Semiparametric regression during 2003-2007

Semiparametric modeling for the cardiometabolic risk index and individual risk factors in the older adult population: A novel proposal

The accurate monitoring of metabolic syndrome in older adults is relevant in terms of its early detection, and its management. This study aimed at proposing a novel semiparametric modeling for a cardiometabolic risk index (CMRI) and individual risk factors in older adults.

METHODS

Multivariate semiparametric regression models were used to study the association between the CMRI with the individual risk factors, which was achieved using secondary analysis the data from the SABE study (Survey on Health, Well-Being, and Aging in Colombia, 2015).

RESULTS

The risk factors were selected through a stepwise procedure. The covariates included showed evidence of non-linear relationships with the CMRI, revealing non-linear interactions between: BMI and age (p< 0.00); arm and calf circumferences (p<0.00); age and females (p<0.00); walking speed and joint pain (p<0.02); and arm circumference and joint pain (p<0.00).

CONCLUSIONS

Semiparametric modeling explained 24.5% of the observed deviance, which was higher than the 18.2% explained by the linear model.

Polar lake microbiomes have distinct evolutionary histories

Toward the poles, life on land is increasingly dominated by microorganisms, yet the evolutionary origin of polar microbiomes remains poorly understood. Here, we use metabarcoding of Arctic, sub-Antarctic, and Antarctic lacustrine benthic microbial communities to test the hypothesis that high-latitude microbiomes are recruited from a globally dispersing species pool through environmental selection. We demonstrate that taxonomic overlap between the regions is limited within most phyla, even at higher-order taxonomic levels, with unique deep-branching phylogenetic clades being present in each region. We show that local and regional taxon richness and net diversification rate of regionally restricted taxa differ substantially between polar regions in both microeukaryotic and bacterial biota. This suggests that long-term evolutionary divergence resulting from low interhemispheric dispersal and diversification in isolation has been a prominent process shaping present-day polar lake microbiomes. Our findings illuminate the distinctive biogeography of polar lake ecosystems and underscore that conservation efforts should include their unique microbiota.

Comparisons of Medical Cost Trajectories Between Non-Hispanic Black and Non-Hispanic White Patients With Newly Diagnosed Localized Prostate Cancer

OBJECTIVES

This study applied a recently developed statistical method to compare the mean cost trajectories between non-Hispanic White (NHW) and non-Hispanic Black (NHB) patients with localized prostate cancer conditioning on patients' survival.

METHODS

In this observational study, we modeled cost trajectories of NHW and NHB patients with localized prostate cancer for 3 survival durations: 24, 48, and 72 months. We also compared the cost trajectories between NHW and NHB, stratified by comorbidities scores.

RESULTS

We find that the mean cost trajectories of NHB were significantly higher than the trajectories of NHW in the last 12 months before death, regardless of the survival duration and patients' baseline comorbidity scores. For patients with comorbidity score ≥2, mean cost trajectories within the first year of diagnosis for NHB were significantly higher than those for NHW, except for the subgroup of patients with comorbidity 2-3 and whose survival length was 72 months.

CONCLUSIONS

Our results suggested that a higher proportion of NHB patients with high comorbidity scores are likely contribute to their higher end-of-life costs than those for NHW patients. To narrow the gap in healthcare-related financial burden between NHB and NHW patients with localized prostate cancer, policy makers need to explore different strategies to better manage comorbidities.

Item Response Thresholds Models: A General Class of Models for Varying Types of Items

A comprehensive class of models is proposed that can be used for continuous, binary, ordered categorical and count type responses. The difficulty of items is described by difficulty functions, which replace the item difficulty parameters that are typically used in item response models. They crucially determine the response distribution and make the models very flexible with regard to the range of distributions that are covered. The model class contains several widely used models as the binary Rasch model and the graded response model as special cases, allows for simplifications, and offers a distribution free alternative to count type items. A major strength of the models is that they can be used for mixed item formats, when different types of items are combined to measure abilities or attitudes. It is an immediate consequence of the comprehensive modeling approach that allows that difficulty functions automatically adapt to the response distribution. Basic properties of the model class are shown. Several real data sets are used to illustrate the flexibility of the models.

Flexible Item Response Models for Count Data: The Count Thresholds Model

A new item response theory model for count data is introduced. In contrast to models in common use, it does not assume a fixed distribution for the responses as, for example, the Poisson count model and extensions do. The distribution of responses is determined by difficulty functions which reflect the characteristics of items in a flexible way. Sparse parameterizations are obtained by choosing fixed parametric difficulty functions, more general versions use an approximation by basis functions. The model can be seen as constructed from binary response models as the Rasch model or the normal-ogive model to which it reduces if responses are dichotomized. It is demonstrated that the model competes well with advanced count data models. Simulations demonstrate that parameters and response distributions are recovered well. An application shows the flexibility of the model to account for strongly varying distributions of responses.

Interpretable modelling of retail demand and price elasticity for passenger flights using booking data

We propose a model of retail demand for air travel and ticket price elasticity at the daily booking and individual flight level. Daily bookings are modelled as a non-homogeneous Poisson process with respect to the time to departure. The booking intensity is a function of booking and flight level covariates, including non-linear effects modelled semi-parametrically using penalized splines. Customer heterogeneity is incorporated using a finite mixture model, where the latent segments have covariate-dependent probabilities. We fit the model to a unique dataset of over one million daily counts of bookings for 9 602 scheduled flights on a short-haul route over two years. A control variate approach with a strong instrument corrects for a substantial level of price endogeneity. A rich latent segmentation is uncovered, along with strong covariate effects. The calibrated model can be used to quantify demand and price elasticity for different flights booked on different days prior to departure and is a step towards continuous pricing; something that is a major objective of airlines. As our model is interpretable, forecasts can be created under different scenarios. For instance, while our model is calibrated on data collected prior to COVID-19, many of the empirical insights are likely to remain valid as air travel recovers in the post-COVID-19 period.

Prediction of Penetration Resistance of a Spherical Penetrometer in Clay Using Multivariate Adaptive Regression Splines Model

This paper presents the technique for solving the penetration resistance factor of a spherical penetrometer in clay under axisymmetric conditions by taking the adhesion factor, the embedded ratio, the normalized unit weight, and the undrained shear strength into account. The finite element limit analysis (FELA) is used to provide the upper bound (UB) or lower bound (LB) solutions, then the multivariate adaptive regression splines (MARS) model is used to train the optimal data between input and output database. The accuracy of MARS equations is confirmed by comparison with the finite element method and the validity of the present solutions was established through comparison to existing results. All numerical results of the penetration resistance factor have significance with three main parameters (i.e., the adhesion factor, the embedded ratio, the normalized unit weight, and the undrained shear strength). The failure mechanisms of spherical penetrometers in clay are also investigated, the contour profiles that occur around the spherical penetrometers also depend on the three parameters. In addition, the proposed technique can be used to estimate the problems that are related or more complicated in soft offshore soils.

Nonparametric Mass Imputation for Data Integration

Data integration combining a probability sample with another nonprobability sample is an emerging area of research in survey sampling. We consider the case when the study variable of interest is measured only in the nonprobability sample, but comparable auxiliary information is available for both data sources. We consider mass imputation for the probability sample using the nonprobability data as the training set for imputation. The parametric mass imputation is sensitive to parametric model assumptions. To develop improved and robust methods, we consider nonparametric mass imputation for data integration. In particular, we consider kernel smoothing for a low-dimensional covariate and generalized additive models for a relatively high-dimensional covariate for imputation. Asymptotic theories and variance estimation are developed. Simulation studies and real applications show the benefits of our proposed methods over parametric counterparts.

High performance implementation of the hierarchical likelihood for generalized linear mixed models: an application to estimate the potassium reference range in massive electronic health records datasets

BACKGROUND

Converting electronic health record (EHR) entries to useful clinical inferences requires one to address the poor scalability of existing implementations of Generalized Linear Mixed Models (GLMM) for repeated measures. The major computational bottleneck concerns the numerical evaluation of multivariable integrals, which even for the simplest EHR analyses may involve millions of dimensions (one for each patient). The hierarchical likelihood (h-lik) approach to GLMMs is a methodologically rigorous framework for the estimation of GLMMs that is based on the Laplace Approximation (LA), which replaces integration with numerical optimization, and thus scales very well with dimensionality.

METHODS

We present a high-performance, direct implementation of the h-lik for GLMMs in the R package TMB. Using this approach, we examined the relation of repeated serum potassium measurements and survival in the Cerner Real World Data (CRWD) EHR database. Analyzing this data requires the evaluation of an integral in over 3 million dimensions, putting this problem beyond the reach of conventional approaches. We also assessed the scalability and accuracy of LA in smaller samples of 1 and 10% size of the full dataset that were analyzed via the a) original, interconnected Generalized Linear Models (iGLM), approach to h-lik, b) Adaptive Gaussian Hermite (AGH) and c) the gold standard for multivariate integration Markov Chain Monte Carlo (MCMC).

RESULTS

Random effects estimates generated by the LA were within 10% of the values obtained by the iGLMs, AGH and MCMC techniques. The H-lik approach was 4-30 times faster than AGH and nearly 800 times faster than MCMC. The major clinical inferences in this problem are the establishment of the non-linear relationship between the potassium level and the risk of mortality, as well as estimates of the individual and health care facility sources of variations for mortality risk in CRWD.

CONCLUSIONS

We found that the direct implementation of the h-lik offers a computationally efficient, numerically accurate approach for the analysis of extremely large, real world repeated measures data via the h-lik approach to GLMMs. The clinical inference from our analysis may guide choices of treatment thresholds for treating potassium disorders in the clinic.

A penalized spline fitting method to optimize geometric parameters of arterial centerlines extracted from medical images

In order to grasp the spatial and temporal evolution of vascular geometry, three-dimensional (3D) arterial bending structure and geometrical changes of arteries and stent grafts (SG) must be quantified using geometrical parameters such as curvature and torsion along the vasculature centerlines extracted from medical images. Here, we develop a robust method for constructing smooth centerlines based on a spline fitting method (SFM) such that the optimized geometric parameters of curvature and torsion can be obtained independently of digitization noise in the images. Conventional SFM consists of the 3rd degree spline basis function and 2nd derivative penalty term. In contrast, the present SFM uses the 5th degree spline basis function and 3rd and 4th derivative penalty terms, the coefficients of which are derived by the Akaike information criterion. The results show that the developed SFM can reduce the errors of curvature and torsion compared to conventional SFM. We then apply the present SFM to the centerline of the SG in an abdominal aortic aneurysm (AAA), and those of bilateral internal carotid arteries (ICA) in 6 cases: 3 cases with aneurysms and 3 cases without any aneurysm. The SG centerlines were obtained from temporal medical images at three scan times. The strong peak of the curvature could be clearly observed in the distal area of the SG, the inversion of the torsion at 0 months in the middle area of SG disappeared over time, and the torsions around the SG bifurcation at the three time periods were inverted. The curvature-torsion graphs along the ICA centerlines superimposing five aneurysmal positions were useful for investigating the relationship between arterial bending structure and aneurysmal positions. Both ICAs had curvature peak values higher than 0.4 within the ICA syphons. The ICA torsion graphs indicated that left and right ICA tended to be a right- and left-handed helix, respectively. In the left ICA syphon, the biggest aneurysm could be observed downstream of the salient torsion inversion. All aneurysms for 3 cases were positioned at the downstream of the inverted torsion.

Assessment of Ordinary Kriging and Inverse Distance Weighting Methods for Modeling Chromium and Cadmium Soil Pollution in E-Waste Sites in Douala, Cameroon

BACKGROUND

Several studies have demonstrated that chromium (Cr) and cadmium (Cd) have adverse impacts on the environment and human health. These elements are present in electronic waste (e-waste) recycling sites. Several interpolation methods have been used to evaluate geographical impacts on humans and the environment.

OBJECTIVES

The aim of the present paper is to compare the accuracy of inverse distance weighting (IDW) and ordinary kriging (OK) in topsoil analysis of e-waste recycling sites in Douala, Cameroon.

METHODS

Selecting the proper spatial interpolation method is crucial for carrying out surface analysis. Ordinary kriging and IDW are interpolation methods used for spatial analysis and surface mapping. Two sets of samples were used and compared. The performances of interpolation methods were evaluated and compared using cross-validation.

RESULTS

The results showed that the OK method performed better than IDW prediction for the spatial distribution of Cr, but the two interpolation methods had the same result for Cd (in the first set of samples). Results from Kolmogorov-Smirnov and Shapiro-Wilk tests showed that the data were normally distributed in the study area. The p value (0.302 and 0.773) was greater than 0.05 for Cr and for Cd (0.267 and 0.712). In the second set of samples, the OK method results (for Cd and Cr) were greatly diminished and the concentrations dropped, looking more like an average on the maps. However, the IDW interpolation gave a better representation of the concentration of Cd and Cr on the maps of the study area. For the second set of samples, OK and IDW for Cd and Cr had more similar results, especially in terms of root mean square error (RMSE).

CONCLUSIONS

Many parameters were better identified from the RMSE statistic obtained from cross-validation after exhaustive testing. Inverse distance weighting appeared more adequate in limited urban areas.

COMPETING INTERESTS

The authors declare no competing financial interests.

More efficient approximation of smoothing splines via space-filling basis selection

We consider the problem of approximating smoothing spline estimators in a nonparametric regression model. When applied to a sample of size [Formula: see text], the smoothing spline estimator can be expressed as a linear combination of [Formula: see text] basis functions, requiring [Formula: see text] computational time when the number [Formula: see text] of predictors is two or more. Such a sizeable computational cost hinders the broad applicability of smoothing splines. In practice, the full-sample smoothing spline estimator can be approximated by an estimator based on [Formula: see text] randomly selected basis functions, resulting in a computational cost of [Formula: see text]. It is known that these two estimators converge at the same rate when [Formula: see text] is of order [Formula: see text], where [Formula: see text] depends on the true function and [Formula: see text] depends on the type of spline. Such a [Formula: see text] is called the essential number of basis functions. In this article, we develop a more efficient basis selection method. By selecting basis functions corresponding to approximately equally spaced observations, the proposed method chooses a set of basis functions with great diversity. The asymptotic analysis shows that the proposed smoothing spline estimator can decrease [Formula: see text] to around [Formula: see text] when [Formula: see text]. Applications to synthetic and real-world datasets show that the proposed method leads to a smaller prediction error than other basis selection methods.

The Association between Sleep Duration and Metabolic Syndrome: The NHANES 2013/2014

BACKGROUND

We aimed to assess the association of sleep with metabolic syndrome in the 2013/2014 National Health and Nutrition Examination Survey (NHANES).

METHODS

Sample size included 2737 out of 2013 and 2014 NHANES surveys. Cross-sectional study of metabolic syndrome and sleep duration was conducted. Metabolic syndrome was defined according to NCEP ATPIII (National Cholesterol Education Program Adult Treatment Panel III) criteria. Metabolic syndrome severity score was calculated based on actual measurement of each component, adjusted for sex and race. The generalized additive model (GAM) was built to assess the smooth relationship between metabolic syndrome/metabolic syndrome severity score and sleep duration. Adjustment of models were done for age, sex, race, and sitting time. The value of effective degree of freedom (EDF) formed by the GAM model shows the degree of curvature of the relationship. A value of 1 for EDF is translated as the linear shape of relationship. Values larger than one denote a more complex relationship between the response variable and the predicting one.

RESULTS

There was a U-shaped association between sleep duration and metabolic syndrome in univariable GAM (EDF = 2.43, p = 0.06) and multivariable GAM (EDF = 2.03, p = 0.20). The lowest risk of metabolic syndrome was observed in people sleeping 7 hours/night. There was a significant U-shaped association between sleep duration and metabolic syndrome severity score in multivariable GAM (EDF = 2.94, p = 0.0004). Similarly, the lowest mean metabolic syndrome severity score was observed in people sleeping 7 hours/night. There was an effect modification of sex and sleep duration indicating strong U-shaped relationship of metabolic syndrome severity score and sleep duration in women (EDF = 3.43, p = 0.00002) and semi-linear association in men (EDF = 1.76, p = 0.04).

CONCLUSION

Short and long sleep duration was associated with higher risk of metabolic syndrome and higher scores of metabolic syndrome severity score in women. Short sleep duration was associated with higher risk of metabolic syndrome and higher scores of metabolic syndrome severity score in men.

A Gaussian Process Model for Color Camera Characterization: Assessment in Outdoor Levantine Rock Art Scenes

In this paper, we propose a novel approach to undertake the colorimetric camera characterization procedure based on a Gaussian process (GP). GPs are powerful and flexible nonparametric models for multivariate nonlinear functions. To validate the GP model, we compare the results achieved with a second-order polynomial model, which is the most widely used regression model for characterization purposes. We applied the methodology on a set of raw images of rock art scenes collected with two different Single Lens Reflex (SLR) cameras. A leave-one-out cross-validation (LOOCV) procedure was used to assess the predictive performance of the models in terms of CIE XYZ residuals and Δ E a b * color differences. Values of less than 3 CIELAB units were achieved for Δ E a b * . The output sRGB characterized images show that both regression models are suitable for practical applications in cultural heritage documentation. However, the results show that colorimetric characterization based on the Gaussian process provides significantly better results, with lower values for residuals and Δ E a b * . We also analyzed the induced noise into the output image after applying the camera characterization. As the noise depends on the specific camera, proper camera selection is essential for the photogrammetric work.

Flexible and structured survival model for a simultaneous estimation of non-linear and non-proportional effects and complex interactions between continuous variables: Performance of this multidimensional penalized spline approach in net survival trend analysis

Cancer survival trend analyses are essential to describe accurately the way medical practices impact patients' survival according to the year of diagnosis. To this end, survival models should be able to account simultaneously for non-linear and non-proportional effects and for complex interactions between continuous variables. However, in the statistical literature, there is no consensus yet on how to build such models that should be flexible but still provide smooth estimates of survival. In this article, we tackle this challenge by smoothing the complex hypersurface (time since diagnosis, age at diagnosis, year of diagnosis, and mortality hazard) using a multidimensional penalized spline built from the tensor product of the marginal bases of time, age, and year. Considering this penalized survival model as a Poisson model, we assess the performance of this approach in estimating the net survival with a comprehensive simulation study that reflects simple and complex realistic survival trends. The bias was generally small and the root mean squared error was good and often similar to that of the true model that generated the data. This parametric approach offers many advantages and interesting prospects (such as forecasting) that make it an attractive and efficient tool for survival trend analyses.

Probing ADP Induced Aggregation Kinetics During Platelet-Nanoparticle Interactions: Functional Dynamics Analysis to Rationalize Safety and Benefits

Platelets, one of the most sensitive blood cells, can be activated by a range of external and internal stimuli including physical, chemical, physiological, and/or non-physiological agents. Platelets need to respond promptly during injury to maintain blood hemostasis. The time profile of platelet aggregation is very complex, especially in the presence of the agonist adenosine 5'-diphosphate (ADP), and it is difficult to probe such complexity using traditional linear dose response models. In the present study, we explored functional analysis techniques to characterize the pattern of platelet aggregation over time in response to nanoparticle induced perturbations. This has obviated the need to represent the pattern of aggregation by a single summary measure and allowed us to treat the entire aggregation profile over time, as the response. The modeling was performed in a flexible manner, without any imposition of shape restrictions on the curve, allowing smooth platelet aggregation over time. The use of a probabilistic framework not only allowed statistical prediction and inference of the aggregation signatures, but also provided a novel method for the estimation of higher order derivatives of the curve, thereby allowing plausible estimation of the extent and rate of platelet aggregation kinetics over time. In the present study, we focused on the estimated first derivative of the curve, obtained from the platelet optical aggregometric profile over time and used it to discern the underlying kinetics as well as to study the effects of ADP dosage and perturbation with gold nanoparticles. In addition, our method allowed the quantification of the extent of inter-individual signature variations. Our findings indicated several hidden features and showed a mixture of zero and first order kinetics interrupted by a metastable zero order ADP dose dependent process. In addition, we showed that the two first order kinetic constants were ADP dependent. However, we were able to perturb the overall kinetic pattern using gold nanoparticles, which resulted in autocatalytic aggregation with a higher aggregate mass and which facilitated the aggregation rate.

The association between sleep duration and lipid profiles: the NHANES 2013-2014

Background

In the current literature, the association between sleep and different lipids is inconsistent. We aimed to assess the association of sleep with HDL cholesterol, triglyceride, and LDL cholesterol in the National Health and Nutrition Examination Survey (NHANES), 2013/2014.

Methods

We included 2705 participants from NHANES, 2013/2014. Cross-sectional information was measured on sleep duration and HDL cholesterol/triglyceride/LDL cholesterol. Generalized additive models (GAM) were constructed to assess the smooth relationship between the HDL cholesterol/triglyceride/LDL cholesterol, and the sleep duration. Models were adjusted for age, sex, race, marital status, household size, sitting time and physical activity. Effective degree of freedom (EDF) value in GAM indicated the amount of non-linearity of the smooth. EDF = 1 was indicative of a linear pattern of association. A value greater than 1 denoted a more complex association between outcome and sleep duration.

Results

The highest mean HDL cholesterol level was observed in participants sleeping 8 h/day. There was a significant non-linear association between sleep duration and HDL cholesterol in unadjusted GAM (EDF = 2.58, P = 0.002) and adjusted GAM (EDF = 1.85, P = 0.003). The lowest mean triglyceride level was observed in people sleeping 6 h/day. There was a significant non-linear association between sleep duration and triglyceride in unadjusted GAM (EDF = 3.05, P = 0.02) and adjusted GAM (EDF = 1.78, P = 0.02). There was no significant non-linear association between sleep duration and LDL cholesterol in either unadjusted GAM (EDF = 1.01, P = 0.2) or adjusted GAM (EDF = 1.01, P = 0.8).

Conclusion

Short sleep duration was associated with low HDL cholesterol/high triglyceride. Further longitudinal studies are warranted to shed extra light on this relationship.

Statistical Methodology in Studies of Prenatal Exposure to Mixtures of Endocrine-Disrupting Chemicals: A Review of Existing Approaches and New Alternatives

BACKGROUND

Prenatal exposures to endocrine-disrupting chemicals (EDCs) during critical developmental windows have been implicated in the etiologies of a wide array of adverse perinatal and pediatric outcomes. Epidemiological studies have concentrated on the health effects of individual chemicals, despite the understanding that EDCs act together via common mechanisms, that pregnant women are exposed to multiple EDCs simultaneously, and that substantial toxicological evidence of adverse developmental effects has been documented. There is a move toward multipollutant models in environmental epidemiology; however, there is no current consensus on appropriate statistical methods.

OBJECTIVES

We aimed to review the statistical methods used in these studies, to identify additional applicable methods, and to determine the strengths and weaknesses of each method for addressing the salient statistical and epidemiological challenges.

METHODS

We searched Embase, MEDLINE, and Web of Science for epidemiological studies of endocrine-sensitive outcomes in the children of mothers exposed to EDC mixtures during pregnancy and identified alternative statistical methods from the wider literature.

DISCUSSION

We identified 74 studies and analyzed the methods used to estimate mixture health effects, identify important mixture components, account for nonmonotonicity in exposure–response relationships, assess interactions, and identify windows of exposure susceptibility. We identified both frequentist and Bayesian methods that are robust to multicollinearity, performing shrinkage, variable selection, dimension reduction, statistical learning, or smoothing, including methods that were not used by the studies included in our review.

CONCLUSIONS

Compelling motivation exists for analyzing EDCs as mixtures, yet many studies make simplifying assumptions about EDC additivity, relative potency, and linearity, or overlook the potential for bias due to asymmetries in chemical persistence. We discuss the potential impacts of these choices and suggest alternative methods to improve analyses of prenatal exposure to EDC mixtures. https://doi.org/10.1289/EHP2207.

Penalized spline smoothing using Kaplan-Meier weights with censored data

In this paper, we consider the problem of nonparametric curve fitting in the specific context of censored data. We propose an extension of the penalized splines approach using Kaplan-Meier weights to take into account the effect of censorship and generalized cross-validation techniques to choose the smoothing parameter adapted to the case of censored samples. Using various simulation studies, we analyze the effectiveness of the censored penalized splines method proposed and show that the performance is quite satisfactory. We have extended this proposal to a generalized additive models (GAM) framework introducing a correction of the censorship effect, thus enabling more complex models to be estimated immediately. A real dataset from Stanford Heart Transplant data is also used to illustrate the methodology proposed, which is shown to be a good alternative when the probability distribution for the response variable and the functional form are not known in censored regression models.

Detection of prostate cancer with multiparametric MRI utilizing the anatomic structure of the prostate

Multiparametric magnetic resonance imaging (mpMRI), which combines traditional anatomic and newer quantitative MRI methods, has been shown to result in improved voxel-wise classification of prostate cancer as compared with any single MRI parameter. While these results are promising, substantial heterogeneity in the mpMRI parameter values and voxel-wise prostate cancer risk has been observed both between and within regions of the prostate. This suggests that classification of prostate cancer can potentially be improved by incorporating structural information into the classifier. In this paper, we propose a novel voxel-wise classifier of prostate cancer that accounts for the anatomic structure of the prostate by Bayesian hierarchical modeling, which can be combined with post hoc spatial Gaussian kernel smoothing to account for residual spatial correlation. Our proposed classifier results in significantly improved area under the ROC curve (0.822 vs 0.729, P < .001) and sensitivity corresponding to 90% specificity (0.599 vs 0.429, P < .001), compared with a baseline model that does not account for the anatomic structure of the prostate. Furthermore, the classifier can also be applied on voxels with missing mpMRI parameters, resulting in similar performance, which is an important practical consideration that cannot be easily accommodated using regression-based classifiers. In addition, our classifier achieved high computational efficiency with a closed-form solution for the posterior predictive cancer probability.

Semiparametric Estimation of Longitudinal Medical Cost Trajectory

Estimating the average monthly medical costs from disease diagnosis to a terminal event such as death for an incident cohort of patients is a topic of immense interest to researchers in health policy and health economics because patterns of average monthly costs over time reveal how medical costs vary across phases of care. The statistical challenges to estimating monthly medical costs longitudinally are multifold; the longitudinal cost trajectory (formed by plotting the average monthly costs from diagnosis to the terminal event) is likely to be nonlinear, with its shape depending on the time of the terminal event, which can be subject to right censoring. The goal of this paper is to tackle this statistically challenging topic by estimating the conditional mean cost at any month t given the time of the terminal event s. The longitudinal cost trajectories with different terminal event times form a bivariate surface of t and s, under the constraint t ≤ s. We propose to estimate this surface using bivariate penalized splines in an Expectation-Maximization algorithm that treats the censored terminal event times as missing data. We evaluate the proposed model and estimation method in simulations and apply the method to the medical cost data of an incident cohort of stage IV breast cancer patients from the Surveillance, Epidemiology and End Results-Medicare Linked Database.

Spatial statistics of hydrography and water chemistry in a eutrophic boreal lake based on sounding and water samples

Spatial variability, an essential characteristic of lake ecosystems, has often been neglected in field research and monitoring. In this study, we apply spatial statistical methods for the key physics and chemistry variables and chlorophyll a over eight sampling dates in two consecutive years in a large (area 103 km²) eutrophic boreal lake in southern Finland. In the four summer sampling dates, the water body was vertically and horizontally heterogenic except with color and DOC, in the two winter ice-covered dates DO was vertically stratified, while in the two autumn dates, no significant spatial differences in any of the measured variables were found. Chlorophyll a concentration was one order of magnitude lower under the ice cover than in open water. The Moran statistic for spatial correlation was significant for chlorophyll a and NO₂+NO₃-N in all summer situations and for dissolved oxygen and pH in three cases. In summer, the mass centers of the chemicals were within 1.5 km from the geometric center of the lake, and the 2nd moment radius ranged in 3.7-4.1 km respective to 3.9 km for the homogeneous situation. The lateral length scales of the studied variables were 1.5-2.5 km, about 1 km longer in the surface layer. The detected spatial "noise" strongly suggests that besides vertical variation also the horizontal variation in eutrophic lakes, in particular, should be considered when the ecosystems are monitored.

Infant body mass index peak and early childhood cardio-metabolic risk markers in a multi-ethnic Asian birth cohort

Background

: Infant body mass index (BMI) peak has received much interest recently as a potential predictor of future obesity and metabolic risk. No studies, however, have examined infant BMI peak in Asian populations, in whom the risk of metabolic disease is higher.

Methods

: We utilized data among 1020 infants from a mother-offspring cohort, who were Singapore citizens or permanent residents of Chinese, Malay or Indian ethnicity with homogeneous parental ethnic backgrounds, and did not receive chemotherapy, psychotropic drugs or have diabetes mellitus. Ethnicity was self-reported at recruitment and later confirmed using genotype analysis. Subject-specific BMI curves were fitted to infant BMI data using natural cubic splines with random coefficients to account for repeated measures in each child. We estimated characteristics of the child's BMI peak [age and magnitude at peak, average pre-peak velocity (aPPV)]. Systolic (SBP) and diastolic blood pressure (DBP), BMI, sum of skinfolds (SSF) and fat-mass index (FMI) were measured during a follow-up visit at age 48 months. Weighted multivariable linear regression was used to assess the predictors (maternal BMI, gestational weight gain, ethnicity, infant sex, gestational age, birthweight-for-gestational age and breastfeeding duration) of infant BMI peak and its associations with outcomes at 48 months. Comparisons between ethnicities were tested using Bonferroni post-hoc correction.

Results

: Of 1020 infants, 80.5% were followed up at the 48-month visit. Mean (SD) BMI, SSF and FMI at 48 months were 15.6 (1.8) kg/m 2 , 16.5 (5.3) mm and 3.8 (1.3) kg/m 2 , respectively. Mean (SD) age at peak BMI was 6.0 (1.6) months, with a magnitude of 17.2 (1.4) kg/m 2 and pre-peak velocity of 0.7 (0.3) kg/m 2 /month. Compared with Chinese infants, the peak occurred later in Malay {B [95% confidence interval (CI): 0.64 mo (0.36, 0.92)]} and Indian infants [1.11 mo (0.76, 1.46)] and was lower in magnitude in Indian infants [-0.45 kg/m 2 (-0.69, -0.20)]. Adjusting for maternal education, BMI, gestational weight gain, ethnicity, infant sex, gestational age, birthweight-for-gestational-age and breastfeeding duration, higher peak and aPPV were associated with greater BMI, SSF and FMI at 48 months. Age at peak was positively associated with BMI at 48 months [0.15 units (0.09, 0.22)], whereas peak magnitude was associated with SBP [0.17 units (0.05, 0.30)] and DBP at 48 months [0.10 units (0.01, 0.22)]. Older age and higher magnitude at peak were associated with increased risk of overweight at 48 months [Relative Risk (95% CI): 1.35 (1.12-1.62) for age; 1.89 (1.60-2.24) for magnitude]. The associations of BMI peak with BMI and SSF at 48 months were stronger in Malay and Indian children than in Chinese children.

Conclusions

: Ethnic-specific differences in BMI peak characteristics, and associations of BMI peak with early childhood cardio-metabolic markers, suggest an important impact of early BMI development on later metabolic outcomes in Asian populations.

Increased motility of mesenchymal stem cells is correlated with inhibition of stimulated peripheral blood mononuclear cells in vitro

Immunomodulatory properties of mesenchymal stem cells (MSC) are key components of their successful applications in clinical setting. However, treatments based on MSC immunomodulation need understanding of cell characteristics before cell transplantation. We used live-imaging to test the suitability of the MSC motility as a parameter for quick prediction of the immunomodulatory potential of human MSC in regulating the activity of stimulated peripheral blood mononuclear cells (PBMC) in vitro. Bone marrow MSC, from various donors and in vitro passages, were cultured with or without stimulated PBMC. After seven days, immunomodulation was assessed by measuring PBMC proliferation, IgG production and cytokine secretion in MSC and PBMC monocultures and co-cultures, and results were correlated to MSC motility. In co-culture, we observed that MSC successfully inhibited PBMC activity, reducing PBMC proliferation and IgG production compared to PBMC monoculture. MSC modulated PBMC to reduce the secretion of TNFα and IL-10, increase IL-6, G-CSF and MCP-1, while GM-CSF was not affected. By live-imaging tracking of cell trajectories, we observed that fast moving MSC were inhibiting more efficiently stimulated PBMC compared to slow ones. In co-culture, fast MSC were more effective in inhibiting IgG production (˜30% less IgG), and secreted higher levels of IL-10 (˜10% increase) and GM-CSF (˜20% increase) compared to slower cells. Furthermore, fast MSC in monocultures produced 2.3-fold more IL-6, 1.5-fold MCP-1 and 1.2-fold G-CSF in comparison to slower cells. In conclusion, live-imaging cell tracking allowed us to develop an indicative assay of the immune-regulatory potential of MSC prior to in vivo administration. Key Words: Human mesenchymal stem cells, Immunomodulatory potential, In vitro cell motility, Stem cell transplantation.

Associations of maternal macronutrient intake during pregnancy with infant BMI peak characteristics and childhood BMI

Background: Infant body mass index (BMI) peak characteristics and early childhood BMI are emerging markers of future obesity and cardiometabolic disease risk, but little is known about their maternal nutritional determinants.Objective: We investigated the associations of maternal macronutrient intake with infant BMI peak characteristics and childhood BMI in the Growing Up in Singapore Towards healthy Outcomes study.Design: With the use of infant BMI data from birth to age 18 mo, infant BMI peak characteristics [age (in months) and magnitude (BMI_peak; in kg/m²) at peak and prepeak velocities] were derived from subject-specific BMI curves that were fitted with the use of mixed-effects model with a natural cubic spline function. Associations of maternal macronutrient intake (assessed by using a 24-h recall during late gestation) with infant BMI peak characteristics (n = 910) and BMI z scores at ages 2, 3, and 4 y were examined with the use of multivariable linear regression.Results: Mean absolute maternal macronutrient intakes (percentages of energy) were 72 g protein (15.6%), 69 g fat (32.6%), and 238 g carbohydrate (51.8%). A 25-g (∼100-kcal) increase in maternal carbohydrate intake was associated with a 0.01/mo (95% CI: 0.0003, 0.01/mo) higher prepeak velocity and a 0.04 (95% CI: 0.01, 0.08) higher BMI_peak These associations were mainly driven by sugar intake, whereby a 25-g increment of maternal sugar intake was associated with a 0.02/mo (95% CI: 0.01, 0.03/mo) higher infant prepeak velocity and a 0.07 (95% CI: 0.01, 0.13) higher BMI_peak Higher maternal carbohydrate and sugar intakes were associated with a higher offspring BMI z score at ages 2-4 y. Maternal protein and fat intakes were not consistently associated with the studied outcomes.Conclusion: Higher maternal carbohydrate and sugar intakes are associated with unfavorable infancy BMI peak characteristics and higher early childhood BMI. This trial was registered at clinicaltrials.gov as NCT01174875.

Spatial regression with covariate measurement error: A semiparametric approach

Spatial data have become increasingly common in epidemiology and public health research thanks to advances in GIS (Geographic Information Systems) technology. In health research, for example, it is common for epidemiologists to incorporate geographically indexed data into their studies. In practice, however, the spatially defined covariates are often measured with error. Naive estimators of regression coefficients are attenuated if measurement error is ignored. Moreover, the classical measurement error theory is inapplicable in the context of spatial modeling because of the presence of spatial correlation among the observations. We propose a semiparametric regression approach to obtain bias-corrected estimates of regression parameters and derive their large sample properties. We evaluate the performance of the proposed method through simulation studies and illustrate using data on Ischemic Heart Disease (IHD). Both simulation and practical application demonstrate that the proposed method can be effective in practice.

Dynamic contrast-enhanced MRI detects acute radiotherapy-induced alterations in mandibular microvasculature: prospective assessment of imaging biomarkers of normal tissue injury

Normal tissue toxicity is an important consideration in the continued development of more effective external beam radiotherapy (EBRT) regimens for head and neck tumors. The ability to detect EBRT-induced changes in mandibular bone vascularity represents a crucial step in decreasing potential toxicity. To date, no imaging modality has been shown to detect changes in bone vascularity in real time during treatment. Based on our institutional experience with multi-parametric MRI, we hypothesized that DCE-MRI can provide in-treatment information regarding EBRT-induced changes in mandibular vascularity. Thirty-two patients undergoing EBRT treatment for head and neck cancer were prospectively imaged prior to, mid-course, and following treatment. DCE-MRI scans were co-registered to dosimetric maps to correlate EBRT dose and change in mandibular bone vascularity as measured by K_trans and V_e. DCE-MRI was able to detect dose-dependent changes in both K_trans and V_e in a subset of patients. One patient who developed ORN during the study period demonstrated decreases in K_trans and V_e following treatment completion. We demonstrate, in a prospective imaging trial, that DCE-MRI can detect dose-dependent alterations in mandibular bone vascularity during chemoradiotherapy, providing biomarkers that are physiological correlates of acute of acute mandibular vascular injury and recovery temporal kinetics.

A Randomized Dose-Escalation Study of the Safety and Anti-Inflammatory Activity of the p38 Mitogen-Activated Protein Kinase Inhibitor Dilmapimod in Severe Trauma Subjects at Risk for Acute Respiratory Distress Syndrome

OBJECTIVES

There are no current pharmacological therapies for the prevention or treatment of acute respiratory distress syndrome. Early dysregulated inflammation likely plays a role in acute respiratory distress syndrome development and possibly acute respiratory distress syndrome outcomes. p38 mitogen-activated protein kinase is central to the regulation of multiple inflammatory mediators implicated in acute organ dysfunction and is the target for a novel class of cytokine-suppressive anti-inflammatory drugs. In preclinical models, p38 inhibitors reduce lung injury following pancreatitis and burn injury.

DESIGN

We conducted a phase IIa, randomized, double-blind, placebo-controlled, parallel-group study to evaluate the safety and tolerability of dilmapimod, a novel p38 mitogen-activated protein kinase inhibitor, in patients at risk for developing acute respiratory distress syndrome admitted with an Injury Severity Score more than 16, excluding head trauma. Enrolled patients received 4- or 24-hour IV dilmapimod infusions at different doses or placebo, daily for 3 days, in four separate cohorts.

SETTING

Multicenter randomized clinical trial of large, academic trauma centers.

MEASUREMENTS AND MAIN RESULTS

Seventy-seven patients were enrolled. Although adverse events were common in this critically ill population, dilmapimod was well tolerated, with no clinically relevant safety findings. Pharmacokinetic models indicated that the higher dose of 10 mg given as continuous infusion over 24 hours had the most favorable plasma concentration profile. Likewise, measures of soluble inflammatory markers including interleukin-6, C-reactive peptide, interleukin-8, and soluble tumor necrosis factor receptor 1 were most different between this dosing arm and placebo. Although the study was not specifically designed with acute respiratory distress syndrome as an outcome, the number of patients who developed acute respiratory distress syndrome was small (2/77).

CONCLUSIONS

The novel p38 mitogen-activated protein kinase inhibitor dilmapimod appears well tolerated and may merit further evaluation for prevention of acute respiratory distress syndrome and other organ injury in larger clinical trials. Furthermore, results of this early-phase trial may aid in design of future studies aimed at prevention of acute respiratory distress syndrome and other organ injury.

Analysis of time to event outcomes in randomized controlled trials by generalized additive models

BACKGROUND

Randomized Controlled Trials almost invariably utilize the hazard ratio calculated with a Cox proportional hazard model as a treatment efficacy measure. Despite the widespread adoption of HRs, these provide a limited understanding of the treatment effect and may even provide a biased estimate when the assumption of proportional hazards in the Cox model is not verified by the trial data. Additional treatment effect measures on the survival probability or the time scale may be used to supplement HRs but a framework for the simultaneous generation of these measures is lacking.

METHODS

By splitting follow-up time at the nodes of a Gauss Lobatto numerical quadrature rule, techniques for Poisson Generalized Additive Models (PGAM) can be adopted for flexible hazard modeling. Straightforward simulation post-estimation transforms PGAM estimates for the log hazard into estimates of the survival function. These in turn were used to calculate relative and absolute risks or even differences in restricted mean survival time between treatment arms. We illustrate our approach with extensive simulations and in two trials: IPASS (in which the proportionality of hazards was violated) and HEMO a long duration study conducted under evolving standards of care on a heterogeneous patient population.

FINDINGS

PGAM can generate estimates of the survival function and the hazard ratio that are essentially identical to those obtained by Kaplan Meier curve analysis and the Cox model. PGAMs can simultaneously provide multiple measures of treatment efficacy after a single data pass. Furthermore, supported unadjusted (overall treatment effect) but also subgroup and adjusted analyses, while incorporating multiple time scales and accounting for non-proportional hazards in survival data.

CONCLUSIONS

By augmenting the HR conventionally reported, PGAMs have the potential to support the inferential goals of multiple stakeholders involved in the evaluation and appraisal of clinical trial results under proportional and non-proportional hazards.