A lightweight neural network with multiscale feature enhancement for liver CT segmentation.

Segmentation of abdominal Computed Tomography (CT) scan is essential for analyzing, diagnosing, and treating visceral organ diseases (e.g., hepatocellular carcinoma). This paper proposes a novel neural network (Res-PAC-UNet) that employs a fixed-width residual UNet backbone and Pyramid Atrous Convolutions, providing a low disk utilization method for precise liver CT segmentation. The proposed network is trained on medical segmentation decathlon dataset using a modified surface loss function. Additionally, we evaluate its quantitative and qualitative performance; the Res16-PAC-UNet achieves a Dice coefficient of 0.950 ± 0.019 with less than half a million parameters. Alternatively, the Res32-PAC-UNet obtains a Dice coefficient of 0.958 ± 0.015 with an acceptable parameter count of approximately 1.2 million.

Mesenchymal stem cells enhance ovarian cancer cell infiltration through IL6 secretion in an amniochorionic membrane based 3D model.

BACKGROUND: The early peritoneal invasion of epithelial ovarian cancer (EOC) by tumoral aggregates presents in ascites is a major concern. The role of the microenvironment seems to be important in this process but the lack of adequate models to study cellular interactions between cancer cells and stromal cells does not allow to uncover the molecular pathways involved. Our goal was to study the interactions between ovarian cancer cells (OCC) and mesenchymal stem cells (MSC) using a 3D model. METHODS: We used millimetric pieces of amniochorionic membrane - referred to as amniotic membrane scaffold (AMS) - to create 3D peritoneal nodules mimicking EOC early invasion. We were able to measure the distribution and the depth of infiltration using confocal microsopy. We extracted MSC from the amniochorionic membrane using the markers CD34-, CD45-, CD73+, CD90+, CD105+ and CD29+ at the Fluorescence Activated Cell Sorting (FACS) analysis. We used transwell and wound healing tests to test OCC migration and invasion in vitro. RESULTS: Here we show that OCC tumors were located in regions rich in MSC (70%). The tumors infiltrated deeper within AMS in regions rich in MSC (p<0.001). In vitro tests revealed that higher IL6 secretion in a context of MSC-OCC co-culture could enhance migration and invasion of OCC. After IL6 receptor antagonism, OCC infiltration was significantly decreased, mostly in regions rich in MSCs, indicating that recruitment and tridimensional invasion of OCC was dependent of IL6 secretion. CONCLUSIONS: The use of tridimensional models using AMS could be a useful tool to decipher early molecular events in ovarian cancer metastasis. Cytokine inhibitors interrupting the cross-talk between OCCs and MSCs such as IL6 should be investigated as a new therapeutic approach in ovarian cancer.