Search

Chronic obstructive pulmonary disease (COPD) is the leading cause of respiratory mortality worldwide. Genetic risk loci provide novel insights into disease pathogenesis. We performed a genome-wide association study in 35,735 cases and 222,076 controls from the UK Biobank and additional studies from the International COPD Genetics Consortium. We identified 82 loci with P-value < 5 × 10−8; 47 were previously described in association with either COPD or population-based lung function. Of the remaining 35 novel loci, 13 were associated with lung function in 79,055 individuals from the SpiroMeta consortium. Using gene expression and regulation data, we identified enrichment for loci in lung tissue, smooth muscle and several lung cell types. We found 14 COPD loci shared with either asthma or pulmonary fibrosis. COPD genetic risk loci clustered into groups of quantitative imaging features and comorbidity associations. Our analyses provide further support to the genetic susceptibility and heterogeneity of COPD., Editorial summary Genome-wide analysis of chronic obstructive pulmonary disease identifies 82 loci, 35 of which are new. Integration of gene expression and genomic annotation data shows enrichment of signals in lung tissue, smooth muscle and several lung cell types.

What follows are the abstracts presented at the Joint Meeting of the International Confocal Group (ICG), the International Dermoscopy Society (IDS), and the International Society for Digital Imaging of the Skin (ISDIS). The meeting was held on March 5, 2016, in Washington, DC, USA, in conjunction with the annual meeting of the American Academy of Dermatology (Figure 1). The abstracts appear in the order in which they were presented., Spectrally encoded confocal microscopy (SECM) is a high-speed reflectance confocal microscopy technology. SECM utilizes a stationary optical element, a diffraction grating, and a broadband light source to illuminate a line on the tissue with multiple spectrally-encoded spots, which enables scan-less line confocal imaging. In order to conduct two-dimensional confocal imaging, however, SECM still needs to use a beam scanning device or needs to mechanically translate the device. When a slit aperture is used, SECM can image a rectangular area of the tissue with multiple spectrally-encoded lines, which enables two-dimensional confocal imaging without using any beam scanning devices. Resulting confocal images are directly projected on a two-dimensional imaging sensor. This new approach, slit-SECM, can uniquely enable development of a smartphone confocal microscope, in which an optics module is attached to a smartphone to conduct confocal imaging and the smartphone imaging sensor is used to acquire two-dimensional confocal images. Due to its low cost, portability, and inherent network connectivity, smart-phone confocal microscopy has a potential to provide an in vivo diagnostic tool in resource-poor countries and also to increase clinical adaptation of confocal imaging in developed countries. We present preliminary results of imaging human skin in vivo with slit-SECM. We developed a slit-SECM bench system with an inexpensive, battery-powered LED ($25) and a low-cost color CMOS sensor ($355). The bench system achieved lateral resolution of 1.3 µm and axial resolution of 6 µm. Confocal images of human skin were acquired at the speed of 10 frames/sec. Acquired confocal images clearly visualized characteristic cellular features of human skin down to the dermal-epidermal junction, including cell nuclei in spinous layer and dermal papilla. Results from this preliminary study show feasibility of conducting skin confocal imaging using inexpensive optical and electrical components and suggest that slit-SECM may be developed into a low-cost smartphone confocal microscope., The differentiation of melanocytic hyperplasia in severe sun-damaged skin (MH) from early stages of Lentigo Maligna (ELM) may be very difficult. Some histopathologic criteria have been proposed helping to classify lesions as malignant or not. But, more distinctive changes of melanoma may not occur for years (pagetoid cells, adnexal spreading, formation of nests of melanocytes and descent of neoplastic cells into the dermis). The exact threshold for malignancy is still controversial and so the management of these lesions remain discussed. We retrospectively evaluated consecutive cases of pigmented lesions with pathology report as MH or ELM from the confocal databases of the main melanoma centre of Sydney, Australia from 2005 and 2015. Only cases with > 1 year follow up and recorded dermoscopic and RCM images were reviewed. We reviewed 11 histopathologic features that were correlated to 13 RCM features and sensitivity and specificity for each RCM criterion were studied. We also analysed the dermoscopic features of each lesion using previously described criteria for LM. 51 cases were retrieved in this period. 40 cases had > 1 year follow up but in 3 cases the dermoscopic images could not be retrieved. There was a significant female predominance (25 cases) and mainly located on upper face (front, peri-oricular and scalp) and lower face (cheeks, peri-oral and chin). The most frequent dermoscopic feature was asymmetric hyperpigmented follicular opening and on RCM most cases showed, Reflectance confocal microscopy (RCM) has advanced from merely a research tool for diagnosis of skin cancer to being used in the clinic to guide patient care. Here we present three technology solutions that will further impact clinical advances: An RCM imaging guided-laser ablation approach has been developed for treatment of basal cell carcinomas (BCCs). Initial testing in 48 lesions using an Er:YAG or an CO2 lasers shows feasibility on patients. Twenty of the lesions, assessed with post-ablation histopathology, show 85% agreement with RCM imaging for detection of residual or clearance of BCCs. Tumor clearance of the remaining 28 lesions has been confirmed with imaging and these are currently being followed-up with additional imaging.An image analysis based automated tool for quantitative analysis of RCM mosaics of melanocytic lesions collected is being developed. The method can distinguish patterns, at dermal-epidermal junction level, of benign (ring, meshwork, clod) from non-specific (potentially malignant) and background non-lesional skin using their textural appearance. Preliminary analysis on 20 RCM mosaics shows classification with 80-67% sensitivity and 99-78% specificity in distinguishing these patterns.An innovative solution has been developed, that integrates a miniature color camera into the objective lens of our microscope, providing simultaneous widefield dermoscopy images of the skin surface and RCM images of the subsurface cellular structure. Initial in vivo testing on 15 volunteers shows feasibility of the approach., The refining of clinico-pathologic correlation using dermoscopy by Scope et al in the area of dermatology has led to “a plea for a combined diagnostic approach of histopathologic and dermoscopic evaluation of melanocytic lesions”. Integration between dermoscopy and histology was only effective in a limited number of scientific works despite its promising premise. We propose that a sub-millimetre skin biopsy device known as the Microbiopsy for minimally invasive and suture-free skin sampling for molecular diagnosis be used for this purpose. To this end we are integrating microbiopsy based sampling into dermoscopy to enable dermoscopy guided microbiopsy for molecular diagnosis. We discuss our recent advances in microbiopsy technology and the future of dermoscopy-guided microbiopsy., Although traditionally used for evaluation of skin tumors, dermoscopy continuously gains appreciation in other fields of dermatology. The dermoscopic patterns of several inflammatory and infectious skin diseases have already been described, and the technique has been shown to improve clinical performance in terms of differential diagnosis in the daily practice. The increasing use of dermoscopy in general dermatology was significantly enhanced by the development of the new generation hand-held dermatoscopes, which can be easily placed in every dermatologist’s pocket and do not require use of immersion fluid. Four main categories of dermoscopic criteria should be considered when applying the technique in inflammatory diseases: 1) vascular features, including purpuric structures (morphology distribution); 2) color variegations; 3) follicular abnormalities and 4) specific features. Nowadays, the dermatoscope should not be regarded a second-level diagnostic equipment, but an irreplaceable diagnostic tool in every-day clinical setting, similarly to the stethoscope in general medicine. We provide an up-to-date summary of data on dermoscopy in general dermatology, attempting to assist clinicians to profitably utilize and apply the available knowledge in the everyday practice., Introduction & Objectives: High Frequency Ultrasound (HFUS) is a non-invasive technique that allows visualizing skin tumors in vivo to obtain size, shape and tumor volume. In this study, we sought to measure the correlation between ex vivo HFUS and histopathology surgical margins of basal cell carcinoma. Material & Methods: This was a prospective, single-blinded study. All patients had been sent for tumor excision. HFUS was performed with a 22 MHz Ultrasound (Taberna Pro-Medicum, Lüneburg, Germany®) before and immediately after excision to determine length and depth. Tumor measurements and surgical margins (SM) were evaluated. SM were established by observing echogenicity differences and tumor shape. The pathologist was blinded to HFUS measurements and SM. Results: A total of 79 basal cell carcinoma (BCC) (56 nodular, 15 superficial and 8 with more than one histological subtype) were included. Out of 79 BCC, 76 (96.2%) had a correspondence between ex vivo HFUS and histology (72 had negative and 4 had positive SM by ex vivo HFUS and histology). Of the remaining 3 tumors, one had uncertain and 2 had negative SM by ex vivo HFUS while positive SM by histology. In this study, ex vivo HFUS allowed correct visualization of negative SM in 72/79 BCC (91.1%). Conclusions: other non invasive techniques such as dermoscopy and confocal microscopy allow an accurate diagnosis in most BCC cases. However, the information related to depth and tumor volume is limited. HFUS devices are portable and give extra valuable information for BCC surgical management. Future studies are needed to compare ex vivo HFUS with other non-invasive technique such optical coherence tomography and confocal microscopy., Overview: The International Skin Imaging Collaboration is an academia/industry partnership designed to facilitate the application of digital skin imaging to help reduce melanoma mortality. The mobile era presents an exciting opportunity for the application of digital photography as an aid to early diagnosis of melanoma. ISIC is designed to address two barriers to the broad successful implementation of mobile solutions for melanoma diagnosis: A lack of standards for dermatologic imaging, and limited access of educators and developers to large numbers of high quality clinically annotated skin lesion images. Standards: There are currently no DICOM standards for dermatologic imaging. Through the efforts of 5 working groups comprised of international melanoma thought leaders from academia and industry we are using the Delphi method to propose standards related to 5 areas of skin cancer imaging: technology, terminology, techniques, privacy, and metadata. The working groups have prepared several publications to date that are in press and/or under review. It is hoped that the efforts of the working groups will serve as the basis for convening a successful DICOM process for dermatologic imaging with an initial focus on melanoma. The ISIC Archive: Current efforts in melanoma education and automated diagnosis typically rely on convenience samples of small numbers of images that vary in quality and annotation. The ISIC archive is a partnership of leading international melanoma centers to provide a large public repository of clinically annotated high quality skin images. The ISIC Archive software is open source and the images in the archive are publicly available through a Creative Commons zero License (CC0). The individual lesion images in the archive are annotated with clinical (e.g., pathology diagnosis) and technical (e.g., EXIF header content) attributes. In addition, a process has been developed to append morphologic annotations to the images at the lesion level (e.g., symmetry) and at the sub-lesion level (e.g., the presence of pigment network in a specific region within a lesion). The initial focus of the archive and annotations has been on dermoscopic images, as these are most diagnostic for melanoma. The archive has been selected for an ongoing 2016–2017 image analysis challenge (for automated lesion segmentation and classification) by the International Symposium of Biomedical Imaging., Mobile technologies are revolutionizing medicine. Smart phones are being leveraged by the ‘quantified self’ movement and major health systems to empower individuals to take a bigger and more direct role in their own health and medical care. Skin cancer is no exception. There is an ongoing proliferation of apps to educate and assist the public in skin cancer recognition. In October 2015, Oregon Health & Science University released an iPhone app “Mole Mapper” designed to advance melanoma research by giving users the ability to accurately measure and monitor moles, and contribute photos of the evolution of their moles over time to Apple’s open source ‘ResearchKit’. Sage Bionetworks, a nonprofit research institute is a key partner in managing and analyzing the melanoma data and images for this project. In this presentation, Drs. Leachman and Webster provide a brief overview of the current state of consumer apps for melanoma detection and the progress of the ResearchKit melanoma project.

Metric learning has the aim to improve classification accuracy by learning a distance measure which brings data points from the same class closer together and pushes data points from different classes further apart. Recent research has demonstrated that metric learning approaches can also be applied to trees, such as molecular structures, abstract syntax trees of computer programs, or syntax trees of natural language, by learning the cost function of an edit distance, i.e. the costs of replacing, deleting, or inserting nodes in a tree. However, learning such costs directly may yield an edit distance which violates metric axioms, is challenging to interpret, and may not generalize well. In this contribution, we propose a novel metric learning approach for trees which we call embedding edit distance learning (BEDL) and which learns an edit distance indirectly by embedding the tree nodes as vectors, such that the Euclidean distance between those vectors supports class discrimination. We learn such embeddings by reducing the distance to prototypical trees from the same class and increasing the distance to prototypical trees from different classes. In our experiments, we show that BEDL improves upon the state-of-the-art in metric learning for trees on six benchmark data sets, ranging from computer science over biomedical data to a natural-language processing data set containing over 300,000 nodes., Comment: Paper at the International Conference of Machine Learning (2018), 2018-07-10 to 2018-07-15 in Stockholm, Sweden

Learning to coordinate between multiple agents is an important problem in many reinforcement learning problems. Key to learning to coordinate is exploiting loose couplings,i.e.,conditional independences between agents. In this paper we study learning in repeated fully cooperative games, multi-agent multi-armed bandits (MAMABs), in which the expected rewards can be expressed as a coordination graph. We propose multi-agent upper confidence exploration (MAUCE), a new algorithm for MAMABs that exploits loose couplings, which enables us to prove a regret bound that is logarithmic in the number of arm pulls and only linear in the number of agents.We empirically compare MAUCE to sparse cooperative Q-learning, and a state-of-the-art combinatorial bandit approach, and show that it performs much better on a variety of settings,including learning control policies for wind farms.

Ordinary differential equations, and in general a dynamical system viewpoint, have seen a resurgence of interest in developing fast optimization methods, mainly thanks to the availability of well-established analysis tools. In this study, we pursue a similar objective and propose a class of hybrid control systems that adopts a 2nd-order differential equation as its continuous flow. A distinctive feature of the proposed differential equation in comparison with the existing literature is a state-dependent, time-invariant damping term that acts as a feedback control input. Given a user-defined scalar α, it is shown that the proposed control input steers the state trajectories to the global optimizer of a desired objective function with a guaranteed rate of convergence O(e−αt). Our framework requires that the objective function satisfies the so called Polyak–{Ł}ojasiewicz inequality. Furthermore, a discretization method is introduced such that the resulting discrete dynamical system possesses an exponential rate of convergence.

Reflectance confocal microscopy (RCM) is an effective, non-invasive pre-screening tool for skin cancer diagnosis, but it requires extensive training and experience to assess accurately. There are few quantitative tools available to standardize image acquisition and analysis, and the ones that are available are not interpretable. In this study, we use a recurrent neural network with attention on convolutional network features. We apply it to delineate skin strata in vertically-oriented stacks of transverse RCM image slices in an interpretable manner. We introduce a new attention mechanism called Toeplitz attention, which constrains the attention map to have a Toeplitz structure. Testing our model on an expert labeled dataset of 504 RCM stacks, we achieve 88.17% image-wise classification accuracy, which is the current state-of-art., Comment: Accepted for ML4H Workshop at NIPS 2017

Climate informatics is an emerging research area that combines the fields of climate science and data science (specifically machine learning, data mining and statistics) to accelerate scientific discovery in climate science. The annual climate informatics workshop, held at NCAR's Mesa Lab since 2012, promotes new collaborations and discusses new methods and directions for this emerging field. This year's proceedings contain 34 peer-reviewed short papers presented at the workshop, which describe many new methods and advances in the field. Making these papers available to all interested researchers is essential to maximize further advances in this important field.