SG-Fusion: A swin-transformer and graph convolution-based multi-modal deep neural network for glioma prognosis.

The integration of morphological attributes extracted from histopathological images and genomic data holds significant importance in advancing tumor diagnosis, prognosis, and grading. Histopathological images are acquired through microscopic examination of tissue slices, providing valuable insights into cellular structures and pathological features. On the other hand, genomic data provides information about tumor gene expression and functionality. The fusion of these two distinct data types is crucial for gaining a more comprehensive understanding of tumor characteristics and progression. In the past, many studies relied on single-modal approaches for tumor diagnosis. However, these approaches had limitations as they were unable to fully harness the information from multiple data sources. To address these limitations, researchers have turned to multi-modal methods that concurrently leverage both histopathological images and genomic data. These methods better capture the multifaceted nature of tumors and enhance diagnostic accuracy. Nonetheless, existing multi-modal methods have, to some extent, oversimplified the extraction processes for both modalities and the fusion process. In this study, we presented a dual-branch neural network, namely SG-Fusion. Specifically, for the histopathological modality, we utilize the Swin-Transformer structure to capture both local and global features and incorporate contrastive learning to encourage the model to discern commonalities and differences in the representation space. For the genomic modality, we developed a graph convolutional network based on gene functional and expression level similarities. Additionally, our model integrates a cross-attention module to enhance information interaction and employs divergence-based regularization to enhance the model's generalization performance. Validation conducted on glioma datasets from the Cancer Genome Atlas unequivocally demonstrates that our SG-Fusion model outperforms both single-modal methods and existing multi-modal approaches in both survival analysis and tumor grading.

Corrigendum: Characterizing the neutrophilic inflammation in chronic rhinosinusitis with nasal polyps.

[This corrects the article DOI: 10.3389/fcell.2021.793073.].

Organic polaritonic light-emitting diodes with high luminance and color purity toward laser displays.

Achieving high-luminescence organic light-emitting devices (OLEDs) with narrowband emission and high color purity is important in various optoelectronic fields. Laser displays exhibit outstanding advantages in next-generation display technologies owing to their ultimate visual experience, but this remains a great challenge. Here, we develop a novel OLED based organic single crystals. By strongly coupling the organic exciton state to an optical microcavity, we obtain polariton electroluminescent (EL) emission from the polariton OLEDs (OPLEDs) with high luminance, narrow-band emission, high color purity, high polarization as well as excellent optically pumped polariton laser. Further, we evaluate the potential for electrically pumped polariton laser through theoretical analysis and provide possible solutions. This work provides a powerful strategy with a material-device combination that paves the way for electrically driven organic single-crystal-based polariton luminescent devices and possibly lasers.

MMD-DTA: A multi-modal deep learning framework for drug-target binding affinity and binding region prediction.

The prediction of drug-target affinity (DTA) plays a crucial role in drug development and the identification of potential drug targets. In recent years, computer-assisted DTA prediction has emerged as a significant approach in this field. In this study, we propose a multi-modal deep learning framework called MMD-DTA for predicting drug-target binding affinity and binding regions. The model can predict DTA while simultaneously learning the binding regions of drug-target interactions through unsupervised learning. To achieve this, MMD-DTA first uses graph neural networks and target structural feature extraction network to extract multi-modal information from the sequences and structures of drugs and targets. It then utilizes the feature interaction and fusion modules to generate interaction descriptors for predicting DTA and interaction strength for binding region prediction. Our experimental results demonstrate that MMD-DTA outperforms existing models based on key evaluation metrics. Furthermore, external validation results indicate that MMD-DTA enhances the generalization capability of the model by integrating sequence and structural information of drugs and targets. The model trained on the benchmark dataset can effectively generalize to independent virtual screening tasks. The visualization of drug-target binding region prediction showcases the interpretability of MMD-DTA, providing valuable insights into the functional regions of drug molecules that interact with proteins.

Risk of coronal imbalance after posterior surgery for adolescent idiopathic scoliosis of type Lenke5C.

PURPOSE: To investigate the incidence, risk factors, and clinical impact of persistent coronal imbalance (PCI) in Lenke5C adolescent idiopathic scoliosis (AIS) undergoing posterior spinal fusion (PSF). METHODS: We analyzed the medical records of 112 Lenke5C AIS patients. They were grouped into PCI (+) group and PCI (-) group according to the occurrence of PCI at 2 years after surgery. Coronal and sagittal parameters were measured and compared between both groups preoperatively, immediately postoperatively, and 2 years postoperatively. Scoliosis Research Society-22 (SRS-22) score was used to evaluate clinical outcomes. RESULTS: Of the 112 patients, 12 had coronal imbalance persisting 2 years after surgery. Logistic regression analysis indicated risk factors including older age [odds ratio (OR) = 1.841, 95% confidence interval (CI) = 1.147-2.132, p = .001], lower preoperative flexibility main thoracic (MT) curve (OR = 1.308, 95% CI = 1.041-2.015, p = .016), greater preoperative apical vertebral translation (AVT) of the thoracolumbar/lumbar (TL/L) curve (AVT-TL/L) (OR = 2.291, 95% CI = 1.120-4.719, p = .001), larger preoperative lowest instrumented vertebra tilt (LIV lilt) (OR = 2.141, 95% CI = 1.491-3.651, p = .011), and postoperative immediate coronal imbalance (OR = 5.512, 95% CI = 4.531-6.891, p = .001). The satisfaction and total score of the SRS-22 scale in the PCI (+) group were lower than those in the PCI (-) group at 2 years after surgery (p <.05). CONCLUSIONS: We found a 10.7% incidence of PCI in patients with Lenke5C AIS undergoing PSF. PCI adversely affects clinical outcomes. Risk factors of PCI included older age, reduced preoperative MT curve flexibility, increased preoperative AVT in the TL/L curve, greater preoperative LIV tilt, and immediate postoperative coronal imbalance.

Smart Response Biomaterials for Pain Management.

The intricate nature of pain classification and mechanism constantly affects the recovery of diseases and the well-being of patients. Key medical challenges persist in devising effective pain management strategies. Therefore, a comprehensive review of relevant methods and research advancements in pain management is conducted. This overview covers the main categorization of pain and its developmental mechanism, followed by a review of pertinent research and techniques for managing pain. These techniques include commonly prescribed medications, invasive procedures, and noninvasive physical therapy methods used in rehabilitation medicine. Additionally, for the first time, a systematic summary of the utilization of responsive biomaterials in pain management is provided, encompassing their response to physical stimuli such as ultrasound, magnetic fields, electric fields, light, and temperature, as well as changes in the physiological environment like reactive oxygen species (ROS) and pH. Even though the application of responsive biomaterials in pain management remains limited and at a fundamental level, recent years have seen the examination and debate of relevant research findings. These profound discussions aim to provide trends and directions for future research in pain management.

Molecular Orientation-Dependent Photonic Polarization Engineering in Organic Single-Crystal-Filled Microcavities.

Designing the polarization degree of freedom of light is crucial in many fields and has widespread application in, for example, all-optical circuits. In this work, we find that in an organic microcavity filled with anisotropic single crystals the cavity modes can be modulated to be elliptically polarized, i.e., partially circularly polarized and partially linearly polarized. The circular polarization component originates from the Rashba-Dresselhaus spin splitting, while the linear polarization component is due to the dislocation of linearly polarized modes. The dislocation of the linear polarizations is ascribed to the orientation of individual molecules and the molecular packing arrangement; hence, the linear polarizations can be controlled by properly structuring the molecular distributions. Our results pave the way for enriching and engineering the polarization properties of individual optical cavity modes in organic microstructures, which may favor the development of polarized lasers with various polarizations.

Straw removal reduces Cd availability and rice Cd accumulation in Cd-contaminated paddy soil: Cd fraction, soil microorganism structure and porewater DOC and Cd.

The impacts of straw removal on rice Cd absorption, behaviour of Cd and microbial community in rhizosphere soil were investigated in paddy fields over two consecutive seasons. The results of the experiments in two fields revealed that straw removal promoted the transformation of soil Cd from acid-extractable and oxidisable fraction to residual fraction and reduced soil DTPA-Cd content with the reduction in DOC and Cd ions in soil porewater, thereby decreasing Cd content in rice. Specifically, the Cd content in brown rice was below 0.2 mg·kg-1 when all rice straw and roots were removed in the slightly Cd-contaminated soils. The α-diversity of soil microbial communities was less influenced by continuous straw removal, ß-diversity was altered and the relative abundances of Anaeromyxobacter, Methylocystis and Mycobacterium microbes were increased. Redundancy analysis and network analysis exhibited that soil pH predominantly influenced the microbial community. Path analysis revealed that the Cd content in brown rice could be directly influenced by the soil Total-Cd and DTPA-Cd, as well as soil pH and OM. Straw removal, including roots removal, is an economical and effective technique to reduce Cd accumulation in rice plants.

Compartmentalized mitochondrial ferroptosis converges with optineurin-mediated mitophagy to impact airway epithelial cell phenotypes and asthma outcomes.

A stable mitochondrial pool is crucial for healthy cell function and survival. Altered redox biology can adversely affect mitochondria through induction of a variety of cell death and survival pathways, yet the understanding of mitochondria and their dysfunction in primary human cells and in specific disease states, including asthma, is modest. Ferroptosis is traditionally considered an iron dependent, hydroperoxy-phospholipid executed process, which induces cytosolic and mitochondrial damage to drive programmed cell death. However, in this report we identify a lipoxygenase orchestrated, compartmentally-targeted ferroptosis-associated peroxidation process which occurs in a subpopulation of dysfunctional mitochondria, without promoting cell death. Rather, this mitochondrial peroxidation process tightly couples with PTEN-induced kinase (PINK)-1(PINK1)-Parkin-Optineurin mediated mitophagy in an effort to preserve the pool of functional mitochondria and prevent cell death. These combined peroxidation processes lead to altered epithelial cell phenotypes and loss of ciliated cells which associate with worsened asthma severity. Ferroptosis-targeted interventions of this process could preserve healthy mitochondria, reverse cell phenotypic changes and improve disease outcomes.

Comparing Artificial Intelligence-Based Versus Conventional Endotracheal Tube Monitoring Systems in Clinical Practice.

Endotracheal tube dislodgement is a common patient safety incident in clinical settings. Current clinical practices, primarily relying on bedside visual inspections and equipment checks, often fail to detect endotracheal tube displacement or dislodgement promptly. This study involved the development of a deep learning, artificial intelligence (AI)-based system for monitoring tube displacement. We also propose a randomized crossover experiment to evaluate the effectiveness of this AI-based monitoring system compared to conventional methods. The assessment will focus on immediacy in detecting and handling of tube anomalies, the completeness and accuracy of shift transitions, and the degree of innovation diffusion. The findings from this research are expected to offer valuable insights into the development and integration of AI in enhancing care provision and facilitating innovation diffusion in medical and nursing research.

EGFR/TKIs induce excessive apoptosis of bladder carcinoma cells by arresting cell cycle and promoting mitochondrial peroxidation damage.

In recent years, bladder carcinoma (BC) has shown an increasing incidence, with poor patient outcomes. In clinical practice, BC is still mainly treated by surgery combined with chemoradiotherapy. However, as chemotherapy resistance of tumor cells becomes more and more obvious, it is urgent to find more effective BC treatment regimes. With the increasing application and growing attention paid to epidermal growth factor receptor-tyrosine kinase inhibitors (EGFR-TKIs) in various neoplastic diseases, EGFR-TKIs have been considered as a new treatment direction in the future. In this study, the research team used AG1478, an EGFR-TKI, to intervene with the BC cell line T24. It was found that the cell activity was statistically decreased, the apoptosis was enhanced, and the cells were dominantly arrested in the G0/G1 phase, confirming the future therapeutic potential of EGFR-TKIs in BC. Besides, the research team further observed that AG1478 also promoted pyroptosis in T24 cells, and its mechanism is related to the induction of mitochondrial oxidative stress damage. The findings lay a more reliable foundation for the future application of EGFR-TKIs in BC.

Deep Canonical Correlation Fusion Algorithm Based on Denoising Autoencoder for ASD Diagnosis and Pathogenic Brain Region Identification.

Autism Spectrum Disorder (ASD) is defined as a neurodevelopmental condition distinguished by unconventional neural activities. Early intervention is key to managing the progress of ASD, and current research primarily focuses on the use of structural magnetic resonance imaging (sMRI) or resting-state functional magnetic resonance imaging (rs-fMRI) for diagnosis. Moreover, the use of autoencoders for disease classification has not been sufficiently explored. In this study, we introduce a new framework based on autoencoder, the Deep Canonical Correlation Fusion algorithm based on Denoising Autoencoder (DCCF-DAE), which proves to be effective in handling high-dimensional data. This framework involves efficient feature extraction from different types of data with an advanced autoencoder, followed by the fusion of these features through the DCCF model. Then we utilize the fused features for disease classification. DCCF integrates functional and structural data to help accurately diagnose ASD and identify critical Regions of Interest (ROIs) in disease mechanisms. We compare the proposed framework with other methods by the Autism Brain Imaging Data Exchange (ABIDE) database and the results demonstrate its outstanding performance in ASD diagnosis. The superiority of DCCF-DAE highlights its potential as a crucial tool for early ASD diagnosis and monitoring.

Knockdown of KBTBD7 attenuates septic lung injury by inhibiting ferroptosis and improving mitochondrial dysfunction.

Lung injury in sepsis is caused by an excessive inflammatory response caused by the entry of pathogenic microorganisms into the body. It is also accompanied by the production of large amounts of ROS. Ferroptosis and mitochondrial dysfunction have also been shown to be related to sepsis. Finding suitable sepsis therapeutic targets is crucial for sepsis research. BTB domain-containing protein 7 (KBTBD7) is involved in regulating inflammatory responses, but its role and mechanism in the treatment of septic lung injury are still unclear. In this study, we evaluated the role and related mechanisms of KBTBD7 in septic lung injury. In in vitro studies, we established an in vitro model by inducing human alveolar epithelial cells with lipopolysaccharide (LPS) and found that KBTBD7 was highly expressed in the in vitro model. KBTBD7 knockdown could reduce the inflammatory response by inhibiting the secretion of pro-inflammatory factors and inhibit the production of ROS, ferroptosis and mitochondrial dysfunction. Mechanistic studies show that KBTBD7 interacts with FOXA1, promotes FOXA1 expression, and indirectly inhibits SLC7A11 transcription. In vivo studies have shown that knocking down KBTBD7 improves lung tissue damage in septic lung injury mice, inhibits inflammatory factors, ROS production and ferroptosis. Taken together, knockdown of KBTBD7 shows an alleviating effect on septic lung injury in vitro and in vivo, providing a potential therapeutic target for the treatment of septic lung injury.

Multi-kernel Learning Fusion Algorithm Based on RNN and GRU for ASD Diagnosis and Pathogenic Brain Region Extraction.

Autism spectrum disorder (ASD) is a complex, severe disorder related to brain development. It impairs patient language communication and social behaviors. In recent years, ASD researches have focused on a single-modal neuroimaging data, neglecting the complementarity between multi-modal data. This omission may lead to poor classification. Therefore, it is important to study multi-modal data of ASD for revealing its pathogenesis. Furthermore, recurrent neural network (RNN) and gated recurrent unit (GRU) are effective for sequence data processing. In this paper, we introduce a novel framework for a Multi-Kernel Learning Fusion algorithm based on RNN and GRU (MKLF-RAG). The framework utilizes RNN and GRU to provide feature selection for data of different modalities. Then these features are fused by MKLF algorithm to detect the pathological mechanisms of ASD and extract the most relevant the Regions of Interest (ROIs) for the disease. The MKLF-RAG proposed in this paper has been tested in a variety of experiments with the Autism Brain Imaging Data Exchange (ABIDE) database. Experimental findings indicate that our framework notably enhances the classification accuracy for ASD. Compared with other methods, MKLF-RAG demonstrates superior efficacy across multiple evaluation metrics and could provide valuable insights into the early diagnosis of ASD.

[Cadmium Phytoremediation Effect of Sweet Sorghum Assisted with Citric Acid on Typical Parent Soil in Southern China].

Sweet sorghum has a large biomass and strong cadmium (Cd) absorption capacity, which has the potential for phytoremediation of Cd-contaminated soil. In order to study the Cd phytoremediation effect of sweet sorghum assisted with citric acid on the typical parent materials in southern China, a field experiment was carried out in two typical parent material farmland areas (neutral purple mud field and jute sand mud field) with Cd pollution in Hunan Province. The results showed that:① Citric acid had no inhibitory effect on the growth of sweet sorghum. After the application of citric acid, the aboveground biomass of sweet sorghum at the maturity stage increased by 10.1%-24.7%. ② Both sweet sorghum planting and citric acid application reduced the soil pH value, and the application of citric acid further reduced the soil pH value at each growth stage of sweet sorghum; this decrease was greater in the neutral purple mud field, which decreased by 0.24-0.72 units. ③ Both sweet sorghum planting and citric acid application reduced the total amount of soil Cd, and the decreases in the neutral purple mud field and jute sand mud field were 23.8%-52.2% and 17.1%-31.8%, respectively. The acid-extractable percentage of soil Cd in both places increased by 38.6%-147.7% and 4.8%-22.7%, respectively. ④ The application of citric acid could significantly increase the Cd content in various tissues of sweet sorghum. The Cd content in the aboveground part of the plant in the neutral purple mud field was higher than that in the jute sand mud field, and the Cd content in stems and leaves was 0.25-1.90 mg·kg-1 and 0.21-0.64 mg·kg-1, respectively. ⑤ After applying citric acid, the Cd extraction amount of sweet sorghum in neutral purple mud soil in the mature stage reached 47.56 g·hm-2. In summary, citric acid could enhance the efficiency of sweet sorghum in the phytoremediation of Cd-contaminated soil, and the effect was better in neutral purple mud fields. This technology has the potential for remediation coupled with agro-production for heavy metal-contaminated farmland.

E3 Ubiquitin Ligase NEDD4L Negatively Regulates Skin Tumorigenesis by Inhibiting IL-6/GP130 Signaling Pathway.

IL-6 signaling plays a crucial role in the survival and metastasis of skin cancer. NEDD4L acts as a suppressor of IL-6 signaling by targeting GP130 degradation. However, the effects of the NEDD4L-regulated IL-6/GP130 signaling pathway on skin cancer remain unclear. In this study, protein expression levels of NEDD4L and GP130 were measured in tumor tissues from patients with cutaneous squamous cell carcinoma. Skin tumors were induced in wild-type and Nedd4l-knockout mice, and activation of the IL-6/GP130/signal transducer and activator of transcription 3 signaling pathway was detected. The results indicated a negative correlation between the protein expression levels of NEDD4L and GP130 in cutaneous squamous cell carcinoma tissues from patients. Nedd4l deficiency significantly promoted 7,12-dimethylbenz[a]anthracene/12-O-tetradecanoylphorbol-13-acetate-induced skin tumorigenesis and benign-to-malignant conversion by activating the IL-6/GP130/signal transducer and activator of transcription 3 signaling pathway, which was abrogated by supplementation with the GP130 inhibitor SC144. Furthermore, our findings suggested that NEDD4L can interact with GP130 and promote its ubiquitination in skin tumors. In conclusion, our results indicate that NEDD4L could act as a tumor suppressor in skin cancer, and inhibition of GP130 could be a potential therapeutic method for treating this disease.

[Effect of Silica Fertilizer(Husk Ash) to Improve Soil Quality and Reduce Cd and As Accumulation in Rice].

A rice pot experiment was conducted to identify the effect of silica fertilizer prepared from husk ash on the soil bioavailability of cadmium (Cd) and arsenic (As), enzyme activities, microbial community structure, and heavy metal content in brown rice at different growth stages. The results showed that the application of 0.1%-1.0% silica fertilizer-husk ash increased the pH value of soil by 0.04-0.24 units and the content of soil available silicon by 44.2%-97.5%. It also decreased the content of available Cd and available As by 16.2%-21.4% and 16.0%-24.9%, respectively. With the increase in application amount, the soil enzyme activities increased at all growth stages, and the sucrase activity and the dehydrogenase activity significantly increased by 6.3%-145.7% and 6.7%-224.1%, respectively. The analysis of the soil microbial community composition structure at mature stages showed that the application of silica fertilizer-husk ash had no effect on microbial α-diversity, but it had a significant effect on microbial ß-diversity and then promoted microbial growth and maintained the stability of the community structure. With the increase in application amount, the contents of Cd in brown rice decreased by 29.3%-89.7%, and the contents of total As and inorganic As in brown rice decreased by 7.8%-42.3% and 17.2%-44.5%, respectively. Under the application of 0.5% and 1.0% silica fertilizer-husk ash, the Cd contents in brown rice were lower than 0.2 mg·kg-1, and the inorganic As contents in brown rice were lower than 0.35 mg·kg-1. In conclusion, the silica fertilizer-husk ash can improve soil quality and reduce the contents of Cd and As in brown rice, and it is eco-friendly and can be used to remedy the paddy soil contaminated with Cd and As.

[Effects of the Application of Irrigation Water Containing Zn at the Key Growth Period on the Uptake and Transport of Cd in Rice].

In this study， a field experiment was conducted to examine the effects of the application of irrigation water containing Zn at the key growth period （booting stage and filling stage） on exchangeable Cd content in the soil， Cd concentration in pore water， and Cd uptake and transport in rice in a Cd-contaminated paddy field in Liuyang City， Hunan Province. The results indicated that： ① the application of irrigation water containing Zn during the key growth period could inhibit the releasing process of exchangeable Cd from the soil into pore water. Compared with that in the control， the content of exchangeable Cd in soil was slightly changed， but the concentration of Cd in soil pore water at the mature stage was significantly reduced by 16.7%-57.6%. ② The application of irrigation water containing Zn at the key growth period could significantly reduce the Cd content in various parts of rice. Cd contents in root， stem， and brown rice with the application of irrigation water containing 20 mg·L-1 Zn before the booting and the filling stage （BF1） were significantly decreased by 56.0%， 83.8%， and 85.2%， respectively. ③ Compared with the application of 100 mg·L-1 irrigation water containing Zn， the application of 20 mg·L-1 irrigation water containing Zn significantly reduced the uptake and transport of Cd in rice， and the translocation factor （TF） of Cd from rice roots to stems was also significantly reduced by 12.5%-56.3%， with the B1 and BF1 treatments reaching significant levels. These results suggested that the application of irrigation water containing Zn could significantly reduce the uptake and accumulation of Cd in rice， and the application of 20 mg·L-1 irrigation water containing Zn before the booting and filling stage could effectively realize the safe production of Cd-contaminated paddy fields.

SARS-CoV-2 Neutralization Capacity in Hemodialysis Patients with and without a Fifth Vaccination with the Updated Comirnaty Original/Omicron BA.4-5 Vaccine.

BACKGROUND: Hemodialysis patients have reduced serologic immunity after SARS-CoV-2 vaccination compared to the general population and an increased risk of morbidity and mortality when exposed to SARS-CoV-2. METHODS: Sixty-six hemodialysis patients immunized four times with the original SARS-CoV-2 vaccines (BNT162b2, mRNA-1273) either received a booster with the adapted Comirnaty Original/Omicron BA.4-5 vaccine 8.3 months after the fourth vaccination and/or experienced a breakthrough infection. Two months before and four weeks after the fifth vaccination, the live-virus neutralization capacities of Omicron variants BA.5, BQ.1.1, and XBB.1.5 were determined, as well as neutralizing and quantitative anti-SARS-CoV-2 spike-specific IgG antibodies. RESULTS: Four weeks after the fifth vaccination with the adapted vaccine, significantly increased neutralizing antibodies and the neutralization of Omicron variants BA.5, BQ.1.1, and XBB.1.5 were observed. The increase was significantly higher than after the fourth vaccination for variants BQ.1.1 and BA.5. Of all analyzed variants, BA.5 was neutralized best after the fifth vaccination. We did not see a difference in humoral immunity between the group with an infection and the group with a vaccination as a fifth spike exposure. Fivefold-vaccinated patients with a breakthrough infection showed a significantly higher neutralization capacity of XBB.1.5. CONCLUSION: A fifth SARS-CoV-2 vaccination with the adapted vaccine improves both wild-type specific antibody titers and the neutralizing capacity of the current Omicron variants BA.5, BQ.1.1, and XBB.1.5 in hemodialysis patients. Additional booster vaccinations with adapted vaccines will likely improve immunity towards current and original SARS-CoV-2 variants and are, therefore, recommended in hemodialysis patients. Further longitudinal studies must show the extent to which this booster vaccination avoids a breakthrough infection.

Deep integrated fusion of local and global features for cervical cell classification.

Cervical cytology image classification is of great significance to the cervical cancer diagnosis and prognosis. Recently, convolutional neural network (CNN) and visual transformer have been adopted as two branches to learn the features for image classification by simply adding local and global features. However, such the simple addition may not be effective to integrate these features. In this study, we explore the synergy of local and global features for cytology images for classification tasks. Specifically, we design a Deep Integrated Feature Fusion (DIFF) block to synergize local and global features of cytology images from a CNN branch and a transformer branch. Our proposed method is evaluated on three cervical cell image datasets (SIPaKMeD, CRIC, Herlev) and another large blood cell dataset BCCD for several multi-class and binary classification tasks. Experimental results demonstrate the effectiveness of the proposed method in cervical cell classification, which could assist medical specialists to better diagnose cervical cancer.