Sialic acid-binding immunoglobulin-like lectin-15 (Siglec-15) is a new immune checkpoint molecule and its role of primary central nervous system lymphoma (PCNSL) tumor microenvironment has been unclear. We explored the Siglec-15 and programed death-ligand 1 (PD-L1) expression in tumor tissues and analyzed the association between the expression of these molecules and overall survival in newly diagnosed PCNSL. A total of 60 patients diagnosed with diffuse large B-cell lymphoma in PCNSL were included in this study. The Siglec-15 and PD-L1 expression on tumor cells, intratumoral macrophages and peritumoral macrophages were immunohistochemically evaluated. The expression of Siglec-15 and PD-L1 was greater in macrophages than in tumor cells. Regarding peritumoral macrophages, the number of Siglec-15-positive samples (n = 24) was greater than the number of PD-L1-positive samples (n = 16). A multivariate Cox analysis showed that the Siglec-15 positivity of peritumoral macrophages and performance of high-dose methotrexate-based chemotherapy were independent predictors of overall survival (hazard ratio: 0.295 and 0.322, respectively). The Kaplan–Meier survival curves showed that patients with Siglec-15-positive peritumoral macrophages had longer overall survival than those with Siglec-15-negative peritumoral macrophages (median overall survival: 3018 days and 746 days, respectively; p = 0.0290). Our findings indicate that the expression of Siglec-15 on peritumoral macrophages induces a favorable outcome in PCNSL patients.

No abstract

In this paper, a cross-domain recommendation (CDR) method based on multi-layer graph analysis with visual information is presented. Although previous graph-based CDR methods have effectively used users' ratings of products and their purchase histories, they lack essential information such as product images that can have an impact on users' decision of whether to purchase products. Then the proposed method newly introduces visual features obtained from product images into the graph-based CDR. For dealing with visual features in multiple domains, we focus on both intra-domain and interdomain relationships through visual features. Specifically, to obtain effective embedding features from users and items in a domain, the proposed method newly introduces visual features into an optimization process of the latest graph neural network that considers only user-item interactions. Consideration of the visual similarity of items within a domain becomes feasible, and embedding features with high representation ability can be estimated. Furthermore, to consider visual information between domains, we construct multi-layer graphs for each domain and introduce visual features into the training of a feature transformer across these graphs. Therefore, consideration of both intra-domain and inter-domain relationships through visual features contributes to the performance improvement. To our best knowledge, this is the first trial to introduce visual features into multi-layer graph-based CDR. The effectiveness of our method is demonstrated by comparing several state-of-the-art methods.

No abstract

No abstract