Automatic X-ray landmark detection and shape segmentation via data-driven joint estimation of image displacements

In this paper, we propose a new method for fully-automatic landmark detection and shape segmentation in X-ray images. To detect landmarks, we estimate the displacements from some randomly sampled image patches to the (unknown) landmark positions, and then we integrate these predictions via a voting scheme. Our key contribution is a new algorithm for estimating these displacements. Different from other methods where each image patch independently predicts its displacement, we jointly estimate the displacements from all patches together in a data driven way, by considering not only the training data but also geometric constraints on the test image. The displacements estimation is formulated as a convex optimization problem that can be solved efficiently. Finally, we use the sparse shape composition model as the a priori information to regularize the landmark positions and thus generate the segmented shape contour. We validate our method on X-ray image datasets of three different anatomical structures: complete femur, proximal femur and pelvis. Experiments show that our method is accurate and robust in landmark detection, and, combined with the shape model, gives a better or comparable performance in shape segmentation compared to state-of-the art methods. Finally, a preliminary study using CT data shows the extensibility of our method to 3D data.

Localization and Segmentation of 3D Intervertebral Discs in MR Images by Data Driven Estimation

This paper addresses the problem of fully-automatic localization and segmentation of 3D intervertebral discs (IVDs) from MR images. Our method contains two steps, where we first localize the center of each IVD, and then segment IVDs by classifying image pixels around each disc center as foreground (disc) or background. The disc localization is done by estimating the image displacements from a set of randomly sampled 3D image patches to the disc center. The image displacements are estimated by jointly optimizing the training and test displacement values in a data-driven way, where we take into consideration both the training data and the geometric constraint on the test image. After the disc centers are localized, we segment the discs by classifying image pixels around disc centers as background or foreground. The classification is done in a similar data-driven approach as we used for localization, but in this segmentation case we are aiming to estimate the foreground/background probability of each pixel instead of the image displacements. In addition, an extra neighborhood smooth constraint is introduced to enforce the local smoothness of the label field. Our method is validated on 3D T2-weighted turbo spin echo MR images of 35 patients from two different studies. Experiments show that compared to state of the art, our method achieves better or comparable results. Specifically, we achieve for localization a mean error of 1.6-2.0 mm, and for segmentation a mean Dice metric of 85%-88% and a mean surface distance of 1.3-1.4 mm.

A complete-pelvis segmentation framework for image-free total hip arthroplasty (THA): methodology and clinical study

Background Complete-pelvis segmentation in antero-posterior pelvic radiographs is required to create a patient-specific three-dimensional pelvis model for surgical planning and postoperative assessment in image-free navigation of total hip arthroplasty. Methods A fast and robust framework for accurately segmenting the complete pelvis is presented, consisting of two consecutive modules. In the first module, a three-stage method was developed to delineate the left hemipelvis based on statistical appearance and shape models. To handle complex pelvic structures, anatomy-specific information processing techniques were employed. As the input to the second module, the delineated left hemi-pelvis was then reflected about an estimated symmetry line of the radiograph to initialize the right hemi-pelvis segmentation. The right hemi-pelvis was segmented by the same three-stage method, Results Two experiments conducted on respectively 143 and 40 AP radiographs demonstrated a mean segmentation accuracy of 1.61±0.68 mm. A clinical study to investigate the postoperative assessment of acetabular cup orientations based on the proposed framework revealed an average accuracy of 1.2°±0.9° and 1.6°±1.4° for anteversion and inclination, respectively. Delineation of each radiograph costs less than one minute. Conclusions Despite further validation needed, the preliminary results implied the underlying clinical applicability of the proposed framework for image-free THA.

Inhibition of HIV-1 multiplication by antisense U7 snRNAs and siRNAs targeting cyclophilin A.

Human immunodeficiency virus 1 (HIV-1) multiplication depends on a cellular protein, cyclophilin A (CyPA), that gets integrated into viral particles. Because CyPA is not required for cell viability, we attempted to block its synthesis in order to inhibit HIV-1 replication. For this purpose, we used antisense U7 small nuclear RNAs (snRNAs) that disturb CyPA pre-mRNA splicing and short interfering RNAs (siRNAs) that target CyPA mRNA for degradation. With dual-specificity U7 snRNAs targeting the 3' and 5' splice sites of CyPA exons 3 or 4, we obtained an efficient skipping of these exons and a strong reduction of CyPA protein. Furthermore, short interfering RNAs targeting two segments of the CyPA coding region strongly reduced CyPA mRNA and protein levels. Upon lentiviral vector-mediated transduction, prolonged antisense effects were obtained for both types of antisense RNAs in the human T-cell line CEM-SS. These transduced CEM-SS cells showed a delayed, and for the siRNAs also reduced, HIV-1 multiplication. Since the two types of antisense RNAs function by different mechanisms, combining the two approaches may result in a synergistic effect.

Triple-peptide receptor targeting in vitro allows detection of all tested gut and bronchial NETs.

UNLABELLED A high proportion of gut and bronchial neuroendocrine tumors (NETs) overexpresses somatostatin receptors, especially the sst2 subtype. It has also recently been observed that incretin receptors, namely glucagonlike peptide 1 (GLP-1) and glucose-dependent insulinotropic peptide (GIP) receptors, can be overexpressed in gut and bronchial NETs. However, because not all tumors can express these receptors in sufficient amounts, in vivo imaging with a single radioligand may not always be successful. We therefore evaluated with in vitro methods whether a cocktail of radioligands targeting these 3 receptors would improve tumor labeling. METHODS In vitro receptor autoradiography was performed on 55 NETs, comparing in each successive section of tumor the binding with a single radioligand, either (125)I-Tyr(3)-octreotide, (125)I-GLP-1(7-36)amide, or (125)I-GIP(1-30), with the binding using a cocktail of all 3 radioligands, given concomitantly under identical experimental conditions. RESULTS Using the cocktail of radioligands, all tumors without exception showed moderate to very high binding, with a receptor density corresponding to 1,000-10,000 dpm/mg of tissue; conversely, single-ligand binding, although identifying most tumors as receptor-positive, failed to detect receptors or measured only a low density of receptors below 1,000 dpm/mg in a significant number of tumors. In addition, the cocktail of radioligands always provided a homogeneous labeling of the whole tumor, whereas single radioligands occasionally showed heterogeneous labeling. CONCLUSION The study suggests that the use of a cocktail of 3 radioligands binding to somatostatin receptors, GLP-1 receptors, and GIP receptors would allow detecting virtually all NETs and labeling them homogeneously in vivo, representing a significant improvement for imaging and therapy in NETs.

Segmentation and Segregation Patterns of Women-Owned High-Tech Firms in Four Metropolitan Regions in the United States

Mayer H. Segmentation and segregation patterns of women-owned high-tech firms in four metropolitan regions in the United States, Regional Studies. The number of women starting and owning a business has increased dramatically and female entrepreneurs are entering non-traditional sectors such as high technology, construction and manufacturing. This paper investigates the trends in high-tech entrepreneurship by women in four US metropolitan regions (Silicon Valley, California; Boston, Massachusetts; Washington, DC; and Portland, Oregon). The research examines the sectoral and spatial segmentation patterns of women-owned high-tech firms. Although women are entering non-traditional sectors, the research finds that women entrepreneurs tend to own businesses in female-typed high-tech sectors. In established high-tech regions like Silicon Valley and Boston, male-typed and female-typed women-owned high-tech firms differ significantly in terms of sectoral and spatial segmentation regardless of firm age. While differences between male-typed and female-typed firms are not significant at the regional level for Washington, DC, the analysis shows significant intra-metropolitan differences for the female-typed high-tech firms. The paper concludes that sectoral and spatial segmentation are powerful dynamics that shape business ownership by women in high technology.

Multivalency effects in Pseudomonas aeruginosa biofilm inhibition and dispersal by glycopeptide dendrimers targeting lectin LecA

The galactose specific lectin LecA partly mediates the formation of antibiotic resistant biofilms by Pseudomonas aeruginosa, an opportunistic pathogen causing lethal airways infections in immunocompromised and cystic fibrosis patients, suggesting that preventing LecA binding to natural saccharides might provide new opportunities for treatment. Here 8-fold (G3) and 16-fold (G4) galactosylated analogs of GalAG2, a tetravalent G2 glycopeptide dendrimer LecA ligand and P. aeruginosa biofilm inhibitor, were obtained by convergent chloroacetyl thioether (ClAc) ligation between 4-fold or 8-fold chloroacetylated dendrimer cores and digalactosylated dendritic arms. Hemagglutination inhibition, isothermal titration calorimetry and biofilm inhibition assays showed that G3 dendrimers bind LecA slightly better than their parent G2 dendrimers and induce complete biofilm inhibition and dispersal of P. aeruginosa biofilms, while G4 dendrimers show reduced binding and no biofilm inhibition. A binding model accounting for the observed saturation of glycopeptide dendrimer galactosyl groups and LecA binding sites is proposed based on the crystal structure of a G3 dendrimer LecA complex.

Fully automatic GBM segmentation in the TCGA-GBM dataset: Prognosis and correlation with VASARI features.

Reproducible definition and quantification of imaging biomarkers is essential. We evaluated a fully automatic MR-based segmentation method by comparing it to manually defined sub-volumes by experienced radiologists in the TCGA-GBM dataset, in terms of sub-volume prognosis and association with VASARI features. MRI sets of 109 GBM patients were downloaded from the Cancer Imaging archive. GBM sub-compartments were defined manually and automatically using the Brain Tumor Image Analysis (BraTumIA). Spearman's correlation was used to evaluate the agreement with VASARI features. Prognostic significance was assessed using the C-index. Auto-segmented sub-volumes showed moderate to high agreement with manually delineated volumes (range (r): 0.4 - 0.86). Also, the auto and manual volumes showed similar correlation with VASARI features (auto r = 0.35, 0.43 and 0.36; manual r = 0.17, 0.67, 0.41, for contrast-enhancing, necrosis and edema, respectively). The auto-segmented contrast-enhancing volume and post-contrast abnormal volume showed the highest AUC (0.66, CI: 0.55-0.77 and 0.65, CI: 0.54-0.76), comparable to manually defined volumes (0.64, CI: 0.53-0.75 and 0.63, CI: 0.52-0.74, respectively). BraTumIA and manual tumor sub-compartments showed comparable performance in terms of prognosis and correlation with VASARI features. This method can enable more reproducible definition and quantification of imaging based biomarkers and has potential in high-throughput medical imaging research.

KRAS and HRAS mutations confer resistance to MET targeting in preclinical models of MET-expressing tumor cells.

The MET receptor tyrosine kinase is often deregulated in human cancers and several MET inhibitors are evaluated in clinical trials. Similarly to EGFR, MET signals through the RAS-RAF-ERK/MAPK pathway which plays key roles in cell proliferation and survival. Mutations of genes encoding for RAS proteins, particularly in KRAS, are commonly found in various tumors and are associated with constitutive activation of the MAPK pathway. It was shown for EGFR, that KRAS mutations render upstream EGFR inhibition ineffective in EGFR-positive colorectal cancers. Currently, there are no clinical studies evaluating MET inhibition impairment due to RAS mutations. To test the impact of RAS mutations on MET targeting, we generated tumor cells responsive to the MET inhibitor EMD1214063 that express KRAS G12V, G12D, G13D and HRAS G12V variants. We demonstrate that these MAPK-activating RAS mutations differentially interfere with MET-mediated biological effects of MET inhibition. We report increased residual ERK1/2 phosphorylation indicating that the downstream pathway remains active in presence of MET inhibition. Consequently, RAS variants counteracted MET inhibition-induced morphological changes as well as anti-proliferative and anchorage-independent growth effects. The effect of RAS mutants was reversed when MET inhibition was combined with MEK inhibitors AZD6244 and UO126. In an in vivo mouse xenograft model, MET-driven tumors harboring mutated RAS displayed resistance to MET inhibition. Taken together, our results demonstrate for the first time in details the role of KRAS and HRAS mutations in resistance to MET inhibition and suggest targeting both MET and MEK as an effective strategy when both oncogenic drivers are expressed.

Impact of MET targeting on tumor-associated angiogenesis and growth of MET mutations-driven models of liver cancer

Deregulated expression of the MET receptor tyrosine kinase has been reported in up to 50% of patients with hepatocellular carcinoma, the most abundant form of liver cancers, and is associated with decreased survival. Consequently, MET is considered as a molecular target in this malignancy, whose progression is highly dependent on extensive angiogenesis. Here we studied the impact of MET small molecule inhibitors on angiogenesis-associated parameters and growth of xenograft liver models consisting of cells expressing MET-mutated variants M1268T and Y1248H, which exhibit constitutive kinase activity. We demonstrate that MET mutations expression is associated with significantly increased production of vascular endothelial growth factor, which is blocked by MET targeting only in cells expressing the M1268T inhibitor-sensitive but not in the Y1248H inhibitor-resistant variant. Decrease in vascular endothelial growth factor production is also associated with reduction of tyrosine phopshorylation of the vascular endothelial growth factor receptor 2 expressed on primary liver sinusoidal endothelial cells and with inhibition of vessel formation. Furthermore, MET inhibition demonstrated an efficient anti-tumor activity and considerable reduction in microvessel density only against the M1268T-derived intrahepatic tumors. Collectively, our data support the role of targeting MET-associated angiogenesis as a major biological determinant for liver tumor growth control.

Dish Detection and Segmentation for Dietary Assessment on Smartphones

Diet-related chronic diseases severely affect personal and global health. However, managing or treating these diseases currently requires long training and high personal involvement to succeed. Computer vision systems could assist with the assessment of diet by detecting and recognizing different foods and their portions in images. We propose novel methods for detecting a dish in an image and segmenting its contents with and without user interaction. All methods were evaluated on a database of over 1600 manually annotated images. The dish detection scored an average of 99% accuracy with a .2s/image run time, while the automatic and semi-automatic dish segmentation methods reached average accuracies of 88% and 91% respectively, with an average run time of .5s/image, outperforming competing solutions.

Development of Drugs Targeting the PI3K Signalling Pathway in Leukaemias and Lymphomas

The phosphoinositide 3-kinase (PI3K) family of signalling enzymes play a key role in the transduction of signals from activated cell surface receptors controlling cell growth and proliferation, survival, metabolism, and migration. The intracellular signalling pathway from activated receptors to PI3K and its downstream targets v-akt murine thymoma viral oncogene homolog (Akt) and mechanistic target of rapamycin (mTOR) is very frequently deregulated by genetic and epigenetic mechanisms in human cancer, including leukaemia and lymphoma. In the past decade, an arsenal of small molecule inhibitors of key enzymes in this pathway has been developed and evaluated in pre-clinical studies and clinical trials in cancer patients. These include pharmacological inhibitors of Akt, mTOR, and PI3K, some of which are approved for the treatment of leukaemia and lymphoma. The PI3K family comprises eight different catalytic isoforms in humans, which have been subdivided into three classes. Class I PI3K isoforms have been extensively studied in the context of human cancer, and the isoforms p110α and p110δ are validated drug targets. The recent approval of a p110δ-specific PI3K inhibitor (idelalisib/Zydelig®) for the treatment of selected B cell malignancies represents the first success in developing these molecules into anti-cancer drugs. In addition to PI3K inhibitors, mTOR inhibitors are intensively studied in leukaemia and lymphoma, and temsirolimus (Torisel®) is approved for the treatment of a type of lymphoma. Based on these promising results it is hoped that additional novel PI3K pathway inhibitors will in the near future be further developed into new drugs for leukaemia and lymphoma.

Sural nerve conduction studies using ultrasound-guided needle positioning: Influence of age and recording location.

INTRODUCTION The aim of this study was to compare orthodromic sural nerve conduction study (NCS) results using ultrasound-guided needle positioning (USNP) to surface electrode recordings. METHODS 51 healthy subjects aged 24 - 80 years, divided into 5 age groups, were examined. Electrical stimuli were applied behind the lateral malleolus. Sensory nerve action potentials (SNAPs) were recorded 8 and 15 cm proximally with surface and needle electrodes. RESULTS Mean SNAP amplitudes in µV (surface/needle electrodes) averaged 12.7 (SD 7.6)/40.6 (SD 20.8), P<0.001, for subjects aged 20-29 years, and 5.0 (SD 2.4)/19.8 (SD 9.8), P<0.01, for subjects aged > 60 years. SNAP amplitudes were smaller at the proximal recording location. DISCUSSION NCS using USNP yield higher amplitude responses than surface electrodes in all age groups at all recording sites. SNAP amplitudes are smaller at proximal recording locations due to sural nerve branching. This article is protected by copyright. All rights reserved.