Accounting for network noise in graph-guided Bayesian modeling of structured high-dimensional data.

There is a growing body of literature on knowledge-guided statistical learning methods for analysis of structured high-dimensional data (such as genomic and transcriptomic data) that can incorporate knowledge of underlying networks derived from functional genomics and functional proteomics. These methods have been shown to improve variable selection and prediction accuracy and yield more interpretable results. However, these methods typically use graphs extracted from existing databases or rely on subject matter expertise, which are known to be incomplete and may contain false edges. To address this gap, we propose a graph-guided Bayesian modeling framework to account for network noise in regression models involving structured high-dimensional predictors. Specifically, we use 2 sources of network information, including the noisy graph extracted from existing databases and the estimated graph from observed predictors in the dataset at hand, to inform the model for the true underlying network via a latent scale modeling framework. This model is coupled with the Bayesian regression model with structured high-dimensional predictors involving an adaptive structured shrinkage prior. We develop an efficient Markov chain Monte Carlo algorithm for posterior sampling. We demonstrate the advantages of our method over existing methods in simulations, and through analyses of a genomics dataset and another proteomics dataset for Alzheimer's disease.

Safety and efficacy of a new high-dose regimen of panobinostat, gemcitabine, busulfan, and melphalan for 1st or 2nd salvage ASCT for refractory/relapsed or high-risk myeloma: Matched-pair comparisons with concurrent control cohorts.

Improvement of autologous stem-cell transplantation (ASCT) for myeloma is needed. Building on our prior work, we prospectively evaluated panobinostat and gemcitabine/busulfan/melphalan (GemBuMel) with ASCT in this population. Patients aged 18-65 years with relapsed/refractory or high-risk myeloma and adequate end-organ function were eligible. Treatment included panobinostat (20 mg/day, days -9 to -2) and GemBuMel (days -8 to -2). Patients were enrolled in 1st (ASCT-1) or 2nd ASCT (ASCT-2) cohorts. We compared their outcomes with all our other concurrent ASCT patients who met eligibility criteria but received melphalan or BuMel off study, matched for age, prior therapy lines, high-risk cytogenetics, and response at ASCT. We enrolled 80 patients, 48 and 32 in the ASCT-1 and ASCT-2 cohorts, respectively; in these two cohorts, high-risk cytogenetics were noted in 33 and 15 patients, respectively; unresponsive disease in 12 and 11 patients, respectively, after a median of 2 and 3 therapy lines, respectively. Transplant-related mortality (TRM) occurred in two ASCT-2 patients. One-year PFS rates were 69% (ASCT-1) and 72% (ASCT-2); 1-year OS rates were 79% (ASCT-1) and 84% (ASCT-2). Minimal residual disease negativity improved after ASCT-1 (8.5%-23%, p < .0001) and ASCT-2 (34%-55%, p = .02), which correlated with improved outcomes. Trial patients and controls (N = 371) had similar TRM and post-ASCT maintenance. Trial patients had better PFS after either a 1st (p = .02) or a 2nd ASCT (p = .04) than matched-paired control patients. In conclusion, panobinostat/GemBuMel is effective for relapsed/refractory or high-risk myeloma patients, with better PFS than concurrent matched controls receiving melphalan or BuMel.

Functional lumen imaging probe topography identifies patients with normal acid exposure and esophageal hypervigilance amongst proton-pump inhibitor non-responders.

BACKGROUND: Multiple factors contribute to symptom generation and treatment response in proton-pump inhibitor non-responders (PPI-NRs). We aimed to test whether PPI-NRs with normal acid exposure have a higher degree of esophageal hypersensitivity and hypervigilance and can be identified using functional lumen imaging probe (FLIP) topography at the time of endoscopy. METHODS: Data from PPI-NRs whom underwent endoscopy, FLIP and wireless 96-h pH-metry were retrospectively analyzed. Patients were grouped according to acid exposure time (AET) as (a) 0 days abnormal (AET > 6%), (b) 1-2 days abnormal, or (c) 3-4 days abnormal. The esophageal hypervigilance and anxiety scale (EHAS) score and other symptom scores were compared between groups. The discriminatory ability of the esophagogastric junction (EGJ) distensibility index (DI) and max EGJ diameter in identifying patients with 0 days abnormal AET was tested via receiver-operating-characteristic (ROC) curve analysis. RESULTS: EHAS score was 38.6 in the 0 days abnormal AET group, 30.4 in the 1-2 days abnormal AET group (p = 0.073 when compared to 0 days abnormal) and 28.2 in the 3-4 days abnormal AET group (p = 0.031 when compared to 0 days abnormal). Area-under-the-curve (AUC) for the DI in association with 0 days AET > 6% was 0.629. A DI of < 2.8 mm2/mmHg had a sensitivity of 83.3%, and negative predictive value of 88% in classifying patients with 0 days abnormal acid exposure (p = 0.004). CONCLUSIONS: FLIP complements prolonged wireless pH-metry in distinguishing the subset of PPI-NRs with completely normal acid exposure and a higher burden of esophageal hypervigilance. Proper identification of patients along the functional heartburn spectrum can improve overall surgical outcomes.

Accounting for temporal variability in functional magnetic resonance imaging improves prediction of intelligence.

Neuroimaging-based prediction methods for intelligence have seen a rapid development. Among different neuroimaging modalities, prediction using functional connectivity (FC) has shown great promise. Most literature has focused on prediction using static FC, with limited investigations on the merits of such analysis compared to prediction using dynamic FC or region-level functional magnetic resonance imaging (fMRI) times series that encode temporal variability. To account for the temporal dynamics in fMRI, we propose a bi-directional long short-term memory (bi-LSTM) approach that incorporates feature selection mechanism. The proposed pipeline is implemented via an efficient algorithm and applied for predicting intelligence using region-level time series and dynamic FC. We compare the prediction performance using different fMRI features acquired from the Adolescent Brain Cognitive Development (ABCD) study involving nearly 7000 individuals. Our detailed analysis illustrates the consistently inferior performance of static FC compared to region-level time series or dynamic FC for single and combined rest and task fMRI experiments. The joint analysis of task and rest fMRI leads to improved intelligence prediction under all models compared to using fMRI from only one experiment. In addition, the proposed bi-LSTM pipeline based on region-level time series identifies several shared and differential important brain regions across fMRI experiments that drive intelligence prediction. A test-retest analysis of the selected regions shows strong reliability across cross-validation folds. Given the large sample size of ABCD study, our results provide strong evidence that superior prediction of intelligence can be achieved by accounting for temporal variations in fMRI.

Bayesian longitudinal tensor response regression for modeling neuroplasticity.

A major interest in longitudinal neuroimaging studies involves investigating voxel-level neuroplasticity due to treatment and other factors across visits. However, traditional voxel-wise methods are beset with several pitfalls, which can compromise the accuracy of these approaches. We propose a novel Bayesian tensor response regression approach for longitudinal imaging data, which pools information across spatially distributed voxels to infer significant changes while adjusting for covariates. The proposed method, which is implemented using Markov chain Monte Carlo (MCMC) sampling, utilizes low-rank decomposition to reduce dimensionality and preserve spatial configurations of voxels when estimating coefficients. It also enables feature selection via joint credible regions which respect the shape of the posterior distributions for more accurate inference. In addition to group level inferences, the method is able to infer individual-level neuroplasticity, allowing for examination of personalized disease or recovery trajectories. The advantages of the proposed approach in terms of prediction and feature selection over voxel-wise regression are highlighted via extensive simulation studies. Subsequently, we apply the approach to a longitudinal Aphasia dataset consisting of task functional MRI images from a group of subjects who were administered either a control intervention or intention treatment at baseline and were followed up over subsequent visits. Our analysis revealed that while the control therapy showed long-term increases in brain activity, the intention treatment produced predominantly short-term changes, both of which were concentrated in distinct localized regions. In contrast, the voxel-wise regression failed to detect any significant neuroplasticity after multiplicity adjustments, which is biologically implausible and implies lack of power.

Integrative learning for population of dynamic networks with covariates.

Although there is a rapidly growing literature on dynamic connectivity methods, the primary focus has been on separate network estimation for each individual, which fails to leverage common patterns of information. We propose novel graph-theoretic approaches for estimating a population of dynamic networks that are able to borrow information across multiple heterogeneous samples in an unsupervised manner and guided by covariate information. Specifically, we develop a Bayesian product mixture model that imposes independent mixture priors at each time scan and uses covariates to model the mixture weights, which results in time-varying clusters of samples designed to pool information. The computation is carried out using an efficient Expectation-Maximization algorithm. Extensive simulation studies illustrate sharp gains in recovering the true dynamic network over existing dynamic connectivity methods. An analysis of fMRI block task data with behavioral interventions reveal sub-groups of individuals having similar dynamic connectivity, and identifies intervention-related dynamic network changes that are concentrated in biologically interpretable brain regions. In contrast, existing dynamic connectivity approaches are able to detect minimal or no changes in connectivity over time, which seems biologically unrealistic and highlights the challenges resulting from the inability to systematically borrow information across samples.

Bayesian generalized biclustering analysis via adaptive structured shrinkage.

Biclustering techniques can identify local patterns of a data matrix by clustering feature space and sample space at the same time. Various biclustering methods have been proposed and successfully applied to analysis of gene expression data. While existing biclustering methods have many desirable features, most of them are developed for continuous data and few of them can efficiently handle -omics data of various types, for example, binomial data as in single nucleotide polymorphism data or negative binomial data as in RNA-seq data. In addition, none of existing methods can utilize biological information such as those from functional genomics or proteomics. Recent work has shown that incorporating biological information can improve variable selection and prediction performance in analyses such as linear regression and multivariate analysis. In this article, we propose a novel Bayesian biclustering method that can handle multiple data types including Gaussian, Binomial, and Negative Binomial. In addition, our method uses a Bayesian adaptive structured shrinkage prior that enables feature selection guided by existing biological information. Our simulation studies and application to multi-omics datasets demonstrate robust and superior performance of the proposed method, compared to other existing biclustering methods.

Scalable Bayesian matrix normal graphical models for brain functional networks.

Recently, there has been an explosive growth in graphical modeling approaches for estimating brain functional networks. In a detailed study, we show that surprisingly, standard graphical modeling approaches for fMRI data may not yield accurate estimates of the brain network due to the inability to suitably account for temporal correlations. We propose a novel Bayesian matrix normal graphical model that jointly models the temporal covariance and the brain network under a separable structure for the covariance to obtain improved estimates. The approach is implemented via an efficient optimization algorithm that computes the maximum-a-posteriori network estimates having desirable theoretical properties and which is scalable to high dimensions. The proposed method leads to substantial gains in network estimation accuracy compared to standard brain network modeling approaches as illustrated via extensive simulations. We apply the method to resting state fMRI data from the Human Connectome Project involving a large number of time scans and brain regions, to study the relationships between fluid intelligence and functional connectivity, where it is not computationally feasible to apply existing matrix normal graphical models. Our proposed approach led to the detection of differences in connectivity between high and low fluid intelligence groups, whereas these differences were less pronounced or absent using the graphical lasso.

A difference degree test for comparing brain networks.

Recently, there has been a proliferation of methods investigating functional connectivity as a biomarker for mental disorders. Typical approaches include massive univariate testing at each edge or comparisons of network metrics to identify differing topological features. Limitations of these methods include low statistical power due to the large number of comparisons and difficulty attributing overall differences in networks to local variation. We propose a method to capture the difference degree, which is the number of edges incident to each region in the difference network. Our difference degree test (DDT) is a two-step procedure for identifying brain regions incident to a significant number of differentially weighted edges (DWEs). First, we select a data-adaptive threshold which identifies the DWEs followed by a statistical test for the number of DWEs incident to each brain region. We achieve this by generating an appropriate set of null networks which are matched on the first and second moments of the observed difference network using the Hirschberger-Qi-Steuer algorithm. This formulation permits separation of the network's true topology from the nuisance topology induced by the correlation measure that alters interregional connectivity in ways unrelated to brain function. In simulations, the proposed approach outperforms competing methods in detecting differentially connected regions of interest. Application of DDT to a major depressive disorder dataset leads to the identification of brain regions in the default mode network commonly implicated in this ruminative disorder.

Rapid Tuberculosis Diagnosis Using Reporter Enzyme Fluorescence.

Tuberculosis is the most frequent cause of death in humans from a single infectious agent. Due to low numbers of bacteria present in sputum during early infection, diagnosis does not usually occur until >3 to 4 months after symptoms develop. We created a new more sensitive diagnostic that can be carried out in 10 min with no processing or technical expertise. This assay utilizes the Mycobacterium tuberculosis-specific biomarker BlaC in reporter enzyme fluorescence (REF) that has been optimized for clinical samples, designated REFtb, along with a more specific fluorogenic substrate, CDG-3. We report the first evaluation of clinical specimens with REFtb assays in comparison to the gold standards for tuberculosis diagnosis, culture and smear microscopy. REFtb assays allowed diagnosis of 160 patients from 16 different countries with a sensitivity of 89% for smear-positive, culture-positive samples and 88% for smear-negative, culture-positive samples with a specificity of 82%. The negative predictive value of REFtb for tuberculosis infection is 93%, and the positive predictive value is 79%. Overall, these data point toward the need for larger accuracy studies by third parties using a commercially available REFtb kit to determine whether incorporation of REFtb into the clinical toolbox for suspected tuberculosis patients would improve case identification. If results similar to our own can be obtained by all diagnostic laboratories, REFtb would allow proper treatment of more than 85% of patients that would be missed during their initial visit to a clinic using current diagnostic strategies, reducing the potential for further spread of disease.

Emergency department imaging superusers.

PURPOSE: To identify and characterize the most frequent users of emergency department (ED) imaging. MATERIALS AND METHODS: All patients with at least one ED visit in 2016 across a four-hospital healthcare system were retrospectively identified and their ED imaging utilization characterized. RESULTS: Overall, 126,940 unique patients underwent 187,603 ED visits (mean 1.5 ± 1.7) and a total of 192,142 imaging examinations (mean 1.7 ± 2.7). Fifty-eight percent of patients were imaged (73,672) and underwent a mean 2.6 ± 2.7 exams. When ranked by ED visits, 1.6% (2007) of patients had ≥ 4 ED visits (mean 6.1 ± 5.4). These ED "clinical superusers" accounted for 7.7% (14,409) of total ED visits and underwent 6.8 ± 5.4 imaging examinations, while non-superusers underwent 1.5 ± 2.2 (p < 0.01). When ranked by ED imaging utilization, 12.3% (15,575) of patients underwent ≥ 4 ED imaging examinations and consumed 49.5% (95,053) of all imaging services. A subset of just 1.3% (1608) of ED patients underwent > 10 annual ED examinations (ED "imaging superusers") and accounted for 12.4% (23,787) of all ED imaging services. Only 0.4% (n = 472) of patients were both clinical and imaging superusers. Despite similar ED visits to clinical superusers (6.0 ± 5.6 vs. 6.1 ± 5.4, p = 0.92), imaging superusers underwent significantly more imaging (14.8 ± 4.8 vs. 6.8 ± 5.4 examinations, p < 0.01). CONCLUSION: Just 12% of ED patients consume 50% of all ED imaging services, and 1.3% consume 12.4%. These ED imaging superusers represent a distinct group from clinical superusers. Prospective identification of this newly described subgroup might permit targeted interventions to control ED imaging volume, restrain costs, and minimize per-patient radiation exposure.

Integrative Bayesian analysis of brain functional networks incorporating anatomical knowledge.

Recently, there has been increased interest in fusing multimodal imaging to better understand brain organization by integrating information on both brain structure and function. In particular, incorporating anatomical knowledge leads to desirable outcomes such as increased accuracy in brain network estimates and greater reproducibility of topological features across scanning sessions. Despite the clear advantages, major challenges persist in integrative analyses including an incomplete understanding of the structure-function relationship and inaccuracies in mapping anatomical structures due to inherent deficiencies in existing imaging technology. This calls for the development of advanced network modeling tools that appropriately incorporate anatomical structure in constructing brain functional networks. We propose a hierarchical Bayesian Gaussian graphical modeling approach which models the brain functional networks via sparse precision matrices whose degree of edge specific shrinkage is a random variable that is modeled using both anatomical structure and an independent baseline component. The proposed approach adaptively shrinks functional connections and flexibly identifies functional connections supported by structural connectivity knowledge. This enables robust brain network estimation even in the presence of misspecified anatomical knowledge, while accommodating heterogeneity in the structure-function relationship. We implement the approach via an efficient optimization algorithm which yields maximum a posteriori estimates. Extensive numerical studies involving multiple functional network structures reveal the clear advantages of the proposed approach over competing methods in accurately estimating brain functional connectivity, even when the anatomical knowledge is misspecified up to a certain degree. An application of the approach to data from the Philadelphia Neurodevelopmental Cohort (PNC) study reveals gender based connectivity differences across multiple age groups, and higher reproducibility in the estimation of network metrics compared to alternative methods.

Estimating dynamic brain functional networks using multi-subject fMRI data.

A common assumption in the study of brain functional connectivity is that the brain network is stationary. However it is increasingly recognized that the brain organization is prone to variations across the scanning session, fueling the need for dynamic connectivity approaches. One of the main challenges in developing such approaches is that the frequency and change points for the brain organization are unknown, with these changes potentially occurring frequently during the scanning session. In order to provide greater power to detect rapid connectivity changes, we propose a fully automated two-stage approach which pools information across multiple subjects to estimate change points in functional connectivity, and subsequently estimates the brain networks within each state phase lying between consecutive change points. The number and positioning of the change points are unknown and learned from the data in the first stage, by modeling a time-dependent connectivity metric under a fused lasso approach. In the second stage, the brain functional network for each state phase is inferred via sparse inverse covariance matrices. We compare the performance of the method with existing dynamic connectivity approaches via extensive simulation studies, and apply the proposed approach to a saccade block task fMRI data.

Scalable Bayesian variable selection for structured high-dimensional data.

Variable selection for structured covariates lying on an underlying known graph is a problem motivated by practical applications, and has been a topic of increasing interest. However, most of the existing methods may not be scalable to high-dimensional settings involving tens of thousands of variables lying on known pathways such as the case in genomics studies. We propose an adaptive Bayesian shrinkage approach which incorporates prior network information by smoothing the shrinkage parameters for connected variables in the graph, so that the corresponding coefficients have a similar degree of shrinkage. We fit our model via a computationally efficient expectation maximization algorithm which scalable to high-dimensional settings ( p ∼ 100 , 000 ). Theoretical properties for fixed as well as increasing dimensions are established, even when the number of variables increases faster than the sample size. We demonstrate the advantages of our approach in terms of variable selection, prediction, and computational scalability via a simulation study, and apply the method to a cancer genomics study.

Number needed to treat for time-to-event data with competing risks.

The number needed to treat is a tool often used in clinical settings to illustrate the effect of a treatment. It has been widely adopted in the communication of risks to both clinicians and non-clinicians, such as patients, who are better able to understand this measure than absolute risk or rate reductions. The concept was introduced by Laupacis, Sackett, and Roberts in 1988 for binary data, and extended to time-to-event data by Altman and Andersen in 1999. However, up to the present, there is no definition of the number needed to treat for time-to-event data with competing risks. This paper introduces such a definition using the cumulative incidence function and suggests non-parametric and semi-parametric inferential methods for right-censored time-to-event data in the presence of competing risks. The procedures are illustrated using the data from a breast cancer clinical trial.

Multi-task Learning with High-Dimensional Noisy Images.

Recent medical imaging studies have given rise to distinct but inter-related datasets corresponding to multiple experimental tasks or longitudinal visits. Standard scalar-on-image regression models that fit each dataset separately are not equipped to leverage information across inter-related images, and existing multi-task learning approaches are compromised by the inability to account for the noise that is often observed in images. We propose a novel joint scalar-on-image regression framework involving wavelet-based image representations with grouped penalties that are designed to pool information across inter-related images for joint learning, and which explicitly accounts for noise in high-dimensional images via a projection-based approach. In the presence of non-convexity arising due to noisy images, we derive non-asymptotic error bounds under non-convex as well as convex grouped penalties, even when the number of voxels increases exponentially with sample size. A projected gradient descent algorithm is used for computation, which is shown to approximate the optimal solution via well-defined non-asymptotic optimization error bounds under noisy images. Extensive simulations and application to a motivating longitudinal Alzheimer's disease study illustrate significantly improved predictive ability and greater power to detect true signals, that are simply missed by existing methods without noise correction due to the attenuation to null phenomenon.

Bayesian Tensor Modeling for Image-based Classification of Alzheimer's Disease.

Tensor-based representations are being increasingly used to represent complex data types such as imaging data, due to their appealing properties such as dimension reduction and the preservation of spatial information. Recently, there is a growing literature on using Bayesian scalar-on-tensor regression techniques that use tensor-based representations for high-dimensional and spatially distributed covariates to predict continuous outcomes. However surprisingly, there is limited development on corresponding Bayesian classification methods relying on tensor-valued covariates. Standard approaches that vectorize the image are not desirable due to the loss of spatial structure, and alternate methods that use extracted features from the image in the predictive model may suffer from information loss. We propose a novel data augmentation-based Bayesian classification approach relying on tensor-valued covariates, with a focus on imaging predictors. We propose two data augmentation schemes, one resulting in a support vector machine (SVM) type of classifier, and another yielding a logistic regression classifier. While both types of classifiers have been proposed independently in literature, our contribution is to extend such existing methodology to accommodate high-dimensional tensor valued predictors that involve low rank decompositions of the coefficient matrix while preserving the spatial information in the image. An efficient Markov chain Monte Carlo (MCMC) algorithm is developed for implementing these methods. Simulation studies show significant improvements in classification accuracy and parameter estimation compared to routinely used classification methods. We further illustrate our method in a neuroimaging application using cortical thickness MRI data from Alzheimer's Disease Neuroimaging Initiative, with results displaying better classification accuracy throughout several classification tasks, including classification on pairs of the three diagnostic groups: normal control, AD patients, and MCI patients; gender classification (males vs females); and cognitive performance based on high and low levels of MMSE scores.

Prognostic value of tumor volume doubling time in lung-metastatic adenoid cystic carcinoma.

OBJECTIVES: Lung metastases in adenoid cystic carcinoma (ACC) usually have indolent growth and the optimal timing to start systemic therapy is not established. We assessed ACC lung metastasis tumor growth dynamics and compared the prognostic value of time to progression (TTP) and tumor volume doubling time (TVDT). METHODS: The study included ACC patients with ≥1 pulmonary metastasis (≥5 mm) and at least 2 chest computed tomography scans. Radiology assessment was performed from the first scan showing metastasis until treatment initiation or death. Up to 5 lung nodules per patient were segmented for TVDT calculation. To assess tumor growth rate (TGR), the correlation coefficient (r) and coefficient of determination (R2) were calculated for measured lung nodules. TTP was assessed per RECIST 1.1; TVDT was calculated using the Schwartz formula. Overall survival was analyzed using the Kaplan-Meier method. RESULTS: The study included 75 patients. Sixty-seven patients (89%) had lung-only metastasis on first CT scan. The TGR was overall constant (median R2 = 0.974). Median TTP and TVDT were 11.2 months and 7.5 months. Shorter TVDT (<6 months) was associated with poor overall survival (HR = 0.48; p = 0.037), but TTP was not associated with survival (HR = 1.02; p = 0.96). Cox regression showed that TVDT but not TTP significantly correlated with OS. TVDT calculated using estimated tumor volume correlated with TVDT obtained by segmentation. CONCLUSION: Most ACC lung metastases have a constant TGR. TVDT may be a better prognostic indicator than TTP in lung-metastatic ACC. TVDT can be estimated by single longitudinal measurement in clinical practice.

Associations Between Radiation Oncologist Demographic Factors and Segmentation Similarity Benchmarks: Insights From a Crowd-Sourced Challenge Using Bayesian Estimation.

PURPOSE: The quality of radiotherapy auto-segmentation training data, primarily derived from clinician observers, is of utmost importance. However, the factors influencing the quality of clinician-derived segmentations are poorly understood; our study aims to quantify these factors. METHODS: Organ at risk (OAR) and tumor-related segmentations provided by radiation oncologists from the Contouring Collaborative for Consensus in Radiation Oncology data set were used. Segmentations were derived from five disease sites: breast, sarcoma, head and neck (H&N), gynecologic (GYN), and GI. Segmentation quality was determined on a structure-by-structure basis by comparing the observer segmentations with an expert-derived consensus, which served as a reference standard benchmark. The Dice similarity coefficient (DSC) was primarily used as a metric for the comparisons. DSC was stratified into binary groups on the basis of structure-specific expert-derived interobserver variability (IOV) cutoffs. Generalized linear mixed-effects models using Bayesian estimation were used to investigate the association between demographic variables and the binarized DSC for each disease site. Variables with a highest density interval excluding zero were considered to substantially affect the outcome measure. RESULTS: Five hundred seventy-four, 110, 452, 112, and 48 segmentations were used for the breast, sarcoma, H&N, GYN, and GI cases, respectively. The median percentage of segmentations that crossed the expert DSC IOV cutoff when stratified by structure type was 55% and 31% for OARs and tumors, respectively. Regression analysis revealed that the structure being tumor-related had a substantial negative impact on binarized DSC for the breast, sarcoma, H&N, and GI cases. There were no recurring relationships between segmentation quality and demographic variables across the cases, with most variables demonstrating large standard deviations. CONCLUSION: Our study highlights substantial uncertainty surrounding conventionally presumed factors influencing segmentation quality relative to benchmarks.

Integrative Bayesian tensor regression for imaging genetics applications.

Identifying biomarkers for Alzheimer's disease with a goal of early detection is a fundamental problem in clinical research. Both medical imaging and genetics have contributed informative biomarkers in literature. To further improve the performance, recently, there is an increasing interest in developing analytic approaches that combine data across modalities such as imaging and genetics. However, there are limited methods in literature that are able to systematically combine high-dimensional voxel-level imaging and genetic data for accurate prediction of clinical outcomes of interest. Existing prediction models that integrate imaging and genetic features often use region level imaging summaries, and they typically do not consider the spatial configurations of the voxels in the image or incorporate the dependence between genes that may compromise prediction ability. We propose a novel integrative Bayesian scalar-on-image regression model for predicting cognitive outcomes based on high-dimensional spatially distributed voxel-level imaging data, along with correlated transcriptomic features. We account for the spatial dependencies in the imaging voxels via a tensor approach that also enables massive dimension reduction to address the curse of dimensionality, and models the dependencies between the transcriptomic features via a Graph-Laplacian prior. We implement this approach via an efficient Markov chain Monte Carlo (MCMC) computation strategy. We apply the proposed method to the analysis of longitudinal ADNI data for predicting cognitive scores at different visits by integrating voxel-level cortical thickness measurements derived from T1w-MRI scans and transcriptomics data. We illustrate that the proposed imaging transcriptomics approach has significant improvements in prediction compared to prediction using a subset of features from only one modality (imaging or genetics), as well as when using imaging and transcriptomics features but ignoring the inherent dependencies between the features. Our analysis is one of the first to conclusively demonstrate the advantages of prediction based on combining voxel-level cortical thickness measurements along with transcriptomics features, while accounting for inherent structural information.