Which is the best TACE agent for patients with different NLR hepatocellular carcinomas? A systematic review and network meta-analysis.

Background: Transarterial chemoembolization (TACE) is a common treatment for hepatocellular carcinoma (HCC), but the best therapeutic agent for TACE treatment has not been determined. The neutrophil/lymphocyte ratio (NLR) is a systemic immune system marker; however, the ability of the NLR to predict the prognosis of patients with HCC is unknown, and no studies have been conducted to determine the most appropriate TACE regimen for HCC patients with different NLRs. Methods: The PubMed, Embase, Web of Science, and CNKI databases were searched through May 28, 2023. Comparisons of overall survival (OS) among cohort studies with different NLRs and different TACE treatment regimens were performed with a random effects model. Findings: Thirty-five studies involving 9210 patients were included in this meta-analysis. The results showed that Group 3-4 (NLR<2.5) patients had a significantly longer OS than Group 1-2 (NLR 2.5-5.0). Among the patients, Group 1-3 (NLR 2.0-5.0) patients had the best survival after treatment with adriamycin (lnHR (95 % CI = 0.48 [0.31, 0.75] and lnHR (95 % CI = 0.41 [0.19, 0.91]). Among the Group 4 patients (NLR<2.0), the best outcome was obtained with platinum + adriamycin (lnHR (95 % CI = 0.59 [0.45, 0.78]), followed by adriamycin. A subgroup analysis of TACE combined with other treatments showed that adriamycin combined with sorafenib was the most effective and superior to the other treatment agents. Interpretation: The NLR can be used to predict the prognosis of HCC patients treated with TACE; the higher the NLR is, the worse the prognosis. Adriamycin may be the best therapeutic agent for HCC patients treated with TACE.

A Novel Set-Based Discrete Particle Swarm Optimization for Wastewater Treatment Process Effluent Scheduling.

With the escalating severity of environmental pollution caused by effluent, the wastewater treatment process (WWTP) has gained significant attention. The wastewater treatment efficiency and effluent quality are significantly impacted by effluent scheduling that adjusts the hydraulic retention time. However, the sequential batch and continuous nature of the effluent pose challenges, resulting in complex scheduling models with strong constraints that are difficult to tackle using existing scheduling methods. To optimize maximum completion time and effluent quality simultaneously, this article proposes a restructured set-based discrete particle swarm optimization (RS-DPSO) algorithm to address the WWTP effluent scheduling problem (WWTP-ESP). First, an effective encoding and decoding method is designed to effectively map solutions to feasible schedules using temporal and spatial information. Second, a restructured set-based discrete particle swarm algorithm is introduced to enhance the searching ability in discrete solution space via restructuring the solution set. Third, a constraint handling strategy based on violation degree ranking is designed to reduce the waste of computational resources. Fourth, a Sobel filter based local search is proposed to guide particle search direction to enhance search efficiency ability. The RS-DPSO provides a novel method for solving WWTP-ESP problems with complex discrete solution space. The comparative experiments indicate that the novel designs are effective and the proposed algorithm has superior performance over existing algorithms in solving the WWTP-ESP.

Oral epithelial dysplasia detection and grading in oral leukoplakia using deep learning.

BACKGROUND: The grading of oral epithelial dysplasia is often time-consuming for oral pathologists and the results are poorly reproducible between observers. In this study, we aimed to establish an objective, accurate and useful detection and grading system for oral epithelial dysplasia in the whole-slides of oral leukoplakia. METHODS: Four convolutional neural networks were compared using the image patches from 56 whole-slide of oral leukoplakia labeled by pathologists as the gold standard. Sequentially, feature detection models were trained, validated and tested with 1,000 image patches using the optimal network. Lastly, a comprehensive system named E-MOD-plus was established by combining feature detection models and a multiclass logistic model. RESULTS: EfficientNet-B0 was selected as the optimal network to build feature detection models. In the internal dataset of whole-slide images, the prediction accuracy of E-MOD-plus was 81.3% (95% confidence interval: 71.4-90.5%) and the area under the receiver operating characteristic curve was 0.793 (95% confidence interval: 0.650 to 0.925); in the external dataset of 229 tissue microarray images, the prediction accuracy was 86.5% (95% confidence interval: 82.4-90.0%) and the area under the receiver operating characteristic curve was 0.669 (95% confidence interval: 0.496 to 0.843). CONCLUSIONS: E-MOD-plus was objective and accurate in the detection of pathological features as well as the grading of oral epithelial dysplasia, and had potential to assist pathologists in clinical practice.

The clinical implication and translational research of OSCC differentiation.

BACKGROUND: The clinical value and molecular characteristics of tumor differentiation in oral squamous cell carcinoma (OSCC) remain unclear. There is a lack of a related molecular classification prediction system based on pathological images for precision medicine. METHODS: Integration of epidemiology, genomics, experiments, and deep learning to clarify the clinical value and molecular characteristics, and develop a novel OSCC molecular classification prediction system. RESULTS: Large-scale epidemiology data (n = 118,817) demonstrated OSCC differentiation was a significant prognosis indicator (p < 0.001), and well-differentiated OSCC was more chemo-resistant than poorly differentiated OSCC. These results were confirmed in the TCGA database and in vitro. Furthermore, we found chemo-resistant related pathways and cell cycle-related pathways were up-regulated in well- and poorly differentiated OSCC, respectively. Based on the characteristics of OSCC differentiation, a molecular grade of OSCC was obtained and combined with pathological images to establish a novel prediction system through deep learning, named ShuffleNetV2-based Molecular Grade of OSCC (SMGO). Importantly, our independent multi-center cohort of OSCC (n = 340) confirmed the high accuracy of SMGO. CONCLUSIONS: OSCC differentiation was a significant indicator of prognosis and chemotherapy selection. Importantly, SMGO could be an indispensable reference for OSCC differentiation and assist the decision-making of chemotherapy.

Conductive imprinted polymeric interfacially modified electrochemical sensors based on covalently bonded layer-by-layer assembly of Gr/Au with flower-like morphology for sensitive detection of 2,4,6-TCP.

Polymeric membrane sensors based on molecular imprinted polymers (MIPs) have been attractive analytical tools for detecting organic species. However, the MIPs in electrochemical sensors developed so far are usually prepared by in situ polymerization of pre-polymers and non-covalent adsorption on the surface of the working electrode. Meanwhile, the MIPs in the electrochemical sensors developed are typically made of a non-conductive polymer film. This results in a relatively low current due to the lack of electron transfer. Additionally, the smoothness of the traditional electrochemical substrate results in a low specific surface area, which reduces the sensitivity of the electrochemical sensor. Here, we describe a novel electrochemical sensor with a conductive interface and MIPs modification. The electrochemical sensor was modified by covalent coupled layer by layer self-assembly with the imprinted polymer film. The incorporation of these two conductive functional materials improves the conductivity of the electrodes and provides interface support materials to obtain high specific surface area. By using 2,4,6-trichlorophenol as the model, the sensitivity of the developed conductive sensor was greatly improved compared to that of the traditional MIPs sensor. We believe that the proposed MIPs-based sensing strategy provides a general and convenient method for making sensitive and selective electrochemical sensors.

Potential toxic effects of perfluorobutanesulfonyl fluoride analysis based on multiple-spectroscopy techniques and molecular modelling analysis.

Perfluorobutanesulfonyl fluoride (PBSF) has been used in the manufacture of fluorochemicals. Since PBSF is not biodegradable, the predicted environmental levels of PBSF are also expected to rise over time. In recent years, there has been a rise in the levels of PBSF in humans. In order to clarify the impact of PBSF on the accumulation of substances in the human body, we examined the interaction mechanism between PBSF and bovine serum albumin (BSA). To investigate the interaction mechanism between PBSF and BSA, we utilized a range of methods including UV-visible spectrophotometry, fluorescence spectroscopy, circular dichroism, molecular docking simulation, and molecular dynamics (MD) simulation. The inherent fluorescence of BSA was effectively suppressed by PBSF through fluorescence quenching analysis, using a static mechanism. The Ka value of 1.34 × 105 mol-1 L indicated a strong binding between PBSF and BSA. Further analysis of the interaction between PBSF and BSA involved examining thermodynamic parameters, fluorescence resonance energy transfer, and conducting other theoretical calculations. These investigations produced results that were in strong accordance with the experimental observations. The participation of hydrophobic interactions between BSA and PBSF was uncovered through molecular docking and MD simulation investigations. Furthermore, this investigation explored the impact of copper ions (Cu2+) and calcium ions (Ca2+) on the interaction between PBSF and BSA, establishing a vital basis for comprehending the mechanism by which PBSF affects proteins in the human surroundings.

Arbuscular mycorrhizal fungi enhanced the drinking water treatment residue-based vertical flow constructed wetlands on the purification of arsenic-containing wastewater.

Arsenic (As) is a toxic metalloid that poses a potential risk to the environment and human health. In this study, drinking water treatment residue (DWTR) and ceramsite-based vertical flow constructed wetlands (VFCWs) were built to purify As-containing wastewater. As a method of bioaugmentation, arbuscular mycorrhizal fungi (AMF) was inoculated to Pteris vittata roots to enhance the As removal of the VFCWs. The results showed that the As removal rates reached 87.82-94.29% (DWTR) and 33.28-58.66% (ceramsite). DWTR and P. vittata contributed 64.33-72.07% and 7.57-29% to the removal of As, while AMF inoculation intensified the As accumulation effect of P. vittata. Proteobacteria, the main As3+ oxidizing bacteria in the aquatic systems, dominated the microbial community, occupying 72.41 ± 7.76%. AMF inoculation increased As-related functional genes abundance in DWTR-based wetlands and provided a reliable means of arsenic resistance in wetlands. These findings indicated that the DWTR-based VFCWs with AMF inoculated P. vittata had a great purification effect on As-containing wastewater, providing a theoretical basis for the application of DWTR and AMF for As removal in constructed wetlands.

Photoinduced Trifluoromethylselenolation of Aryl Halides with [Me4N][SeCF3].

A visible-light-induced metal-free trifluoromethylselenolation of aryl iodides and bromides with [Me4N][SeCF3] is described. The reaction was conducted at ambient temperature by successfully harnessing the light-sensitive SeCF3 reagent. Mechanistically, the EDA complexes between aryl halide and the -SeCF3 anion or the base might be formed and excited by light, which subsequently undergo intracomplex SET processes to generate aryl and •SeCF3 radicals as key intermediates, allowing a convenient and green access to various aryl trifluoromethyl selenoethers.

Precursor template-induced egg white-derived peptides self-assembly for the enhancement of curcumin: Structure, environmental stability, and bioavailability.

Natural multicomponent peptides with abundant bioactivity, varied sizes, and tunable interaction potential are available for rational designing novel self-assembled delivery carriers. Herein, we exploited zein-hyaluronic acid nanoparticles (Z-HA NPs) with a predetermined ordered structure as precursor templates to induce the self-assembly of egg white-derived peptides (EWDP) to generate stable spherical architectures for the enhancement of curcumin (Cur). The resulting Z-EWDP-HA NPs encapsulated hydrophobic Cur through robust hydrogen bonding and hydrophobic interactions with high encapsulation efficiency (97.38% at pH 7.0). The NPs presented superior Cur aqueous solubility, redispersibility, and photothermal stability. More importantly, the self-assembled EWDP could exert synergistic antioxidant activity with Cur and enhance the bioaccessibility of Cur. Meanwhile, the favorable biocompatibility and membrane affinity of EWDP further prolonged residence and time-controlled release feature of Cur in the small intestine. Precursor template-induced multicomponent peptides' self-assembly provides an efficient and controllable strategy for co-enhanced bioactivity and self-assembly capacity of peptides, which could dramatically broaden the functionalization of multicomponent peptides hydrolyzed from natural food proteins.

Supramolecular Polymer Ion Conductor with Weakened Li Ion Solvation Enables Room Temperature All-Solid-State Lithium Metal Batteries.

Improved durability, enhanced interfacial stability, and room temperature applicability are desirable properties for all-solid-state lithium metal batteries (ASSLMBs), yet these desired properties are rarely achieved simultaneously. Here, in this work, it is noticed that the huge resistance at Li metal/electrolyte interface dominantly impeded the normal cycling of ASSLMBs especially at around room temperature (<30 °C). Accordingly, a supramolecular polymer ion conductor (SPC) with "weak solvation" of Li+ was prepared. Benefiting from the halogen-bonding interaction between the electron-deficient iodine atom (on 1,4-diiodotetrafluorobenzene) and electron-rich oxygen atoms (on ethylene oxide), the O-Li+ coordination was significantly weakened. Therefore, the SPC achieves rapid Li+ transport with high Li+ transference number, and importantly, derives a unique Li2 O-rich SEI with low interfacial resistance on lithium metal surface, therefore enabling stable cycling of ASSLMBs even down to 10 °C. This work is a new exploration of halogen-bonding chemistry in solid polymer electrolyte and highlights the importance of "weak solvation" of Li+ in the solid-state electrolyte for room temperature ASSLMBs.

ATHENA: an independently validated autophagy-related epigenetic prognostic prediction model of head and neck squamous cell carcinoma.

The majority of these existing prognostic models of head and neck squamous cell carcinoma (HNSCC) have unsatisfactory prediction accuracy since they solely utilize demographic and clinical information. Leveraged by autophagy-related epigenetic biomarkers, we aim to develop a better prognostic prediction model of HNSCC incorporating CpG probes with either main effects or gene-gene interactions. Based on DNA methylation data from three independent cohorts, we applied a 3-D analysis strategy to develop An independently validated auTophagy-related epigenetic prognostic prediction model of HEad and Neck squamous cell carcinomA (ATHENA). Compared to prediction models with only demographic and clinical information, ATHENA has substantially improved discriminative ability, prediction accuracy and more clinical net benefits, and shows robustness in different subpopulations, as well as external populations. Besides, epigenetic score of ATHENA is significantly associated with tumor immune microenvironment, tumor-infiltrating immune cell abundances, immune checkpoints, somatic mutation and immunity-related drugs. Taken together these results, ATHENA has the demonstrated feasibility and utility of predicting HNSCC survival ( http://bigdata.njmu.edu.cn/ATHENA/ ).

Continuous ammonia electrosynthesis using physically interlocked bipolar membrane at 1000 mA cm-2.

Electrosynthesis of ammonia from nitrate reduction receives extensive attention recently for its relatively mild conditions and clean energy requirements, while most existed electrochemical strategies can only deliver a low yield rate and short duration for the lack of stable ion exchange membranes at high current density. Here, a bipolar membrane nitrate reduction process is proposed to achieve ionic balance, and increasing water dissociation sites is delivered by constructing a three-dimensional physically interlocked interface for the bipolar membrane. This design simultaneously boosts ionic transfer and interfacial stability compared to traditional ones, successfully reducing transmembrane voltage to 1.13 V at up to current density of 1000 mA cm-2. By combining a Co three-dimensional nanoarray cathode designed for large current and low concentration utilizations, a continuous and high yield bipolar membrane reactor for NH3 electrosynthesis realized a stable electrolysis at 1000 mA cm-2 for over 100 h, Faradaic efficiency of 86.2% and maximum yield rate of 68.4 mg h-1 cm-2 with merely 2000 ppm NO3- alkaline electrolyte. These results show promising potential for artificial nitrogen cycling in the near future.

A Fluorinated Covalent Organic Framework with Accelerated Oxygen Transfer Nanochannels for High-Performance Zinc-Air Batteries.

The establishment of abundant three-phase interfaces with accelerated mass transfer in air cathodes is highly desirable for the development of high-rate and long-cycling rechargeable zinc-air batteries (ZABs). Covalent organic frameworks (COFs) exhibit tailored nanopore structures, facilitating the rational tuning of their specific properties. Here, by finely tuning the fluorinated nanopores of a COF, a novel air cathode for rechargeable ZABs is unprecedentedly designed and synthesized. COF nanosheets are decorated with fluorinated alkyl chains, which shows high affinity to oxygen (O2 ), in its nanopores (fluorinated COF). The fluorinated COF nanosheets are stacked into well-defined O2 -transport channels, which are then assembled into aerophilic "nano-islands" on the hydrophilic FeNi layered-double-hydroxide (FeNi LDH) electrocatalyst surface. Therefore, the mass-transport "highway" for O2 and water is segregated on the nanoscale, which significantly enlarges the area of three-phase boundaries and greatly promotes the mass-transfer therein. ZABs based on the COF-modified air cathode deliver a small charge/discharge voltage gap (0.64 V at 5 mA cm-2 ), a peak power density (118 mW cm-2 ), and a stable cyclability. This work provides a feasible approach for the design of the air cathodes for high-performance ZABs, and will expand the new application of COFs.

Correlation Analysis of Surgical Efficacy and Risk Factors of Cervical Spondylotic Myelopathy with High Signal Intensity on MRI-T2WI.

OBJECTIVES: The purpose of this study was to investigate the surgical efficacy and risk factors of cervical spondylotic myelopathy (CSM) patients with increased signal intensity (ISI) on T2-weighted magnetic resonance imaging (MRI-T2WI). METHODS: We compared the surgical outcomes of CSM patients with and without ISI. In addition, we compared the efficacy of anterior and posterior cervical decompression in CSM patients with ISI. We also analyzed the risk factors of MRI-T2WI ISI in CSM patients. RESULTS: The incidence of ISI among 153 CSM patients was 71.89 %. The JOA score and JOA remission rate were better in the ISI-free than in the ISI group. The postoperative JOA score and JOA remission rate were better in the posterior than the anterior approach surgery group. The disease duration and vertebral canal volume were found to be risk factors for ISI in CSM patients. CONCLUSION: Among patients with CSM, the prognosis is worse for those with ISI than those without ISI. Posterior cervical decompression surgery produces a better curative effect than anterior cervical decompression surgery in CSM patients with ISI. CSM patients with longer disease duration and small vertebral canal volume should undergo surgical treatment as early as possible.

Oriented intergrowth of the catalyst layer in membrane electrode assembly for alkaline water electrolysis.

The application of membrane electrode assemblies is considered a promising approach for increasing the energy efficiency of conventional alkaline water electrolysis. However, previous investigations have mostly focused on improving membrane conductivity and electrocatalyst activity. This study reports an all-in-one membrane electrode assembly obtained by de novo design. The introduction of a porous membrane readily enables the oriented intergrowth of ordered catalyst layers using solvothermal methods, leading to the formation of an all-in-one MEA for alkaline water electrolysis. This all-in-one MEA features ordered catalyst layers with large surface areas, a low-tortuosity pore structure, integrated catalyst layer/membrane interfaces, and a well-ordered OH- transfer channel. Owing to this design, a high current density of 1000 mA cm-2 is obtained at 1.57 V in 30 wt% KOH, resulting in a 94% energy efficiency. This work highlights the prospects of all-in-one membrane electrode assemblies in designing next-generation high-performance alkaline water electrolysis.

Dual-Stream Subspace Clustering Network for revealing gene targets in Alzheimer's disease.

The rapid development of scRNA-seq technology in recent years has enabled us to capture high-throughput gene expression profiles at single-cell resolution, reveal the heterogeneity of complex cell populations, and greatly advance our understanding of the underlying mechanisms in human diseases. Traditional methods for gene co-expression clustering are limited to discovering effective gene groups in scRNA-seq data. In this paper, we propose a novel gene clustering method based on convolutional neural networks called Dual-Stream Subspace Clustering Network (DS-SCNet). DS-SCNet can accurately identify important gene clusters from large scales of single-cell RNA-seq data and provide useful information for downstream analysis. Based on the simulated datasets, DS-SCNet successfully clusters genes into different groups and outperforms mainstream gene clustering methods, such as DBSCAN and DESC, across different evaluation metrics. To explore the biological insights of our proposed method, we applied it to real scRNA-seq data of patients with Alzheimer's disease (AD). DS-SCNet analyzed the single-cell RNA-seq data with 10,850 genes, and accurately identified 8 optimal clusters from 6673 cells. Enrichment analysis of these gene clusters revealed functional signaling pathways including the ILS signaling, the Rho GTPase signaling, and hemostasis pathways. Further analysis of gene regulatory networks identified new hub genes such as ELF4 as important regulators of AD, which indicates that DS-SCNet contributes to the discovery and understanding of the pathogenesis in Alzheimer's disease.

Uncovering the dynamic effects of DEX treatment on lung cancer by integrating bioinformatic inference and multiscale modeling of scRNA-seq and proteomics data.

Lung cancer is one of the leading causes of cancer-related death, with a five-year survival rate of 18%. It is a priority for us to understand the underlying mechanisms affecting lung cancer therapeutics' implementation and effectiveness. In this study, we combine the power of Bioinformatics and Systems Biology to comprehensively uncover functional and signaling pathways of drug treatment using bioinformatics inference and multiscale modeling of both scRNA-seq data and proteomics data. Based on a time series of lung adenocarcinoma derived A549 cells after DEX treatment, we first identified the differentially expressed genes (DEGs) in those lung cancer cells. Through the interrogation of regulatory network of those DEGs, we identified key hub genes including TGFß, MYC, and SMAD3 varied underlie DEX treatment. Further gene set enrichment analysis revealed the TGFß signaling pathway as the top enriched term. Those genes involved in the TGFß pathway and their crosstalk with the ERBB pathway presented a strong survival prognosis in clinical lung cancer samples. With the basis of biological validation and literature-based curation, a multiscale model of tumor regulation centered on both TGFß-induced and ERBB-amplified signaling pathways was developed to characterize the dynamic effects of DEX therapy on lung cancer cells. Our simulation results were well matched to available data of SMAD2, FOXO3, TGFß1, and TGFßR1 over the time course. Moreover, we provided predictions of different doses to illustrate the trend and therapeutic potential of DEX treatment. The innovative and cross-disciplinary approach can be further applied to other computational studies in tumorigenesis and oncotherapy. We released the approach as a user-friendly tool named BIMM (Bioinformatic Inference and Multiscale Modeling), with all the key features available at https://github.com/chenm19/BIMM.

Epigenome-wide gene-age interaction study reveals reversed effects of MORN1 DNA methylation on survival between young and elderly oral squamous cell carcinoma patients.

DNA methylation serves as a reversible and prognostic biomarker for oral squamous cell carcinoma (OSCC) patients. It is unclear whether the effect of DNA methylation on OSCC overall survival varies with age. As a result, we performed a two-phase gene-age interaction study of OSCC prognosis on an epigenome-wide scale using the Cox proportional hazards model. We identified one CpG probe, cg11676291 MORN1 , whose effect was significantly modified by age (HRdiscovery = 1.018, p = 4.07 × 10-07, FDR-q = 3.67 × 10-02; HRvalidation = 1.058, p = 8.09 × 10-03; HR combined = 1.019, p = 7.36 × 10-10). Moreover, there was an antagonistic interaction between hypomethylation of cg11676291 MORN1 and age (HRinteraction = 0.284; 95% CI, 0.135-0.597; p = 9.04 × 10-04). The prognosis of OSCC patients was well discriminated by the prognostic score incorporating cg11676291 MORN1 -age interaction (HR high vs. low = 3.66, 95% CI: 2.40-5.60, p = 1.93 × 10-09). By adding 24 significant gene-age interactions using a looser criterion, we significantly improved the area under the receiver operating characteristic curve (AUC) of the model at 3- and 5-year prognostic prediction (AUC3-year = 0.80, AUC5-year = 0.79, C-index = 0.75). Our study identified a significant interaction between cg11676291 MORN1 and age on OSCC survival, providing a potential therapeutic target for OSCC patients.

Automatic Detection of Image-Based Features for Immunosuppressive Therapy Response Prediction in Oral Lichen Planus.

Oral lichen planus (OLP) is a chronic inflammatory disease, and the common management focuses on controlling inflammation with immunosuppressive therapy. While the response to the immunosuppressive therapy is heterogeneous, exploring the mechanism and prediction of the response gain greater importance. Here, we developed a workflow for prediction of immunosuppressive therapy response prediction in OLP, which could automatically acquire image-based features. First, 38 features were acquired from 208 OLP pathological images, and 6 features were subsequently obtained which had a significant impact on the effect of OLP immunosuppressive therapy. By observing microscopic structure and integrated with the corresponding transcriptome, the biological implications of the 6 features were uncovered. Though the pathway enrichment analysis, three image-based features which advantageous to therapy indicated the different lymphocytes infiltration, and the other three image-based features which bad for therapy respectively indicated the nicotinamide adenine dinucleotide (NADH) metabolic pathway, response to potassium ion pathway and adenosine monophosphate (AMP) activated protein kinase pathway. In addition, prediction models for the response to immunosuppressive therapy, were constructed with above image-based features. The best performance prediction model built by logistic regression showed an accuracy of 90% and the area under the receiver operating characteristic curve (AUROC) reached 0.947. This study provided a novel approach to automatically obtain biological meaningful image-based features from unannotated pathological images, which could indicate the immunosuppressive therapy in OLP. Besides, the novel and accurate prediction model may be useful for the OLP clinical management.

High-Accuracy Oral Squamous Cell Carcinoma Auxiliary Diagnosis System Based on EfficientNet.

It is important to diagnose the grade of oral squamous cell carcinoma (OSCC), but the current evaluation of the biopsy slide still mainly depends on the manual operation of pathologists. The workload of manual evaluation is large, and the results are greatly affected by the subjectivity of the pathologists. In recent years, with the development and application of deep learning, automatic evaluation of biopsy slides is gradually being applied to medical diagnoses, and it has shown good results. Therefore, a new OSCC auxiliary diagnostic system was proposed to automatically and accurately evaluate the patients' tissue slides. This is the first study that compared the effects of different resolutions on the results. The OSCC tissue slides from The Cancer Genome Atlas (TCGA, n=697) and our independent datasets (n=337) were used for model training and verification. In the test dataset of tiles, the accuracy was 93.1% at 20x resolution (n=306,134), which was higher than that at 10x (n=154,148, accuracy=90.9%) and at 40x (n=890,681, accuracy=89.3%). The accuracy of the new system based on EfficientNet, which was used to evaluate the tumor grade of the biopsy slide, reached 98.1% [95% confidence interval (CI): 97.1% to 99.1%], and the area under the receiver operating characteristic curve (AUROC) reached 0.998 (95%CI: 0.995 to 1.000) in the TCGA dataset. When verifying the model on the independent image dataset, the accuracy still reached 91.4% (95% CI: 88.4% to 94.4%, at 20x) and the AUROC reached 0.992 (95%CI: 0.982 to 1.000). It may benefit oral pathologists by reducing certain repetitive and time-consuming tasks, improving the efficiency of diagnosis, and facilitating the further development of computational histopathology.