Dual ensemble system for polyp segmentation with submodels adaptive selection ensemble.

Colonoscopy is one of the main methods to detect colon polyps, and its detection is widely used to prevent and diagnose colon cancer. With the rapid development of computer vision, deep learning-based semantic segmentation methods for colon polyps have been widely researched. However, the accuracy and stability of some methods in colon polyp segmentation tasks show potential for further improvement. In addition, the issue of selecting appropriate sub-models in ensemble learning for the colon polyp segmentation task still needs to be explored. In order to solve the above problems, we first implement the utilization of multi-complementary high-level semantic features through the Multi-Head Control Ensemble. Then, to solve the sub-model selection problem in training, we propose SDBH-PSO Ensemble for sub-model selection and optimization of ensemble weights for different datasets. The experiments were conducted on the public datasets CVC-ClinicDB, Kvasir, CVC-ColonDB, ETIS-LaribPolypDB and PolypGen. The results show that the DET-Former, constructed based on the Multi-Head Control Ensemble and the SDBH-PSO Ensemble, consistently provides improved accuracy across different datasets. Among them, the Multi-Head Control Ensemble demonstrated superior feature fusion capability in the experiments, and the SDBH-PSO Ensemble demonstrated excellent sub-model selection capability. The sub-model selection capabilities of the SDBH-PSO Ensemble will continue to have significant reference value and practical utility as deep learning networks evolve.

CMNet: deep learning model for colon polyp segmentation based on dual-branch structure.

Purpose: Colon cancer is one of the top three diseases in gastrointestinal cancers, and colon polyps are an important trigger of colon cancer. Early diagnosis and removal of colon polyps can avoid the incidence of colon cancer. Currently, colon polyp removal surgery is mainly based on artificial-intelligence (AI) colonoscopy, supplemented by deep-learning technology to help doctors remove colon polyps. With the development of deep learning, the use of advanced AI technology to assist in medical diagnosis has become mainstream and can maximize the doctor's diagnostic time and help doctors to better formulate medical plans. Approach: We propose a deep-learning model for segmenting colon polyps. The model adopts a dual-branch structure, combines a convolutional neural network (CNN) with a transformer, and replaces ordinary convolution with deeply separable convolution based on ResNet; a stripe pooling module is introduced to obtain more effective information. The aggregated attention module (AAM) is proposed for high-dimensional semantic information, which effectively combines two different structures for the high-dimensional information fusion problem. Deep supervision and multi-scale training are added in the model training process to enhance the learning effect and generalization performance of the model. Results: The experimental results show that the proposed dual-branch structure is significantly better than the single-branch structure, and the model using the AAM has a significant performance improvement over the model not using the AAM. Our model leads 1.1% and 1.5% in mIoU and mDice, respectively, when compared with state-of-the-art models in a fivefold cross-validation on the Kvasir-SEG dataset. Conclusions: We propose and validate a deep learning model for segmenting colon polyps, using a dual-branch network structure. Our results demonstrate the feasibility of complementing traditional CNNs and transformer with each other. And we verified the feasibility of fusing different structures on high-dimensional semantics and successfully retained the high-dimensional information of different structures effectively.

Tri-branch feature pyramid network based on federated particle swarm optimization for polyp segmentation.

Deep learning technology has shown considerable potential in various domains. However, due to privacy issues associated with medical data, legal and ethical constraints often result in smaller datasets. The limitations of smaller datasets hinder the applicability of deep learning technology in the field of medical image processing. To address this challenge, we proposed the Federated Particle Swarm Optimization algorithm, which is designed to increase the efficiency of decentralized data utilization in federated learning and to protect privacy in model training. To stabilize the federated learning process, we introduced Tri-branch feature pyramid network (TFPNet), a multi-branch structure model. TFPNet mitigates instability during the aggregation model deployment and ensures fast convergence through its multi-branch structure. We conducted experiments on four different public datasets：CVC-ClinicDB, Kvasir, CVC-ColonDB and ETIS-LaribPolypDB. The experimental results show that the Federated Particle Swarm Optimization algorithm outperforms single dataset training and the Federated Averaging algorithm when using independent scattered data, and TFPNet converges faster and achieves superior segmentation accuracy compared to other models.

Manipulation of Temperature-Dependent Luminescence Behaviors of Mn4+-Activated Fluoride Phosphors.

In this paper, the giant tunability of thermal behaviors, i.e., from thermal deterioration to substantial growth, is firmly demonstrated for the vibronic luminescence of Mn4+ ions in fluoride phosphors. Such peculiar behavior is uncovered to be associated with the thermal excitation of a low-frequency phonon bath, and a theoretical model involving the excitation-wavelength-dependent populations of vibronic levels and the temperature-dependent nonradiative recombination processes is successfully constructed. Two main governing parameters, namely, the thermal activation energy Ea and the involved average phonon energy ΔE, are thus determined for the distinct thermal behaviors of Mn4+-ion luminescence. This demonstration may pave the way for manipulating the thermal behaviors of vibronic luminescence in solids to some extent.

Monolayer MXene Nanoelectromechanical Piezo-Resonators with 0.2 Zeptogram Mass Resolution.

2D materials-based nanoelectromechanical resonant systems with high sensitivity can precisely trace quantities of ultra-small mass molecules and therefore are broadly applied in biological analysis, chemical sensing, and physical detection. However, conventional optical and capacitive transconductance schemes struggle to measure high-order mode resonant effectively, which is the scientific key to further achieving higher accuracy and lower noise. In the present study, the different vibrations of monolayer Ti3 C2 Tx MXene piezo-resonators are investigated, and achieve a high-order f2,3 resonant mode with a ≈234.59 ± 0.05 MHz characteristic peak due to the special piezoelectrical structure of the Ti3 C2 Tx MXene layer. The effective measurements of signals have a low thermomechanical motion spectral density (9.66 ± 0.01  fmHz$\frac{{fm}}{{\sqrt {Hz} }}$ ) and an extensive dynamic range (118.49 ± 0.42 dB) with sub-zeptograms resolution (0.22 ± 0.01 zg) at 300 K temperature and 1 atm. Furthermore, the functional groups of the Ti3 C2 Tx MXene with unique adsorption properties enable a high working range ratio of ≈3100 and excellent repeatability. This Ti3 C2 Tx MXene device demonstrates encouraging performance advancements over other nano-resonators and will lead the related engineering applications including high-sensitivity mass detectors.

Recent advances in MXene-based force sensors: a mini-review.

As an emerging two-dimensional (2D) material, MXene has excellent conductivity and abundant surface functional groups. Its unique layered structure, large surface area, and prominent hydrophilicity show remarkable performances, which allow abundant possibilities to work as the sensing element alone or combined with other auxiliary materials. As a senior member of MXenes, Ti3C2T x has shown great potential in the development of force sensors. The research development of force sensors based on Ti3C2T x MXene is reviewed in this paper, presenting the advanced development of force sensors in various forms and summaring their current preparation strategies and characteristics. In addition, the corresponding challenges and prospects of the MXene-based sensors are also discussed for future research.

Mendelian randomization study of telomere length and lung cancer risk in East Asian population.

Associations between telomere length and cancer risk have been investigated in many epidemiological studies, but the results are controversial. These associations may be biased by reverse causation or confounded by environmental exposures. To avoid potential biases, we used Mendelian randomization method to evaluate whether TL is the causal risk factor for lung cancer. We conducted Mendelian randomization analysis in two published East Asian GWAS studies (7127 cases and 6818 controls). We used both weighted genetic risk score and inverse-variance weighting method to estimate the relationship between TL and lung cancer risk. Nonlinear test also used to detect potential association trends. We observed that increased weight GRS was associated with increased risk of lung cancer (OR = 2.25, 95%CI: 1.81-2.78, P = 1.18 × 10-13 ). In different subtypes, weight GRS was significantly associated with lung adenocarcinoma risk (OR = 2.69, 95% CI: 2.11-3.42, P = 7.20 × 10-16 ); while lung squamous cell carcinoma showed a marginal association (OR = 1.45, 95% CI = 1.01-2.10, P = .047). Nonlinear analysis suggested a log-linear dose-response relationship between increased weight GRS and lung cancer risk. Our results indicated that longer TL increases lung cancer risk. Those biological mechanisms changes caused by long TL may play an important role in lung carcinogenesis.

Correlation between secreted protein acidic and rich in cysteine protein expression and the prognosis of postoperative patients exhibiting esophageal squamous cell carcinoma.

The aim of the present study was to investigate the association between the expression level of secreted protein acidic and rich in cysteine (SPARC) and the prognosis of postoperative patients with esophageal squamous cell carcinoma (ESCC). The expression level of SPARC was detected in the 89 ESCC tissue cases and 100 healthy esophageal mucosa cases, which served as the controls. Immunohistochemistry and reverse transcription­polymerase chain reaction (RT­PCR) were employed to evaluate the SPARC expression in cases with ESCC. RT­PCR demonstrated that the positive rates of SPARC mRNA expression in ESCC were 71.91% (64/89). The positive rates of normal esophageal mucosa mRNA expression were 15.00% (15/100), which were significantly lower than that in the ESCC tissue samples. The difference was statistically significant (P<0.001). Immunohistochemical staining indicated that the positive expression rate of SPARC protein in the ESCC tissue samples was significantly higher than that in the esophageal mucosa tissue samples (65.17 vs. 8.00%; P<0.001). The expression of SPARC protein was negatively correlated with lymph node metastasis (P<0.05), which was not associated with the pathologic gross morphology, tumor differentiation degree or other clinical features. The survival of patients with ESCC was not associated with the expression level of SPARC protein (P>0.05), but was associated with the tumor location (P<0.05), differentiation (P<0.001) and staging (P<0.05). Thus, SPARC mRNA and protein were highly expressed in ESCC, and negatively correlated with lymph node metastasis, which was not associated with postoperative survival of ESCC patients. Thus, detection of SPARC mRNA and protein expression levels may facilitate early diagnosis and prognosis assessment of ESCC.

Insights into ordered microstructures and ordering mechanisms of ABC star terpolymers by integrating dynamic self-consistent field theory and variable cell shape methods.

A theoretical approach coupling dynamic self-consistent field (SCF) theory for inhomogeneous polymeric fluids and variable cell shape (VCS) method for automatically adjusting cell shape and size is developed to investigate ordered microstructures and the ordering mechanisms of block copolymer melts. Using this simulation method, we first re-examined the microphase separation of the simplest AB diblock copolymers, and tested the validity and efficiency of the novel method by comparing the results with those obtained from the dynamic SCF theory. An appropriate relaxation parameter of the VCS method effectively accelerates the system towards a stable morphology without distortions or defects. The dynamic SCF/VCS method is then applied to identify the richness morphologies of ABC star terpolymers and explore the ordering mechanisms of star terpolymer melts quenched from homogenous states. A diverse range of ordered microstructures, including two-dimensional tiling patterns, hierarchical structures and ordinary microstructures, are predicted. Three types of ordering mechanisms, namely, one-step, quick-slow and step-wise procedures, are discovered in the disorder-to-order transition of ABC star terpolymers. The procedures of microphase separation in the ABC star terpolymer melts are remarkably affected by the composition of star terpolymers and the strength of interaction parameters.

Polypeptide-based "smart" micelles for dual-drug delivery: a combination study of experiments and simulations.

A dual-drug-loaded micelle is designed and constructed from a mixture of poly(propylene oxide)-b-poly(γ-benzyl-l-glutamate)-b-poly(ethylene glycol) (PPO-b-PBLG-b-PEG) triblock terpolymers and two model drugs, doxorubicin (DOX) and naproxen (Nap). In the micelles, the DOX is chemically linked to the PBLG backbones through an acid-cleavable hydrazone bond, whereas the Nap is physically encapsulated in the cores. The drug loading and releasing behaviors of the dual-drug-loaded micelles as well as single drug-loaded micelles (DOX-conjugated or Nap-loaded micelles) are studied. The structures of micelles are characterized by means of microscopies and dynamic light scattering, and further examined by dissipative particle dynamics (DPD) simulations. It is revealed that the micelles possess a core-shell-corona structure in which the PPO/Nap, PBLG/DOX, and PEG aggregate to form the core, shell, and corona, respectively. In vitro studies reveal that the release of DOX and Nap is pH- and thermosensitive. Such drug releasing behaviors are also examined by DPD simulations, and more information regarding the mechanism is obtained. In addition, the bio-related properties such as cellular uptake of the micelles and biocompatibility of the deliveries are evaluated. The results show that the dual-drug-loaded micelles are biocompatible at normal physiological conditions and retain the anti-cancer efficiency.

[Effects of transection of cervical spinal cord on lipopolysaccharide induced acute lung injury in rat].

OBJECTIVE: To investigate the effects of transection of cervical spinal cord (TCSC) on acute lung injury (ALI) and its potential mechanism in rat. METHODS: Seventy-two rats were randomly divided into three groups: normal control group (NC, n=8), endotoxemia group (ET, n=32) and endotoxemia with TCSC group (TCSC, n=32), and the latter two groups were divided into four subgroups respectively according to different time intervals (n=8). Endotoxemia model was established by injecting lipopolysaccharide (LPS, 10 mg/kg) intravenously, and the spinal cord at 7th cervical spine of rats was transected in TCSC group. Samples of blood and lung were collected at different time intervals. The plasma levels of norepinephrine (NE) and interleukin-6 (IL-6) were measured by high performance liquid chromatography (HPLC) and enzyme-lined immunosorbent assay (ELISA) respectively, and arterial blood oxygen pressure (PaO2) was determined by blood-gas analyser. The changes in histopathology and lung wet/dry weight (W/D) ratio were also observed in every group. RESULTS: The changes in the levels of NE and lung W/D ratio of the TCSC group was significantly decreased than those of ET group, but PaO2 of TCSC group was increased obviously than that of ET group (all P<0.05), and the degree of lung injury was less intensive in the TCSC group. At all the time points, the level of IL-6 of TCSC group was lower compared with ET group (all P<0.05). The results of correlation analysis suggested that there was a positive correlation between plasma NE and IL-6 concentration (r=0.458, P<0.05), a negative correlation between NE and PaO2 (r= -0.528, P<0.05). CONCLUSION: Sympathectomy as a result of TCSC at 7th cervical spine may palliate the degree of ALI and improve oxygenation in rats with endotoxemia by inhibiting excessive activation of adrenoceptor.