A partial convolution generative adversarial network for lesion synthesis and enhanced liver tumor segmentation.

Lesion segmentation is critical for clinicians to accurately stage the disease and determine treatment strategy. Deep learning based automatic segmentation can improve both the segmentation efficiency and accuracy. However, training a robust deep learning segmentation model requires sufficient training examples with sufficient diversity in lesion location and lesion size. This study is to develop a deep learning framework for generation of synthetic lesions with various locations and sizes that can be included in the training dataset to enhance the lesion segmentation performance. The lesion synthesis network is a modified generative adversarial network (GAN). Specifically, we innovated a partial convolution strategy to construct a U-Net-like generator. The discriminator is designed using Wasserstein GAN with gradient penalty and spectral normalization. A mask generation method based on principal component analysis (PCA) was developed to model various lesion shapes. The generated masks are then converted into liver lesions through a lesion synthesis network. The lesion synthesis framework was evaluated for lesion textures, and the synthetic lesions were used to train a lesion segmentation network to further validate the effectiveness of the lesion synthesis framework. All the networks are trained and tested on the LITS public dataset. Our experiments demonstrate that the synthetic lesions generated by our approach have very similar distributions for the two parameters, GLCM-energy and GLCM-correlation. Including the synthetic lesions in the segmentation network improved the segmentation dice performance from 67.3% to 71.4%. Meanwhile, the precision and sensitivity for lesion segmentation were improved from 74.6% to 76.0% and 66.1% to 70.9%, respectively. The proposed lesion synthesis approach outperforms the other two existing approaches. Including the synthetic lesion data into the training dataset significantly improves the segmentation performance.

ABCD1 as a Novel Diagnostic Marker for Solid Pseudopapillary Neoplasm of the Pancreas.

The diagnosis of solid pseudopapillary neoplasm of the pancreas (SPN) can be challenging due to potential confusion with other pancreatic neoplasms, particularly pancreatic neuroendocrine tumors (NETs), using current pathological diagnostic markers. We conducted a comprehensive analysis of bulk RNA sequencing data from SPNs, NETs, and normal pancreas, followed by experimental validation. This analysis revealed an increased accumulation of peroxisomes in SPNs. Moreover, we observed significant upregulation of the peroxisome marker ABCD1 in both primary and metastatic SPN samples compared with normal pancreas and NETs. To further investigate the potential utility of ABCD1 as a diagnostic marker for SPN via immunohistochemistry staining, we conducted verification in a large-scale patient cohort with pancreatic tumors, including 127 SPN (111 primary, 16 metastatic samples), 108 NET (98 nonfunctional pancreatic neuroendocrine tumor, NF-NET, and 10 functional pancreatic neuroendocrine tumor, F-NET), 9 acinar cell carcinoma (ACC), 3 pancreatoblastoma (PB), 54 pancreatic ductal adenocarcinoma (PDAC), 20 pancreatic serous cystadenoma (SCA), 19 pancreatic mucinous cystadenoma (MCA), 12 pancreatic ductal intraepithelial neoplasia (PanIN) and 5 intraductal papillary mucinous neoplasm (IPMN) samples. Our results indicate that ABCD1 holds promise as an easily applicable diagnostic marker with exceptional efficacy (AUC=0.999, sensitivity=99.10%, specificity=100%) for differentiating SPN from NET and other pancreatic neoplasms through immunohistochemical staining.

Highly efficient, selective, and stable photocatalytic methane coupling to ethane enabled by lattice oxygen looping.

Light-driven oxidative coupling of methane (OCM) for multi-carbon (C2+) product evolution is a promising approach toward the sustainable production of value-added chemicals, yet remains challenging due to its low intrinsic activity. Here, we demonstrate the integration of bismuth oxide (BiOx) and gold (Au) on titanium dioxide (TiO2) substrate to achieve a high conversion rate, product selectivity, and catalytic durability toward photocatalytic OCM through rational catalytic site engineering. Mechanistic investigations reveal that the lattice oxygen in BiOx is effectively activated as the localized oxidant to promote methane dissociation, while Au governs the methyl transfer to avoid undesirable overoxidation and promote carbon─carbon coupling. The optimal Au/BiOx-TiO2 hybrid delivers a conversion rate of 20.8 millimoles per gram per hour with C2+ product selectivity high to 97% in the flow reactor. More specifically, the veritable participation of lattice oxygen during OCM is chemically looped by introduced dioxygen via the Mars-van Krevelen mechanism, endowing superior catalyst stability.

A nonmetallic plasmonic catalyst for photothermal CO2 flow conversion with high activity, selectivity and durability.

The meticulous design of active sites and light absorbers holds the key to the development of high-performance photothermal catalysts for CO2 hydrogenation. Here, we report a nonmetallic plasmonic catalyst of Mo2N/MoO2-x nanosheets by integrating a localized surface plasmon resonance effect with two distinct types of active sites for CO2 hydrogenation. Leveraging the synergism of dual active sites, H2 and CO2 molecules can be simultaneously adsorbed and activated on N atom and O vacancy, respectively. Meanwhile, the plasmonic effect of this noble-metal-free catalyst signifies its promising ability to convert photon energy into localized heat. Consequently, Mo2N/MoO2-x nanosheets exhibit remarkable photothermal catalytic performance in reverse water-gas shift reaction. Under continuous full-spectrum light irradiation (3 W·cm-2) for a duration of 168 h, the nanosheets achieve a CO yield rate of 355 mmol·gcat-1·h-1 in a flow reactor with a selectivity exceeding 99%. This work offers valuable insights into the precise design of noble-metal-free active sites and the development of plasmonic catalysts for reducing carbon footprints.

Identification of Potential Key Genes and Prognostic Biomarkers of Lung Cancer Based on Bioinformatics.

Objective: To analyze and identify the core genes related to the expression and prognosis of lung cancer including lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) by bioinformatics technology, with the aim of providing a reference for clinical treatment. Methods: Five sets of gene chips, GSE7670, GSE151102, GSE33532, GSE43458, and GSE19804, were obtained from the Gene Expression Omnibus (GEO) database. After using GEO2R to analyze the differentially expressed genes (DEGs) between lung cancer and normal tissues online, the common DEGs of the five sets of chips were obtained using a Venn online tool and imported into the Database for Annotation, Visualization, and Integrated Discovery (DAVID) database for Gene Ontology (GO) enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. The protein-protein interaction (PPI) network was constructed by STRING online software for further study, and the core genes were determined by Cytoscape software and KEGG pathway enrichment analysis. The clustering heat map was drawn by Excel software to verify its accuracy. In addition, we used the University of Alabama at Birmingham Cancer (UALCAN) website to analyze the expression of core genes in P53 mutation status, confirmed the expression of crucial core genes in lung cancer tissues with Gene Expression Profiling Interactive Analysis (GEPIA) and GEPIA2 online software, and evaluated their prognostic value in lung cancer patients with the Kaplan-Meier online plotter tool. Results: CHEK1, CCNB1, CCNB2, and CDK1 were selected. The expression levels of these four genes in lung cancer tissues were significantly higher than those in normal tissues. Their increased expression was negatively correlated with lung cancer patients (including LUAD and LUSC) prognosis and survival rate. Conclusion: CHEK1, CCNB1, CCNB2, and CDK1 are the critical core genes of lung cancer and are highly expressed in lung cancer. They are negatively correlated with the prognosis of lung cancer patients (including LUAD and LUSC) and closely related to the formation and prediction of lung cancer. They are valuable predictors and may be predictive biomarkers of lung cancer.

Key genes involved with prognosis were identified in lung adenocarcinoma by integrated bioinformatics analysis.

Objective: By screening the core genes in lung adenocarcinoma (LUAD) with bioinformatics, our study evaluated its prognosis value and role in infiltration process of immune cells. Methods: Using GEO database, we screened 5 gene chips, including GSE11072, GSE32863, GSE43458, GSE115002, and GSE116959. Then, we obtained the corresponding differentially expressed genes by analyzed 5 gene chips online by GEO2R (P < 0.05, |logFC| > 1). Then, through DAVID online platform, Cytoscape 3.6.1 software and PPI network analysis, the network was visualized and obtain the final core genes. Next, we plan to use the GEPIA, UALCAN, Kaplan-Meier plotter and Time 2.0 database for corresponding analysis. The GEPIA database was used to verify the expression of core genes in LUAD and normal lung tissues, and survival analysis was used to evaluate the value of core genes in the prognosis of LUAD patients. UALCAN was used to verify the expression of the LUAD core gene and promoter methylation status, and the predictive value of core genes was evaluated in LUAD patients by the Kaplan-Meier plotter online tool. Then, we used the Time 2.0 database to identify the relationship to immune infiltration in LUAD. Finally, we used the human protein atlas (HPA) database for online immunohistochemical analysis of the expressed proteins. Results: The expression of CCNB2 and CDC20 in LUAD were higher than those in normal lung tissues, their increased expression was negatively correlated with the overall survival rate of LUAD, and they were involved in cell cycle signal transduction, oocyte meiosis signal transduction as well as the infiltration process of immune cells in LUAD. The expression proteins of CCNB2 and CDC20 were also different in lung cancer tissue and normal lung tissue. Therefore, CCNB2 and CDC20 were identified as the vital core genes. Conclusion: CCNB2 and CDC20 are essential genes that may constitute prognostic biomarkers in LUAD, they also participate the immune infiltration process and protein expression process of LUAD, and might provides basis for clinical anti-tumor drug research.

Multi-site, Multi-domain Airway Tree Modeling.

Open international challenges are becoming the de facto standard for assessing computer vision and image analysis algorithms. In recent years, new methods have extended the reach of pulmonary airway segmentation that is closer to the limit of image resolution. Since EXACT'09 pulmonary airway segmentation, limited effort has been directed to the quantitative comparison of newly emerged algorithms driven by the maturity of deep learning based approaches and extensive clinical efforts for resolving finer details of distal airways for early intervention of pulmonary diseases. Thus far, public annotated datasets are extremely limited, hindering the development of data-driven methods and detailed performance evaluation of new algorithms. To provide a benchmark for the medical imaging community, we organized the Multi-site, Multi-domain Airway Tree Modeling (ATM'22), which was held as an official challenge event during the MICCAI 2022 conference. ATM'22 provides large-scale CT scans with detailed pulmonary airway annotation, including 500 CT scans (300 for training, 50 for validation, and 150 for testing). The dataset was collected from different sites and it further included a portion of noisy COVID-19 CTs with ground-glass opacity and consolidation. Twenty-three teams participated in the entire phase of the challenge and the algorithms for the top ten teams are reviewed in this paper. Both quantitative and qualitative results revealed that deep learning models embedded with the topological continuity enhancement achieved superior performance in general. ATM'22 challenge holds as an open-call design, the training data and the gold standard evaluation are available upon successful registration via its homepage (https://atm22.grand-challenge.org/).

Label-free coronavirus disease 2019 lesion segmentation based on synthetic healthy lung image subtraction.

PURPOSE: Coronavirus disease 2019 (COVID-19) has become a global pandemic and is still posing a severe health risk to the public. Accurate and efficient segmentation of pneumonia lesions in computed tomography (CT) scans is vital for treatment decision-making. We proposed a novel unsupervised approach using a cycle consistent generative adversarial network (cycle-GAN) which automates and accelerates the process of lesion delineation. METHOD: The workflow includes lung volume segmentation, healthy lung image synthesis, infected and healthy image subtraction, and binary lesion mask generation. The lung volume was first delineated using a pre-trained U-net and worked as the input for the following network. A cycle-GAN was trained to generate synthetic healthy lung CT images from infected lung images. After that, the pneumonia lesions were extracted by subtracting the synthetic healthy lung CT images from the infected lung CT images. A median filter and k-means clustering were then applied to contour the lesions. The auto segmentation approach was validated on three different datasets. RESULTS: The average Dice coefficient reached 0.666 ± 0.178 on the three datasets. Especially, the dice reached 0.748 ± 0.121 and 0.730 ± 0.095, respectively, on two public datasets Coronacases and Radiopedia. Meanwhile, the average precision and sensitivity for lesion segmentation on the three datasets were 0.679 ± 0.244 and 0.756 ± 0.162. The performance is comparable to existing supervised segmentation networks and outperforms unsupervised ones. CONCLUSION: The proposed label-free segmentation method achieved high accuracy and efficiency in automatic COVID-19 lesion delineation. The segmentation result can serve as a baseline for further manual modification and a quality assurance tool for lesion diagnosis. Furthermore, due to its unsupervised nature, the result is not influenced by physicians' experience which otherwise is crucial for supervised methods.