Polygenic and transcriptional risk scores identify chronic obstructive pulmonary disease subtypes.

Rationale: Genetic variants and gene expression predict risk of chronic obstructive pulmonary disease (COPD), but their effect on COPD heterogeneity is unclear. Objectives: Define high-risk COPD subtypes using both genetics (polygenic risk score, PRS) and blood gene expression (transcriptional risk score, TRS) and assess differences in clinical and molecular characteristics. Methods: We defined high-risk groups based on PRS and TRS quantiles by maximizing differences in protein biomarkers in a COPDGene training set and identified these groups in COPDGene and ECLIPSE test sets. We tested multivariable associations of subgroups with clinical outcomes and compared protein-protein interaction networks and drug repurposing analyses between high-risk groups. Measurements and Main Results: We examined two high-risk omics-defined groups in non-overlapping test sets (n=1,133 NHW COPDGene, n=299 African American (AA) COPDGene, n=468 ECLIPSE). We defined "High activity" (low PRS/high TRS) and "severe risk" (high PRS/high TRS) subgroups. Participants in both subgroups had lower body-mass index (BMI), lower lung function, and alterations in metabolic, growth, and immune signaling processes compared to a low-risk (low PRS, low TRS) reference subgroup. "High activity" but not "severe risk" participants had greater prospective FEV 1 decline (COPDGene: -51 mL/year; ECLIPSE: - 40 mL/year) and their proteomic profiles were enriched in gene sets perturbed by treatment with 5-lipoxygenase inhibitors and angiotensin-converting enzyme (ACE) inhibitors. Conclusions: Concomitant use of polygenic and transcriptional risk scores identified clinical and molecular heterogeneity amongst high-risk individuals. Proteomic and drug repurposing analysis identified subtype-specific enrichment for therapies and suggest prior drug repurposing failures may be explained by patient selection.

Explainable deep learning for insights in El Niño and river flows.

The El Niño Southern Oscillation (ENSO) is a semi-periodic fluctuation in sea surface temperature (SST) over the tropical central and eastern Pacific Ocean that influences interannual variability in regional hydrology across the world through long-range dependence or teleconnections. Recent research has demonstrated the value of Deep Learning (DL) methods for improving ENSO prediction as well as Complex Networks (CN) for understanding teleconnections. However, gaps in predictive understanding of ENSO-driven river flows include the black box nature of DL, the use of simple ENSO indices to describe a complex phenomenon and translating DL-based ENSO predictions to river flow predictions. Here we show that eXplainable DL (XDL) methods, based on saliency maps, can extract interpretable predictive information contained in global SST and discover SST information regions and dependence structures relevant for river flows which, in tandem with climate network constructions, enable improved predictive understanding. Our results reveal additional information content in global SST beyond ENSO indices, develop understanding of how SSTs influence river flows, and generate improved river flow prediction, including uncertainty estimation. Observations, reanalysis data, and earth system model simulations are used to demonstrate the value of the XDL-CN based methods for future interannual and decadal scale climate projections.

Machine Learning-Based Determination of Sampling Depth for Complex Environmental Systems: Case Study with Single-Cell Raman Spectroscopy Data in EBPR Systems.

Rapid progress in various advanced analytical methods, such as single-cell technologies, enable unprecedented and deeper understanding of microbial ecology beyond the resolution of conventional approaches. A major application challenge exists in the determination of sufficient sample size without sufficient prior knowledge of the community complexity and, the need to balance between statistical power and limited time or resources. This hinders the desired standardization and wider application of these technologies. Here, we proposed, tested and validated a computational sampling size assessment protocol taking advantage of a metric, named kernel divergence. This metric has two advantages: First, it directly compares data set-wise distributional differences with no requirements on human intervention or prior knowledge-based preclassification. Second, minimal assumptions in distribution and sample space are made in data processing to enhance its application domain. This enables test-verified appropriate handling of data sets with both linear and nonlinear relationships. The model was then validated in a case study with Single-cell Raman Spectroscopy (SCRS) phenotyping data sets from eight different enhanced biological phosphorus removal (EBPR) activated sludge communities located across North America. The model allows the determination of sufficient sampling size for any targeted or customized information capture capacity or resolution level. Promised by its flexibility and minimal restriction of input data types, the proposed method is expected to be a standardized approach for sampling size optimization, enabling more comparable and reproducible experiments and analysis on complex environmental samples. Finally, these advantages enable the extension of the capability to other single-cell technologies or environmental applications with data sets exhibiting continuous features.

Machine Learning Prediction of Progression in Forced Expiratory Volume in 1 Second in the COPDGene® Study.

Background: The heterogeneous nature of chronic obstructive pulmonary disease (COPD) complicates the identification of the predictors of disease progression. We aimed to improve the prediction of disease progression in COPD by using machine learning and incorporating a rich dataset of phenotypic features. Methods: We included 4496 smokers with available data from their enrollment and 5-year follow-up visits in the COPD Genetic Epidemiology (COPDGene®) study. We constructed linear regression (LR) and supervised random forest models to predict 5-year progression in forced expiratory in 1 second (FEV1) from 46 baseline features. Using cross-validation, we randomly partitioned participants into training and testing samples. We also validated the results in the COPDGene 10-year follow-up visit. Results: Predicting the change in FEV1 over time is more challenging than simply predicting the future absolute FEV1 level. For random forest, R-squared was 0.15 and the area under the receiver operator characteristic (ROC) curves for the prediction of participants in the top quartile of observed progression was 0.71 (testing) and respectively, 0.10 and 0.70 (validation). Random forest provided slightly better performance than LR. The accuracy was best for Global initiative for chronic Obstructive Lung Disease (GOLD) grades 1-2 participants, and it was harder to achieve accurate prediction in advanced stages of the disease. Predictive variables differed in their relative importance as well as for the predictions by GOLD. Conclusion: Random forest, along with deep phenotyping, predicts FEV1 progression with reasonable accuracy. There is significant room for improvement in future models. This prediction model facilitates the identification of smokers at increased risk for rapid disease progression. Such findings may be useful in the selection of patient populations for targeted clinical trials.

Skin strata delineation in reflectance confocal microscopy images using recurrent convolutional networks with attention.

Reflectance confocal microscopy (RCM) is an effective non-invasive tool for cancer diagnosis. However, acquiring and reading RCM images requires extensive training and experience, and novice clinicians exhibit high discordance in diagnostic accuracy. Quantitative tools to standardize image acquisition could reduce both required training and diagnostic variability. To perform diagnostic analysis, clinicians collect a set of RCM mosaics (RCM images concatenated in a raster fashion to extend the field view) at 4-5 specific layers in skin, all localized in the junction between the epidermal and dermal layers (dermal-epidermal junction, DEJ), necessitating locating that junction before mosaic acquisition. In this study, we automate DEJ localization using deep recurrent convolutional neural networks to delineate skin strata in stacks of RCM images collected at consecutive depths. Success will guide to automated and quantitative mosaic acquisition thus reducing inter operator variability and bring standardization in imaging. Testing our model against an expert labeled dataset of 504 RCM stacks, we achieved [Formula: see text] classification accuracy and nine-fold reduction in the number of anatomically impossible errors compared to the previous state-of-the-art.

Semantic segmentation of reflectance confocal microscopy mosaics of pigmented lesions using weak labels.

Reflectance confocal microscopy (RCM) is a non-invasive imaging tool that reduces the need for invasive histopathology for skin cancer diagnoses by providing high-resolution mosaics showing the architectural patterns of skin, which are used to identify malignancies in-vivo. RCM mosaics are similar to dermatopathology sections, both requiring extensive training to interpret. However, these modalities differ in orientation, as RCM mosaics are horizontal (parallel to the skin surface) while histopathology sections are vertical, and contrast mechanism, RCM with a single (reflectance) mechanism resulting in grayscale images and histopathology with multi-factor color-stained contrast. Image analysis and machine learning methods can potentially provide a diagnostic aid to clinicians to interpret RCM mosaics, eventually helping to ease the adoption and more efficiently utilizing RCM in routine clinical practice. However standard supervised machine learning may require a prohibitive volume of hand-labeled training data. In this paper, we present a weakly supervised machine learning model to perform semantic segmentation of architectural patterns encountered in RCM mosaics. Unlike more widely used fully supervised segmentation models that require pixel-level annotations, which are very labor-demanding and error-prone to obtain, here we focus on training models using only patch-level labels (e.g. a single field of view within an entire mosaic). We segment RCM mosaics into "benign" and "aspecific (nonspecific)" regions, where aspecific regions represent the loss of regular architecture due to injury and/or inflammation, pre-malignancy, or malignancy. We adopt Efficientnet, a deep neural network (DNN) proven to accurately accomplish classification tasks, to generate class activation maps, and use a Gaussian weighting kernel to stitch smaller images back into larger fields of view. The trained DNN achieved an average area under the curve of 0.969, and Dice coefficient of 0.778 showing the feasibility of spatial localization of aspecific regions in RCM images, and making the diagnostics decision model more interpretable to the clinicians.

Predicting and Monitoring Upper-Limb Rehabilitation Outcomes Using Clinical and Wearable Sensor Data in Brain Injury Survivors.

OBJECTIVE: Rehabilitation specialists have shown considerable interest for the development of models, based on clinical data, to predict the response to rehabilitation interventions in stroke and traumatic brain injury survivors. However, accurate predictions are difficult to obtain due to the variability in patients' response to rehabilitation interventions. This study aimed to investigate the use of wearable technology in combination with clinical data to predict and monitor the recovery process and assess the responsiveness to treatment on an individual basis. METHODS: Gaussian Process Regression-based algorithms were developed to estimate rehabilitation outcomes (i.e., Functional Ability Scale scores) using either clinical or wearable sensor data or a combination of the two. RESULTS: The algorithm based on clinical data predicted rehabilitation outcomes with a Pearson's correlation of 0.79 compared to actual clinical scores provided by clinicians but failed to model the variability in responsiveness to the intervention observed across individuals. In contrast, the algorithm based on wearable sensor data generated rehabilitation outcome estimates with a Pearson's correlation of 0.91 and modeled the individual responses to rehabilitation more accurately. Furthermore, we developed a novel approach to combine estimates derived from the clinical data and the sensor data using a constrained linear model. This approach resulted in a Pearson's correlation of 0.94 between estimated and clinician-provided scores. CONCLUSION: This algorithm could enable the design of patient-specific interventions based on predictions of rehabilitation outcomes relying on clinical and wearable sensor data. SIGNIFICANCE: This is important in the context of developing precision rehabilitation interventions.

Segmentation of cellular patterns in confocal images of melanocytic lesions in vivo via a multiscale encoder-decoder network (MED-Net).

In-vivo optical microscopy is advancing into routine clinical practice for non-invasively guiding diagnosis and treatment of cancer and other diseases, and thus beginning to reduce the need for traditional biopsy. However, reading and analysis of the optical microscopic images are generally still qualitative, relying mainly on visual examination. Here we present an automated semantic segmentation method called "Multiscale Encoder-Decoder Network (MED-Net)" that provides pixel-wise labeling into classes of patterns in a quantitative manner. The novelty in our approach is the modeling of textural patterns at multiple scales (magnifications, resolutions). This mimics the traditional procedure for examining pathology images, which routinely starts with low magnification (low resolution, large field of view) followed by closer inspection of suspicious areas with higher magnification (higher resolution, smaller fields of view). We trained and tested our model on non-overlapping partitions of 117 reflectance confocal microscopy (RCM) mosaics of melanocytic lesions, an extensive dataset for this application, collected at four clinics in the US, and two in Italy. With patient-wise cross-validation, we achieved pixel-wise mean sensitivity and specificity of 74% and 92%, respectively, with 0.74 Dice coefficient over six classes. In the scenario, we partitioned the data clinic-wise and tested the generalizability of the model over multiple clinics. In this setting, we achieved pixel-wise mean sensitivity and specificity of 77% and 94%, respectively, with 0.77 Dice coefficient. We compared MED-Net against the state-of-the-art semantic segmentation models and achieved better quantitative segmentation performance. Our results also suggest that, due to its nested multiscale architecture, the MED-Net model annotated RCM mosaics more coherently, avoiding unrealistic-fragmented annotations.

Machine Learning Characterization of COPD Subtypes: Insights From the COPDGene Study.

COPD is a heterogeneous syndrome. Many COPD subtypes have been proposed, but there is not yet consensus on how many COPD subtypes there are and how they should be defined. The COPD Genetic Epidemiology Study (COPDGene), which has generated 10-year longitudinal chest imaging, spirometry, and molecular data, is a rich resource for relating COPD phenotypes to underlying genetic and molecular mechanisms. In this article, we place COPDGene clustering studies in context with other highly cited COPD clustering studies, and summarize the main COPD subtype findings from COPDGene. First, most manifestations of COPD occur along a continuum, which explains why continuous aspects of COPD or disease axes may be more accurate and reproducible than subtypes identified through clustering methods. Second, continuous COPD-related measures can be used to create subgroups through the use of predictive models to define cut-points, and we review COPDGene research on blood eosinophil count thresholds as a specific example. Third, COPD phenotypes identified or prioritized through machine learning methods have led to novel biological discoveries, including novel emphysema genetic risk variants and systemic inflammatory subtypes of COPD. Fourth, trajectory-based COPD subtyping captures differences in the longitudinal evolution of COPD, addressing a major limitation of clustering analyses that are confounded by disease severity. Ongoing longitudinal characterization of subjects in COPDGene will provide useful insights about the relationship between lung imaging parameters, molecular markers, and COPD progression that will enable the identification of subtypes based on underlying disease processes and distinct patterns of disease progression, with the potential to improve the clinical relevance and reproducibility of COPD subtypes.

Utilizing Machine Learning for Image Quality Assessment for Reflectance Confocal Microscopy.

In vivo reflectance confocal microscopy (RCM) enables clinicians to examine lesions' morphological and cytological information in epidermal and dermal layers while reducing the need for biopsies. As RCM is being adopted more widely, the workflow is expanding from real-time diagnosis at the bedside to include a capture, store, and forward model with image interpretation and diagnosis occurring offsite, similar to radiology. As the patient may no longer be present at the time of image interpretation, quality assurance is key during image acquisition. Herein, we introduce a quality assurance process by means of automatically quantifying diagnostically uninformative areas within the lesional area by using RCM and coregistered dermoscopy images together. We trained and validated a pixel-level segmentation model on 117 RCM mosaics collected by international collaborators. The model delineates diagnostically uninformative areas with 82% sensitivity and 93% specificity. We further tested the model on a separate set of 372 coregistered RCM-dermoscopic image pairs and illustrate how the results of the RCM-only model can be improved via a multimodal (RCM + dermoscopy) approach, which can help quantify the uninformative regions within the lesional area. Our data suggest that machine learning-based automatic quantification offers a feasible objective quality control measure for RCM imaging.

COPDGene® 2019: Redefining the Diagnosis of Chronic Obstructive Pulmonary Disease.

BACKGROUND: Chronic obstructive pulmonary disease (COPD) remains a major cause of morbidity and mortality. Present-day diagnostic criteria are largely based solely on spirometric criteria. Accumulating evidence has identified a substantial number of individuals without spirometric evidence of COPD who suffer from respiratory symptoms and/or increased morbidity and mortality. There is a clear need for an expanded definition of COPD that is linked to physiologic, structural (computed tomography [CT]) and clinical evidence of disease. Using data from the COPD Genetic Epidemiology study (COPDGene®), we hypothesized that an integrated approach that includes environmental exposure, clinical symptoms, chest CT imaging and spirometry better defines disease and captures the likelihood of progression of respiratory obstruction and mortality. METHODS: Four key disease characteristics - environmental exposure (cigarette smoking), clinical symptoms (dyspnea and/or chronic bronchitis), chest CT imaging abnormalities (emphysema, gas trapping and/or airway wall thickening), and abnormal spirometry - were evaluated in a group of 8784 current and former smokers who were participants in COPDGene® Phase 1. Using these 4 disease characteristics, 8 categories of participants were identified and evaluated for odds of spirometric disease progression (FEV1 > 350 ml loss over 5 years), and the hazard ratio for all-cause mortality was examined. RESULTS: Using smokers without symptoms, CT imaging abnormalities or airflow obstruction as the reference population, individuals were classified as Possible COPD, Probable COPD and Definite COPD. Current Global initiative for obstructive Lung Disease (GOLD) criteria would diagnose 4062 (46%) of the 8784 study participants with COPD. The proposed COPDGene® 2019 diagnostic criteria would add an additional 3144 participants. Under the new criteria, 82% of the 8784 study participants would be diagnosed with Possible, Probable or Definite COPD. These COPD groups showed increased risk of disease progression and mortality. Mortality increased in patients as the number of their COPD characteristics increased, with a maximum hazard ratio for all cause-mortality of 5.18 (95% confidence interval [CI]: 4.15-6.48) in those with all 4 disease characteristics. CONCLUSIONS: A substantial portion of smokers with respiratory symptoms and imaging abnormalities do not manifest spirometric obstruction as defined by population normals. These individuals are at significant risk of death and spirometric disease progression. We propose to redefine the diagnosis of COPD through an integrated approach using environmental exposure, clinical symptoms, CT imaging and spirometric criteria. These expanded criteria offer the potential to stimulate both current and future interventions that could slow or halt disease progression in patients before disability or irreversible lung structural changes develop.

Intelligent Labeling Based on Fisher Information for Medical Image Segmentation Using Deep Learning.

Deep convolutional neural networks (CNN) have recently achieved superior performance at the task of medical image segmentation compared to classic models. However, training a generalizable CNN requires a large amount of training data, which is difficult, expensive, and time-consuming to obtain in medical settings. Active Learning (AL) algorithms can facilitate training CNN models by proposing a small number of the most informative data samples to be annotated to achieve a rapid increase in performance. We proposed a new active learning method based on Fisher information (FI) for CNNs for the first time. Using efficient backpropagation methods for computing gradients together with a novel low-dimensional approximation of FI enabled us to compute FI for CNNs with a large number of parameters. We evaluated the proposed method for brain extraction with a patch-wise segmentation CNN model in two different learning scenarios: universal active learning and active semi-automatic segmentation. In both scenarios, an initial model was obtained using labeled training subjects of a source data set and the goal was to annotate a small subset of new samples to build a model that performs well on the target subject(s). The target data sets included images that differed from the source data by either age group (e.g. newborns with different image contrast) or underlying pathology that was not available in the source data. In comparison to several recently proposed AL methods and brain extraction baselines, the results showed that FI-based AL outperformed the competing methods in improving the performance of the model after labeling a very small portion of target data set (<0.25%).

Active Deep Learning with Fisher Information for Patch-wise Semantic Segmentation.

Deep learning with convolutional neural networks (CNN) has achieved unprecedented success in segmentation, however it requires large training data, which is expensive to obtain. Active Learning (AL) frameworks can facilitate major improvements in CNN performance with intelligent selection of minimal data to be labeled. This paper proposes a novel diversified AL based on Fisher information (FI) for the first time for CNNs, where gradient computations from backpropagation are used for efficient computation of FI on the large CNN parameter space. We evaluated the proposed method in the context of newborn and adolescent brain extraction problem under two scenarios: (1) semi-automatic segmentation of a particular subject from a different age group or with a pathology not available in the original training data, where starting from an inaccurate pre-trained model, we iteratively label small number of voxels queried by AL until the model generates accurate segmentation for that subject, and (2) using AL to build a universal model generalizable to all images in a given data set. In both scenarios, FI-based AL improved performance after labeling a small percentage (less than 0.05%) of voxels. The results showed that FI-based AL significantly outperformed random sampling, and achieved accuracy higher than entropy-based querying in transfer learning, where the model learns to extract brains of newborn subjects given an initial model trained on adolescents.

Identification of Chronic Obstructive Pulmonary Disease Axes That Predict All-Cause Mortality: The COPDGene Study.

Chronic obstructive pulmonary disease (COPD) is a syndrome caused by damage to the lungs that results in decreased pulmonary function and reduced structural integrity. Pulmonary function testing (PFT) is used to diagnose and stratify COPD into severity groups, and computed tomography (CT) imaging of the chest is often used to assess structural changes in the lungs. We hypothesized that the combination of PFT and CT phenotypes would provide a more powerful tool for assessing underlying morphologic differences associated with pulmonary function in COPD than does PFT alone. We used factor analysis of 26 variables to classify 8,157 participants recruited into the COPDGene cohort between January 2008 and June 2011 from 21 clinical centers across the United States. These factors were used as predictors of all-cause mortality using Cox proportional hazards modeling. Five factors explained 80% of the covariance and represented the following domains: factor 1, increased emphysema and decreased pulmonary function; factor 2, airway disease and decreased pulmonary function; factor 3, gas trapping; factor 4, CT variability; and factor 5, hyperinflation. After more than 46,079 person-years of follow-up, factors 1 through 4 were associated with mortality and there was a significant synergistic interaction between factors 1 and 2 on death. Considering CT measures along with PFT in the assessment of COPD can identify patients at particularly high risk for death.

Longitudinal Modeling of Lung Function Trajectories in Smokers with and without Chronic Obstructive Pulmonary Disease.

RATIONALE: The relationship between longitudinal lung function trajectories, chest computed tomography (CT) imaging, and genetic predisposition to chronic obstructive pulmonary disease (COPD) has not been explored. OBJECTIVES: 1) To model trajectories using a data-driven approach applied to longitudinal data spanning adulthood in the Normative Aging Study (NAS), and 2) to apply these models to demographically similar subjects in the COPDGene (Genetic Epidemiology of COPD) Study with detailed phenotypic characterization including chest CT. METHODS: We modeled lung function trajectories in 1,060 subjects in NAS with a median follow-up time of 29 years. We assigned 3,546 non-Hispanic white males in COPDGene to these trajectories for further analysis. We assessed phenotypic and genetic differences between trajectories and across age strata. MEASUREMENTS AND MAIN RESULTS: We identified four trajectories in NAS with differing levels of maximum lung function and rate of decline. In COPDGene, 617 subjects (17%) were assigned to the lowest trajectory and had the greatest radiologic burden of disease (P < 0.01); 1,283 subjects (36%) were assigned to a low trajectory with evidence of airway disease preceding emphysema on CT; 1,411 subjects (40%) and 237 subjects (7%) were assigned to the remaining two trajectories and tended to have preserved lung function and negligible emphysema. The genetic contribution to these trajectories was as high as 83% (P = 0.02), and membership in lower lung function trajectories was associated with greater parental histories of COPD, decreased exercise capacity, greater dyspnea, and more frequent COPD exacerbations. CONCLUSIONS: Data-driven analysis identifies four lung function trajectories. Trajectory membership has a genetic basis and is associated with distinct lung structural abnormalities.

Lobar Emphysema Distribution Is Associated With 5-Year Radiological Disease Progression.

BACKGROUND: Emphysema has considerable variability in its regional distribution. Craniocaudal emphysema distribution is an important predictor of the response to lung volume reduction. However, there is little consensus regarding how to define upper lobe-predominant and lower lobe-predominant emphysema subtypes. Consequently, the clinical and genetic associations with these subtypes are poorly characterized. METHODS: We sought to identify subgroups characterized by upper-lobe or lower-lobe emphysema predominance and comparable amounts of total emphysema by analyzing data from 9,210 smokers without alpha-1-antitrypsin deficiency in the Genetic Epidemiology of COPD (COPDGene) cohort. CT densitometric emphysema was measured in each lung lobe. Random forest clustering was applied to lobar emphysema variables after regressing out the effects of total emphysema. Clusters were tested for association with clinical and imaging outcomes at baseline and at 5-year follow-up. Their associations with genetic variants were also compared. RESULTS: Three clusters were identified: minimal emphysema (n = 1,312), upper lobe-predominant emphysema (n = 905), and lower lobe-predominant emphysema (n = 796). Despite a similar amount of total emphysema, the lower-lobe group had more severe airflow obstruction at baseline and higher rates of metabolic syndrome compared with subjects with upper-lobe predominance. The group with upper-lobe predominance had greater 5-year progression of emphysema, gas trapping, and dyspnea. Differential associations with known COPD genetic risk variants were noted. CONCLUSIONS: Subgroups of smokers defined by upper-lobe or lower-lobe emphysema predominance exhibit different functional and radiological disease progression rates, and the upper-lobe predominant subtype shows evidence of association with known COPD genetic risk variants. These subgroups may be useful in the development of personalized treatments for COPD.

A Probabilistic Active Learning Algorithm Based on Fisher Information Ratio.

The task of labeling samples is demanding and expensive. Active learning aims to generate the smallest possible training data set that results in a classifier with high performance in the test phase. It usually consists of two steps of selecting a set of queries and requesting their labels. Among the suggested objectives to score the query sets, information theoretic measures have become very popular. Yet among them, those based on Fisher information (FI) have the advantage of considering the diversity among the queries and tractable computations. In this work, we provide a practical algorithm based on Fisher information ratio to obtain query distribution for a general framework where, in contrast to the previous FI-based querying methods, we make no assumptions over the test distribution. The empirical results on synthetic and real-world data sets indicate that this algorithm gives competitive results.

Automated video-mosaicking approach for confocal microscopic imaging in vivo: an approach to address challenges in imaging living tissue and extend field of view.

We describe a computer vision-based mosaicking method for in vivo videos of reflectance confocal microscopy (RCM). RCM is a microscopic imaging technique, which enables the users to rapidly examine tissue in vivo. Providing resolution at cellular-level morphology, RCM imaging combined with mosaicking has shown to be highly sensitive and specific for non-invasively guiding skin cancer diagnosis. However, current RCM mosaicking techniques with existing microscopes have been limited to two-dimensional sequences of individual still images, acquired in a highly controlled manner, and along a specific predefined raster path, covering a limited area. The recent advent of smaller handheld microscopes is enabling acquisition of videos, acquired in a relatively uncontrolled manner and along an ad-hoc arbitrarily free-form, non-rastered path. Mosaicking of video-images (video-mosaicking) is necessary to display large areas of tissue. Our video-mosaicking methods addresses this need. The method can handle unique challenges encountered during video capture such as motion blur artifacts due to rapid motion of the microscope over the imaged area, warping in frames due to changes in contact angle and varying resolution with depth. We present test examples of video-mosaics of melanoma and non-melanoma skin cancers, to demonstrate potential clinical utility.

A Bayesian Nonparametric Model for Disease Subtyping: Application to Emphysema Phenotypes.

We introduce a novel Bayesian nonparametric model that uses the concept of disease trajectories for disease subtype identification. Although our model is general, we demonstrate that by treating fractions of tissue patterns derived from medical images as compositional data, our model can be applied to study distinct progression trends between population subgroups. Specifically, we apply our algorithm to quantitative emphysema measurements obtained from chest CT scans in the COPDGene Study and show several distinct progression patterns. As emphysema is one of the major components of chronic obstructive pulmonary disease (COPD), the third leading cause of death in the United States [1], an improved definition of emphysema and COPD subtypes is of great interest. We investigate several models with our algorithm, and show that one with age , pack years (a measure of cigarette exposure), and smoking status as predictors gives the best compromise between estimated predictive performance and model complexity. This model identified nine subtypes which showed significant associations to seven single nucleotide polymorphisms (SNPs) known to associate with COPD. Additionally, this model gives better predictive accuracy than multiple, multivariate ordinary least squares regression as demonstrated in a five-fold cross validation analysis. We view our subtyping algorithm as a contribution that can be applied to bridge the gap between CT-level assessment of tissue composition to population-level analysis of compositional trends that vary between disease subtypes.

A Marked Poisson Process Driven Latent Shape Model for 3D Segmentation of Reflectance Confocal Microscopy Image Stacks of Human Skin.

Segmenting objects of interest from 3D data sets is a common problem encountered in biological data. Small field of view and intrinsic biological variability combined with optically subtle changes of intensity, resolution, and low contrast in images make the task of segmentation difficult, especially for microscopy of unstained living or freshly excised thick tissues. Incorporating shape information in addition to the appearance of the object of interest can often help improve segmentation performance. However, the shapes of objects in tissue can be highly variable and design of a flexible shape model that encompasses these variations is challenging. To address such complex segmentation problems, we propose a unified probabilistic framework that can incorporate the uncertainty associated with complex shapes, variable appearance, and unknown locations. The driving application that inspired the development of this framework is a biologically important segmentation problem: the task of automatically detecting and segmenting the dermal-epidermal junction (DEJ) in 3D reflectance confocal microscopy (RCM) images of human skin. RCM imaging allows noninvasive observation of cellular, nuclear, and morphological detail. The DEJ is an important morphological feature as it is where disorder, disease, and cancer usually start. Detecting the DEJ is challenging, because it is a 2D surface in a 3D volume which has strong but highly variable number of irregularly spaced and variably shaped "peaks and valleys." In addition, RCM imaging resolution, contrast, and intensity vary with depth. Thus, a prior model needs to incorporate the intrinsic structure while allowing variability in essentially all its parameters. We propose a model which can incorporate objects of interest with complex shapes and variable appearance in an unsupervised setting by utilizing domain knowledge to build appropriate priors of the model. Our novel strategy to model this structure combines a spatial Poisson process with shape priors and performs inference using Gibbs sampling. Experimental results show that the proposed unsupervised model is able to automatically detect the DEJ with physiologically relevant accuracy in the range 10- 20 µm .