Weakly supervised bilayer convolutional network in segmentation of HER2 related cells to guide HER2 targeted therapies.

Overexpression of human epidermal growth factor receptor 2 (HER2/ERBB2) is identified as a prognostic marker in metastatic breast cancer and a predictor to determine the effects of ERBB2-targeted drugs. Accurate ERBB2 testing is essential in determining the optimal treatment for metastatic breast cancer patients. Brightfield dual in situ hybridization (DISH) was recently authorized by the United States Food and Drug Administration for the assessment of ERRB2 overexpression, which however is a challenging task due to a variety of reasons. Firstly, the presence of touching clustered and overlapping cells render it difficult for segmentation of individual HER2 related cells, which must contain both ERBB2 and CEN17 signals. Secondly, the fuzzy cell boundaries make the localization of each HER2 related cell challenging. Thirdly, variation in the appearance of HER2 related cells is large. Fourthly, as manual annotations are usually made on targets with high confidence, causing sparsely labeled data with some unlabeled HER2 related cells defined as background, this will seriously confuse fully supervised AI learning and cause poor model outcomes. To deal with all issues mentioned above, we propose a two-stage weakly supervised deep learning framework for accurate and robust assessment of ERBB2 overexpression. The effectiveness and robustness of the proposed deep learning framework is evaluated on two DISH datasets acquired at two different magnifications. The experimental results demonstrate that the proposed deep learning framework achieves an accuracy of 96.78 ± 1.25, precision of 97.77 ± 3.09, recall of 84.86 ± 5.83 and Dice Index of 90.77 ± 4.1 and an accuracy of 96.43 ± 2.67, precision of 97.82 ± 3.99, recall of 87.14 ± 10.17 and Dice Index of 91.87 ± 6.51 for segmentation of ERBB2 overexpression on the two experimental datasets, respectively. Furthermore, the proposed deep learning framework outperforms 15 state-of-the-art benchmarked methods by a significant margin (P<0.05) with respect to IoU on both datasets.

Suppressing the Initial Growth of Sidewall GaN by Modifying Micron-Sized Patterned Sapphire Substrate with H3PO4-Based Etchant.

Micron-sized patterned sapphire substrates (PSS) are used to improve the performance of GaN-based light-emitting diodes (LEDs). However, the growth of GaN is initiated not only from the bottom c-plane but also from the sidewall of the micron-sized patterns. Therefore, the coalescence of these GaN crystals creates irregular voids. In this study, two kinds of nucleation layers (NL)-ex-situ AlN NL and in-situ GaN NL-were used, and the growth of sidewall GaN was successfully suppressed in both systems by modifying the micron-sized PSS surface.

High Performance and Low power Monolithic Three-Dimensional Sub-50 nm Poly Si Thin film transistor (TFTs) Circuits.

Development of manufacture trend for TFTs technologies has focused on improving electrical properties of films with the cost reduction to achieve commercialization. To achieve this goal, high-performance sub-50 nm TFTs-based MOSFETs with ON-current (Ion)/subthreshold swing (S.S.) of 181 µA/µm/107 mV/dec and 188 µA/µm/98 mV/dec for NMOSFETs and PMOSFETs in a monolithic 3D circuit were demonstrated by a low power with low thermal budget process. In addition, a stackable static random access memory (SRAM) integrated with TFTs-based MOSFET with static noise margins (SNM) equals to 390 mV at VDD = 1.0 V was demonstrated. Overall processes include a low thermal budget via ultra-flat and ultra-thin poly-Si channels by solid state laser crystallization process, chemical-mechanical polishing (CMP) planarization, plasma-enhanced atomic layer deposition (ALD) gate stacking layers and infrared laser activation with a low thermal budget. Detailed material and electrical properties were investigated. The advanced 3D architecture with closely spaced inter-layer dielectrics (ILD) enables high-performance stackable MOSFETs and SRAM for power-saving IoT/mobile products at a low cost or flexible substrate.

Synthesis of Pt nanopetals on highly ordered silicon nanocones for enhanced methanol electrooxidation activity.

Platinum (Pt) nanopetals were electrodeposited on highly ordered silicon nanocones (SiNCs) and explored as the electrocatalyst for methanol oxidation reaction (MOR) for direct methanol fuel cells applications. Highly ordered SiNCs array fabricated using the porous anodic aluminum oxide as the template had a high surface area. Well-dispersed Pt nanopetals possessing high electrocatalytic surface area was synthesized by pulse-electrodeposition on the SiNCs. Pt nanopetals loaded on highly ordered SiNC support exhibited very good catalytic activity for MOR and a high tolerance against CO poisoning, as compared to Pt nanoflowers/flat Si, Pt nanoparticles/flat Si, and many previously reported works. The abundance of a large surface area for facile transport of methanol, SiO(2) sites in the vicinity of the SiNCs, as well as less contact area between the Pt nanopetals catalyst and SiNCs are suggested to be the major factors enhancing the electrocatalytic performance of the Pt nanopetal/SiNC electrode. Moreover, we believe this new nanostructure (Pt nanopetals/SiNCs) will enable many new advances in nanotechnology.

A promising approach to the synthesis of 3D nanoporous graphitic carbon as a unique electrocatalyst support for methanol oxidation.

A 3D nanoporous graphitic carbon (g-C) material is synthesized by using an adamantane (C(10)H(16)) flame, and utilized to support a Pt(50)-Ru(50) alloy catalyst. The physico-chemical properties of the Pt(50)-Ru(50)/3D nanoporous g-C electrode are examined by a range of spectroscopy techniques as well as Brunauer-Emmett-Teller surface area analysis. Cyclic voltammetry measurements are used for electrochemical characterization of the Pt(50)-Ru(50)/3D nanoporous g-C electrode. The electrochemical investigations show that the supported Pt(50)-Ru(50) has excellent activity and stability towards methanol electro-oxidation. Good CO tolerance is also shown, and considered to be due to the presence of Ru nanoparticles. It is proposed that Ru is able to promote the oxidation of strongly adsorbed CO on Pt by supplying an oxygen source: Ru(OH)(ad). Moreover, the presence of 3D nanopores in the g-C support may also contribute to the observed higher current density by virtue of the easy transport of methanol and the oxidation products through these nanopores.