Search

ObjectivesThe diagnosis and treatment of brain tumors have greatly benefited from extensive research in traditional radiomics, leading to improved efficiency for clinicians. With the rapid development of cutting-edge technologies, especially deep learning, further improvements in accuracy and automation are expected. In this study, we explored a hybrid deep learning scheme that integrates several advanced techniques to achieve reliable diagnosis of primary brain tumors with enhanced classification performance and interpretability.MethodsThis study retrospectively included 230 patients with primary brain tumors, including 97 meningiomas, 66 gliomas and 67 pituitary tumors, from the First Affiliated Hospital of Yangtze University. The effectiveness of the proposed scheme was validated by the included data and a commonly used data. Based on super-resolution reconstruction and dynamic learning rate annealing strategies, we compared the classification results of several deep learning models. The multi-classification performance was further improved by combining feature transfer and machine learning. Classification performance metrics included accuracy (ACC), area under the curve (AUC), sensitivity (SEN), and specificity (SPE).ResultsIn the deep learning tests conducted on two datasets, the DenseNet121 model achieved the highest classification performance, with five-test accuracies of 0.989 ± 0.006 and 0.967 ± 0.013, and AUCs of 0.999 ± 0.001 and 0.994 ± 0.005, respectively. In the hybrid deep learning tests, LightGBM, a promising classifier, achieved accuracies of 0.989 and 0.984, which were improved from the original deep learning scheme of 0.987 and 0.965. Sensitivities for both datasets were 0.985, specificities were 0.988 and 0.984, respectively, and relatively desirable receiver operating characteristic (ROC) curves were obtained. In addition, model visualization studies further verified the reliability and interpretability of the results.ConclusionsThese results illustrated that deep learning models combining several advanced technologies can reliably improve the performance, automation, and interpretability of primary brain tumor diagnosis, which is crucial for further brain tumor diagnostic research and individualized treatment.

Abstract Nanocomposite modification has attracted much attention in improving properties of bio-based polymer coating material for coated fertilizer. Herein two comparable polyhedral oligomeric silsesquioxanes (POSS), with eight poly(ethylene glycol) (PEG) and octaphenyl groups attached to the cage, respectively, were successfully incorporated into thin castor oil-based polyurethane coatings via in-situ polymerization on the urea surface. The nanostructure coatings are environmentally friendly, easy to prepare, and property-tunable. The results show that the vertex group of POSS had a pronounced influence on dispersion level and interaction between polyurethane and POSS that well-tuned the release pattern and period of coated urea, even at the coating rate as low as of 2 wt%. The liquid POSS with long and flexible PEG groups had better compatibility and dispersibility in polyurethane matrix than the solid POSS with rigid octaphenyl groups, as evidenced by SEM/EDS. The unique properties were resulted from the different extents of physical crosslinkings. This modification of bio-based polyurethane coating with POSS provided an alternative method of regulating and controlling the properties of coated fertilizer.

The synthesis of branched-chain alkyl polyglucosides (APG) from 2-butyl-1-octanol, a Guerbet alcohol, in an acid/phase transfer catalyst (PTC) system was studied. Under optimized conditions the conversion rate of this reaction was 98.4% and the selectivity 16.5:1 (APG to polysaccharide, w/o). The APG was characterized by HPLC, LC/MS and 1HNMR. The surfactant properties showed that the wetting time was 17.9 s, and the emulsification time of a 0.15% aqueous solution with soybean oil was 22 min 44 s and with liquid paraffin 11 min 05 s. The critical micelle concentration (CMC) was 5 · 10-4 mol/L, and the surface tension was 25.87 mN/m with a 1% aqueous solution. It is a low foaming surfactant. [ABSTRACT FROM AUTHOR]

This study reports on what we believe are novel mechanism(s) of the vascular protective action of adiponectin. We used intravital microscopy to measure leukocyte-endothelium interactions in adiponectin-deficient (Ad(-/-)) mice and found that adiponectin deficiency was associated with a 2-fold increase in leukocyte rolling and a 5-fold increase in leukocyte adhesion in the microcirculation. Measurement of endothelial NO (eNO) revealed that adiponectin deficiency drastically reduced levels of eNO in the vascular wall. Immunohistochemistry demonstrated increased expression of E-selectin and VCAM-1 in the vascular endothelium of Ad(-/-) mice. Systemic administration of the recombinant globular adiponectin domain (gAd) to Ad(-/-) mice significantly attenuated leukocyte-endothelium interactions and adhesion molecule expression in addition to restoring physiologic levels of eNO. Importantly, prior administration of gAd also protected WT mice against TNF-alpha-induced leukocyte-endothelium interactions, indicating a pharmacologic action of gAd. Mechanistically, blockade of eNOS with N(omega)-nitro-L-arginine methyl ester ( L-NAME) abolished the inhibitory effect of gAd on leukocyte adhesion, demonstrating the obligatory role of eNOS signaling in the antiinflammatory action of gAd. We believe this is the first demonstration that gAd protects the vasculature in vivo via increased NO bioavailability with suppression of leukocyte-endothelium interactions. Overall, we provide evidence that loss of adiponectin induces a primary state of endothelial dysfunction with increased leukocyte-endothelium adhesiveness. [ABSTRACT FROM AUTHOR]