Risk factors and prognostic analysis of right ventricular dysfunction after lung resection for NSCLC.

Objectives: Lung cancer is the leading cause of cancer death, and 80-85% of all lung cancer cases are non-small cell lung cancer (NSCLC). Surgical resection is the standard treatment for early-stage NSCLC. However, lung resection, a surgical procedure, can result in complications and increased mortality. Recent studies have shown a significant correlation between complications after lung resection and right ventricular dysfunction. Methods: Transthoracic echocardiography-derived right ventricular-pulmonary artery coupling (RV-PAC) was utilized to assess right ventricular function in these patients. Multivariate logistic regression analysis was also conducted to assess risk factors independently associated with RV-PA uncoupling. The 3- and 5-year cumulative survival rates were estimated with Kaplan-Meier curves, and differences between groups were analyzed using the Mantel-Cox log-rank test. Results: RV-PA uncoupling was defined as a TAPSE/PASP value < 0.67 mm/mm Hg according to spline analysis. The results of multivariable logistic regression analysis indicated that diabetes is an independent risk factor for right ventricular dysfunction after lung resection in patients with NSCLC. Kaplan-Meier analysis revealed a significant decrease in the survival rate of patients with RV-PA uncoupling at both the 3-year follow-up (73% vs 40%, p < 0.001) and 5-year follow-up (64% vs 37%, p < 0.001). Conclusions: After lung resection for NSCLC, the patient's right ventricular function predicts prognosis. Patients with right ventricular dysfunction, particularly those with diabetes mellitus, have a worse prognosis. It is crucial to actively prevent and correct risk factors to reduce the mortality rate in these patients.

Class-Irrelevant Feature Removal for Few-Shot Image Classification.

Most existing few-shot image classification methods employ global pooling to aggregate class-relevant local features in a data-drive manner. Due to the difficulty and inaccuracy in locating class-relevant regions in complex scenarios, as well as the large semantic diversity of local features, the class-irrelevant information could reduce the robustness of the representations obtained by performing global pooling. Meanwhile, the scarcity of labeled images exacerbates the difficulties of data-hungry deep models in identifying class-relevant regions. These issues severely limit deep models' few-shot learning ability. In this work, we propose to remove the class-irrelevant information by making local features class relevant, thus bypassing the big challenge of identifying which local features are class irrelevant. The resulting class-irrelevant feature removal (CIFR) method consists of three phases. First, we employ the masked image modeling strategy to build an understanding of images' internal structures that generalizes well. Second, we design a semantic-complementary feature propagation module to make local features class relevant. Third, we introduce a weighted dense-connected similarity measure, based on which a loss function is raised to fine-tune the entire pipeline, with the aim of further enhancing the semantic consistency of the class-relevant local features. Visualization results show that CIFR achieves the removal of class-irrelevant information by making local features related to classes. Comparison results on four benchmark datasets indicate that CIFR yields very promising performance.

Investigation of Lithium-Ion Battery Performance Utilizing Magnetic Controllable Superionic Conductor Li3(V1-x Fe x )2(PO4)3/C (x = 0.05 and 0.10).

Lithium-ion batteries with Li3V2(PO4)3/C as the cathode have been a popular research topic in recent years; however, studies of the effects of external magnetic fields on them are less common. This study investigates the effects of an external magnetic field applied parallel to the direction of the anode and cathode on the ion transport through iron-doped Li3(V1-x Fe x )2(PO4)3, the outer carbon coating, the film/electrolyte/separator, and up to the lithium metal electrode on a microscopic level. The results reveal that for the x = 0.05 sample with lower doping, the magnetostriction expansion of Li3(V1-x Fe x )2(PO4)3 and the magnetostrictive contraction effect of the outer ordered carbon layer cancel each other out, resulting in no significant enhancement of the battery's energy and power density due to the external magnetic field. In contrast, the x = 0.1 sample, lacking magnetostrictive contraction in the outer ordered carbon layer, shows that its energy and power density can be influenced by the magnetic field. Under zero magnetic field, the cyclic performance exhibits superior average capacity performance in the x = 0.05 sample, while the x = 0.1 sample shows a lower decay rate. Both samples are affected by the magnetic field; however, the x = 0.1 sample performs better under magnetic conditions. In particular, in the C-rate tests under a magnetic field, the sample with x = 0.1 showed a significant relative reduction in capacity decay rate by 20.18% compared to the sample with x = 0.05.

Elevating nanomaterial optical sensor arrays through the integration of advanced machine learning techniques for enhancing visual inspection of food quality and safety.

Food quality and safety problems caused by inefficient control in the food chain have significant implications for human health, social stability, and economic progress and optical sensor arrays (OSAs) can effectively address these challenges. This review aims to summarize the recent applications of nanomaterials-based OSA for food quality and safety visual monitoring, including colourimetric sensor array (CSA) and fluorescent sensor array (FSA). First, the fundamental properties of various advanced nanomaterials, mainly including metal nanoparticles (MNPs) and nanoclusters (MNCs), quantum dots (QDs), upconversion nanoparticles (UCNPs), and others, were described. Besides, the diverse machine learning (ML) and deep learning (DL) methods of high-dimensional data obtained from the responses between different sensing elements and analytes were presented. Moreover, the recent and representative applications in pesticide residues, heavy metal ions, bacterial contamination, antioxidants, flavor matters, and food freshness detection were comprehensively summarized. Finally, the challenges and future perspectives for nanomaterials-based OSAs are discussed. It is believed that with the advancements in artificial intelligence (AI) techniques and integrated technology, nanomaterials-based OSAs are expected to be an intelligent, effective, and rapid tool for food quality assessment and safety control.

Biomimetic hierarchical porous high entropy alloy for significantly enhancing overall seawater splitting.

Herein, a biomimetic hierarchical porous high entropy alloy (BHP-HEA) is prepared by a strategy combining selective laser melting and selective phase dissolution. It exhibited excellent seawater splitting performance, which only needs a low potential of 1.53 V to realize a current density of 100 mA cm-2, with exceptional stability.

Characterization of metabolic features and potential anti-osteoporosis mechanism of pinoresinol diglucoside using metabolite profiling and network pharmacology.

RATIONALE: Eucommia cortex is the core herb in traditional Chinese medicine preparations for the treatment of osteoporosis. Pinoresinol diglucoside (PDG), the quality control marker and the key pharmacodynamic component in Eucommia cortex, has attracted global attention because of its definite effects on osteoporosis. However, the in vivo metabolic characteristics of PDG and its anti-osteoporotic mechanism are still unclear, restricting its development and application. METHODS: Ultra-high-performance liquid chromatography coupled with quadrupole time-of-flight mass spectrometry was used to analyze the metabolic characteristics of PDG in rats, and its anti-osteoporosis targets and mechanism were predicted using network pharmacology. RESULTS: A total of 51 metabolites were identified or tentatively characterized in rats after oral administration of PDG (10 mg/kg/day), including 9 in plasma, 28 in urine, 13 in feces, 10 in liver, 4 in heart, 3 in spleen, 11 in kidneys, and 5 in lungs. Furan-ring opening, dimethoxylation, glucuronidation, and sulfation were the main metabolic characteristics of PDG in vivo. The potential mechanism of PDG against osteoporosis was predicted using network pharmacology. PDG and its metabolites could regulate BCL2, MARK3, ALB, and IL6, involving PI3K-Akt signaling pathway, estrogen signaling pathway, and so on. CONCLUSIONS: This study was the first to demonstrate the metabolic characteristics of PDG in vivo and its potential anti-osteoporosis mechanism, providing the data for further pharmacological validation of PDG in the treatment of osteoporosis.

Emerging Atomistic Modeling Methods for Heterogeneous Electrocatalysis.

Heterogeneous electrocatalysis lies at the center of various technologies that could help enable a sustainable future. However, its complexity makes it challenging to accurately and efficiently model at an atomic level. Here, we review emerging atomistic methods to simulate the electrocatalytic interface with special attention devoted to the components/effects that have been challenging to model, such as solvation, electrolyte ions, electrode potential, reaction kinetics, and pH. Additionally, we review relevant computational spectroscopy methods. Then, we showcase several examples of applying these methods to understand and design catalysts relevant to green hydrogen. We also offer experimental views on how to bridge the gap between theory and experiments. Finally, we provide some perspectives on opportunities to advance the field.

A simple and efficient gene functional analysis method for studying the growth and development of peach seedlings.

Stable genetic transformation of peach [Prunus persica (L.) Batsch] still faces many technical challenges, and existing transient expression methods are limited by tissue type or developmental stage, making it difficult to conduct functional analysis of genes regulating shoot growth. To overcome this dilemma, we developed a three-step method for efficient analysis of gene functions during peach seedling growth and development. This method resulted in transformation frequencies ranging from 48 to 87%, depending on the gene. From transformation of germinating seeds to phenotyping of young saplings took just 1.5 months and can be carried out any time of year. To test the applicability of this method, the function of three tree architecture-related genes, namely PpPDS, PpMAX4, and PpWEEP, and two lateral root-related genes, PpIAA14-1 and -2, were confirmed. Since functional redundancy can challenge gene functional analyses, tests were undertaken with the growth-repressor DELLA, which has three homologous genes, PpDGYLA (DG), PpDELLA1 (D1), and -2 (D2), in peach that are functionally redundant. Silencing using a triple-target vector (TRV2-DG-D1-D2) resulted in transgenic plants taller than those carrying just TRV2-DG or TRV2. Simultaneously silencing the three DELLA genes also attenuated the stature of two dwarf genotypes, 'FHSXT' and 'HSX', which normally accumulate DELLA proteins. Our study provides a method for the functional analysis of genes in peach and can be used for the study of root, stem, and leaf development. We believe this method can be replicated in other woody plants.

[Change Characteristics of Carbonaceous Aerosol in Chengdu from 2018 to 2021].

Carbonaceous aerosol is an important component of atmospheric fine particulates （PM2.5） that has an important effect on global climate change, atmospheric visibility, regional air quality, and human health. In order to investigate the long-term change characteristics of carbonaceous aerosols under the background of emission reduction, the concentrations of organic carbon （OC）, elemental carbon （EC） in PM2.5 samples, and volatile organic compounds （VOCs） in Chengdu from 2018 to 2021 and the corresponding meteorological factors were obtained through real-time online monitoring. The results showed that the average ρ（OC） and ρ（EC） during the monitoring period were （10.9 ±5.7） µg·m-3 and （2.6 ±1.9） µg·m-3, accounting for 25.2% and 6.0% of PM2.5, respectively, and the average ρ（SOC） was （5.7 ±3.3） µg·m-3, accounting for 52.9% of OC. The concentrations of OC, EC, and PM2.5 showed a downward trend from 2018 to 2020 [PM2.5： The concentration of average annual decrease was -7.1 µg·ï¼ˆm3·a） -1, with an average annual decrease of -14.6 %·a-1； OC： -1.7 µg·ï¼ˆm3·a）-1, -14.2 %·a-1； EC： -0.1 µg·ï¼ˆm3·a）-1, -4.4 %·a-1], and the concentrations of each pollutant in 2021 rebounded in different ranges compared with those in 2020. The concentrations of PM2.5 and OC were as follows： winter > spring > autumn > summer, and the concentrations of EC were as follows： winter > autumn > spring > summer. The proportions of OC and EC were higher in summer and autumn than in other seasons, with the average proportions of 26.8% and 6.9%, respectively. With the aggravation of the pollution level, OC, EC, and SOC concentrations gradually increased, but the proportions in PM2.5 showed a gradual downtrend, indicating that the control factor of PM2.5 pollution in Chengdu was not the carbon component. Source apportionment results showed that carbonaceous aerosols in Chengdu were mainly affected by motor vehicles, industrial sources, biomass combustion sources, and VOCs secondary reaction. From 2019 to 2021, EC was affected by the characteristic components of motor vehicles and decreased yearly. OC and EC were affected by VOCs more in spring and autumn than in other seasons. VOCs emission management should be increased in spring and autumn to reduce the impact of secondary reaction.

Analysis of the associations between moderate-to-vigorous physical activity and screen time on psychological symptoms among university students: a cross-sectional survey based on six geographic regions in China.

BACKGROUND: Declining physical activity and increasing screen time (ST) among Chinese adolescents have become major concerns shared by scholars, while mental health issues are also on the rise. Previous studies have confirmed the association between physical activity and screen time and psychological symptoms, but it is unclear how their psychological symptoms, especially for Chinese university students who have a high proportion of psychological symptoms, and no research evidence has been found. METHODS: This study investigated physical activity, screen time, and psychological symptoms in 11,173 university students aged 19-22 years in six regions of China. A binary logistic regression analysis was used to analyze the association between moderate-to-vigorous physical activity (MVPA) and screen time and psychological symptoms. And the generalize linear model (GLM) analysis was used to further analyze the association between MVPA and screen time and psychological symptoms. RESULTS: The detection rate of psychological symptoms among Chinese university students was 16.3%, with a higher percentage of female students (17.5%) than male students (14.7%). The proportion of male students (8.2%) with MVPA > 60 min/d was higher than that of female students (2.3%), and the proportion of male students (33.8%) and female students (34.5%) with screen time > 2 h/d was basically the same. The generalize linear model (GLM) analysis showed that university students with MVPA < 30 min/d and screen time > 2 h/d (OR = 1.59, 95% CI: 1.10-2.31) had the highest risk of psychological symptoms (OR = 1.59, 95% CI: 1.10-2.31) compared to university students with MVPA > 60 min/d and screen time < 1 h/d as the reference group. The risk of psychological symptoms was the highest among those with MVPA < 30 min/d and screen time > 2 h/d (OR = 1.59,95% CI: 1.10-2.31). In addition, university students with MVPA > 60 min/d and a screen time of 1-2 h/d (OR = 0.09, 95% CI: 0.03-0.25) had the lowest risk of psychological symptoms (P < 0.001). The same trend was observed for both male and female students. CONCLUSION: Chinese university students have a certain proportion of psychological symptom problems, and there is a significant between MVPA and screen time and psychological symptoms, and the same trend exists for both male and female students. Chinese university students should perform MVPA for not less than 60 min a day, and at the same time control the duration of screen time, and screen time should be controlled between 1 and 2 h a day, which has a better promotion effect on psychological health.

Effects of oxygen concentration on the corrosion behavior of high entropy alloy AlCoCrFeNi in simulated deep sea.

In light of the low dissolved oxygen concentration in the deep sea, the corrosion mechanisms of the high entropy alloy (HEA) AlCoCrFeNi in artificial seawater with varying oxygen concentrations (2.0, 4.0, 7.0 mg/L) were studied. As the oxygen concentration decreases, the alloy's free corrosion potential decreases, and at 2.0 mg/L, the corrosion rate is 421 times higher than that at 7.0 mg/L. The corrosion form transforms from pitting to uniform corrosion. The primary reasons for this are the passivation film is thin under low oxygen concentration conditions, as well as the preferential dissolution of the alloy elements Al and Ni due to their high activity and "local acidizing" properties, respectively. In designing a super corrosion-resistant high entropy alloy for use in the deep sea, it is advisable to avoid the use of element Al and to add Ni with caution.

Viral shedding pattern of severe fever with thrombocytopenia syndrome virus in severely ill patients: A prospective, Multicenter cohort study.

Background: Severe fever with thrombocytopenia syndrome (SFTS) is spreading rapidly in Asia. The pathway of SFTS virus shedding from patient and specific use of personal protective equipments (PPEs) against viral transmission have rarely been reported. The study was to determine SFTS virus (SFTSV) shedding pattern from the respiratory, digestive and urinary tract to outside in patients. Methods: Patients were divided into mild and severe groups in three sentinel hospitals for SFTS in Anhui province from April 2020 to October 2022. SFTSV level from blood, throat swabs, fecal/anal swabs, urine and bedside environment swabs of SFTS patients were detected by qRT-PCR. Specific PPEs were applied in healthcare workers contacting with the patients who had oropharyngeal virus shedding and hemorrhagic signs. Results: A total of 189 SFTSV-confirmed patients were included in the study, 54 patients died (case fatality rate, 28.57 %). Positive SFTSV in throat swabs (T-SFTSV), fecal/anal swabs (F-SFTSV) and urine (U-SFTSV) were detected in 121 (64.02 %), 91 (48.15 %) and 65 (34.4 %) severely ill patients, respectively. The levels of T-SFTSV, F-SFTSV and U-SFTSV were positively correlated with the load of SFTSV in blood. We firstly revealed that SFTSV positive rate of throat swabs were correlated with occurrence of pneumonia and case fatality rate of patients (P < 0.0001). Specific precaution measures were applied by healthcare workers in participating cardiopulmonary resuscitation and orotracheal intubation for severely ill patients with positive T-SFTSV, no event of SFTSV human-to-human transmission occurred after application of effective PPEs. Conclusions: Our research demonstrated SFTSV could shed out from blood, oropharynx, feces and urine in severely ill patients. The excretion of SFTSV from these parts was positively correlated with viral load in the blood. Effective prevention measures against SFTSV human-to-human transmission are needed.

Assessment of left ventricular ejection fraction in artificial intelligence based on left ventricular opacification.

Background: Left ventricular opacification (LVO) improves the accuracy of left ventricular ejection fraction (LVEF) by enhancing the visualization of the endocardium. Manual delineation of the endocardium by sonographers has observer variability. Artificial intelligence (AI) has the potential to improve the reproducibility of LVO to assess LVEF. Objectives: The aim was to develop an AI model and evaluate the feasibility and reproducibility of LVO in the assessment of LVEF. Methods: This retrospective study included 1305 echocardiography of 797 patients who had LVO at the Department of Ultrasound Medicine, Union Hospital, Huazhong University of Science and Technology from 2013 to 2021. The AI model was developed by 5-fold cross validation. The validation datasets included 50 patients prospectively collected in our center and 42 patients retrospectively collected in the external institution. To evaluate the differences between LV function determined by AI and sonographers, the median absolute error (MAE), spearman correlation coefficient, and intraclass correlation coefficient (ICC) were calculated. Results: In LVO, the MAE of LVEF between AI and manual measurements was 2.6% in the development cohort, 2.5% in the internal validation cohort, and 2.7% in the external validation cohort. Compared with two-dimensional echocardiography (2DE), the left ventricular (LV) volumes and LVEF of LVO measured by AI correlated significantly with manual measurements. AI model provided excellent reliability for the LV parameters of LVO (ICC > 0.95). Conclusions: AI-assisted LVO enables more accurate identification of the LV endocardium and reduces observer variability, providing a more reliable way for assessing LV function.

The Role of Ferroptosis in Amyotrophic Lateral Sclerosis Treatment.

Amyotrophic lateral sclerosis (ALS) is a rare neurodegenerative disease with a challenging treatment landscape, due to its complex pathogenesis and limited availability of clinical drugs. Ferroptosis, an iron-dependent form of programmed cell death (PCD), stands distinct from apoptosis, necrosis, autophagy, and other cell death mechanisms. Recent studies have increasingly highlighted the role of iron deposition, reactive oxygen species (ROS) accumulation, oxidative stress, as well as systemic Xc- and glutamate accumulation in the antioxidant system in the pathogenesis of amyotrophic lateral sclerosis. Therefore, targeting ferroptosis emerges as a promising strategy for amyotrophic lateral sclerosis treatment. This review introduces the regulatory mechanism of ferroptosis, the relationship between amyotrophic lateral sclerosis and ferroptosis, and the drugs used in the clinic, then discusses the current status of amyotrophic lateral sclerosis treatment, hoping to provide new directions and targets for its treatment.

Zinc-indium-sulfide favors efficient C - H bond activation by concerted proton-coupled electron transfer.

C - H bond activation is a ubiquitous reaction that remains a major challenge in chemistry. Although semiconductor-based photocatalysis is promising, the C - H bond activation mechanism remains elusive. Herein, we report value-added coupling products from a wide variety of biomass and fossil-derived reagents, formed via C - H bond activation over zinc-indium-sulfides (Zn-In-S). Contrary to the commonly accepted stepwise electron-proton transfer pathway (PE-ET) for semiconductors, our experimental and theoretical studies evidence a concerted proton-coupled electron transfer (CPET) pathway. A pioneering microkinetic study, considering the relevant elementary steps of the surface chemistry, reveals a faster C - H activation with Zn-In-S because of circumventing formation of a charged radical, as it happens in PE-ET where it retards the catalysis due to strong site adsorption. For CPET over Zn-In-S, H abstraction, forming a neutral radical, is rate-limiting, but having lower energy barriers than that of PE-ET. The rate expressions derived from the microkinetics provide guidelines to rationally design semiconductor catalysis, e.g., for C - H activation, that is based on the CPET mechanism.

Tuberculosis infection among close contacts of patients with pulmonary tuberculosis in China: a population-based, multicentered study.

OBJECTIVES: Limited information is currently available on the prevalence of and risk factors for tuberculosis infection (TBI) among close contacts of patients with pulmonary TB (PTB) in China. In this study, we estimated the burden of TBI among close contacts using QuantiFERON-TB Gold In-Tube assay (QFT) and identified factors associated with TB transmission among this high-risk population. METHODS: From January 1, 2018 to August 31, 2020, we identified laboratory-confirmed patients with PTB from a population-based, multicentered, cluster-randomized control trial for tuberculosis preventive treatment. Close contacts of these patients were identified, interviewed, and tested using the QFT assay. We estimated TBI prevalence and calculated ORs and 95% CIs for TBI risk factors. RESULTS: A total of 3138 index cases and 8117 close contacts were identified. Of these contacts, 36 had PTB (a prevalence of 443.51 cases/100 000 population). Among the remaining 7986 close contacts; 3124 (39.12%) reported a positive QFT result. QFT positivity was significantly associated with older age (adjusted OR, 1.77; [95% CI, 1.27-2.47], 2.20; [95% CI, 1.59-3.05], and 2.74; [95% CI, 1.96-3.82]) for age groups: 35-44, 45-54, and 55-64, respectively) when compared with a younger age group: 5-14; longer contact duration (adjusted OR, 1.44; 95% CI, 1.22-1.69); and sharing of a bedroom (adjusted OR, 1.39; 95% CI, 1.18-1.65). DISCUSSION: Our findings indicate a high TBI burden among the close contacts of PTB. The results also highlighted that contact tracing and investigation for TBI are necessary and beneficial, particularly for those who are older, have had a longer contact duration, and share a bedroom.

High-Energy Earth-Abundant Cathodes with Enhanced Cationic/Anionic Redox for Sustainable and Long-Lasting Na-Ion Batteries.

Layered iron/manganese-based oxides are a class of promising cathode materials for sustainable batteries due to their high energy densities and earth abundance. However, the stabilization of cationic and anionic redox reactions in these cathodes during cycling at high voltage remain elusive. Here, an electrochemically/thermally stable P2-Na0.67Fe0.3Mn0.5Mg0.1Ti0.1O2 cathode material with zero critical elements is designed for sodium-ion batteries (NIBs) to realize a highly reversible capacity of ≈210 mAh g-1 at 20 mA g-1 and good cycling stability with a capacity retention of 74% after 300 cycles at 200 mA g-1, even when operated with a high charge cut-off voltage of 4.5 V versus sodium metal. Combining a suite of cutting-edge characterizations and computational modeling, it is shown that Mg/Ti co-doping leads to stabilized surface/bulk structure at high voltage and high temperature, and more importantly, enhances cationic/anionic redox reaction reversibility over extended cycles with the suppression of other undesired oxygen activities. This work fundamentally deepens the failure mechanism of Fe/Mn-based layered cathodes and highlights the importance of dopant engineering to achieve high-energy and earth-abundant cathode material for sustainable and long-lasting NIBs.

The lncRNA1-miR6288b-3p-PpTCP4-PpD2 module regulates peach branch number by affecting brassinosteroid biosynthesis.

Branch number is one of the most important agronomic traits of fruit trees such as peach. Little is known about how LncRNA and/or miRNA modules regulate branching through transcription factors. Here, we used molecular and genetic tools to clarify the molecular mechanisms underlying brassinosteroid (BR) altering plant branching. We found that the number of sylleptic branch and BR content in pillar peach ('Zhaoshouhong') was lower than those of standard type ('Okubo'), and exogenous BR application could significantly promote branching. PpTCP4 expressed great differentially comparing 'Zhaoshouhong' with 'Okubo'. PpTCP4 could directly bind to DWARF2 (PpD2) and inhibited its expression. PpD2 was the only one differentially expressed key gene in the path of BR biosynthesis. At the same time, PpTCP4 was identified as a target of miR6288b-3p. LncRNA1 could act as the endogenous target mimic of miR6288b-3p and repress expression of miR6288b-3p. Three deletions and five SNP sites of lncRNA1 promoter were found in 'Zhaoshouhong', which was an important cause of different mRNA level of PpTCP4 and BR content. Moreover, overexpressed PpTCP4 significantly inhibited branching. A novel mechanism in which the lncRNA1-miR6288b-3p-PpTCP4-PpD2 module regulates peach branching number was proposed.

MACG-Net: Multi-axis cross gating network for deformable medical image registration.

Deformable Image registration is a fundamental yet vital task for preoperative planning, intraoperative information fusion, disease diagnosis and follow-ups. It solves the non-rigid deformation field to align an image pair. Latest approaches such as VoxelMorph and TransMorph compute features from a simple concatenation of moving and fixed images. However, this often leads to weak alignment. Moreover, the convolutional neural network (CNN) or the hybrid CNN-Transformer based backbones are constrained to have limited sizes of receptive field and cannot capture long range relations while full Transformer based approaches are computational expensive. In this paper, we propose a novel multi-axis cross grating network (MACG-Net) for deformable medical image registration, which combats these limitations. MACG-Net uses a dual stream multi-axis feature fusion module to capture both long-range and local context relationships from the moving and fixed images. Cross gate blocks are integrated with the dual stream backbone to consider both independent feature extractions in the moving-fixed image pair and the relationship between features from the image pair. We benchmark our method on several different datasets including 3D atlas-based brain MRI, inter-patient brain MRI and 2D cardiac MRI. The results demonstrate that the proposed method has achieved state-of-the-art performance. The source code has been released at https://github.com/Valeyards/MACG.