Improving generalized regression analysis for the spatial prediction of forest communities

Aim This study used data from temperate forest communities to assess: (1) five different stepwise selection methods with generalized additive models, (2) the effect of weighting absences to ensure a prevalence of 0.5, (3) the effect of limiting absences beyond the environmental envelope defined by presences, (4) four different methods for incorporating spatial autocorrelation, and (5) the effect of integrating an interaction factor defined by a regression tree on the residuals of an initial environmental model. Location State of Vaud, western Switzerland. Methods Generalized additive models (GAMs) were fitted using the grasp package (generalized regression analysis and spatial predictions, http://www.cscf.ch/grasp). Results Model selection based on cross-validation appeared to be the best compromise between model stability and performance (parsimony) among the five methods tested. Weighting absences returned models that perform better than models fitted with the original sample prevalence. This appeared to be mainly due to the impact of very low prevalence values on evaluation statistics. Removing zeroes beyond the range of presences on main environmental gradients changed the set of selected predictors, and potentially their response curve shape. Moreover, removing zeroes slightly improved model performance and stability when compared with the baseline model on the same data set. Incorporating a spatial trend predictor improved model performance and stability significantly. Even better models were obtained when including local spatial autocorrelation. A novel approach to include interactions proved to be an efficient way to account for interactions between all predictors at once. Main conclusions Models and spatial predictions of 18 forest communities were significantly improved by using either: (1) cross-validation as a model selection method, (2) weighted absences, (3) limited absences, (4) predictors accounting for spatial autocorrelation, or (5) a factor variable accounting for interactions between all predictors. The final choice of model strategy should depend on the nature of the available data and the specific study aims. Statistical evaluation is useful in searching for the best modelling practice. However, one should not neglect to consider the shapes and interpretability of response curves, as well as the resulting spatial predictions in the final assessment.

Solving higher-dimensional continuous time stochastic control problems by value function regression

The paper develops a method to solve higher-dimensional stochasticcontrol problems in continuous time. A finite difference typeapproximation scheme is used on a coarse grid of low discrepancypoints, while the value function at intermediate points is obtainedby regression. The stability properties of the method are discussed,and applications are given to test problems of up to 10 dimensions.Accurate solutions to these problems can be obtained on a personalcomputer.

Weighted Kernel regression

In the fixed design regression model, additional weights areconsidered for the Nadaraya--Watson and Gasser--M\"uller kernel estimators.We study their asymptotic behavior and the relationships between new andclassical estimators. For a simple family of weights, and considering theIMSE as global loss criterion, we show some possible theoretical advantages.An empirical study illustrates the performance of the weighted estimatorsin finite samples.

How wide is the gap? An investigation of gender wage differences using quantile regression

In this paper we examine the determinants of wages and decompose theobserved differences across genders into the "explained by differentcharacteristics" and "explained by different returns components"using a sample of Spanish workers. Apart from the conditionalexpectation of wages, we estimate the conditional quantile functionsfor men and women and find that both the absolute wage gap and thepart attributed to different returns at each of the quantiles, farfrom being well represented by their counterparts at the mean, aregreater as we move up in the wage range.

Label-free cytotoxicity screening assay by digital holographic microscopy

Abstract We introduce a label-free technology based on digital holographic microscopy (DHM) with applicability for screening by imaging, and we demonstrate its capability for cytotoxicity assessment using mammalian living cells. For this first high content screening compatible application, we automatized a digital holographic microscope for image acquisition of cells using commercially available 96-well plates. Data generated through both label-free DHM imaging and fluorescence-based methods were in good agreement for cell viability identification and a Z'-factor close to 0.9 was determined, validating the robustness of DHM assay for phenotypic screening. Further, an excellent correlation was obtained between experimental cytotoxicity dose-response curves and known IC values for different toxic compounds. For comparable results, DHM has the major advantages of being label free and close to an order of magnitude faster than automated standard fluorescence microscopy.

Validation procedures in radiological diagnostic models. Neural network and logistic regression

The objective of this paper is to compare the performance of twopredictive radiological models, logistic regression (LR) and neural network (NN), with five different resampling methods. One hundred and sixty-seven patients with proven calvarial lesions as the only known disease were enrolled. Clinical and CT data were used for LR and NN models. Both models were developed with cross validation, leave-one-out and three different bootstrap algorithms. The final results of each model were compared with error rate and the area under receiver operating characteristic curves (Az). The neural network obtained statistically higher Az than LR with cross validation. The remaining resampling validation methods did not reveal statistically significant differences between LR and NN rules. The neural network classifier performs better than the one based on logistic regression. This advantage is well detected by three-fold cross-validation, but remains unnoticed when leave-one-out or bootstrap algorithms are used.

Fixed and random effects in Classical and Bayesian regression

This paper proposes a common and tractable framework for analyzingdifferent definitions of fixed and random effects in a contant-slopevariable-intercept model. It is shown that, regardless of whethereffects (i) are treated as parameters or as an error term, (ii) areestimated in different stages of a hierarchical model, or whether (iii)correlation between effects and regressors is allowed, when the sameinformation on effects is introduced into all estimation methods, theresulting slope estimator is also the same across methods. If differentmethods produce different results, it is ultimately because differentinformation is being used for each methods.

The importance of individual heterogeneity in the decomposition of measures of socioeconomic inequality in health: An approach based on quantile regression

This paper shows how recently developed regression-based methods for thedecomposition of health inequality can be extended to incorporateindividual heterogeneity in the responses of health to the explanatoryvariables. We illustrate our method with an application to the CanadianNPHS of 1994. Our strategy for the estimation of heterogeneous responsesis based on the quantile regression model. The results suggest that thereis an important degree of heterogeneity in the association of health toexplanatory variables which, in turn, accounts for a substantial percentageof inequality in observed health. A particularly interesting finding isthat the marginal response of health to income is zero for healthyindividuals but positive and significant for unhealthy individuals. Theheterogeneity in the income response reduces both overall health inequalityand income related health inequality.

Random measurement error and regression dilution bias.

Summary points: - The bias introduced by random measurement error will be different depending on whether the error is in an exposure variable (risk factor) or outcome variable (disease) - Random measurement error in an exposure variable will bias the estimates of regression slope coefficients towards the null - Random measurement error in an outcome variable will instead increase the standard error of the estimates and widen the corresponding confidence intervals, making results less likely to be statistically significant - Increasing sample size will help minimise the impact of measurement error in an outcome variable but will only make estimates more precisely wrong when the error is in an exposure variable

Intralesional adenovirus-mediated interleukin-2 gene transfer for advanced solid cancers and melanoma.

Numerous preclinical and clinical studies have shown that interleukin-2 (IL-2) induces regression of metastatic tumors. We have conducted a phase I/II, multicenter, open-label, dose-escalating study to evaluate the safety, efficacy, and biological effects of repeated intratumoral injections of adenovirus-IL-2 (TG1024) in patients with advanced solid tumors and melanoma. Thirty five patients (twenty-five with metastatic melanoma and ten with other solid tumors) were treated in eight successive cohorts at dose levels ranging from 3 x 10(8) to 3 x 10(11) viral particles (vp). Intratumoral TG1024 injections in combination with dacarbazine (DTIC) were tested in metastatic melanoma in one cohort. No clinical responses were observed at doses below 3 x 10(11) vp. Six local objective responses were recorded in patients receiving 3 x 10(11) vp per treatment [five in metastatic melanoma and one in metastatic squamous cell carcinoma (SCC) of the skin], of which two were complete responses (CRs). Most of the common side effects were injection site reactions and flu-like syndrome. TG1024 dose intensification across cohorts resulted in increased serum IL-2 levels after the injection. Intratumoral TG1024 injection induced pronounced inflammation of the treated lesion, with predominant CD8(+), TIA+ lymphocytic infiltrate. Our results show that intratumoral injections of TG1024 are safe and well tolerated. The clinical activity of TG1024 observed in this study warrants further investigations.

Panitumumab plus radiotherapy versus chemoradiotherapy in patients with unresected, locally advanced squamous-cell carcinoma of the head and neck (CONCERT-2): a randomised, controlled, open-label phase 2 trial.

BACKGROUND: We aimed to compare panitumumab, a fully human monoclonal antibody against EGFR, plus radiotherapy with chemoradiotherapy in patients with unresected, locally advanced squamous-cell carcinoma of the head and neck. METHODS: In this international, open-label, randomised, controlled, phase 2 trial, we recruited patients with locally advanced squamous-cell carcinoma of the head and neck from 22 sites in eight countries worldwide. Patients aged 18 years and older with stage III, IVa, or IVb, previously untreated, measurable (≥10 mm for at least one dimension), locally advanced squamous-cell carcinoma of the head and neck (non-nasopharygeal) and an Eastern Cooperative Oncology Group performance status of 0-1 were randomly assigned (2:3) by an independent vendor to open-label chemoradiotherapy (two cycles of cisplatin 100 mg/m(2) during radiotherapy) or to radiotherapy plus panitumumab (three cycles of panitumumab 9 mg/kg every 3 weeks administered with radiotherapy) using a stratified randomisation with a block size of five. All patients received 70-72 Gy to gross tumour and 54 Gy to areas of subclinical disease with accelerated fractionation radiotherapy. The primary endpoint was local-regional control at 2 years, analysed in all randomly assigned patients who received at least one dose of their assigned protocol-specific treatment (chemotherapy, radiation, or panitumumab). The trial is closed and this is the final analysis. This study is registered with ClinicalTrials.gov, number NCT00547157. FINDINGS: Between Nov 30, 2007, and Nov 16, 2009, 152 patients were enrolled, and 151 received treatment (61 in the chemoradiotherapy group and 90 in the radiotherapy plus panitumumab group). Local-regional control at 2 years was 61% (95% CI 47-72) in the chemoradiotherapy group and 51% (40-62) in the radiotherapy plus panitumumab group. The most frequent grade 3-4 adverse events were mucosal inflammation (25 [40%] of 62 patients in the chemoradiotherapy group vs 37 [42%] of 89 patients in the radiotherapy plus panitumumab group), dysphagia (20 [32%] vs 36 [40%]), and radiation skin injury (seven [11%] vs 21 [24%]). Serious adverse events were reported in 25 (40%) of 62 patients in the chemoradiotherapy group and in 30 (34%) of 89 patients in the radiotherapy plus panitumumab group. INTERPRETATION: Panitumumab cannot replace cisplatin in the combined treatment with radiotherapy for unresected stage III-IVb squamous-cell carcinoma of the head and neck, and the role of EGFR inhibition in locally advanced squamous-cell carcinoma of the head and neck needs to be reassessed. FUNDING: Amgen.

Comparative effects of teriparatide and ibandronate on spine bone mineral density (BMD) and microarchitecture (TBS) in postmenopausal women with osteoporosis: a 2-year open-label study.

Treatment effects over 2 years of teriparatide vs. ibandronate in postmenopausal women with osteoporosis were compared using lumbar spine bone mineral density (BMD) and trabecular bone score (TBS). Teriparatide induced larger increases in BMD and TBS compared to ibandronate, suggesting a more pronounced effect on bone microarchitecture of the bone anabolic drug. INTRODUCTION: The trabecular bone score (TBS) is an index of bone microarchitecture, independent of bone mineral density (BMD), calculated from anteroposterior spine dual X-ray absorptiometry (DXA) scans. The potential role of TBS for monitoring treatment response with bone-active substances is not established. The aim of this study was to compare the effects of recombinant human 1-34 parathyroid hormone (teriparatide) and the bisphosphonate ibandronate (IBN), on lumbar spine (LS) BMD and TBS in postmenopausal women with osteoporosis. METHODS: Two patient groups with matched age, body mass index (BMI), and baseline LS BMD, treated with either daily subcutaneous teriparatide (N = 65) or quarterly intravenous IBN (N = 122) during 2 years and with available LS BMD measurements at baseline and 2 years after treatment initiation were compared. RESULTS: Baseline characteristics (overall mean ± SD) were similar between groups in terms of age 67.9 ± 7.4 years, body mass index 23.8 ± 3.8 kg/m(2), BMD L1-L4 0.741 ± 0.100 g/cm(2), and TBS 1.208 ± 0.100. Over 24 months, teriparatide induced a significantly larger increase in LS BMD and TBS than IBN (+7.6 % ± 6.3 vs. +2.9 % ± 3.3 and +4.3 % ± 6.6 vs. +0.3 % ± 4.1, respectively; P < 0.0001 for both). LS BMD and TBS were only weakly correlated at baseline (r (2) = 0.04) with no correlation between the changes in BMD and TBS over 24 months. CONCLUSIONS: In postmenopausal women with osteoporosis, a 2-year treatment with teriparatide led to a significantly larger increase in LS BMD and TBS than IBN, suggesting that teriparatide had more pronounced effects on bone microarchitecture than IBN.

Concomitant chemoradiotherapy versus acceleration of radiotherapy with or without concomitant chemotherapy in locally advanced head and neck carcinoma (GORTEC 99-02): an open-label phase 3 randomised trial.

BACKGROUND: Concomitant chemoradiotherapy and accelerated radiotherapy independently improve outcomes for patients with locally advanced head and neck squamous-cell carcinoma (HNSCC). We aimed to assess the efficacy and safety of a combination of these approaches. METHODS: In our open-label phase 3 randomised trial, we enrolled patients with locally advanced, stage III and IV (non-metastatic) HNSCC and an Eastern Cooperative Oncology Group performance status of 0-2. We randomly allocated patients centrally with a computer program (with centre, T stage, N stage, and localisation as minimisation factors) in a 1:1:1 ratio to receive conventional chemoradiotherapy (70 Gy in 7 weeks plus three cycles of 4 days' concomitant carboplatin-fluorouracil), accelerated radiotherapy-chemotherapy (70 Gy in 6 weeks plus two cycles of 5 days' concomitant carboplatin-fluorouracil), or very accelerated radiotherapy alone (64·8 Gy [1·8 Gy twice daily] in 3·5 weeks). The primary endpoint, progression-free survival (PFS), was assessed in all enrolled patients. This trial is completed. The trial is registered with ClinicalTrials.gov, number NCT00828386. FINDINGS: Between Feb 29, 2000, and May 9, 2007, we randomly allocated 279 patients to receive conventional chemoradiotherapy, 280 to accelerated radiotherapy-chemotherapy, and 281 to very accelerated radiotherapy. Median follow-up was 5·2 years (IQR 4·9-6·2); rates of chemotherapy and radiotherapy compliance were good in all groups. Accelerated radiotherapy-chemotherapy offered no PFS benefit compared with conventional chemoradiotherapy (HR 1·02, 95% CI 0·84-1·23; p=0·88) or very accelerated radiotherapy (0·83, 0·69-1·01; p=0·060); conventional chemoradiotherapy improved PFS compared with very accelerated radiotherapy (0·82, 0·67-0·99; p=0·041). 3-year PFS was 37·6% (95% CI 32·1-43·4) after conventional chemoradiotherapy, 34·1% (28·7-39·8) after accelerated radiotherapy-chemotherapy, and 32·2% (27·0-37·9) after very accelerated radiotherapy. More patients in the very accelerated radiotherapy group had RTOG grade 3-4 acute mucosal toxicity (226 [84%] of 268 patients) compared with accelerated radiotherapy-chemotherapy (205 [76%] of 271 patients) or conventional chemoradiotherapy (180 [69%] of 262; p=0·0001). 158 (60%) of 265 patients in the conventional chemoradiotherapy group, 176 (64%) of 276 patients in the accelerated radiotherapy-chemotherapy group, and 190 (70%) of 272 patients in the very accelerated radiotherapy group were intubated with feeding tubes during treatment (p=0·045). INTERPRETATION: Chemotherapy has a substantial treatment effect given concomitantly with radiotherapy and acceleration of radiotherapy cannot compensate for the absence of chemotherapy. We noted the most favourable outcomes for conventional chemoradiotherapy, suggesting that acceleration of radiotherapy is probably not beneficial in concomitant chemoradiotherapy schedules. FUNDING: French Ministry of Health.

Cognitive efficacy of quetiapine in early-onset first-episode psychosis: a 12-week open label trial.

Twenty-three adolescents with psychotic disorders, aged from 13 to 18 years, participated in a 12-week open label trial (17 adolescents completed the study) in order to examine the impact of quetiapine on clinical status and cognitive functions (encompassing processing speed, attention, short-term memory, long-term memory and executive function). An improvement in Clinical Global Impression and Positive and Negative Symptom Scale (P's ≤ 0.001) was observed. In addition, after controlling for amelioration of symptoms, a significant improvement was observed on one executive function (P = 0.044; Trail Making Part B). The remaining cognitive abilities showed stability. In addition, we observed an interaction between quetiapine doses (>300 mg/day or <300 mg/day) and time, where lower doses showed more improvement in verbal short-term memory (P = 0.048), inhibition abilities (P = 0.038) and positive symptoms (P = 0.020). The neuropsychological functioning of adolescents with psychotic disorders remained mainly stable after 12 weeks of treatment with quetiapine. However, lower doses seemed to have a better impact on two components of cognition (inhibition abilities and verbal short-term memory) and on positive symptoms.