Deep learning-based automated lesion segmentation on mouse stroke magnetic resonance images.

Magnetic resonance imaging (MRI) is widely used for ischemic stroke lesion detection in mice. A challenge is that lesion segmentation often relies on manual tracing by trained experts, which is labor-intensive, time-consuming, and prone to inter- and intra-rater variability. Here, we present a fully automated ischemic stroke lesion segmentation method for mouse T2-weighted MRI data. As an end-to-end deep learning approach, the automated lesion segmentation requires very little preprocessing and works directly on the raw MRI scans. We randomly split a large dataset of 382 MRI scans into a subset (n = 293) to train the automated lesion segmentation and a subset (n = 89) to evaluate its performance. We compared Dice coefficients and accuracy of lesion volume against manual segmentation, as well as its performance on an independent dataset from an open repository with different imaging characteristics. The automated lesion segmentation produced segmentation masks with a smooth, compact, and realistic appearance that are in high agreement with manual segmentation. We report dice scores higher than the agreement between two human raters reported in previous studies, highlighting the ability to remove individual human bias and standardize the process across research studies and centers.

Components of high-resolution manometry that change surgical decisions.

BACKGROUND: High-resolution manometry (HRM) is vital in evaluating patients for surgery at the gastroesophageal (GE) junction. Previously, we reported manometry alters surgery choices at the GE junction over 50% of the time, and its components, i.e., abnormal motility and distal contractile integral (DCI), are vital in decision-making. This single-institution retrospective study examines how HRM characteristics, reported with the Chicago classification, can alter the intended surgical plans for foregut surgery. METHODS: We collected data on pre-operative symptoms for patients undergoing HRM studies from 2012 to 2016, i.e., Upper GI X-rays, 48-h pH studies, DeMeester scores, upper endoscopy, and biopsy reports. HRM results were further categorized via Chicago classification (i.e., normal or abnormal motility). The DCI was determined; Patients not seen by a surgeon were excluded. Then a single surgeon, blinded to patient identity and HRM results, determined the planned procedure. The reviewer was then exposed to the HRM results; procedural plans were revised if needed. HRM results were then evaluated to determine which factors most influenced the surgical decisions. RESULTS: 298 HRM studies were initially identified; 114 met search criteria. Overall, HRM altered the planned procedure in 50.9% of cases (n = 58), with abnormal motility in 54.4% (62/114) cases. Abnormal motility findings corresponded to 70.6% (41/58) of the patients in which HRM changed the surgery decision. A DCI of < 1000 was identified in only 31.6% (36/114) of all patients, but 39.7% (23/58) of cases where the surgical decision was altered. A DCI of > 5000 was identified in only 10.5% (12/114) of all patients but 10.3% (6/58) of cases with altered surgical decisions. A DCI < 1000 and abnormal motility were generally associated with a partial fundoplication. CONCLUSIONS: This study demonstrates the impact of identifying abnormal motility via the Chicago classification and factors like DCI on surgical choice at the GE junction.

webKnossos: efficient online 3D data annotation for connectomics.

We report webKnossos, an in-browser annotation tool for 3D electron microscopic data. webKnossos provides flight mode, a single-view egocentric reconstruction method enabling trained annotator crowds to reconstruct at a speed of 1.5 ± 0.6 mm/h for axons and 2.1 ± 0.9 mm/h for dendrites in 3D electron microscopic data from mammalian cortex. webKnossos accelerates neurite reconstruction for connectomics by 4- to 13-fold compared with current state-of-the-art tools, thus extending the range of connectomes that can realistically be mapped in the future.

Alpha 1,3 fucosyltransferases are master regulators of prostate cancer cell trafficking.

How cancer cells bind to vascular surfaces and extravasate into target organs is an underappreciated, yet essential step in metastasis. We postulate that the metastatic process involves discrete adhesive interactions between circulating cancer cells and microvascular endothelial cells. Sialyl Lewis X (sLe(X)) on prostate cancer (PCa) cells is thought to promote metastasis by mediating PCa cell binding to microvascular endothelial (E)-selectin. Yet, regulation of sLe(X) and related E-selectin ligand expression in PCa cells is a poorly understood factor in PCa metastasis. Here, we describe a glycobiological mechanism regulating E-selectin-mediated adhesion and metastatic potential of PCa cells. We demonstrate that alpha1,3 fucosyltransferases (FT) 3, 6, and 7 are markedly elevated in bone- and liver-metastatic PCa and dictate synthesis of sLe(X) and E-selectin ligands on metastatic PCa cells. Upregulated FT3, FT6, or FT7 expression induced robust PCa PC-3 cell adhesion to bone marrow (BM) endothelium and to inflamed postcapillary venules in an E-selectin-dependent manner. Membrane proteins, CD44, carcinoembryonic antigen (CEA), podocalyxin-like protein (PCLP), and melanoma cell adhesion molecule (MCAM) were major scaffolds presenting E-selectin-binding determinants on FT-upregulated PC-3 cells. Furthermore, elevated FT7 expression promoted PC-3 cell trafficking to and retention in BM through an E-selectin dependent event. These results indicate that alpha1,3 FTs could enhance metastatic efficiency of PCa by triggering an E-selectin-dependent trafficking mechanism.

Analysis of physiologic E-selectin-mediated leukocyte rolling on microvascular endothelium.

E-selectin is a type-1 membrane protein on microvascular endothelial cells that helps initiate recruitment of circulating leukocytes to cutaneous, bone and inflamed tissues. E-selectin expression is constitutive on dermal and bone microvessels and is inducible by pro-inflammatory cytokines, such as IL-1alpha/ and TNF-alpha, on microvessels in inflamed tissues. This lectin receptor mediates weak binding interactions with carbohydrate counter-receptor ligands on circulating leukocytes, which results in a characteristic rolling behavior. Because these interactions precede more stable adhesive events and diapedesis activity, characterization of leukocyte rolling activity and identification of leukocyte E-selectin ligands have been major goals in studies of leukocyte trafficking and inflammation and in the development of anti-inflammatory therapeutics (1-5). The intent of this report is to provide a visual, comprehensive description of the most widely-used technology for studying E-selectin E-selectin ligand interactions under physiologic blood flow conditions. Our laboratory in conjunction with the Harvard Skin Disease Research Center uses a state-of-the-art parallel-plate flow chamber apparatus accompanied by digital visualization and new recording software, NIS-Elements. This technology allows us to analyze adhesion events in real time for onscreen visualization as well as record rolling activity in a video format. Cell adhesion parameters, such as rolling frequency, shear resistance and binding/tethering efficiency, are calculated with NIS-Elements software, exported to an Excel spreadsheet and subjected to statistical analysis. In the demonstration presented here, we employed the parallel-plate flow chamber to investigate E-selectin-dependent leukocyte rolling activity on live human bone marrow endothelial cells (hBMEC). Human hematopoietic progenitor KG1a cells, which express a high level of E-selectin ligand, were used as our leukocyte model, while an immortalized hBMEC cell line, HBMEC-60 cells, was used as our endothelial cell model (6). To induce and simulate native E-selectin expression in the flow chamber, HBMEC-60 cells were first activated with IL-1 . Our video presentation showed that parallel-plate flow analysis is a suitable method for studying physiologic E-selectin-mediated leukocyte rolling activities and that functional characterization of leukocyte E-selectin ligand(s) in the flow chamber can be ascertained by implementing protease or glycosidase digestions.

Analysis of glycosyltransferase expression in metastatic prostate cancer cells capable of rolling activity on microvascular endothelial (E)-selectin.

Prostate cancer (PCa) cell tethering and rolling on microvascular endothelium has been proposed to promote the extravasation of PCa cells. We have shown that these adhesive events are mediated through binding interactions between endothelial (E)-selectin and Lewis carbohydrates on PCa cells. Prior data indicate that E-selectin-mediated rolling of bone-metastatic PCa MDA PCa 2b (MDA) cells is dependent on sialyl Lewis X (sLe(X))-bearing glycoproteins. To explore the molecular basis of sLe(X) synthesis and E-selectin ligand (ESL) activity on PCa cells, we compared and contrasted the expression level of glycosyltransferases, characteristically involved in sLe(X) and ESL synthesis, in ESL(+) MDA cells among other ESL(-) metastatic PCa cell lines. We also created and examined ESL(hi) and ESL(lo) variants of MDA cells to provide a direct comparison of the glycosyltransferase expression level. We found that normal prostate tissue and all metastatic PCa cell lines expressed glycosyltransferases required for sialo-lactosamine synthesis, including N-acetylglucosaminyl-, galactosyl-, and sialyltransferases. However, compared with expression in normal prostate tissue, ESL(+) MDA cells expressed a 31- and 10-fold higher level of alpha1,3 fucosyltransferases (FT) 3 and 6, respectively. Moreover, FT3 and FT6 were expressed at 2- to 354-fold lower levels in ESL(-) PCa cell lines. Consistent with these findings, ESL(hi) MDA cells expressed a 131- and 51-fold higher level of FT3 and FT6, respectively, compared with expression in ESL(lo) MDA cells. We also noted that alpha1,3 FT7 was expressed at a 5-fold greater level in ESL(hi) MDA cells. Furthermore, ESL(lo) MDA cells did not display sLe(X) on glycoproteins capable of bearing sLe(X), notably P-selectin glycoprotein ligand-1. These results implicate the importance of alpha1,3 FT3, FT6, and/or FT7 in sLe(X) and ESL synthesis on metastatic PCa cells.