No abstract

No abstract

Breast cancer is the most common malignancy and the leading cause of cancer-related deaths in women. Recent studies have investigated the prognostic value of the tumor microenvironment (TME)-related genes in breast cancer. The purpose of this research is to identify the immune-associated prognostic signature for breast cancer evaluate the probability of their prognostic value and compare the current staging system. In this study, we comprehensively evaluated the infiltration patterns of TME in 1,077 breast cancer patients downloaded from TCGA by applying the ssGSEA method to the transcriptome of these patients. Thus, generated two groups of immune cell infiltration. Based on two groups of low infiltration and high infiltration immune cell groups, 983 common differentially expressed genes were found using the limma algorithm. In addition, studying potential mechanisms, the GSEA method was used to indicate some pathways with remarkable enrichment in two clusters of immune cell infiltration. Finally, the seven immune-associated hub genes with survival as prognostic signatures were identified by using univariate Cox, survival, and LASSO analyses and constructed a TME score. The prognostic value of the TME score was self-validated in the TCGA cohort and further validated in an external independent set from METABRIC and GEO database by time-dependent survival receiver operation. Univariate and multivariate analyses of clinicopathological characteristics indicated that the TME score was an independent prognostic factor. In conclusion, the proposed TME score model should be considered as a prognostic factor, similar to the current TNM stage, and the seven immune-related genes can be a valuable potential biomarker for breast cancer.

Salient object detection is a hot spot of current computer vision. The emergence of the convolutional neural network (CNN) greatly improves the existing detection methods. In this paper, we present 3MNet, which is based on the CNN, to make the utmost of various features of the image and utilize the contour detection task of the salient object to explicitly model the features of multi-level structures, multiple tasks and multiple channels, so as to obtain the final saliency map of the fusion of these features. Specifically, we first utilize contour detection task for auxiliary detection and then utilize use multi-layer network structure to extract multi-scale image information. Finally, we introduce a unique module into the network to model the channel information of the image. Our network has produced good results on five widely used datasets. In addition, we also conducted a series of ablation experiments to verify the effectiveness of some components in the network.Xinghe Yan and Zhenxue Chen have contributed equally.