BACKGROUND.Many liver staging systems have been proposed for patients with hepatocellular carcinoma after locoregional therapy; however, controversies persist regarding which system is the best. In this study, the authors compared the performance of 7 staging systems in a cohort of patients with hepatocellular carcinoma who underwent transarterial chemoembolization.METHODS.In total, 131 patients with hepatocellular carcinoma who underwent transarterial chemoembolization between August 1998 and February 2005 were included in the study. Demographic, laboratory, and tumor characteristics were determined at diagnosis and before therapy. At the time of censorship, 109 patients had died (83.2%). Predictors of survival were identified by using the Cox proportional hazards model. The likelihood‐ratio chi‐square statistic and the Akaike Information Criterion were calculated for 7 prognostic systems to evaluate their discriminatory ability. Comparisons of the survival rate between each stage were performed to evaluate the monotonicity of the gradients using Kaplan‐Meier estimation and the log‐rank test.RESULTS.The 5‐year survival rate for the entire cohort was 13.6%. The independent predictors of survival were serum albumin level (≤3.4 g/dL), the presence of ascites, serum α‐fetoprotein level (>60 ng/mL), and portal or hepatic vein tumor thrombosis (P = .001, P = .001, P = .004, and P = .000, respectively). The Cancer of the Liver Italian Program classification system was superior to the other 6 prognostic systems regarding discriminatory ability and the monotonicity of the gradients.CONCLUSIONS.In this comparison of many staging systems, the Cancer of Liver Italian Program system provided the best prognostic stratification for a cohort the patients with hepatocellular carcinoma who underwent transarterial chemoembolization. Cancer 2008. © 2007 American Cancer Society.

Inspired by the human somatosensory system, pressure applied to multiple pressure sensors is received in parallel and combined into a representative signal pattern, which is subsequently processed using machine learning. The pressure signals are combined using a wireless system, where each sensor is assigned a specific resonant frequency on the reflection coefficient (S11) spectrum, and the applied pressure changes the magnitude of the S11 pole with minimal frequency shift. This allows the differentiation and identification of the pressure applied to each sensor. The pressure sensor consists of polypyrrole‐coated microstructured poly(dimethylsiloxane) placed on top of electrodes, operating as a capacitive sensor. The high dielectric constant of polypyrrole enables relatively high pressure‐sensing performance. The coils are vertically stacked to enable the reader to receive the signals from all of the sensors simultaneously at a single location, analogous to the junction between neighboring primary neurons to a secondary neuron. Here, the stacking order is important to minimize the interference between the coils. Furthermore, convolutional neural network (CNN)‐based machine learning is utilized to predict the applied pressure of each sensor from unforeseen S11 spectra. With increasing training, the prediction accuracy improves (with mean squared error of 0.12), analogous to humans' cognitive learning ability.

Bit-depth is the number of bits for each color channel of a pixel in an image. Although many modern displays support unprecedented higher bit-depth to show more realistic and natural colors with a high dynamic range, most media sources are still in bit-depth of 8 or lower. Since insufficient bit-depth may generate annoying false contours or lose detailed visual appearance, bit-depth expansion (BDE) from low bit-depth (LBD) images to high bit-depth (HBD) images becomes more and more important. In this paper, we adopt a learning-based approach for BDE and propose a novel CNN-based bit-depth expansion network (BitNet) that can effectively remove false contours and restore visual details at the same time. We have carefully designed our BitNet based on an encoder-decoder architecture with dilated convolutions and a novel multi-scale feature integration. We have performed various experiments with four different datasets including MIT-Adobe FiveK, Kodak, ESPL v2, and TESTIMAGES, and our proposed BitNet has achieved state-ofthe-art performance in terms of PSNR and SSIM among other existing BDE methods and famous CNN-based image processing networks. Unlike previous methods that separately process each color channel, we treat all RGB channels at once and have greatly improved color restoration. In addition, our network has shown the fastest computational speed in near real-time.

No abstract

No abstract