Tumor initiating cells (TICs) possessing cancer stemness were shown to be enriched after therapy, resulting in the relapse and metastasis of head and neck squamous cell carcinomas (HNC). An effective therapeutic approach suppressing the HNC-TICs would be a potential method to improve the treatments for HNC. We observed that the treatment of silibinin (SB) dose dependently down-regulated the ALDH1 activity, CD133 positivity, stemness signatures expression, self-renewal property, and chemoresistance in ALDH1+CD44+ HNC-TICs. Using miRNA-microarray and mechanistic studies, SB increased the expression of microRNA-494 (miR-494) and both Bmi1 and ADAM10 were identified as the novel targets of miR-494. Moreover, overexpression of miR-494 results in a reduction in cancer stemness. However, knockdown of miR-494 in CD44−ALDH1−non-HNC-TICs enhanced cancer stemness and oncogenicity, while co-knockdown of Bmi1 and ADAM10 effectively reversed these phenomena. Mice model showed that SB treatment by oral gavage to xenograft tumors reduced tumor growth and prolonged the survival time of tumor-bearing mice by activation of miR-494-inhibiting Bmi1/ADAM10 expression. Survival analysis indicated that a miR494highBmi1lowADAM10low phenotype predicted a favourable clinical outcome. We conclude that the inhibition of tumor aggressiveness in HNC-TICs by SB was mediated by up-regulation miR-494, suggesting that SB would be a valuable anti-cancer drug for treatment of HNC.

Finite element method (FEM) analysis has become a common method to analyze the lesion formation during temperature-controlled radiofrequency (RF) cardiac ablation. We present a process of FEM modeling a system including blood, myocardium, and an ablation catheter with a thermistor embedded at the tip. The simulation used a simple proportional-integral (PI) controller to control the entire process operated in temperature-controlled mode. Several factors affect the lesion size such as target temperature, blood flow rate, and application time. We simulated the time response of RF ablation at different locations by using different target temperatures. The applied sites were divided into two groups each with a different convective heat transfer coefficient. The first group was high-flow such as the atrioventricular (AV) node and the atrial aspect of the AV annulus, and the other was low-flow such as beneath the valve or inside the coronary sinus. Results showed the change of lesion depth and lesion width with time, under different conditions. We collected data for all conditions and used it to create a database. We implemented a user-interface, the lesion size estimator, where the user enters set temperature and location. Based on the database, the software estimated lesion dimensions during different applied durations. This software could be used as a first-step predictor to help the electrophysiologist choose treatment parameters.

No abstract

Generative models based on deep neural networks often have a high-dimensional latent space, ranging sometimes to a few hundred dimensions or even higher, which typically makes them hard for a user to explore directly. We propose differential subspace search to allow efficient iterative user exploration in such a space, without relying on domain- or data-specific assumptions. We develop a general framework to extract low-dimensional subspaces based on a local differential analysis of the generative model, such that a small change in such a subspace would provide enough change in the resulting data. We do so by applying singular value decomposition to the Jacobian of the generative model and forming a subspace with the desired dimensionality spanned by a given number of singular vectors stochastically selected on the basis of their singular values, to maintain ergodicity. We use our framework to present 1D subspaces to the user via a 1D slider interface. Starting from an initial location, the user finds a new candidate in the presented 1D subspace, which is in turn updated at the new candidate location. This process is repeated until no further improvement can be made. Numerical simulations show that our method can better optimize synthetic black-box objective functions than the alternatives that we tested. Furthermore, we conducted a user study using complex generative models and the results show that our method enables more efficient exploration of high-dimensional latent spaces than the alternatives.

SUMMARYThis paper surveys major techniques in 3D communications area, which covers the whole pipeline of the 3D video communication framework, including 3D content creation, data representation, compression, delivery, decompression, post-processing, and 3D scene rendering stages. Both the current state-of-the-art, stereo 3D, and future trend, free-viewpoint 3D, are demonstrated in details. On the other hand, the paper highlights a few features in the emerging 4G wireless systems that are critical for 3D communications system design. At the end, the topics with potential but challenges, for example 3D over 4G networks, distributed 3D video coding, 3D multi-user communication, scalability and universal 3D access, are discussed and pointed out to audiences for further investigation.