Human umbilical cord blood derived-mesenchymal stem cells (hUCMSCs) offer an attractive alternative to bone marrow-derived MSCs (BMMSCs) for cell-based therapy as it is a less invasive source of biological material. However, limited studies have been conducted with hUCMSCs as compared to BMMSCs. The present study was conducted to evaluate the effects of hUCMSCs in esophageal carcinoma (EC). hUCMSCs together with EC cells were transplanted subcutaneously into BALB/c nude mice to observe the effects of hUCMSCs on tumor establishment. hUCMSCs injected through the caudal vein to the mice with pre-established EC to observe the effects of hUCMSCs on tumor outgrowth. In order to elucidate the underlying mechanisms, we also performed in vitro experiments including directly co-culture, transwell assay, proliferation assay and western blotting analysis. hUCMSCs promoted EC formation in nude mice. In the in vivo model of pre-established EC, intravenously injected hUCMSCs potently promoted tumor growth. When in vitro co-cultured with hUCMSCs, EC cells proliferation increased. After co-cultured with hUCMSCs through transwell system, EC cells showed increased proliferation. Through transwell assay, we also observed that EC cells recruited MSCs, and MSCs promoted EC cells migration and invasion. Western blotting data showed that the expressions of proliferation related proteins Bcl-2, survivin and metastasis related proteins MMP-2 and MMP-9 were up-regulated in the EC cells transwell co-cultured with hUCMSCs. Our results indicated that hUCMSCs could favor tumor growth in vivo and in vitro. Thus, the exploitation of hUCMSCs in new therapeutic strategies should be cautious under the malignant conditions.
Brain tissue segmentation plays a critical role in the diagnosis, treatment, and study of brain diseases. Accurately identifying these boundaries is essential for improving segmentation accuracy. However, distinguishing boundaries between different brain tissues can be challenging, as they often overlap. Existing deep learning methods primarily calculate the overall segmentation results without adequately addressing local regions, leading to error propagation and mis-segmentation along boundaries. In this study, we propose a novel mis-segmentation-focused loss function based on a two-stage nnU-Net framework. Our approach aims to enhance the model’s ability to handle ambiguous boundaries and overlapping anatomical structures, thereby achieving more accurate brain tissue segmentation results. Specifically, the first stage targets the identification of mis-segmentation regions using a global loss function, while the second stage involves defining a mis-segmentation loss function to adaptively adjust the model, thus improving its capability to handle ambiguous boundaries and overlapping anatomical structures. Experimental evaluations on two datasets demonstrate that our proposed method outperforms existing approaches both quantitatively and qualitatively.