No abstract

Recent advances in saliency detection have utilized deep learning to obtain high level features to detect salient regions in a scene. These advances have demonstrated superior results over previous works that utilize hand-crafted low level features for saliency detection. In this paper, we demonstrate that hand-crafted features can provide complementary information to enhance performance of saliency detection that utilizes only high level features. Our method utilizes both high level and low level features for saliency detection under a unified deep learning framework. The high level features are extracted using the VGG-net, and the low level features are compared with other parts of an image to form a low level distance map. The low level distance map is then encoded using a convolutional neural network(CNN) with multiple 1 × 1 convolutional and ReLU layers. We concatenate the encoded low level distance map and the high level features, and connect them to a fully connected neural network classifier to evaluate the saliency of a query region. Our experiments show that our method can further improve the performance of stateof-the-art deep learning-based saliency detection methods.

Acute myelogenous leukemia (AML) is a heterogeneous disorder characterized by clonal proliferation of stem cell-like blasts in bone marrow (BM); however, their unique cellular interaction within the BM microenvironment and its functional significance remain unclear. Here, we assessed the BM microenvironment of AML patients and demonstrate that the leukemia stem cells induce a change in the transcriptional programming of the normal mesenchymal stromal cells (MSC). The modified leukemic niche alters the expressions of cross-talk molecules (i.e., CXCL12 and JAG1) in MSCs to provide a distinct cross-talk between normal and leukemia cells, selectively suppressing normal primitive hematopoietic cells while supporting leukemogenesis and chemoresistance. Of note, AML patients exhibited distinct heterogeneity in the alteration of mesenchymal stroma in BM. The distinct pattern of stromal changes in leukemic BM at initial diagnosis was associated with a heterogeneous posttreatment clinical course with respect to the maintenance of complete remission for 5 to 8 years and early or late relapse. Thus, remodeling of mesenchymal niche by leukemia cells is an intrinsic self-reinforcing process of leukemogenesis that can be a parameter for the heterogeneity in the clinical course of leukemia and hence serve as a potential prognostic factor. Cancer Res; 75(11); 2222-31. Ó2015 AACR.

Hydrogen peroxide (H2O2) is one of essential oxygen metabolites in living organisms, but is generated in large amounts during inflammatory responses. Therefore, H2O2 has great potential as diagnostic and therapeutic markers of several inflammatory and life‐threatening diseases. Here, chemiluminescent and antioxidant micelles are reported as novel theranostic agents for H2O2‐associated inflammatory diseases. The chemiluminescent micelles composed of amphiphilic block copolymer Pluronic F‐127, hydroxybenzyl alcohol‐incorporated copolyoxalate (HPOX) and fluorescent dyes perform peroxalate chemiluminescence reactions to detect H2O2 as low as 100 nM and image H2O2 generated in inflamed mouse ankles. The micelles encapsulating HPOX reduce the generation of reactive oxygen species in lipopolysaccharide (LPS)‐activated macrophages by scavenging overproduced H2O2 and releasing antioxidant hydroxybenzyl alcohol (HBA). They also exert inhibitory effects on H2O2‐induced apoptosis. HPOX‐based chemiluminescent and antioxidant micelles have great potential as a theranostic agent for H2O2‐associated inflammatory diseases.