Genome-wide identification, stress- and hormone-responsive expression characteristics, and regulatory pattern analysis of Scutellaria baicalensis SbSPLs.

SQUAMOSA PROMOTER BINDING PROTEIN-LIKEs (SPLs) encode plant-specific transcription factors that regulate plant growth and development, stress response, and metabolite accumulation. However, there is limited information on Scutellaria baicalensis SPLs. In this study, 14 SbSPLs were identified and divided into 8 groups based on phylogenetic relationships. SbSPLs in the same group had similar structures. Abscisic acid-responsive (ABRE) and MYB binding site (MBS) cis-acting elements were found in the promoters of 8 and 6 SbSPLs. Segmental duplications and transposable duplications were the main causes of SbSPL expansion. Expression analysis based on transcriptional profiling showed that SbSPL1, SbSPL10, and SbSPL13 were highly expressed in roots, stems, and flowers, respectively. Expression analysis based on quantitative real-time polymerase chain reaction (RT‒qPCR) showed that most SbSPLs responded to low temperature, drought, abscisic acid (ABA) and salicylic acid (SA), among which the expression levels of SbSPL7/9/10/12 were significantly upregulated in response to abiotic stress. These results indicate that SbSPLs are involved in the growth, development and stress response of S. baicalensis. In addition, 8 Sba-miR156/157 s were identified, and SbSPL1-5 was a potential target of Sba-miR156/157 s. The results of target gene prediction and coexpression analysis together indicated that SbSPLs may be involved in the regulation of L-phenylalanine (L-Phe), lignin and jasmonic acid (JA) biosynthesis. In summary, the identification and characterization of the SbSPL gene family lays the foundation for functional research and provides a reference for improved breeding of S. baicalensis stress resistance and quality traits.

Learning representation for multiple biological networks via a robust graph regularized integration approach.

Learning node representation is a fundamental problem in biological network analysis, as compact representation features reveal complicated network structures and carry useful information for downstream tasks such as link prediction and node classification. Recently, multiple networks that profile objects from different aspects are increasingly accumulated, providing the opportunity to learn objects from multiple perspectives. However, the complex common and specific information across different networks pose challenges to node representation methods. Moreover, ubiquitous noise in networks calls for more robust representation. To deal with these problems, we present a representation learning method for multiple biological networks. First, we accommodate the noise and spurious edges in networks using denoised diffusion, providing robust connectivity structures for the subsequent representation learning. Then, we introduce a graph regularized integration model to combine refined networks and compute common representation features. By using the regularized decomposition technique, the proposed model can effectively preserve the common structural property of different networks and simultaneously accommodate their specific information, leading to a consistent representation. A simulation study shows the superiority of the proposed method on different levels of noisy networks. Three network-based inference tasks, including drug-target interaction prediction, gene function identification and fine-grained species categorization, are conducted using representation features learned from our method. Biological networks at different scales and levels of sparsity are involved. Experimental results on real-world data show that the proposed method has robust performance compared with alternatives. Overall, by eliminating noise and integrating effectively, the proposed method is able to learn useful representations from multiple biological networks.

Forkhead Box S1 mediates epithelial-mesenchymal transition through the Wnt/ß-catenin signaling pathway to regulate colorectal cancer progression.

BACKGROUND: Recent studies have shown that the fox family plays a vital role in tumorigenesis and progression. Forkhead Box S1 (FOXS1), as a newly identified subfamily of the FOX family, is overexpressed in certain types of malignant tumors and closely associated with patient's prognosis. However, the role and mechanism of the FOXS1 in colorectal cancer (CRC) remain unclear. METHOD: FOXS1 level in CRC tissues and cell lines was analyzed by western blot and quantitative real-time polymerase chain reaction (qRT-PCR). Immunohistochemistry (IHC) was used to detect the relationship between FOXS1 expression and clinicopathological features in 136 patients in our unit. The expression of FOXS1 was knocked down in CRC cells using small interfering RNA (siRNA) technology. Cell proliferation was assessed by CCK8 assay, colony formation, and 5-Ethynyl-20-deoxyuridine (EdU) incorporation assay. Flow cytometry detected apoptosis and wound healing, and Transwell assays determined cell migration and invasion. Western blotting was used to detect the levels of proteins associated with the Wnt/ß-catenin signaling pathway. Then, we used short hairpin RNA (shRNA) to knock down FOXS1 to see the effect of FOXS1 on the proliferation, migration, invasion, and metastasis of CRC cells in vivo. Finally, we investigated the impact of Wnt activator LiCl on the proliferation, migration, invasion, and metastasis of CRC cells after FOXS1 knockdown. RESULT: Compared to those in normal groups, FOXS1 overexpressed in CRC tissues and CRC cells (P < 0.05). Upregulation of FOXS1 association with poor prognosis of CRC patients. si-FOXS1 induced apoptosis and inhibited proliferation, migration, invasion, the epithelial-mesenchymal transition (EMT), and the Wnt/ß-catenin signaling pathway in vitro; sh-FOXS1 inhibited the volume and weight of subcutaneous xenografts and the number of lung metastases in vivo. LiCl, an activator of Wnt signaling, partially reversed the effect of FOXS1 overexpression on CRC cells. CONCLUSION: FOXS1 could function as an oncogene and promote CRC cell proliferation, migration, invasion and metastasis through the Wnt/ßcatenin signaling pathway, FOXS1 may be a potential target for CRC treatment.

Biomechanical study on the stability and strain conduction of intertrochanteric fracture fixed with proximal femoral nail antirotation versus triangular supporting intramedullary nail.

OBJECTIVES: Based on the features of the three-dimensional spatial structure of the proximal femoral trabeculae, we developed a bionic triangular supporting intramedullary nail (TSIN) for the treatment of the femoral intertrochanteric fracture. The current study aimed to compare the mechanical stability and restoration of mechanical conduction between proximal femoral nail antirotation (PFNA) and TSIN to fix the intertrochanteric fractures. METHODS: Firstly, five sets of PFNA and TSIN with the same size were selected and fixed on a biomechanical testing machine, and strain gauges were pasted on the main nail, lag screw, and supporting screw to load to the vertical load to 600 N, and the displacement and strain values were recorded. Secondly, formalin-preserved femurs were selected, and the left and right femurs of the same cadaver were randomly divided into two groups to prepare intertrochanteric femur fractures (AO classification 31-A1), which were fixed with PFNA (n = 15) and TSIN (n = 15), respectively. Sixteen sites around the fracture line were chosen to paste strain gauges and loaded vertically to 600 N, and then, the fracture fragment displacement and strain values were recorded. Finally, a 10,000-cycle test ranging from 10 to 600 N was conducted, and the cycle number and displacement value were recorded. RESULTS: The overall displacement of PFNA was 2.17 ± 0.18 mm, which was significantly greater than the displacement of the TSIN group (1.66 ± 0.05 mm, P < 0.05) under a vertical load of 600 N. The strain below the PFNA lag screw was 868.29 ± 147.85, which was significantly greater than that of the TSIN (456.02 ± 35.06, P < 0.05); the strain value at the medial side of the PFNA nail was 444.00 ± 34.23, which was significantly less than that of the TSIN (613.57 ± 108.00, P < 0.05). Under the vertical load of 600 N, the displacement of the fracture fragments of the PFNA group was 0.95 ± 0.25 mm, which was significantly greater than that of the TSIN group (0.41 ± 0.09 mm, P < 0.05). The femoral specimens in the PFNA group showed significantly greater strains at the anterior (1, 2, and 4), lateral (7, 9, and 10), posterior (11), and medial (15 and 16) sites than those in the TSIN group (all P < 0.05). In the cyclic compression experiment, the displacements of the PFNA group at 2000, 4000, 6000, 8000, and 10,000 cycles were 1.38 mm, 1.81 mm, 2.07 mm, 2.64 mm, and 3.58 mm, respectively, which were greater than the corresponding displacements of the TSIN group: 1.01 mm, 1.48 mm, 1.82 mm, 2.05 mm, and 2.66 mm (P8000 = 0.012, P10000 = 0.006). CONCLUSIONS: The current study showed that TSIN had apparent advantages in stability and stress conduction. TSIN enhanced the stability of intertrochanteric fractures, particularly in superior fracture fragments, improved stress conduction, reduced the stress in the anterior and medial femur, and restored the biomechanical properties of the femur.

Finite element analysis of biomechanical effects of residual varus/valgus malunion after femoral fracture on knee joint.

OBJECTIVE: Post-operative femoral shaft fractures are often accompanied by a residual varus/valgus deformity, which can result in osteoarthritis in severe cases. The purpose of this study was to investigate the biomechanical effects of residual varus/valgus deformities after middle and lower femoral fracture on the stress distribution and contact area of knee joint. METHODS: Thin-slice CT scanning of lower extremities and MRI imaging of knee joints were obtained from a healthy adult male to establish normal lower limb model (neutral position). Then, the models of 3°, 5°, and 10° of varus/valgus were established respectively by modifying middle and lower femur of normal model. To validate the modifying, a patient-specific model, whose BMI was same to former and had 10° of varus deformity of tibia, was built and simulated under the same boundary conditions. RESULT: The contact area and maximum stress of modified models were similar to those of patient-specific model. The contact area and maximum stress of medial tibial cartilage in normal neutral position were 244.36 mm2 and 0.64 MPa, while those of lateral were 196.25 mm2 and 0.76 MPa. From 10° of valgus neutral position to 10° of varus, the contact area and maximum stress of medial tibial cartilage increased, and the lateral gradually decreased. The contact area and maximum stress of medial meniscus in normal neutral position were 110.91 mm2 and 3.24 MPa, while those of lateral were 135.83 mm2 and 3.45 MPa. The maximum stress of medial tibia subchondral bone in normal neutral position was 1.47 MPa, while that of lateral was 0.65 MPa. The variation trend of medial/lateral meniscus and subchondral bone was consistent with that of tibial plateau cartilage in the contact area and maximum stress. CONCLUSION: This study suggested that varus/valgus deformity of femur had an obvious effect on the contact area and stress distribution of knee joint, providing biomechanical evidence and deepening understanding when performing orthopedic trauma surgery or surgical correction of the already existing varus/valgus deformity.

The renal protect function of erythropoietin after release of bilateral ureteral obstruction in a rat model.

Congenital urinary tract obstruction is one of the most frequent malformations in fetuses or neonates, which usually causes profound impairment of renal function including reductions in both glomerular filtration rate (GFR) and tubular handling of water and solutes. Although obstruction can be released by surgical operation, the child will suffer from diuresis for sometime. It has been reported that erythropoietin (EPO) could prevent the down-regulation of aquaporin-2 (AQP2) and urinary-concentrating defects induced by renal ischemia/reperfusion (I/R) injury. However, whether EPO could promote the recovery of renal function and AQP2 expression after releasing of ureteral obstruction has not been reported yet. The purposes of the present study were to investigate the effects of EPO on renal function and AQP2 expression after release of bilateral ureteral obstruction (BUO-R) in rats. The results showed that EPO could promote interleukin (IL) 10 (IL-10) expression; inhibit tumor necrosis factor-α (TNF-α), IL-6, and inducible nitric oxide synthase (iNOS) expressions; reduce the fractional excretion of sodium (FENa) and plasma creatinine (CREA) and urea; and promote the recovery of water and salt handling and AQP2 expression in BUO-R rats, especially in the high dose of EPO-treated group rats. In conclusion, EPO could promote the recovery of renal function and AQP2 expression in BUO-R rats, which may partially associate with its anti-inflammation effect.

Decreased expression of SOX17 is associated with tumor progression and poor prognosis in breast cancer.

The SOX17 (SRY-related HMG-box) transcription factor is involved in a variety of biological processes and is related to the tumorigenesis and progression of multiple tumors. However, the clinical application of SOX17 for breast cancer prognosis is currently limited. The aim of this study was to investigate the clinicopathologic and prognostic significance of SOX17 expression in human breast cancer. qPCR and western blot assays were performed to measure the expression of SOX17 in breast cancer cell lines and 30 matched pairs of breast cancer and corresponding noncancerous tissues. A SOX17 overexpression cell model was used to examine changes in cell growth in vitro. Immunohistochemical analyses were performed to retrospectively examine the prognostic impact of SOX17 expression in 187 additional breast cancer patients. Our results showed that SOX17 expression was decreased at both the messenger RNA (mRNA) and protein levels in the breast cancer cell lines and tissues, and that SOX17 overexpression could strongly suppress cell growth in vitro. Furthermore, the lack of SOX17 protein expression was strongly correlated with higher tumor grade (P = 0.002), lymph node metastasis (P < 0.001), and tumor node metastasis (TNM) stage (P = 0.001) and had poorer disease-free survival (DFS) and overall survival (OS) compared to normal SOX17 expression (P = 0.002 and 0.001, respectively). Univariate and multivariate analyses indicated that lower SOX17 expression was an independent prognostic factor for DFS (P = 0.007; HR = 2.854; 95 % CI 1.326-6.147) and OS (P = 0.005; HR = 5.035; 95 % CI 1.648-15.385) for breast cancer. Our findings indicate that SOX17 expression is a useful prognostic biomarker for breast cancer.

Highly enantioselective [3+2] cycloaddition of vinylcyclopropane with nitroalkenes catalyzed by palladium(0) with a chiral bis(tert-amine) ligand.

An enantioselective [3+2] cycloaddition of vinyl cyclopropane derived from 1,3-indanedione with nitroalkenes catalyzed by palladium(0) with a chiral bis(tert-amine) ligand was developed in high yields with good diastereoselectivities and excellent enantioselectivities. The resulting bis(tert-amine)-palladium complex proved to be a highly efficient catalyst for this cycloaddition.

Fe(III)-Catalyzed Cross-Dehydrogenative Arylation (CDA) between Oxindoles and Arenes under an Air Atmosphere.

An efficient Fe(III)-catalyzed cross-dehydrogenative arylation (CDA) of 3-substituted oxindoles with activated arenes under an air atmosphere was developed to provide 3,3'-disubstituted oxindoles in good yields.

Copper(II)-catalyzed highly regio- and stereo-selective hydrosilylation of unactivated internal alkynes with silylborate in water.

The highly regio- and stereoselective hydrosilylation of internal alkynes with silylborate catalyzed by Cu(OTf)2 with 1,10-phenanthroline as the ligand in the presence of Cs2CO3 in water is developed. This protocol was applied efficiently in the aqueous synthesis of multi-substituted vinylsilanes.

Susceptibility-weighted imaging of the venous networks around the brain stem.

INTRODUCTION: The venous network of the brainstem is complex and significant. Susceptibility-weighted imaging (SWI) is a practical technique which is sensitive to veins, especially tiny veins. Our purpose of this study was to evaluate the visualization of the venous network of brainstem by using SWI at 3.0 T. METHODS: The occurrence rate of each superficial veins of brainstem was evaluated by using SWI on a 3 T MR imaging system in 60 volunteers. The diameter of the lateral mesencephalic vein and peduncular vein were measured by SWI using the reconstructed mIP images in the sagittal view. And the outflow of the veins of brainstem were studied and described according to the reconstructed images. RESULTS: The median anterior pontomesencephalic vein, median anterior medullary vein, peduncular vein, right vein of the pontomesencephalic sulcus, and right lateral anterior pontomesencephalic vein were detected in all the subjects (100%). The outer diameter of peduncular vein was 1.38 ± 0.26 mm (range 0.8-1.8 mm). The lateral mesencephalic vein was found in 75% of the subjects and the mean outer diameter was 0.81 ± 0.2 mm (range 0.5-1.2 mm). The inner veins of mesencephalon were found by using SWI. CONCLUSION: The venous networks around the brain stem can be visualized by SWI clearly. This result can not only provide data for anatomical study, but also may be available for the surgical planning in the infratentorial region.

Telerehabilitation Approaches for Stroke Patients: Systematic Review and Meta-analysis of Randomized Controlled Trials.

BACKGROUND: Stroke remains one of the most common causes of adult disability in the world. In recent years, diverse telerehabilitation programs have been conceived and studied to improve the abilities of the activities of daily living and increased independence of stroke patients living at home. The systematic review was conducted to determine whether telerehabilitation leads to an improvement in abilities of activities of daily living for stroke patients. METHODS: Randomized controlled trials (RCTs) evaluating the effects of telerehabilitation in stroke survivors living at home were identified by searching 7 electronic databases from inception to March 2015, and by hand searching for conference literatures between 2000 and 2015. Assessments of risk bias and data extraction were conducted independently by 2 reviews. RESULTS: The search strategy identified 2587 records, of which 11 studies were thought to be eligible. Pooled results from 7 studies showed no significant differences in abilities of activities of daily living (Barthel Index scale: standardized mean difference [SMD] -.05, 95% confidence interval [CI] -.24 to .13; Berg Balance Scale: SMD -.05, 95% CI -.7 to .37) and motor function (Fugl-Meyer Extremity: SMD .05, 95% CI -.09 to 1.09) between groups. CONCLUSIONS: This review provides limited, moderate evidence that telerehabilitation of all approaches has equal effects with conventional rehabilitation in improving abilities of activities of daily living and motor function for stroke survivors. Further research of RCTs in this area (rehabilitation field of telemedicine) is ungently required to extend the evidence base.

Identification of the immunoproteome of the meningococcus by cell surface immunoprecipitation and MS.

Most healthy adults are protected from meningococcal disease by the presence of naturally acquired anti-meningococcal antibodies; however, the identity of the target antigens of this protective immunity remains unclear, particularly for protection against serogroup B disease. To identify the protein targets of natural protective immunity we developed an immunoprecipitation and proteomics approach to define the immunoproteome of the meningococcus. Sera from 10 healthy individuals showing serum bactericidal activity against both a meningococcal C strain (L91543) and the B strain MC58, together with commercially available pooled human sera, were used as probe antisera. Immunoprecipitation was performed with each serum sample and live cells from both meningococcal strains. Immunoprecipitated proteins were identified by MS. Analysis of the immunoproteome from each serum demonstrated both pan-reactive antigens that were recognized by most sera as well as subject-specific antigens. Most antigens were found in both meningococcal strains, but a few were strain-specific. Many of the immunoprecipitated proteins have been characterized previously as surface antigens, including adhesins and proteases, several of which have been recognized as vaccine candidate antigens, e.g. factor H-binding protein, NadA and neisserial heparin-binding antigen. The data demonstrate clearly the presence of meningococcal antibodies in healthy individuals with no history of meningococcal infection and a wide diversity of immune responses. The identification of the immunoreactive proteins of the meningococcus provides a basis for understanding the role of each antigen in the natural immunity associated with carriage and may help to design vaccination strategies.

Dynamic changes and clinical significance of serum tryptase levels in STEMI patients treated with primary PCI.

OBJECTIVE: To investigate the dynamic changes in serum tryptase levels and their association with clinical data in patients with acute ST-elevation myocardial infarction (STEMI) treated with primary percutaneous coronary intervention (PCI). METHODS: Serum tryptase levels were measured in 99 STEMI patients and 25 control subjects. RESULTS: Tryptase levels were significantly increased at admission, and descended after primary PCI. Tryptase levels at 0.5, 2 and 6 h after PCI were negatively correlated with the percentage of ST-segment resolution (STR) as well as left ventricular ejection fraction (LVEF). CONCLUSIONS: High tryptase levels after PCI were associated with poor myocardial reperfusion and poor cardiac function.

Highly enantioselective reaction of 2-oxindoles with (3-indolyl)methanols by cooperative catalysis of a Lewis acid and organocatalyst.

An efficient cooperative biscinchona alkaloid and Lewis acid catalytic system was developed in the enantioselective α-alkylation of 2-oxindoles with (3-indolyl)(phenyl)methanols to provide (2-oxindole)-linker-indole derivatives in good yields (70-83%) with high enantioselectivities (81%-92%).

Free voiding patterns in preterm and full-term newborn infants are different between males and females.

AIM: The neonatal period is critical in bladder development, encompassing the transition from foetal bladder contractions to voluntary infant urination. The aim of this study was to investigate different voiding parameters between male and female newborn infants. METHODS: We studied 102 healthy, single birth newborn infants - 54 preterm and 48 full-term - without lower urinary tract diseases, hospitalised in the neonatal intensive care unit from March 2011 to March 2012. Free voiding was observed from 9 a.m. to 9 p.m., and the free voiding parameters and fluid intake were recorded and compared between male and female newborn infants using the Student's t-test and chi-square test. RESULTS: Male preterm newborns demonstrated larger mean postvoid residual volumes and lower bladder emptying rates than female preterm newborns (p < 0.05), and male full-term newborns had lower bladder emptying rates than female full-term newborns (p < 0.05). The bladder emptying rates of newborns defecating simultaneously with voiding were not statistically different between males and females of the same gestational age (p > 0.05). CONCLUSION: Male newborns were more likely to have larger postvoid residual volumes than females, and defecating simultaneously with voiding may promote bladder emptying in male newborns.

When Invariant Representation Learning Meets Label Shift: Insufficiency and Theoretical Insights.

As a crucial step toward real-world learning scenarios with changing environments, dataset shift theory and invariant representation learning algorithm have been extensively studied to relax the identical distribution assumption in classical learning setting. Among the different assumptions on the essential of shifting distributions, generalized label shift (GLS) is the latest developed one which shows great potential to deal with the complex factors within the shift. In this paper, we aim to explore the limitations of current dataset shift theory and algorithm, and further provide new insights by presenting a comprehensive understanding of GLS. From theoretical aspect, two informative generalization bounds are derived, and the GLS learner are proved to be sufficiently close to optimal target model from the Bayesian perspective. The main results show the insufficiency of invariant representation learning, and prove the sufficiency and necessity of GLS correction for generalization, which provide theoretical supports and innovations for exploring generalizable model under dataset shift. From methodological aspect, we provide a unified view of existing shift correction frameworks, and propose a kernel embedding-based correction algorithm (KECA) to minimize the generalization error and achieve successful knowledge transfer. Both theoretical results and extensive experiment evaluations demonstrate the sufficiency and necessity of GLS correction for addressing dataset shift and the superiority of proposed algorithm.

Geometric Understanding of Discriminability and Transferability for Visual Domain Adaptation.

To overcome the restriction of identical distribution assumption, invariant representation learning for unsupervised domain adaptation (UDA) has made significant advances in computer vision and pattern recognition communities. In UDA scenario, the training and test data belong to different domains while the task model is learned to be invariant. Recently, empirical connections between transferability and discriminability have received increasing attention, which is the key to understand the invariant representations. However, theoretical study of these abilities and in-depth analysis of the learned feature structures are unexplored yet. In this work, we systematically analyze the essentials of transferability and discriminability from the geometric perspective. Our theoretical results provide insights into understanding the co-regularization relation and prove the possibility of learning these abilities. From methodology aspect, the abilities are formulated as geometric properties between domain/cluster subspaces (i.e., orthogonality and equivalence) and characterized as the relation between the norms/ranks of multiple matrices. Two optimization-friendly learning principles are derived, which also ensure some intuitive explanations. Moreover, a feasible range for the co-regularization parameters is deduced to balance the learning of geometric structures. Based on the theoretical results, a geometry-oriented model is proposed for enhancing the transferability and discriminability via nuclear norm optimization. Extensive experiment results validate the effectiveness of the proposed model in empirical applications, and verify that the geometric abilities can be sufficiently learned in the derived feasible range.

A Controlled Variable Study of the Biomechanical Properties of the Proximal Femur before and after Cancellous Bone Removal.

OBJECTIVE: The biomechanical characteristics of proximal femoral trabeculae are closely related to the occurrence and treatment of proximal femoral fractures. Therefore, it is of great significance to study its biomechanical effects of cancellous bone in the proximal femur. This study examines the biomechanical effects of the cancellous bone in the proximal femur using a controlled variable method, which provide a foundation for further research into the mechanical properties of the proximal femur. METHODS: Seventeen proximal femoral specimens were selected to scan by quantitative computed tomography (QCT), and the gray values of nine regions were measure to evaluated bone mineral density (BMD) using Mimics software. Then, an intact femur was fixed simulating unilateral standing position. Vertical compression experiments were then performed again after removing cancellous bone in the femoral head, femoral neck, and intertrochanteric region, and data were recorded. According to the controlled variable method, the femoral head, femoral neck, and intertrochanteric trabeculae were sequentially removed based on the axial loading of the intact femur, and the displacement and strain changes of the femur samples under axial loading were recorded. Gom software was used to measure and record displacement and strain maps of the femoral surface. RESULTS: There was a statistically significant difference in anteroposterior displacement of cancellous bone destruction in the proximal femur (p < 0.001). Proximal femoral bone mass explained 77.5% of the strength variation, in addition proximal femoral strength was mainly affected by bone mass at the level of the upper outer, lower inner, lower greater trochanter, and lesser trochanter of the femoral head. The normal stress conduction of the proximal femur was destroyed after removing cancellous bone, the stress was concentrated in the femoral head and lateral femoral neck, and the femoral head showed a tendency to subside after destroying cancellous bone. CONCLUSION: The trabecular removal significantly altered the strain distribution and biomechanical strength of the proximal femur, demonstrating an important role in supporting and transforming bending moment under the vertical load. In addition, the strength of the proximal femur mainly depends on the bone density of the femoral head and intertrochanteric region.

Cross-site prognosis prediction for nasopharyngeal carcinoma from incomplete multi-modal data.

Accurate prognosis prediction for nasopharyngeal carcinoma based on magnetic resonance (MR) images assists in the guidance of treatment intensity, thus reducing the risk of recurrence and death. To reduce repeated labor and sufficiently explore domain knowledge, aggregating labeled/annotated data from external sites enables us to train an intelligent model for a clinical site with unlabeled data. However, this task suffers from the challenges of incomplete multi-modal examination data fusion and image data heterogeneity among sites. This paper proposes a cross-site survival analysis method for prognosis prediction of nasopharyngeal carcinoma from domain adaptation viewpoint. Utilizing a Cox model as the basic framework, our method equips it with a cross-attention based multi-modal fusion regularization. This regularization model effectively fuses the multi-modal information from multi-parametric MR images and clinical features onto a domain-adaptive space, despite the absence of some modalities. To enhance the feature discrimination, we also extend the contrastive learning technique to censored data cases. Compared with the conventional approaches which directly deploy a trained survival model in a new site, our method achieves superior prognosis prediction performance in cross-site validation experiments. These results highlight the key role of cross-site adaptability of our method and support its value in clinical practice.