Extremal properties of certain risk models

Cette thèse s'intéresse à étudier les propriétés extrémales de certains modèles de risque d'intérêt dans diverses applications de l'assurance, de la finance et des statistiques. Cette thèse se développe selon deux axes principaux, à savoir: Dans la première partie, nous nous concentrons sur deux modèles de risques univariés, c'est-à- dire, un modèle de risque de déflation et un modèle de risque de réassurance. Nous étudions le développement des queues de distribution sous certaines conditions des risques commun¬s. Les principaux résultats sont ainsi illustrés par des exemples typiques et des simulations numériques. Enfin, les résultats sont appliqués aux domaines des assurances, par exemple, les approximations de Value-at-Risk, d'espérance conditionnelle unilatérale etc. La deuxième partie de cette thèse est consacrée à trois modèles à deux variables: Le premier modèle concerne la censure à deux variables des événements extrême. Pour ce modèle, nous proposons tout d'abord une classe d'estimateurs pour les coefficients de dépendance et la probabilité des queues de distributions. Ces estimateurs sont flexibles en raison d'un paramètre de réglage. Leurs distributions asymptotiques sont obtenues sous certaines condi¬tions lentes bivariées de second ordre. Ensuite, nous donnons quelques exemples et présentons une petite étude de simulations de Monte Carlo, suivie par une application sur un ensemble de données réelles d'assurance. L'objectif de notre deuxième modèle de risque à deux variables est l'étude de coefficients de dépendance des queues de distributions obliques et asymétriques à deux variables. Ces distri¬butions obliques et asymétriques sont largement utiles dans les applications statistiques. Elles sont générées principalement par le mélange moyenne-variance de lois normales et le mélange de lois normales asymétriques d'échelles, qui distinguent la structure de dépendance de queue comme indiqué par nos principaux résultats. Le troisième modèle de risque à deux variables concerne le rapprochement des maxima de séries triangulaires elliptiques obliques. Les résultats théoriques sont fondés sur certaines hypothèses concernant le périmètre aléatoire sous-jacent des queues de distributions. -- This thesis aims to investigate the extremal properties of certain risk models of interest in vari¬ous applications from insurance, finance and statistics. This thesis develops along two principal lines, namely: In the first part, we focus on two univariate risk models, i.e., deflated risk and reinsurance risk models. Therein we investigate their tail expansions under certain tail conditions of the common risks. Our main results are illustrated by some typical examples and numerical simu¬lations as well. Finally, the findings are formulated into some applications in insurance fields, for instance, the approximations of Value-at-Risk, conditional tail expectations etc. The second part of this thesis is devoted to the following three bivariate models: The first model is concerned with bivariate censoring of extreme events. For this model, we first propose a class of estimators for both tail dependence coefficient and tail probability. These estimators are flexible due to a tuning parameter and their asymptotic distributions are obtained under some second order bivariate slowly varying conditions of the model. Then, we give some examples and present a small Monte Carlo simulation study followed by an application on a real-data set from insurance. The objective of our second bivariate risk model is the investigation of tail dependence coefficient of bivariate skew slash distributions. Such skew slash distributions are extensively useful in statistical applications and they are generated mainly by normal mean-variance mixture and scaled skew-normal mixture, which distinguish the tail dependence structure as shown by our principle results. The third bivariate risk model is concerned with the approximation of the component-wise maxima of skew elliptical triangular arrays. The theoretical results are based on certain tail assumptions on the underlying random radius.

Beclin 1 mitigates motor and neuropathological deficits in genetic mouse models of Machado-Joseph disease.

Machado-Joseph disease or spinocerebellar ataxia type 3, the most common dominantly-inherited spinocerebellar ataxia, results from translation of the polyglutamine-expanded and aggregation prone ataxin 3 protein. Clinical manifestations include cerebellar ataxia and pyramidal signs and there is no therapy to delay disease progression. Beclin 1, an autophagy-related protein and essential gene for cell survival, is decreased in several neurodegenerative disorders. This study aimed at evaluating if lentiviral-mediated beclin 1 overexpression would rescue motor and neuropathological impairments when administered to pre- and post-symptomatic lentiviral-based and transgenic mouse models of Machado-Joseph disease. Beclin 1-mediated significant improvements in motor coordination, balance and gait with beclin 1-treated mice equilibrating longer periods in the Rotarod and presenting longer and narrower footprints. Furthermore, in agreement with the improvements observed in motor function beclin 1 overexpression prevented neuronal dysfunction and neurodegeneration, decreasing formation of polyglutamine-expanded aggregates, preserving Purkinje cell arborization and immunoreactivity for neuronal markers. These data show that overexpression of beclin 1 in the mouse cerebellum is able to rescue and hinder the progression of motor deficits when administered to pre- and post-symptomatic stages of the disease.

Validation of Logistic Regression Models for Landslide Susceptibility Maps

A wide range of numerical models and tools have been developed over the last decades to support the decision making process in environmental applications, ranging from physical models to a variety of statistically-based methods. In this study, a landslide susceptibility map of a part of Three Gorges Reservoir region of China was produced, employing binary logistic regression analyses. The available information includes the digital elevation model of the region, geological map and different GIS layers including land cover data obtained from satellite imagery. The landslides were observed and documented during the field studies. The validation analysis is exploited to investigate the quality of mapping.

A CMR study of the effects of tissue edema and necrosis on left ventricular dyssynchrony in acute myocardial infarction: implications for cardiac resynchronization therapy.

ABSTRACT: BACKGROUND: In acute myocardial infarction (AMI), both tissue necrosis and edema are present and both might be implicated in the development of intraventricular dyssynchrony. However, their relative contribution to transient dyssynchrony is not known. Cardiovascular magnetic resonance (CMR) can detect necrosis and edema with high spatial resolution and it can quantify dyssynchrony by tagging techniques. METHODS: Patients with a first AMI underwent percutaneous coronary interventions (PCI) of the infarct-related artery within 24 h of onset of chest pain. Within 5-7 days after the event and at 4 months, CMR was performed. The CMR protocol included the evaluation of intraventricular dyssynchrony by applying a novel 3D-tagging sequence to the left ventricle (LV) yielding the CURE index (circumferential uniformity ratio estimate; 1 = complete synchrony). On T2-weighted images, edema was measured as high-signal (>2 SD above remote tissue) along the LV mid-myocardial circumference on 3 short-axis images (% of circumference corresponding to the area-at-risk). In analogy, on late-gadolinium enhancement (LGE) images, necrosis was quantified manually as percentage of LV mid-myocardial circumference on 3 short-axis images. Necrosis was also quantified on LGE images covering the entire LV (expressed as %LV mass). Finally, salvaged myocardium was calculated as the area-at-risk minus necrosis (expressed as % of LV circumference). RESULTS: After successful PCI (n = 22, 2 female, mean age: 57 ± 12y), peak troponin T was 20 ± 36ug/l and the LV ejection fraction on CMR was 41 ± 8%. Necrosis mass was 30 ± 10% and CURE was 0.91 ± 0.05. Edema was measured as 58 ± 14% of the LV circumference. In the acute phase, the extent of edema correlated with dyssynchrony (r2 = -0.63, p < 0.01), while extent of necrosis showed borderline correlation (r2 = -0.19, p = 0.05). PCI resulted in salvaged myocardium of 27 ± 14%. LV dyssynchrony (=CURE) decreased at 4 months from 0.91 ± 0.05 to 0.94 ± 0.03 (p < 0.004, paired t-test). At 4 months, edema was absent and scar %LV slightly shrunk to 23.7 ± 10.0% (p < 0.002 vs baseline). Regression of LV dyssynchrony during the 4 months follow-up period was predicted by both, the extent of edema and its necrosis component in the acute phase. CONCLUSIONS: In the acute phase of infarction, LV dyssynchrony is closely related to the extent of edema, while necrosis is a poor predictor of acute LV dyssynchrony. Conversely, regression of intraventricular LV dyssynchrony during infarct healing is predicted by the extent of necrosis in the acute phase.

New genetic models and mesenchymal cells as tools to ameliorate anti-tumor immunity

Résumé Le but final de ce projet est d'utiliser des cellules T ou des cellules souches mésenchymateuses modifiées génétiquement afin de surexprimer localement les deux chémokines CXCL13 et CCL2 ensemble ou chacune séparément à l'intérieur d'une tumeur solide. CXCL13 est supposé induire des structures lymphoïdes ectopiques. Un niveau élevé de CCL2 est présumé initier une inflammation aiguë. La combinaison des deux effets amène à un nouveau modèle d'étude des mécanismes régulateur de la tolérance périphérique et de l'immunité tumorale. Les connaissances acquises grâce à ce modèle pourraient permettre le développement ou l'amélioration des thérapies immunes du cancer. Le but premier de ce travail a été l'établissement d'un modèle génétique de la souris permettant d'exprimer spécifiquement dans la tumeur les deux chémokines d'intérêt à des niveaux élevés. Pour accomplir cette tâche, qui est en fait une thérapie génétique de tumeurs solides, deux types de cellules porteuses potentielles ont été évaluées. Des cellules CD8+ T et des cellules mésenchymateuses de la moelle osseuse transférées dans des receveurs portant une tumeur. Si on pouvait répondre aux besoins de la thérapie génétique, indépendamment de la thérapie immune envisagée, on posséderait là un outil précieux pour bien d'autres approches thérapeutiques. Plusieurs lignées de souris transgéniques ont été générées comme source de cellules CD8+ T modifiées afin d'exprimer les chémokines d'intérêt. Dans une approche doublement transgénique les propriétés de deux promoteurs spécifiques de cellules T ont été combinées en utilisant la technologie Cre-loxP. Le promoteur de granzyme B confère une dépendance d'activation et le promoteur distal de lck assure une forte expression constitutive dès que les cellules CD8+ T ont été activées. Les transgènes construits ont montré une bonne performance in vivo et des souris qui expriment CCL2 dans des cellules CD8+ T activées ont été obtenues. Ces cellules peuvent maintenant être utilisées avec différents protocoles pour transférer des cellules T cytotoxiques (CTL) dans des receveurs porteur d'une tumeur, permettant ainsi d'évaluer leur capacité en tant que porteuse de chémokine d'infiltrer la tumeur. L'établissement de souris transgéniques, qui expriment pareillement CXCL13 est prévu dans un avenir proche. L'évaluation de cellules mésenchymateuses de la moelle osseuse a démontré que ces cellules se greffent efficacement dans le stroma tumoral suite à la co-injection avec des cellules tumorales. Cela représente un outil précieux pour la recherche, vu qu'il permet d'introduire des cellules manipulées dans un modèle tumoral. Les résultats confirment partiellement d'autres résultats rapportés dans un modèle amélioré. Cependant, l'efficacité et la spécificité suggérées de la migration systémique de cellules mésenchymateuses de la moelle osseuse dans une tumeur n'ont pas été observées dans notre modèle, ce qui indique, que ces cellules ne se prêtent pas à une utilisation thérapeutique. Un autre résultat majeur de ce travail est l'établissement de cultures de cellules mésenchymateuses de la moelle osseuse in vitro conditionnées par des tumeurs, ce qui a permis à ces cellules de s'étendre plus rapidement en gardant leur capacité de migration et de greffe. Cela offre un autre outil précieux, vu que la culture in vitro est un pas nécessaire pour une manipulation thérapeutique. Abstract The ultimate aim of the presented project is to use genetically modified T cells or mesenchymal stem cells to locally overexpress the two chemokines CXCL13 and CCL2 together or each one alone inside a solid tumor. CXCL13 is supposed to induce ectopic lymphoid structures and a high level of CCL2 is intended to trigger acute inflamation. The combination of these two effects represents a new model for studying mechanisms that regulate peripheral tolerance and tumor immunity. Gained insights may help developing or improving immunotherapy of cancer. The primary goal of the executed work was the establishment of a genetic mouse model that allows tumor-specific expression of high levels of the two chemokines of interest. For accomplishing this task, which represents gene therapy of solid tumors, two types of potentially useful carrier cells were evaluated. CD8+ T cells and mesenchymal bone marrow cells to be used in adoptive cell transfers into tumor-bearing mice. Irrespectively of the envisaged immunotherapy, satisfaction of so far unmet needs of gene therapy would be a highly valuable tool that may be employed by many other therapeutic approaches, too. Several transgenic mouse lines were generated as a source of CD8+ T cells modified to express the chemokines of interest. In a double transgenic approach the properties of two T cell-specific promoters were combined using Cre-loxP technology. The granzyme B promoter confers activation-dependency and the lck distal promoter assures strong constitutive expression once the CD8+ T cell has been activated. The constructed transgenes showed a good performance in vivo and mice expressing CCL2 in activated CD8+ T cells were obtained. These cells can now be used with different protocols for adoptively transferring cytotoxic T cells (CTL) into tumor-bearing recipients, thus allowing to study their capacity as tumor-infiltrating chemokine carrier. The establishment of transgenic mice likewisely expressing CXCL13 is expected in the near future. In addition, T cells from generated single transgenic mice that have high expression of an EGFP reporter in both CD4+ and CD8+ cells can be easily traced in vivo when setting up adoptive transfer conditions. The evaluation of mesenchymal bone marrow cells demonstrated that these cells can efficiently engraft into tumor stroma upon local coinjection with tumor cells. This represents a valuable tool for research purposes as it allows to introduce manipulated stromal cells into a tumor model. Therefore, the established engraftment model is suited for studying the envisaged immunotherapy. These results confirm to some extend previously reported results in an improved model, however, the suggested systemic tumor homing efficiency and specificity of mesenchymal bone marrow cells was not observed in our model indicating that these cells may not be suited for therapeutic use. Another major result of the presented work is the establishment oftumor-conditioned in vitro culture of mesenchymal bone marrow cells, which allowed to more rapidly expand these cells while maintaining their tumor homing and engrafting capacities. This offers another valuable tool as in vitro culture is a necessary step for therapeutic manipulations.

Linezolid alone or combined with rifampin against methicillin-resistant Staphylococcus aureus in experimental foreign-body infection.

We investigated the activity of linezolid, alone and in combination with rifampin (rifampicin), against a methicillin-resistant Staphylococcus aureus (MRSA) strain in vitro and in a guinea pig model of foreign-body infection. The MIC, minimal bactericidal concentration (MBC) in logarithmic phase, and MBC in stationary growth phase were 2.5, >20, and >20 microg/ml, respectively, for linezolid; 0.01, 0.08, and 2.5 microg/ml, respectively, for rifampin; and 0.16, 0.63, >20 microg/ml, respectively, for levofloxacin. In time-kill studies, bacterial regrowth and the development of rifampin resistance were observed after 24 h with rifampin alone at 1x or 4x the MIC and were prevented by the addition of linezolid. After the administration of single intraperitoneal doses of 25, 50, and 75 mg/kg of body weight, linezolid peak concentrations of 6.8, 12.7, and 18.1 microg/ml, respectively, were achieved in sterile cage fluid at approximately 3 h. The linezolid concentration remained above the MIC of the test organism for 12 h with all doses. Antimicrobial treatments of animals with cage implant infections were given twice daily for 4 days. Linezolid alone at 25, 50, and 75 mg/kg reduced the planktonic bacteria in cage fluid during treatment by 1.2 to 1.7 log(10) CFU/ml; only linezolid at 75 mg/kg prevented bacterial regrowth 5 days after the end of treatment. Linezolid used in combination with rifampin (12.5 mg/kg) was more effective than linezolid used as monotherapy, reducing the planktonic bacteria by >or=3 log(10) CFU (P < 0.05). Efficacy in the eradication of cage-associated infection was achieved only when linezolid was combined with rifampin, with cure rates being between 50% and 60%, whereas the levofloxacin-rifampin combination demonstrated the highest cure rate (91%) against the strain tested. The linezolid-rifampin combination is a treatment option for implant-associated infections caused by quinolone-resistant MRSA.

Inflammatory Markers and Incident Heart Failure Risk in Older Adults: The Health, Aging, and Body Composition Study

Background: Inflammation is associated with heart failure (HF) risk factors and also directly affects myocardial function. However, the association between inflammation and HF risk in older adults has not been adequately evaluated. Methods: The association of baseline serum concentrations of interleukin-6 (IL-6), tumor necrosis factor alpha (TNF- ), and C-reactive protein (CRP) with incident HF was assessed with Cox proportional hazards models among 2610 older persons without prevalent HF enrolled in the Health, Aging, and Body Composition (Health ABC) Study (age, 73.6±2.9 years; 48.3% men; 59.6% white). Results: Median (interquartile range) baseline concentrations of IL-6, TNF- , and CRP were 1.80 (1.23, 2.76) pg/mL, 3.14 (2.41, 4.06) pg/mL, and 1.64 (0.99, 3.04) µg/mL, respectively. On follow-up (median, 9.4 years), 311 participants (11.9%) developed HF. In models controlling for clinical predictors of HF and incident coronary heart disease, doubling of IL-6, TNF- , and CRP concentrations was associated with 34% (95% CI, 18 -52%; P<.001), 33% (95% CI, 9 - 63%; P=.006), and 13% (95% CI, 3-24%; P=.01) increase in HF risk, respectively. In models including all 3 markers, IL-6 and TNF- , but not CRP, remained significant. Findings were similar across sex and race. Post-HF ejection fraction (EF) was available in 239 (76.8%) cases. When only cases with preserved EF were considered (n=105), IL-6 (HR per doubling, 1.57; 95% CI, 1.28 -1.94; P<.001), TNF- (HR per doubling, 1.59; 95% CI, 1.12-2.26; P=.01), and CRP (HR per doubling, 1.23; 95% CI, 1.05-1.44; P=.01) were all associated with HF risk in adjusted models. In contrast, when only cases with reduced EF (n=134) were considered, only IL-6 attained marginal significance in adjusted models (HR per doubling, 1.20; 95% CI, 0.99 -1.46; P=.06). Participants with 2 or 3 markers above median had pronounced HF risk in adjusted models (HR, 1.66; 95% CI, 1.12-2.46; P=.01; and HR, 1.76; 95% CI, 1.16 -2.65; P=.007, respectively). Addition of IL-6 to the clinical Health ABC HF model improved discrimination (C index from 0.717 to 0.734; P=.001) and fit (decreased Bayes information criterion by 17.8; P<.001). Conclusions: Inflammatory markers are associated with HF risk among older adults and may improve HF risk stratification.

Reduced ejection fraction after myocardial infarction: is it sufficient to justify implantation of a defibrillator ?

Introduction : La diminution de la fraction d'éjection après infarctus du myocarde est-elle suffisante pour justifier l'implantation prophylactique de défibrillateur ? Introduction: La diminution de la fraction d'éjection (FE) est un facteur prédictif majeur de mortalité cardiaque chez les patients avec ancien infarctus du myocarde (IM). Les plus récentes études, pour la plupart conduites à la fin des années 90, montrent une diminution de la mortalité post-hospitalière à 10 à 20% par année. Une amélioration de la survie par implantation prophylactique de défibrillateurs a été démontrée chez les patients avec dysfonction myocardique avancée après ancien IM dans la cadre de l'étude MADIT II (Multicenter Automatic Defibrillator Implantation Trial) et est depuis lors, considérée comme une thérapie recommandée. Le bénéfice de l'implantation prophylactique de défibrillateurs en terme de nombre nécessaire de patients à traiter pour prévenir un évènement est d'un intérêt primordial en raison des complications potentielles liées à la procédure et l'appareillage, des implications psychologiques ainsi qu'en raison de l'analyse coût-efficacité. Le nombre de patients à traiter est corrélé au risque actuel de mortalité rythmique, probablement influencé par les récentes avancées thérapeutiques dans la prise en charge aigue de l'IM (notamment par angioplastie primaire) et de l'insuffisance cardiaque. Or, dans '' l'étude MADIT II, le recrutement des patients était effectué entre 1997 et 2001 et l'intervalle moyen entre l'IM et le recrutement était de plus de 6 ans. Le but de l'étude était donc d'évaluer, dans la pratique générale, la mortalité actuelle de patients ayant survécu à un IM avec diminution consécutive significative de la FE. La stratification du risque sur la base des variables hospitalières a de même été explorée. Méthode: Une analyse de cohorte de patients avec infarctus aigu du myocarde admis de 1999 à 2000 a été effectuée dans 2 centres hospitaliers (un universitaire et un cantonal). Tous les dossiers cliniques des patients sortis d'hôpital avec FE documentée ≤ 0.40 ont été inclus. Les caractéristiques des patients, les procédures diagnostiques et thérapeutiques ainsi que la médication étaient reportées. La période de suivi débutait à la sortie d'hôpital. Les end-points primaires étaient la mortalité globale, la mortalité cardiaque et les morts subites. Une analyse bivariée et multivariée a été effectuée. Résultats: 165 patients ont été inclus. Durant un suivi médian de 30 mois (interquartile range 22-36), 18 patients sont décédés (5 morts classifiées comme cardiaques, 8 comme subites, 5 de causes non cardiaques). Les taux de mortalité à un et deux ans selon analyse de Kaplan-Meier étaient de 6.7 et 8.6%, respectivement. Les variables reflétant l'atteinte coronarienne et sa prise en charge (antécédent d'IM, reperfusion aigue, revascularisation complète) étaient plus significativement associées à la mortalité comparées aux variables reflétant la dysfonction myocardique (FE, classe Killip à l'admission). Conclusion : La mortalité des patients ayant survécu à un IM avec dysfonction myocardique avancée a diminué de façon substantielle depuis la fin des années 90. La diminution de la mortalité rythmique implique une augmentation proportionnelle du nombre de patients à traiter par implantation de défibrillateur pour prévenir un évènement. Par ailleurs, une stratification du risque après IM basée uniquement sur la FE semble être inappropriée au vu de son impact mineur sur la mortalité comparé aux autres facteurs de risque identifiés dans notre étude. Les patients sans antécédents d'infarctus et ceux ayant bénéficié d'une revascularisation aigue ou complète avait une mortalité cumulée à 2 ans de, respectivement, 2.5%, 3.9% et 2.3%. Pour ce type de patients à bas risque le bénéfice de l'implantation prophylactique de défibrillateur doit être remis en question. Au vu de cette étude, une stratification additionnelle du risque avant implantation de défibrillateur est à considérer.

Modern statistical models for forensic fingerprint examinations : a critical review

Over the last decade, the development of statistical models in support of forensic fingerprint identification has been the subject of increasing research attention, spurned on recently by commentators who claim that the scientific basis for fingerprint identification has not been adequately demonstrated. Such models are increasingly seen as useful tools in support of the fingerprint identification process within or in addition to the ACE-V framework. This paper provides a critical review of recent statistical models from both a practical and theoretical perspective. This includes analysis of models of two different methodologies: Probability of Random Correspondence (PRC) models that focus on calculating probabilities of the occurrence of fingerprint configurations for a given population, and Likelihood Ratio (LR) models which use analysis of corresponding features of fingerprints to derive a likelihood value representing the evidential weighting for a potential source.

Using 3D surface datasets to understand landslide evolution: From analogue models to real case study

Early detection of landslide surface deformation with 3D remote sensing techniques, as TLS, has become a great challenge during last decade. To improve our understanding of landslide deformation, a series of analogue simulation have been carried out on non-rigid bodies coupled with 3D digitizer. All these experiments have been carried out under controlled conditions, as water level and slope angle inclination. We were able to follow 3D surface deformation suffered by complex landslide bodies from precursory deformation still larger failures. These experiments were the basis for the development of a new algorithm for the quantification of surface deformation using automatic tracking method on discrete points of the slope surface. To validate the algorithm, comparisons were made between manually obtained results and algorithm surface displacement results. Outputs will help in understanding 3D deformation during pre-failure stages and failure mechanisms, which are fundamental aspects for future implementation of 3D remote sensing techniques in early warning systems.

Do pseudo-absence selection strategies influence species distribution models and their predictions? An information-theoretic approach based on simulated data

Background Multiple logistic regression is precluded from many practical applications in ecology that aim to predict the geographic distributions of species because it requires absence data, which are rarely available or are unreliable. In order to use multiple logistic regression, many studies have simulated "pseudo-absences" through a number of strategies, but it is unknown how the choice of strategy influences models and their geographic predictions of species. In this paper we evaluate the effect of several prevailing pseudo-absence strategies on the predictions of the geographic distribution of a virtual species whose "true" distribution and relationship to three environmental predictors was predefined. We evaluated the effect of using a) real absences b) pseudo-absences selected randomly from the background and c) two-step approaches: pseudo-absences selected from low suitability areas predicted by either Ecological Niche Factor Analysis: (ENFA) or BIOCLIM. We compared how the choice of pseudo-absence strategy affected model fit, predictive power, and information-theoretic model selection results. Results Models built with true absences had the best predictive power, best discriminatory power, and the "true" model (the one that contained the correct predictors) was supported by the data according to AIC, as expected. Models based on random pseudo-absences had among the lowest fit, but yielded the second highest AUC value (0.97), and the "true" model was also supported by the data. Models based on two-step approaches had intermediate fit, the lowest predictive power, and the "true" model was not supported by the data. Conclusion If ecologists wish to build parsimonious GLM models that will allow them to make robust predictions, a reasonable approach is to use a large number of randomly selected pseudo-absences, and perform model selection based on an information theoretic approach. However, the resulting models can be expected to have limited fit.

Trabecular bone score (TBS) the new parameter of 2D texture analysis for the evaluation of 3D bone micro architecture status

X-ray is a technology that is used for numerous applications in the medical field. The process of X-ray projection gives a 2-dimension (2D) grey-level texture from a 3- dimension (3D) object. Until now no clear demonstration or correlation has positioned the 2D texture analysis as a valid indirect evaluation of the 3D microarchitecture. TBS is a new texture parameter based on the measure of the experimental variogram. TBS evaluates the variation between 2D image grey-levels. The aim of this study was to evaluate existing correlations between 3D bone microarchitecture parameters - evaluated from μCT reconstructions - and the TBS value, calculated on 2D projected images. 30 dried human cadaveric vertebrae were acquired on a micro-scanner (eXplorer Locus, GE) at isotropic resolution of 93 μm. 3D vertebral body models were used. The following 3D microarchitecture parameters were used: Bone volume fraction (BV/TV), Trabecular thickness (TbTh), trabecular space (TbSp), trabecular number (TbN) and connectivity density (ConnD). 3D/2D projections has been done by taking into account the Beer-Lambert Law at X-ray energy of 50, 100, 150 KeV. TBS was assessed on 2D projected images. Correlations between TBS and the 3D microarchitecture parameters were evaluated using a linear regression analysis. Paired T-test is used to assess the X-ray energy effects on TBS. Multiple linear regressions (backward) were used to evaluate relationships between TBS and 3D microarchitecture parameters using a bootstrap process. BV/TV of the sample ranged from 18.5 to 37.6% with an average value at 28.8%. Correlations' analysis showedthat TBSwere strongly correlatedwith ConnD(0.856≤r≤0.862; p<0.001),with TbN (0.805≤r≤0.810; p<0.001) and negatively with TbSp (−0.714≤r≤−0.726; p<0.001), regardless X-ray energy. Results show that lower TBS values are related to "degraded" microarchitecture, with low ConnD, low TbN and a high TbSp. The opposite is also true. X-ray energy has no effect onTBS neither on the correlations betweenTBS and the 3Dmicroarchitecture parameters. In this study, we demonstrated that TBS was significantly correlated with 3D microarchitecture parameters ConnD and TbN, and negatively with TbSp, no matter what X-ray energy has been used. This article is part of a Special Issue entitled ECTS 2011. Disclosure of interest: None declared.