, with 37 of them exhibited hepatoprotective effects. The molecular mechanism of the active compounds mainly involves antioxidative stress, anti-inflammation and autophagy through Kelch-like ECH-associating protein 1/nuclear factor erythroid 2 related factor 2/antioxidant response element (Keap1/Nrf2/ARE), nuclear factor kappa B (NF-кB), and transforming growth factor ß (TGF-ß)/Smad 2/3 signaling pathways. This review is expected to provide scientific ideas for future research related to developing and utilizing the dibenzocyclooctadiene lignans from Schisandraceae family.

A Novel Spatial Position Prediction Navigation System Makes Surgery More Accurate.

During intravascular interventional surgery, the 3D surgical navigation system can provide doctors with 3D spatial information of the vascular lumen, reducing the impact of missing dimension caused by digital subtraction angiography (DSA) guidance and further improving the success rate of surgeries. Nevertheless, this task often comes with the challenge of complex registration problems due to vessel deformation caused by respiratory motion and high requirements for the surgical environment because of the dependence on external electromagnetic sensors. This article proposes a novel 3D spatial predictive positioning navigation (SPPN) technique to predict the real-time tip position of surgical instruments. In the first stage, we propose a trajectory prediction algorithm integrated with instrumental morphological constraints to generate the initial trajectory. Then, a novel hybrid physical model is designed to estimate the trajectory's energy and mechanics. In the second stage, a point cloud clustering algorithm applies multi-information fusion to generate the maximum probability endpoint cloud. Then, an energy-weighted probability density function is introduced using statistical analysis to achieve the prediction of the 3D spatial location of instrument endpoints. Extensive experiments are conducted on 3D-printed human artery and vein models based on a high-precision electromagnetic tracking system. Experimental results demonstrate the outstanding performance of our method, reaching 98.2% of the achievement ratio and less than 3 mm of the average positioning accuracy. This work is the first 3D surgical navigation algorithm that entirely relies on vascular interventional robot sensors, effectively improving the accuracy of interventional surgery and making it more accessible for primary surgeons.

Learning Skill Characteristics From Manipulations.

Percutaneous coronary intervention (PCI) has increasingly become the main treatment for coronary artery disease. The procedure requires high experienced skills and dexterous manipulations. However, there are few techniques to model PCI skill so far. In this study, a learning framework with local and ensemble learning is proposed to learn skill characteristics of different skill-level subjects from their PCI manipulations. Ten interventional cardiologists (four experts and six novices) were recruited to deliver a medical guidewire to two target arteries on a porcine model for in vivo studies. Simultaneously, translation and twist manipulations of thumb, forefinger, and wrist are acquired with electromagnetic (EM) and fiber-optic bend (FOB) sensors, respectively. These behavior data are then processed with wavelet packet decomposition (WPD) under 1-10 levels for feature extraction. The feature vectors are further fed into three candidate individual classifiers in the local learning layer. Furthermore, the local learning results from different manipulation behaviors are fused in the ensemble learning layer with three rule-based ensemble learning algorithms. In subject-dependent skill characteristics learning, the ensemble learning can achieve 100% accuracy, significantly outperforming the best local result (90%). Furthermore, ensemble learning can also maintain 73% accuracy in subject-independent schemes. These promising results demonstrate the great potential of the proposed method to facilitate skill learning in surgical robotics and skill assessment in clinical practice.

Primary repair of esophageal atresia gross type C via thoracoscopic magnetic compression anastomosis: A case report.

BACKGROUND: Esophageal atresia (EA) is a life-threatening congenital malformation in newborns, and the traditional repair approaches pose technical challenges and are extremely invasive. Therefore, surgeons have been actively investigating new minimally invasive techniques to address this issue. Magnetic compression anastomosis has been reported in several studies for its potential in repairing EA. In this paper, the primary repair of EA with magnetic compression anastomosis under thoracoscopy was reported. CASE SUMMARY: A full-term male weighing 3500 g was diagnosed with EA gross type C. The magnetic devices used in this procedure consisted of two magnetic rings and several catheters. Tracheoesophageal fistula ligation and two purse strings were performed. The magnetic compression anastomosis was then completed thoracoscopically. After the primary repair, no additional operation was conducted. A patent anastomosis was observed on the 15th day postoperatively, and the magnets were removed on the 23rd day. No leakage existed when the transoral feeding started. CONCLUSION: Thoracoscopic magnetic compression anastomosis may be a promising minimally invasive approach for repairing EA.

Endoscopic magnetic compression stricturoplasty for congenital esophageal stenosis: A case report.

BACKGROUND: Congenital esophageal stenosis (CES) is a rare malformation of the digestive tract. Endoscopic dilation and thoracotomy have been the main treatments for CES. However, there is no well-defined management protocol. Magnetic compression stricturoplasty (MCS) has been used in refractory esophageal stricture in children after esophageal atresia. CASE SUMMARY: We describe the first case of MCS for CES in one female child patient. The child (aged 3 years and 1 mo) was admitted due to frequent vomiting and choking after eating complementary food since 7 mo old. Esophagography and gastroendoscopy showed that there was stenosis in the lower esophagus, suggesting a diagnosis of CES. The patient did not receive any treatment for esophageal stricture including surgery or endoscopic dilation procedures before MCS. MCS procedure was smoothly conducted without complications. At 24 mo after MCS, durable esophageal patency without dysphagia was achieved. CONCLUSION: MCS may serve as an alternative and efficient method for patients with CES.

Esophageal magnetic compression anastomosis in dogs.

BACKGROUND: Magnetic compression anastomosis (MCA) is a novel suture-free reconstruction of the digestive tract. It has been used in gastrointestinal anastomosis, jejunal anastomosis, cholangioenteric anastomosis and so on. The traditional operative outcomes of congenital esophageal atresia and benign esophageal stricture are poor, and there are too many complications postoperatively. AIM: To test MCA technology to reconstruct the esophagus in dogs, prior to studying the feasibility and safety of MCA in humans. METHODS: Thirty-six dogs were randomized into either the study or control group (n = 18 per group). The dogs in the study group were subjected to end-to-end esophageal anastomosis with the magnetic compression device, while those in the control group underwent hand-sewn anastomosis with 4-0 absorbable multifilament Vicryl. We used interrupted single-layer inverting sutures. The anastomosis time, gross appearance, weight and pathology of the anastomosis were evaluated at one month, three months and six months postoperatively. RESULTS: The anastomosis time of the MCA group was shorter than that of the hand-sewn group (7.5 ± 1.0 min vs 12.5 ± 1.8 min, P < 0.01). In the MCA group, X-ray examination was performed every day to locate the magnetic device in the esophagus before the magnetic device fell off from the esophagus. In the hand-sewn group, dogs did not undergo X-ray examination. One month after the surgeries, the mean weight of the dogs in the hand-sewn group had decreased more than that of the dogs in the MCA group (11.63 ± 0.71 kg vs 12.73 ± 0.80 kg, P < 0.05). At 3 mo and 6 mo after the operation, the dogs' weights were similar between the two groups (13.75 ± 0.84 kg vs 14.03 ± 0.82 kg, 14.93 ± 0.80 kg vs 15.44 ± 0.47 kg). The number of inflammatory cells in MCA group was lower than that in hand-sewn group on 1 mo after operation. CONCLUSION: MCA is an effective and safe method for esophageal reconstruction. The anastomosis time of the MCA group was less than that of the hand-sewn group. This study shows that MCA technology may be applied to human esophageal reconstruction, provided these favorable results are confirmed by more publications.

EDDM3A drives gastric cancer progression by promoting HIF-1α-dependent aerobic glycolysis.

Epididymal protein 3A (EDDM3A) is a protein involved in sperm maturation. It has been demonstrated that EDDM3A expression is upregulated and promotes cell proliferation in non-small cell lung cancer (NSCLC). However, the role of EDDM3A in other types of human cancers, including gastric cancer (GC), is still unexplored. Here, we show that the expression of EDDM3A is significantly upregulated in gastric cancer (GC) tissues and its upregulation correlates with poorer survival in patients with gastric cancer. Knockdown of EDDM3A inhibited growth and metastasis of GC cells, whereas overexpression of EDDM3A exhibited the opposite effect. Mechanistically, enhanced aerobic glycolysis mediated by upregulation of HIF-1α and subsequently increased target glycolytic genes and decreased mitochondrial biogenesis was found to contribute to the promotion of tumor growth and metastasis by EDDM3A in GC cells. Additionally, upregulation of EDDM3A in GC is at least partially mediated by downregulation of miR-618. In conclusion, elevated EDDM3A plays a pivotal oncogenic role in gastric carcinogenesis, suggesting it as a potential therapeutic target for treatment of GC.

A Multilayer and Multimodal-Fusion Architecture for Simultaneous Recognition of Endovascular Manipulations and Assessment of Technical Skills.

The clinical success of the percutaneous coronary intervention (PCI) is highly dependent on endovascular manipulation skills and dexterous manipulation strategies of interventionalists. However, the analysis of endovascular manipulations and related discussion for technical skill assessment are limited. In this study, a multilayer and multimodal-fusion architecture is proposed to recognize six typical endovascular manipulations. The synchronously acquired multimodal motion signals from ten subjects are used as the inputs of the architecture independently. Six classification-based and two rule-based fusion algorithms are evaluated for performance comparisons. The recognition metrics under the determined architecture are further used to assess technical skills. The experimental results indicate that the proposed architecture can achieve the overall accuracy of 96.41%, much higher than that of a single-layer recognition architecture (92.85%). In addition, the multimodal fusion brings significant performance improvement in comparison with single-modal schemes. Furthermore, the K -means-based skill assessment can obtain an accuracy of 95% to cluster the attempts made by different skill-level groups. These hopeful results indicate the great possibility of the architecture to facilitate clinical skill assessment and skill learning.

A Unified Framework for Multi-Guidewire Endpoint Localization in Fluoroscopy Images.

OBJECTIVE: In this paper, Keypoint Localization Region-based CNN (KL R-CNN) is proposed, which can simultaneously accomplish the guidewire detection and endpoint localization in a unified model. METHODS: KL R-CNN modifies Mask R-CNN by replacing the mask branch with a novel keypoint localization branch. Besides, some settings of Mask R-CNN are also modified to generate the keypoint localization results at a higher detail level. At the same time, based on the existing metrics of Average Precision (AP) and Percentage of Correct Keypoints (PCK), a new metric named APPCK is proposed to evaluate the overall performance on the multi-guidewire endpoint localization task. Compared with existing metrics, APPCK is easy to use and its results are more intuitive. RESULTS: Compared with existing methods, KL R-CNN has better performance when the threshold is loose, reaching a mean APPCK of 90.65% when the threshold is 9 pixels. CONCLUSION: KL R-CNN achieves the state-of-the-art performance on the multi-guidewire endpoint localization task and has application potentials. SIGNIFICANCE: KL R-CNN can achieve the localization of guidewire endpoints in fluoroscopy images, which is a prerequisite for computer-assisted percutaneous coronary intervention. KL R-CNN can also be extended to other multi-instrument localization tasks.

Machine Learning for Structure Determination in Single-Particle Cryo-Electron Microscopy: A Systematic Review.

Recently, single-particle cryo-electron microscopy (cryo-EM) has become an indispensable method for determining macromolecular structures at high resolution to deeply explore the relevant molecular mechanism. Its recent breakthrough is mainly because of the rapid advances in hardware and image processing algorithms, especially machine learning. As an essential support of single-particle cryo-EM, machine learning has powered many aspects of structure determination and greatly promoted its development. In this article, we provide a systematic review of the applications of machine learning in this field. Our review begins with a brief introduction of single-particle cryo-EM, followed by the specific tasks and challenges of its image processing. Then, focusing on the workflow of structure determination, we describe relevant machine learning algorithms and applications at different steps, including particle picking, 2-D clustering, 3-D reconstruction, and other steps. As different tasks exhibit distinct characteristics, we introduce the evaluation metrics for each task and summarize their dynamics of technology development. Finally, we discuss the open issues and potential trends in this promising field.

An Improved Networked Predictive Controller for Vascular Robot Using 5G Networks.

Percutaneous coronary intervention (PCI) has gradually become the most common treatment of coronary artery disease (CAD) in clinical practice due to its advantages of small trauma and quick recovery. However, the availability of hospitals with cardiac catheterization facilities and trained interventionalists is extremely limited in remote and underdeveloped areas. Remote vascular robotic system can assist interventionalists to complete operations precisely, and reduce occupational health hazards occurrence. In this paper, a bionic remote vascular robot is introduced in detail from three parts: mechanism, communication architecture, and controller model. Firstly, human finger-like mechanisms in vascular robot enable the interventionalists to advance, retract and rotate the guidewires or balloons. Secondly, a 5G-based communication system is built to satisfy the end-to-end requirements of strong data transmission and packet priority setting in remote robot control. Thirdly, a generalized predictive controller (GPC) is developed to suppress the effect of time-varying network delay and parameter identification error, while adding a designed polynomial compensation module to reduce tracking error and improve system responsiveness. Then, the simulation experiment verifies the system performance in comparison with different algorithms, network delay, and packet loss rate. Finally, the improved control system conducted PCI on an experimental pig, which reduced the delivery integral absolute error (IAE) by at least 20% compared with traditional methods.

Design and Performance Evaluation of a Novel Vascular Robotic System for Complex Percutaneous Coronary Interventions.

The robotic-assisted percutaneous coronary intervention is an emerging technology with great potential to solve the shortcomings of existing treatments. However, the current robotic systems can not manipulate two guidewires or ballons/stents simultaneously for coronary bifurcation lesions. This paper presents VasCure, a novel bio-inspired vascular robotic system, to deliver two guidewires and stents into the main branch and side branch of bifurcation lesions in sequence. The system is designed in master-slave architecture to reduce occupational hazards of radiation exposure and orthopedic injury to interventional surgeons. The slave delivery device has one active roller and two passive rollers to manipulate two interventional devices. The performance of the VasCure was verified by in vitro and in vivo animal experiments. In vitro results showed the robotic system has good accuracy to deliver guidewires and the maximum error is 0.38mm. In an animal experiment, the interventional surgeon delivered two guidewires and balloons to the left circumflex branch and the left anterior descending branch of the pig, which confirmed the feasibility of the vascular robotic system.

A highly conductive self-assembled multilayer graphene nanosheet film for electronic tattoos in the applications of human electrophysiology and strain sensing.

Highly conductive, conformable and gel-free electrodes are desirable in human electrophysiology. Besides, intimately coupling with human skin, wearable strain sensors can detect numerous physiological signals, such as wrist pulse and breath. In this study, a multilayer graphene nanosheet film (MGNF) with high conductivity was prepared by the Marangoni self-assembly for using in tattoo dry electrodes (TDEs) and in a graphene tattoo strain sensor (GTSS). Compared to commercial Ag/AgCl gel electrodes, TDEs have lower skin-electrode contact impedance and could detect human electrocardiogram for 24-hour wearing more accurately as well as electromyogram. Through designing a slim serpentine ribbon structure, a resistance-type GTSS, without deterioration even after 2000 cycles, is well demonstrated for human wrist pulse and breath sensing. With the advantages of high conductivity and conformability, MGNF provides support to fabricate low-cost, customizable, and high-performance electronic tattoos for human electrophysiology and strain sensing.

Real-Time Multi-Guidewire Endpoint Localization in Fluoroscopy Images.

The real-time localization of the guidewire endpoints is a stepping stone to computer-assisted percutaneous coronary intervention (PCI). However, methods for multi-guidewire endpoint localization in fluoroscopy images are still scarce. In this paper, we introduce a framework for real-time multi-guidewire endpoint localization in fluoroscopy images. The framework consists of two stages, first detecting all guidewire instances in the fluoroscopy image, and then locating the endpoints of each single guidewire instance. In the first stage, a YOLOv3 detector is used for guidewire detection, and a post-processing algorithm is proposed to refine the guidewire detection results. In the second stage, a Segmentation Attention-hourglass (SA-hourglass) network is proposed to predict the endpoint locations of each single guidewire instance. The SA-hourglass network can be generalized to the keypoint localization of other surgical instruments. In our experiments, the SA-hourglass network is applied not only on a guidewire dataset but also on a retinal microsurgery dataset, reaching the mean pixel error (MPE) of 2.20 pixels on the guidewire dataset and the MPE of 5.30 pixels on the retinal microsurgery dataset, both achieving the state-of-the-art localization results. Besides, the inference rate of our framework is at least 20FPS, which meets the real-time requirement of fluoroscopy images (6-12FPS).

Magnetic compression anastomosis for rectal atresia following necrotizing enterocolitis: A case report.

RATIONALE: Rectal atresia caused by necrotizing enterocolitis (NEC) is a serious and rare complication in children. Magnetic compression anastomosis (MCA) has been effectively applied in children with congenital oesophageal atresia and biliary atresia. Herein, we reported a case of successfully application of MCA in an infant with rectal atresia following NEC. PATIENT CONCERNS: A 30 weeks premature birth female fetal infant was transferred to our neonatal intensive care unit due to premature delivery, low birth weight, and neonatal respiratory distress. On postpartum day 11, the infant developed abdominal distension and mucosanguineous feces. This infant was then clinically diagnosed as NEC. She underwent anesthesia and intestinal fistula operation on postpartum day 11 because of NEC. DIAGNOSIS: After 3 months, radiographic examination revealed rectal atresia and stricture. INTERVENTIONS: This infant was successfully treated with MCA following a cecum-rectal anastomosis and ileocecal valve was reserved. OUTCOMES: On postoperative day 9, she passed the 2 magnets per rectum. In addition, there were no difficult defecation or fecal incontinence or other short-term complications. After the 7-month follow-up, the patient had an excellent clinical outcome. LESSONS: MCA is a feasible and effective method for treating rectal atresia in infants.

Magnetic compression for anastomosis in treating an infant born with long-gap oesophageal atresia: A case report.

RATIONALE: Neonatal long-gap esophageal atresia (LGEA) with tracheoesophageal fistula (TEF) is an uncommon but serious congenital malformation of the esophagus in newborns, and it remains challenging for pediatric surgeons. Magnetic compress has been shown to be effective for the treatment of LGEA in children and adults. However, the implementation of this unique technique for neonatal LGEA has not been evaluated. PATIENT CONCERNS: A female infant was born at 37 weeks of gestation. Prenatal ultrasound imaging revealed signs of esophageal atresia, including the absence of the gastric bubble and polyhydramnios. DIAGNOSES: A diagnosis of LGEA with TEF was confirmed at birth by contrast X-ray. INTERVENTIONS: She was treated with magnetic compression anastomosis (MCA) following an esophago-esophagostomy. Two magnetic rings were customized, and the MCA was conducted during the same stage surgery of ligating the TEF. Under the magnetic force, the 2 magnet rings pulled along the gastric tube to achieve anastomosis. The postoperative permanent suction of these 2 pouches was instituted, and spontaneous growth was awaited. Magnet removal was performed at 36 days, and enteral nutrition was continued via a gastric tube for 4 weeks at post-operation. OUTCOMES: The upper gastrointestinal contrast confirmed the anastomotic patency perfectly after 3 months. The patient was followed up for 18 months, and exhibited durable esophageal patency without dysphagia. LESSONS: These results suggest that MCA is feasible and effective for treating LGEA in infants.

Pyramid attention recurrent networks for real-time guidewire segmentation and tracking in intraoperative X-ray fluoroscopy.

In endovascular and cardiovascular surgery, real-time and accurate segmentation and tracking of interventional instruments can aid in reducing radiation exposure, contrast agent and processing time. Nevertheless, this task often comes with the challenges of the elongated deformable structures with low contrast in noisy X-ray fluoroscopy. To address these issues, a novel efficient network architecture, termed pyramid attention recurrent networks (PAR-Net), is proposed for real-time guidewire segmentation and tracking. The proposed PAR-Net contains three major modules, namely pyramid attention module, recurrent residual module and pre-trained MobileNetV2 encoder. Specifically, a hybrid loss function of both reinforced focal loss and dice loss is proposed to better address the issues of class imbalance and misclassified examples. Quantitative and qualitative evaluations on clinical intraoperative images demonstrate that the proposed approach significantly outperforms simpler baselines as well as the best previously published result for this task, achieving the state-of-the-art performance.