Accurate glioma grading before surgery is of the utmost importance in treatment planning and prognosis prediction. But previous studies on magnetic resonance imaging (MRI) images were not effective ...enough. According to the remarkable performance of convolutional neural network (CNN) in medical domain, we hypothesized that a deep learning algorithm can achieve high accuracy in distinguishing the World Health Organization (WHO) low grade and high grade gliomas.
One hundred and thirteen glioma patients were retrospectively included. Tumor images were segmented with a rectangular region of interest (ROI), which contained about 80% of the tumor. Then, 20% data were randomly selected and leaved out at patient-level as test dataset. AlexNet and GoogLeNet were both trained from scratch and fine-tuned from models that pre-trained on the large scale natural image database, ImageNet, to magnetic resonance images. The classification task was evaluated with five-fold cross-validation (CV) on patient-level split.
The performance measures, including validation accuracy, test accuracy and test area under curve (AUC), averaged from five-fold CV of GoogLeNet which trained from scratch were 0.867, 0.909, and 0.939, respectively. With transfer learning and fine-tuning, better performances were obtained for both AlexNet and GoogLeNet, especially for AlexNet. Meanwhile, GoogLeNet performed better than AlexNet no matter trained from scratch or learned from pre-trained model.
In conclusion, we demonstrated that the application of CNN, especially trained with transfer learning and fine-tuning, to preoperative glioma grading improves the performance, compared with either the performance of traditional machine learning method based on hand-crafted features, or even the CNNs trained from scratch.
Herein, it is reported the influence of solution processing and treatments, such as adding marginal solvent, ultrasonication, and UV treatment, on the resulting perovskite (CsPbBr3) quantum dot ...(QD)/poly(3‐hexylthiophene) (P3HT) composite nanofibril films (CNFs) to improve the charge dissociation and photonic synaptic performance. A photonic synaptic transistor with CNFs can perform fundamental functions, including short‐term plasticity, long‐term plasticity, spike‐number‐dependent, and spike‐time‐dependent plasticity, to mimic sensing, computing, and memory functions. Notably, a synaptic device with CNFs presents an ultralow energy consumption of 0.18 fJ and zero‐gate operation. The superior performance of synaptic devices with CNFs can be attributed to two factors: (i) homogeneous axial distribution of the QDs and (ii) the formation of P3HT nanofibrils and co‐aggregates. Therefore, enhanced interfacial charge transfer between QDs and P3HT, ensuring decent carrier transport capability, is achieved. Collectively, the composite artificial synapse successfully provides an effective guide that offers a new perspective for the fabrication of one‐dimensional self‐assembled nanostructure‐based artificial synapses emulating human‐like memory, neuromorphic computing, and artificial intelligent systems.
Semiconducting self‐assembled composite nanostructures via solution processing is a promising strategy to improve the charge dissociation and photonic synaptic performance. In this study, quantum dot/poly(3‐hexylthiophene) nanofibrils are studied to understand the morphology/optoelectronic relation. The composite artificial synapse exhibits fundamental functions, including short‐term plasticity, long‐term plasticity, and spike‐number‐dependent and spike‐time‐dependent plasticity with ultralow energy consumption of 0.18 fJ and zero‐gate operation.
Despite rapid advances in modern medical technology and significant improvements in survival rates of many cancers, pancreatic cancer is still a highly lethal gastrointestinal cancer with a low ...5-year survival rate and difficulty in early detection. At present, the incidence and mortality of pancreatic cancer are increasing year by year worldwide, no matter in the United States, Europe, Japan, or China. Globally, the incidence of pancreatic cancer is projected to increase to 18.6 per 100000 in 2050, with the average annual growth of 1.1%, meaning that pancreatic cancer will pose a significant public health burden. Due to the special anatomical location of the pancreas, the development of pancreatic cancer is usually diagnosed at a late stage with obvious clinical symptoms. Therefore, a comprehensive understanding of the risk factors for pancreatic cancer is of great clinical significance for effective prevention of pancreatic cancer. In this paper, the epidemiological characteristics, developmental trends, and risk factors of pancreatic cancer are reviewed and analyzed in detail.
Two‐dimensional (2D) metal–organic framework nanosheets (MOF NSs) play a vital role in catalysis, but the most preparation is ultrasonication or solvothermal. Herein, a liquid–liquid interfacial ...synthesis method has been developed for the efficient fabrication of a series of 2D Ni MOF NSs. The active sites could be modulated by readily tuning the ratios of metal precursors and organic linkers (RM/L). The Ni MOF NSs display highly RM/L dependent activities towards 2e oxygen reduction reaction (ORR) to hydrogen peroxide (H2O2), where the Ni MOF NSs with the RM/L of 6 exhibit the optimal near‐zero overpotential, ca. 98 % H2O2 selectivity and production rate of ca. 80 mmol gcat−1 h−1 in 0.1 M KOH. As evidenced by X‐ray absorption fine structure spectroscopy, the coordination environment of active sites changed from saturation to unsaturation, and the partially unsaturated metal atoms are crucial to create optimal sites for enhancing the electrocatalysis.
2D Ni metal–organic framework nanosheets (MOF NSs) with controlled coordination mode were carefully created through a liquid‐liquid interfacial synthesis strategy for the first time and adopted as efficient electrocatalysts for hydrogen peroxide (H2O2) synthesis. The optimized partially unsaturated Ni MOF NSs‐6 exhibits near‐zero overpotential as well as ca. 98 % H2O2 selectivity in 0.1 M KOH, exceeding most electrocatalysts up to date.
Herein, interfacial engineering is demonstrated to improve the thermal stability of non‐fullerene bulk‐heterojunction (BHJ) OPVs to a practical level. An amphiphilic dendritic block copolymer (DBC) ...is developed through a facile coupling method and employed as the surface modifier of ZnO electron‐transporting layer in inverted OPVs. Besides showing distinct self‐assembly behavior, the synthesized DBC possesses high compatibility with plasmonic gold nanoparticles (NPs) due to the constituent malonamide and ethylene oxide units. The hybrid DBC@AuNPs interlayer is shown to improve device's performance from 14.0% to 15.4% because it enables better energy‐level alignment and improves interfacial compatibility at the ZnO/BHJ interface. Moreover, the DBC@AuNPs interlayer not only improves the interfacial thermal stability at the ZnO/BHJ interface but also endows a more ideal BHJ morphology with an enhanced thermal robustness. The derived device reserves 77% of initial PCE after thermal aging at 65 °C for 3000 h and yields an extended T80 lifetime of >1100 h when stored at a constant thermal condition at 65 °C, outperforming the control device. Finally, the device is evaluated to possess a T80 lifetime of over 1.79 years at room temperature (298 K) when stored in an inert condition, showing great potential for commercialization.
An amphiphilic dendritic block copolymer is developed to serve as an efficient surface modifier of ZnO electron‐transporting layer in an organic photovoltaic device. When using an interlayer based on its hybridization with gold nanoparticles, the device can deliver improved performance and possess a lifetime of over 1.79 years when stored at room temperature in inert conditions.
Detecting visually salient regions in images is one of the fundamental problems in computer vision. We propose a novel method to decompose an image into large scale perceptually homogeneous elements ...for efficient salient region detection, using a soft image abstraction representation. By considering both appearance similarity and spatial distribution of image pixels, the proposed representation abstracts out unnecessary image details, allowing the assignment of comparable saliency values across similar regions, and producing perceptually accurate salient region detection. We evaluate our salient region detection approach on the largest publicly available dataset with pixel accurate annotations. The experimental results show that the proposed method outperforms 18 alternate methods, reducing the mean absolute error by 25.2% compared to the previous best result, while being computationally more efficient.
A novel approach for using conjugated rod–coil materials as a floating gate in the fabrication of nonvolatile photonic transistor memory devices, consisting of n‐type Sol‐PDI and p‐type C10‐DNTT, is ...presented. Sol‐PDI and C10‐DNTT are used as dual functions of charge‐trapping (conjugated rod) and tunneling (insulating coil), while n‐type BPE‐PDI and p‐type DNTT are employed as the corresponding transporting layers. By using the same conjugated rod in the memory layer and transporting channel with a self‐assembled structure, both n‐type and p‐type memory devices exhibit a fast response, a high current contrast between “Photo‐On” and “Electrical‐Off” bistable states over 105, and an extremely low programing driving force of 0.1 V. The fabricated photon‐driven memory devices exhibit a quick response to different wavelengths of light and a broadband light response that highlight their promising potential for light‐recorder and synaptic device applications.
High‐performance photonic transistor memory devices are fabricated using conjugated rod–coil materials as a photoactive floating gate, in which the conjugated rods and side‐chain coils act as charge‐trapping and tunneling moieties, respectively. By inheriting their self‐assembled structure, both n‐type and p‐type memory devices exhibit a fast response, a current contrast over 105, and an extremely low programing driving force of 0.1 V.
Background
Accurate glioma grading plays an important role in the clinical management of patients and is also the basis of molecular stratification nowadays.
Purpose/Hypothesis
To verify the ...superiority of radiomics features extracted from multiparametric MRI to glioma grading and evaluate the grading potential of different MRI sequences or parametric maps.
Study Type
Retrospective; radiomics.
Population
A total of 153 patients including 42, 33, and 78 patients with Grades II, III, and IV gliomas, respectively.
Field Strength/Sequence
3.0T MRI/T1‐weighted images before and after contrast‐enhanced, T2‐weighted, multi‐b‐value diffusion‐weighted and 3D arterial spin labeling images.
Assessment
After multiparametric MRI preprocessing, high‐throughput features were derived from patients' volumes of interests (VOIs). The support vector machine‐based recursive feature elimination was adopted to find the optimal features for low‐grade glioma (LGG) vs. high‐grade glioma (HGG), and Grade III vs. IV glioma classification tasks. Then support vector machine (SVM) classifiers were established using the optimal features. The accuracy and area under the curve (AUC) was used to assess the grading efficiency.
Statistical Tests
Student's t‐test or a chi‐square test were applied on different clinical characteristics to confirm whether intergroup significant differences exist.
Results
Patients' ages between LGG and HGG groups were significantly different (P < 0.01). For each patient, 420 texture and 90 histogram parameters were derived from 10 VOIs of multiparametric MRI. SVM models were established using 30 and 28 optimal features for classifying LGGs from HGGs and grades III from IV, respectively. The accuracies/AUCs were 96.8%/0.987 for classifying LGGs from HGGs, and 98.1%/0.992 for classifying grades III from IV, which were more promising than using histogram parameters or using the single sequence MRI.
Data Conclusion
Texture features were more effective for noninvasively grading gliomas than histogram parameters. The combined application of multiparametric MRI provided a higher grading efficiency. The proposed radiomic strategy could facilitate clinical decision‐making for patients with varied glioma grades.
Level of Evidence: 3
Technical Efficacy: Stage 2
J. Magn. Reson. Imaging 2018;48:1518–1528
In this note, a novel adaptive compensation control scheme is proposed for a class of nonlinear systems with unknown control direction and a possibly infinite number of unknown actuator failures. By ...introducing a bound estimation approach, high-order Lyapunov functions and a Nussbaum function with faster growth rate, the obstacle caused by unknown failures and unknown control direction is successfully circumvented and all signals of the closed-loop system are proved to be globally uniformly bounded. Moreover, the proposed scheme is able to steer the tracking error into a predefined small residue set. Simulation results are presented to illustrate the effectiveness of the proposed scheme.
A key challenge in feature correspondence is the difficulty in differentiating true and false matches at a local descriptor level. This forces adoption of strict similarity thresholds that discard ...many true matches. However, if analyzed at a global level, false matches are usually randomly scattered while true matches tend to be coherent (clustered around a few dominant motions), thus creating a coherence based separability constraint. This paper proposes a non-linear regression technique that can discover such a coherence based separability constraint from highly noisy matches and embed it into a correspondence likelihood model. Once computed, the model can filter the entire set of nearest neighbor matches (which typically contains over 90 percent false matches) for true matches. We integrate our technique into a full feature correspondence system which reliably generates large numbers of good quality correspondences over wide baselines where previous techniques provide few or no matches.