Search

Abstract: The aim of this paper is to develop a spatial Gaussian predictive process (SGPP) framework for accurately predicting neuroimaging data by using a set of covariates of interest, such as age and diagnostic status, and an existing neuroimaging data set. To achieve a better prediction, we not only delineate spatial association between neuroimaging data and covariates, but also explicitly model spatial dependence in neuroimaging data. The SGPP model uses a functional principal component model to capture medium-to-long-range (or global) spatial dependence, while SGPP uses a multivariate simultaneous autoregressive model to capture short-range (or local) spatial dependence as well as cross-correlations of different imaging modalities. We propose a three-stage estimation procedure to simultaneously estimate varying regression coefficients across voxels and the global and local spatial dependence structures. Furthermore, we develop a predictive method to use the spatial correlations as well as the cross-correlations by employing a cokriging technique, which can be useful for the imputation of missing imaging data. Simulation studies and real data analysis are used to evaluate the prediction accuracy of SGPP and show that SGPP significantly outperforms several competing methods, such as voxel-wise linear model, in prediction. Although we focus on the morphometric variation of lateral ventricle surfaces in a clinical study of neurodevelopment, it is expected that SGPP is applicable to other imaging modalities and features. [Copyright &y& Elsevier]

Survival data can contain an unknown fraction of subjects who are “cured” in the sense of not being at risk of failure. We describe such data with cure-mixture models, which separately model cure status and the hazard of failure among non-cured subjects. No diagnostic currently exists for evaluating the fit of such models; the popular Schoenfeld residual (Schoenfeld, 1982. Partial residuals for the proportional hazards regression-model. Biometrika 69, 239–241) is not applicable to data with cures. In this article, we propose a pseudo-residual, modeled on Schoenfeld's, to assess the fit of the survival regression in the non-cured fraction. Unlike Schoenfeld's approach, which tests the validity of the proportional hazards (PH) assumption, our method uses the full hazard and is thus also applicable to non-PH models. We derive the asymptotic distribution of the residuals and evaluate their performance by simulation in a range of parametric models. We apply our approach to data from a smoking cessation drug trial. [ABSTRACT FROM PUBLISHER]

Twin imaging studies have been valuable for understanding the relative contribution of the environment and genes on brain structures and their functions. Conventional analyses of twin imaging data include three sequential steps: spatially smoothing imaging data, independently fitting a structural equation model at each voxel, and finally correcting for multiple comparisons. However, conventional analyses are limited due to the same amount of smoothing throughout the whole image, the arbitrary choice of smoothing extent, and the decreased power in detecting environmental and genetic effects introduced by smoothing raw images. The goal of this paper is to develop a two-stage multiscale adaptive regression method (TwinMARM) for spatial and adaptive analysis of twin neuroimaging and behavioral data. The first stage is to establish the relationship between twin imaging data and a set of covariates of interest, such as age and gender. The second stage is to disentangle the environmental and genetic influences on brain structures and their functions. In each stage, TwinMARM employs hierarchically nested spheres with increasing radii at each location and then captures spatial dependence among imaging observations via consecutively connected spheres across all voxels. Simulation studies show that our TwinMARM significantly outperforms conventional analyses of twin imaging data. Finally, we use our method to detect statistically significant effects of genetic and environmental variations on white matter structures in a neonatal twin study. [ABSTRACT FROM PUBLISHER]

ABSTRACT Aims Smoking cessation trials generally record information on daily smoking behavior, but base analyses on measures of smoking status at the end of treatment (EOT). We present an alternative approach that analyzes the entire sequence of daily smoking status observations. Methods We analyzed daily abstinence data from a smoking cessation trial, using two longitudinal logistic regression methods: a mixed-effects (ME) model and a generalized estimating equations (GEE) model. We compared results to a standard analysis that takes abstinence status at EOT as outcome. We evaluated time-varying covariates (smoking history and time-varying drug effect) in the longitudinal analysis and compared ME and GEE approaches. Results We observed some differences in the estimated treatment effect odds ratios across models, with narrower confidence intervals under the longitudinal models. GEE yields similar results to ME when only baseline factors appear in the model, but gives biased results when one includes time-varying covariates. The longitudinal models indicate that the quit probability declines and the drug effect varies over time. Both the previous day's smoking status and recent smoking history predict quit probability, independently of the drug effect. Conclusion When analysing outcomes of studies from smoking cessation interventions, longitudinal models with multiple outcome data points, rather than just end of treatment, can makes efficient use of the data and incorporate time-varying covariates. The generalized estimating equations approach should be avoided when using time-varying predictors. [ABSTRACT FROM AUTHOR]

Summary: Panel‐count data arise when each study subject is observed only at discrete time points in a recurrent event study, and only the numbers of the event of interest between observation time points are recorded (Sun and Zhao, 2013). However, sometimes the exact number of events between some observation times is unknown and what we know is only whether the event of interest has occurred. In this article, we will refer this type of data to as mixed panel‐count data and propose a likelihood‐based semiparametric regression method for their analysis by using the nonhomogeneous Poisson process assumption. However, we establish the asymptotic properties of the resulting estimator by employing the empirical process theory and without using the Poisson assumption. Also, we conduct an extensive simulation study, which suggests that the proposed method works well in practice. Finally, the method is applied to a Childhood Cancer Survivor Study that motivated this study. [ABSTRACT FROM AUTHOR]

Simple Summary: Limited effective therapies exist for the treatment of neurofibromatosis type 1 (NF1)-associated plexiform neurofibroma (PN). The aim of our prospective clinical trial was to evaluate the activity and safety of vinblastine and methotrexate (conventional cytotoxic chemotherapies) given on a metronomic schedule in children and young adults with NF1 and PN. Of the 23 participants evaluable for treatment outcomes and toxicities, 14 completed all protocol therapy. There were no participants that discontinued therapy due to dose-limiting toxicities. In addition, there were no participants that demonstrated a partial response. VBL/MTX was well-tolerated but did not result in objective volumetric responses. Limited therapies exist for neurofibromatosis type 1 (NF1)-associated plexiform neurofibroma (PN). For this reason, the activity of vinblastine (VBL) and methotrexate (MTX) was evaluated in children and young adults with NF1 and PN. Patients ≤ 25 years of age with progressive and/or inoperable NF1-PN received VBL 6 mg/m2 and MTX 30 mg/m2 weekly for 26 weeks, followed by every 2 weeks for 26 weeks. Objective response rate was the primary endpoint. Of 25 participants enrolled, 23 were evaluable. The median age of participants was 6.6 years (range 0.3–20.7). The most frequent toxicities were neutropenia and elevation of transaminases. On two-dimensional (2D) imaging, 20 participants (87%) had stable tumor, with a median time to progression of 41.5 months (95% confidence interval 16.9, 64.9). Two of eight participants (25%) with airway involvement demonstrated functional improvements including decreased positive pressure requirements and apnea-hypopnea index. A post hoc three-dimensional (3D) analysis of PN volumes was completed on 15 participants with amenable imaging; 7 participants (46%) had progressive disease on or by the end of therapy. VBL/MTX was well-tolerated but did not result in objective volumetric response. Furthermore, 3D volumetric analysis highlighted the lack of sensitivity of 2D imaging for PN response evaluation. [ABSTRACT FROM AUTHOR]

Purpose : Accurate detection of leptomeningeal metastasis (LM) is critical for risk stratification and treatment of pediatric brain tumors. Poor-quality staging MRI has been associated with decreased survival in this population, but technical factors differentiating good from poor quality screening MRIs remain undefined. To test the hypothesis that key technical factors are associated with accurate MRI diagnosis of leptomeningeal metastasis in children with leptomeningeal seeding brain tumors. Methods: MRIs acquired at outside facilities and repeated in our institution within 35 days for 75 children with leptomeningeal seeding tumors were assessed for slice thickness and gap; use of T2 FLAIR + Contrast, acquisition plane of 3DT1WI + Contrast (brain); axial T1 + Contrast sequence, and use of pre-contrast T1 images (spine). Reported findings were recorded as positive, negative, or equivocal for LM and classified as true positive (TP; unequivocal metastasis), false negative (FN; not reported), false positive (FP; resolved without treatment), or true negative. Wilcoxon signed-rank and Fisher's exact test were used to assess technical differences between TP and FN MRIs. Results: Rate of LM detection was greater with smaller interslice gap in brain (P = 0.003) and spine (P = 0.002); use of T2 FLAIR + Contrast (P = 0.005) and sagittal plane for 3DT1WI + Contrast (P = 0.028) in brain; and use of alternatives to axial TSE/FSE in spine (P = 0.048). Slice thickness was not significant. Pre-contrast T1WI did not contribute to LM diagnosis in spine. Conclusion: Using post-contrast T2 FLAIR and sagittal 3DT1 in brain, small/no interslice gap, and avoiding TSE/FSE axials in spine may facilitate leptomeningeal metastasis detection in children with brain tumors. [ABSTRACT FROM AUTHOR]

Introduction: In diffuse intrinsic pontine gliomas (DIPG), subtracting pre-contrast from post-contrast T1-weighted images (T1WI) occasionally reveals subtle, 'occult' enhancement. We hypothesized that this represents intravascular enhancement related to angiogenesis and hence that these tumors should have greater blood volume fractions than do non-enhancing tumors. Methods: We retrospectively screened MR images of 66 patients initially diagnosed with DIPG and analyzed pretreatment conventional and dynamic susceptibility contrast (DSC) perfusion MRI studies of 61 patients. To determine the incidence of occult enhancement, cerebral blood volume (CBV) values were compared in areas of occult enhancement (OcE), no enhancement (NE), and normal-appearing deep cerebellar white matter (DCWM). Results: Tumors of 10 patients (16.4 %) had occult enhancement; those of 6 patients (9.8 %) had no enhancement at all. The average CBV in areas of occult enhancement was significantly higher than that in non-enhancing areas of the same tumor ( P = .03), within DCWM in the same patient ( P = .03), and when compared to anatomically paired/similar regions of interest (ROI) in patients with non-enhancing tumors ( P = .005). Conclusion: Areas of OcE correspond to areas of higher CBV in DIPG, which may be an MRI marker for angiogenesis, but larger scale studies may be needed to determine its potential relevance to grading by imaging, treatment stratification, biopsy guidance, and evaluation of response to targeted therapy. [ABSTRACT FROM AUTHOR]

We consider longitudinal discrete data that may be unequally spaced in time and may exhibit overdispersion, so that the variance of the outcome variable is inflated relative to its assumed distribution. We implement an approach that extends generalized linear models for analysis of longitudinal data and is likelihood based, in contrast to generalized estimating equations (GEE) that are semiparametric. The method assumes independence between subjects; first‐order antedependence within subjects; exponential family distributions for the first outcome on each subject and for the subsequent conditional distributions; and linearity of the expectations of the conditional distributions. We demonstrate application of the method in an analysis of seizure counts and in a study to evaluate the performance of transplant centers. Simulations for both studies demonstrate the benefits of the proposed likelihood based approach; however, they also demonstrate better than anticipated performance for GEE. [ABSTRACT FROM AUTHOR]

Introduction: Sickle cell disease (SCD) research is hampered by disparities in participation due in part to mistrust of research among racial/ethnic minorities. Beyond the historic context of research misconduct, little is known about the associations of social ecologic factors with mistrust and of mistrust with SCD clinical trials enrollment. This study evaluated proximal (age, gender, disease severity, perceived stress, SES) and distal (religious beliefs, social support, instrumental support) factors related to mistrust of research among caregivers of children with SCD and adolescents and young adults (AYAs) with SCD.Methods: Over an 18-month period (2009-2010), participants completed questionnaires of perceived barriers and benefits to clinical trials enrollment, perceived stress, and self-reported demographic and disease-related information. Analyses (January-June 2015) used multivariable linear regressions to evaluate predictors of mistrust.Results: Data were analyzed for 154 caregivers (mean age, 38.75 years; SD=9.56 years; 90.30% female) and 88 AYAs (mean age, 24.76 years; SD=7.25 years; 46.40% female). Among caregivers (full model, R(2)=0.14, p≤0.001), greater mistrust was explained by higher perceived stress (β=0.04, p=0.052); religious beliefs (β=0.61, p≤0.001); and greater instrumental support (β=0.07, p=0.044). Among AYAs (full model, R(2)=0.18, p≤0.001), higher mistrust was explained by being male (β=-0.56, p≤0.001) and lower instrumental support (β=-0.11, p=0.016). Mistrust was significantly greater among caregivers that reported no prior involvement in medical research (p=0.003).Conclusions: By understanding the complexity through which social ecologic factors contribute to mistrust, researchers may create targeted strategies to address mistrust and increase engagement in SCD research for caregivers and AYAs. [ABSTRACT FROM AUTHOR]