RESUMO Objetivo: Investigar fatores associados à recidiva, retratamento por abandono e óbito entre retratados por abandono para pacientes com TB pulmonar, de ambos os sexos, maiores que treze anos, residentes em área periférica do município de São Paulo. Material e Métodos: Estudos de caso-controle não pareado aninhado a uma coorte prospectiva de pacientes com TB pulmonar confirmada por cultura, selecionados entre 2001 e 2002 e acompanhados até 2006. Os casos foram pacientes que apresentaram recidivas, retratamento por abandono e óbito entre retratados por abandono; os controles foram pacientes com cura sem retratamento. Os dados foram obtidos mediante aplicação de questionários estruturados aplicados à época do ingresso no estudo e por entrevista domiciliar em 2004 e 2006, complementados por informação da Vigilância de TB. Na investigação dos fatores associados à recidiva, ao retratamento por abandono e óbito entre retratados por abandono, esses três desfechos foram tomados como variáveis dependentes e como variáveis independentes, as exposições de interesse. As odds ratio (OR) brutas e ajustadas foram estimadas com os respectivos intervalos de 95% de confiança pela regressão logística multivariada não condicional. A importância das variáveis para o modelo final foi avaliada através do teste da razão de verossimilhança, utilizando-se p<0,05. Resultados: As variáveis associadas à recidiva e ajustadas para sexo e idade independentemente das demais foram: co-infecção com HIV (OR=9,3; IC95%= 1,6 - 54,1), contato domiciliar prévio (OR= 2,2; IC95% = 0,8 - 6,2), caso de TB no domicílio após o paciente (OR= 3,8; IC95%= 1,2 - 12,8); diabetes (OR=1,6; IC95%= 0,3 - 7,7), MDR (OR=15,5; IC95%= 1,2 - 200,5). As variáveis associadas ao retratamento por abandono ajustadas para sexo e idade independentes das demais variáveis foram: TBMDR (OR=38,7; IC95%= 2,9 - 515,3), co-infecção com HIV (OR=24,8; IC95%= 3,8 - 163,0), história de alcoolismo (OR= 4,2; IC95%=1,1 - 17,5) e internação por complicações de TB (OR= 7,2; IC95% =2,5 - 21,0). As variáveis associadas ao óbito entre retratados por abandono ajustadas para sexo e idade independente das demais e foram: TBMDR (OR=152,4; IC95%= 9,8 - 237,4), co-infecção com HIV (OR=29,0; IC95%=7,1 -1 19,1), alcoolismo (OR=11,8; IC95%=1,3-102,8) e regime prisional (OR=5,0; IC95%= 1,0-25,8). Conclusões: Os resultados apresentados apontam grupos de maior risco para retratamento por TB por recidiva, abandono de tratamento e óbito que devem ser considerados no aperfeiçoamento do DOTS em nosso país. / ABSTRACT Objective: Investigate factors associated to relapse, re-treatment by default and death among patients retreated by default for patients with pulmonary tuberculosis, of both sexes, more than thirteen years old, living in the periphery of the city of São Paulo. Material and Methods: A nested case-control study not paired to a prospective cohort of pulmonary tuberculosis patients confirmed by culture, selected between 2001 e 2002 and followed up until 2006. Cases were patients who had relapses, re-treatment by default and death among patients retreated by default; controls were patients with cure without re-treatment. The data were obtained through application of structured questionnaires applied at the time of entry in the study and interview at home in 2004 and 2006, supplemented by information from the surveillance of TB and Information System Mortality of the Foundation SEADE. Relapse, re-treatment by default and death among patients retreated by default were taken as dependent variables and the variables of interest as independent. Crude and adjusted odds ratio (OR) and its 95% confidence interval were calculated by multiple logistic regression not conditional. Statistical significance was assessed by the likelihood ratio test with p < 0.05. Results: The variables associated independently with relapse and adjusted for sex and age were: case of MDR-TB (OR=15.5; 95% CI: 1.2-200.5), co-infection with HIV (OR=9.3 ,95%CI: 1.6-54.1), diabetes (OR=1.6, 95% CI: 0.3-7.7), previous household contact with TB (OR=2.2, 95% CI: 0.8-6.2) and TB at home after patient studied (OR=3,8, 95% CI: 1,2-12,8). The variables associated independently with the retreatment of default and adjusted for sex and age were: case of MDR-TB (OR = 38.7, 95 % CI: 2.9-515.3), co-infection with HIV (OR=24.8, 95% CI: 3.8-163.0), alcohol abuse (OR=4.2, 95% CI: 1.1-17.5) and hospitalization to TB complications (OR=7.2, 95% CI: 2.5-21.0). The variables associated independently with death among patients retreated by default and adjusted for sex and age were: case of MDR-TB (OR=152.4, 95% CI: 9.8-237.4), co-infection with HIV (OR=29.0, 95%CI: 7.1-119.1), alcohol (OR=11.8, 95% CI: 1.3-102.8) and prison system (OR=5.0, 95%CI: 1.0-25.8). Conclusions: The results presented here show groups of higher risk among TB patients for re-treatment by relapse, treatment default and death among re-treated by default to be considered in the improvement of DOTS in our country.

A estabilidade pós-tratamento sempre foi um assunto delicado da ortodontia. A maioria dos pacientes e até mesmo os ortodontistas, julgam o sucesso do tratamento ortodôntico pela estabilidade dos resultados em longo prazo. Por este motivo a literatura é repleta de estudos acerca deste assunto, principalmente da recidiva do apinhamento ântero -inferior. Já o interesse pelo apinhamento ânterosuperior e sua recidiva é crescente por conta de uma maior preocupação com a estética do sorriso. Desta forma, o propósito deste estudo é comparar a recidiva do apinhamento ântero-superior em pacientes com má oclusão de Classe I e Classe II de Angle. A amostra consiste em 70 pacientes divididos em 3 grupos. O Grupo 1 apresenta 30 pacientes (12 do gênero masculino e 18 do gênero feminino) com má oclusão de Classe I, idade média inicial de 13,16 anos e tratados com extrações dos 4 primeiros pré -molares. O Grupo 2 apresenta 20 pacientes (11 do gênero masculino e 9 do gênero feminino) com má oclusão de Classe II divisão 1, idade média inicial de 12,95 anos e também tratados com extrações dos 4 primeiros pré-molares. O Grupo 3 apresenta 20 pacientes (11 do gênero masculino e 9 do gênero feminino) com má oclusão de Classe II divisão 1, idade média inicial de 13,09 anos e tratados com extrações dos 2 primeiros pré -molares superiores. Foram avaliados os modelos de estudo nas fases inicial (T1), final (T2) e no mínimo 5 anos pós -tratamento (T3) de cada pac iente. As variáveis do arco superior avaliadas e comparadas estatisticamente pela Análise de Variância (ANOVA) foram: Índice de Irregularidade de Little superior (IRLS), comprimento do arco (CAS), distância intercaninos (DICS), interpré-molares (DI2PMS) e intermolares (DIMS). Como os resultados entre os grupos não mostraram diferenças estatisticamente significantes nas fases avaliadas, a amostra foi unificada. O Teste de Correlação de Pearson e o Teste de Regressão Linear Múltipla foram utilizados para veri ficar se alguma variável estudada teria influência sobre o apinhamento nas três fases (IRLS1, IRLS2, IRLS3). Os resultados mostraram que a recidiva do apinhamento superior (IRLS3-2) é influenciada pelo apinhamento inicial (IRLS1) e que os dentes tendem a voltar à posição original. O gênero feminino apresentou mais recidiva do apinhamento ântero -superior que o gênero masculino. / The posttreatment stability was always a delicate issue in Orthodontics. Most of the orthodontic patients and even the orthodontists judge their treatment as successful based on the treatment outcomes stability in the long-term. Because of this reason the literature has a lot of studies about this issue, mostly on relapse of mandibular anterior crowding. However the interest on maxillary anterior crowding and its relapse has been growing because of the greater conscious on smile esthetic among patients. Therefore, the purpose of this study was to compare the relapse of maxillary anterior crowding in cases presenting Angles Class I and Class II malocclusions. The experimental sample consisted of 70 patients divided into 3 groups. Group 1 comprised 30 patients (12 male; 18 female) at a mean initial age of 13.16 years, with Class I malocclusion, treated with all first premolars extraction. Group 2 comprised 20 patients (11 male; 9 female) at a mean initial age of 12.95 years, with Class II division 1 malocclusion, also treated with all first premolars extraction. Group 3 comprised 20 patients (10 male; 10 female) at a mean initial age of 13.09 years, with Class II division 1 malocclusion, treated with extraction of two maxillary first premolars. Dental casts measurements were obtained at three stages (pretreatment, posttreatment and postretention) and the variables assessed were Little Irregularity Index, maxillary arch length, intercanine, interpremolar and intermolar widths. The statistical analysis was performed by one-way ANOVA and Tukey tests if necessary (intragroup comparison) and by independent t-tests (intergroup comparison). As the results among the groups did not show statistically significant difference, the experimental sample was unified in order to inve stigate, using Pearson correlation coefficient and Multiple linear regression, if some studied variable would have influence the crowding in the three stages (IRLS1, IRLS2, IRLS3). The results showed that the maxillary crowding relapse (IRLS3-2) is influenced by the initial (IRLS1), and the teeth tend to return to their pretreatment position. The females presented more maxillary anterior crowding relapse than males.

One of the main difficulties of addiction treatment is the high risk of relapse even after a longabstinence and fully detoxification. Therefore, discovering the underlying molecular principlesof relapse is essential. The metabotropic glutamate receptor, mGluR5, is considered to beinvolved in this aspect. One of the brain structures expressing mGluR5 is the striatum, an areawith well-established role in addiction which is largely composed of medium-sized spinyneurons (MSNs). These neurons are basically divided into two major subpopulationscharacterized based on their projections and protein properties. It is known that the mGluR5receptor is expressed on both subpopulations of MSNs. Consequently, it can be used to establishthe proportional contribution of each of MSNs subpopulations in relapse to addiction. In ourconstellation, we have generated a mouse line designed to have a selective mGluR5 knock-downin one of these subpopulations – the dopamine D1 receptor (D1R) expressing neurons. It hashowever been unclear if the expression of the transgene is indeed limited to only D1R-expressingneurons. By immunofluorescence technique, I here show that the construct is expressed only inMSNs and is restricted to the D1R-expressing cell population in the striatum. Thus the transgenicmouse line is a good tool for the study of mGluR5 selectively in D1R expressing neurons.

Abstract Diffuse large B-cell lymphoma (DLBCL) is the most common type of Non-Hodgkin´s Lymphoma (NHL). The standard treatment for DLBCL is R-CHOP chemoimmunotherapy (rituximab, cyclophosphamide, vincristine, doxorubicin and prednisone). About one -third of patients have refractory disease or the lymphoma relapses. Prognosis after relapse of refractory disease is poor. Fitter and younger patients are recommended new intensive salvage chemotherapy followed by autologous stem cell transplantation. Central nervous system (CNS) relapse is the most feared complication with dismal prognosis in DLBCL. High dose methotrexate intravenously administered concurrently with R-CHOP treatment has shown to be most promising to prevent CNS relapses. Primary CNS lymphoma (PCNSL) is a rare aggressive lymphoma limited to the CNS and eyes. PCNSL is a chemo-and radiosensitive disease, but long-term response is rare since the blood brain barrier (BBB) limits access of many drugs to the CNS. BBB disruption (BBBD) is a treatment modality where the BBB is opened by hypertonic mannitol infusion. Administration of chemotherapeutics will achieve over ten-fold concentrations in the CNS and eradicate microscopic disease involvement. This study retrospectively analyses patients who treated as first line with Bonn/Bonn-like treatment (study I), with BBBD treatment followed by high-dose treatment/autologous stem cell transplantation (HDT/ASCT) in first- or second-line (study II) or those treated with primary R-CHOP or its derivatives with or without concurrent CNS-targeted treatment (study III). HD-MTX-based multichemotherapy is an effective induction treatment in CNS lymphoma, but long-lasting responses are rare. BBBD-treatment is well-tolerated and a promising method to attain high drug concentrations in the CNS to eradicate microscopic disease involvement in first- and second-line. CNS-prophylaxis with HD-MTX prevents CNS events in high risk DLBCL. PCNSL is agressive disease despite excellent primary response with HD-MTX based multichemotherapy. BBBD-treatment is a promising method to eradicate microscopic disease in the CNS and achieve a long-term response and cure rate. Fatal CNS relapses can be avoided using CNS-targeted treatment. / Tiivistelmä Diffuusi suurisoluinen B-solulymfooma (DLBCL) on yleisin non-Hodgkin lymfooma (NHL), jonka standardihoitona toimii R-CHOP (rituksimabi, syklofosfamidi, vinkristiini, doksorubisiini, prednisoloni). Noin kolmasosalla potilaista tautii etenee hoidosta huolimatta tai uusii hoidon päätyttyä. Relapoituneen tai refraktaarin taudin ennuste on huono. Hyväkuntoisilla ja nuoremmilla potilailla pyritään etenemään uuteen induktiohoitoon ja korkea-annoshoitoon autologisen kantasolusiirteen turvin. Keskushermostouusiutuma on huonoennusteisin DLBCL:n komplikaatio. Suuriannosmetotreksaattihoito liitettynä R-CHOP-hoitoon estää keskushermostouusiutumia. Primaari aivolymfooma (PCNSL) on harvinainen keskushermoston ja silmien alueelle rajautuva lymfooma. PCNSL on herkkä sytostaatti-ja sädehoidolle, mutta pitkäkestoisia vasteita nähdään harvoin. Veriaivoeste estää useimpien tehokkaiden sytostaattien pääsyn keskushermostoon. Veriaivoesteen aukaisuhoidossa veriaivoeste avataan hypertonisella mannitoli-infuusiolla. Toimenpiteen jälkeisellä sytostaatti-infuusiolla saavutetaan kymmenkertaiset lääkeainepitoisuudet keskushermostossa ja voidaan hoitaa mikroskooppista veriaivoesteen takana sijaitsevaa tautia. Väitöskirjatyön tutkimukset ovat retrospektiivisiä. Ensimmäisessä osatyössä analysoitiin PCNSL potilaat, jotka saivat ensilinjassa Bonnin tai Bonnin kaltaista hoitoa. Toisessa osatyössä potilaat hoidettiin joko ensi- tai toisessa linjassa BBBD-hoidolla, päättyen konsolidaatiohoitona annettavaan korkea-annoshoitoon autologisen kantasolusiirteen turvin. Kolmannessa osatyössä analysoitiin suuren aivouusiutumariskin potilaita, joko yhdessä tai ilman keskushermostoon suunnattua hoitoa samanaikaisesti R-CHOP-hoidon kanssa. Suuriannosmetotreksaatti-pohjainen yhdistelmäsolunsalpaajahoito on tehokas induktiohoito aivolymfoomassa pitkäkestoisten vasteiden ollessa harvinaisia. BBBD-hoito on hyvin siedetty ja lupaava hoitomuoto, jolla keskushermostossa voidaan saavuttaa suuret lääkeainepitoisuudet, jotka riittävät hoitamaan mikroskooppisen taudin sekä ensi että toisessa linjassa. Keskushermostoprofylaksia suuriannosmetotreksaatilla estää keskushermosto-uusiutumia suuren riskin DLBCL-potilailla. PCNSL on agressiivinen tauti huolimatta erinomaisista metotreksaattipohjaisilla hoidoilla saavutetuista ensilinjan vasteista. BBBD-hoito on lupaava keino eradikoida mikroskooppinen tauti keskushermostosta ja saavuttaa pitkäaikaisia hoitovasteita, sekä pysyvä paraneminen aivolymfoomassa. Suuriannosmetotreksaattia sisältävällä sytostaattihoidolla voidaan estää fataaleja aivorelapseja DLBCL:ssä.

La qualité de vie (QV) liée à la sclérose en plaques (SEP) a été très étudiée lors de ces dernières années et plusieurs échelles d'évaluation ont ainsi été élaborées. Cependant, peu d'échelles de QV liée à la SEP validées sont disponibles en version française, et aucune ne s'attache à évaluer spécifiquement les périodes de poussées ni les perspectives temporelles. C'est l'objet de ce travail. La première étape a permis d'élaborer une première version d'une échelle d'évaluation de la perception de la QV liée à la SEP et à ses poussées (PERSEPP) en se basant sur la réalisation d'entretiens semi-directifs auprès de patients atteints de formes rémittentes de SEP et de soignants. A partir d'une analyse de contenu thématique de l'ensemble des entretiens individuels et de groupes réalisés, des items spécifiques du concept de QV liée à la SEP et à ses poussées ont ainsi été sélectionnés afin d'élaborer l'échelle PERSEPP constituée d'un questionnaire principal et de modules complémentaires. Une étude pilote a ensuite été réalisée afin d'évaluer différents critères d'acceptabilité et de faisabilité de l'échelle PERSEPP auprès de 40 patients. A la suite de cette étape, une étude multicentrique a permis d'inclure des patients atteints de forme rémittente de SEP, en poussée ou hors poussée, et suivis à l'hôpital ou chez des neurologues libéraux. L'analyse des données recueillies auprès de 305 patients a permis de compléter plusieurs étapes nécessaires à la validation d'une échelle de QV liée à la santé et d'obtenir une version définitive de l'échelle PERSEPP. De par sa méthodologie d'élaboration, cette échelle est spécifique de la forme rémittente de SEP. / Quality of life (QoL) related to multiple sclerosis (MS) has been widely studied in recent years and several rating scales have been developed. However, few QoL scales validated in MS are available in French, and none specifically focuses on assessing periods of relapses or time perspective. This is the purpose of this work. The first step consisted in establishing a first version of a rating scale of QoL perceived in MS and its relapses (PERSEPP) based on semi-structured interviews with patients affected by relapsing-remitting forms of MS and with health care providers. From a thematic content analysis of ail individual interviews and focus groups conducted, specific items of the concept of QoL related to MS and its relapses were selected to develop the PERSEPP scale consisting of a main questionnaire and additional modules. A pilot study with 40 patients was then conducted to evaluate various criteria for acceptability and feasibility of the PERSEPP scale. Following this step, a multicenter study included patients with a relapsing-remitting form of MS, in relapse or not in relapse, and followed by hospital or private practice neurologists. Analysis of the data collected from 305 patients has enabled us to complete several steps necessary for the validation of a health­related QoL scale and obtain a final version of the PERSEPP scale. This scale is specific for relapsing­remitting form of MS due to the methodology used.

South Africa has witnessed radical policy reforms since the advent of the new democratic dispensation in 1994. As provincial, district and local practitioners developed responses to those new national education policies, implementation issues were revealed in all their complexity. Policy implementation became one of the most difficult challenges South Africa had to contend with. It necessitated the development of the capacity of the state and its people to implement policy. The complexities of the people development environment in South Africa create a range of challenges for development. The lack of multi-level empirical data on training effects continues to exacerbate this state of affairs. This context requires that comprehensive assessment mechanisms be put in place for training programmes aimed at the realisation of policy goals. The broad purpose of this inquiry, therefore, is to determine the impact of an education management development training intervention as it passes through different levels of the education system in South Africa - national, provincial, district and local. The following key questions guided this investigation. Firstly, how do stakeholder understandings of “education management development” transfer from one level to another in a cascade model of training? And secondly, what is the operational impact of an education management development-training programme at the different levels (i.e., province, district and school) of the education system? Questionnaires, free attitude interviews and observations were used as key data collection instruments. Data was analysed using a combination of quantitative and qualitative strategies for making sense of the training information. Data was interpreted against the backdrop of the literature on the “transfer of training”, and these findings are represented in the last three chapters of this thesis. This research generated four major findings, with a variety of subsidiary findings, which deepen our insights into policy implementation as it relates to education management development in South Africa. Some of the major findings of this study are that: a) The organisers did not in the first place conduct the baseline study on training needs and secondly, they did not take into consideration the organisational requirements of the anticipated training. This anomaly could ultimately compromise the outcomes of this intervention. b) The policy deployment processes of the North West Department of Education (of ensuring that the governmental policies for quality, cost and service delivery are understood from the highest to the lowest level of the organisation) are fraught with problems that undermined basic understandings of the Education Management Development training programme which was conducted between 1998 – 2000. c) Successful policy implementation depends crucially on resource allocation and, in the case of the Education Management Development Training Programme, the nature and magnitude of allocated resources e.g., transport provision could not guarantee positive training results; and d) The overly rationalistic view adopted by the trainers of the Education Management development-training programme did not take into account the complex contexts within which change unfolds. Resultantly, the changing of the roles of facilitators could not be planned-for in advance. The significance of this study is that it identifies the barriers to learning in training events, and sheds new light on the transfer of training problem that continues to undermine organisational change and human resource development. Some of the unique findings of the study can be found in the fact that the Expert Trainers could only recall a few intentions of the EMD modular training. The District Facilitators displayed limited knowledge of the disciplinary procedures in their areas of operation and there was limited conceptualisation of the EMD by the principals of schools. / Thesis (PhD (Education Policy Studies))--University of Pretoria, 2005. / Education Management and Policy Studies / unrestricted

The purpose of this study is to describe the subjective psychological experience of relapse in cocaine/crack and heroin users with the aim of identifying the significant cognitive, emotional and social themes involved in relapse. A better understanding of relapse may aid in providing more effective treatment for substance users. Both the intra- and interpsychic factors involved in relapse that emerge from the study are viewed from within a broad systems theory approach. In this study, not only the whole system is of relevance, but also the subsystems. Various sub systems are also identified to allow for the recognition of patterns, functions and recursive feedback loops that maintain substance-using behaviour. Due to the qualitative nature of the study, the context surrounding substance abuse and the substance users assumes vital importance. The interrelationships between the various intrapsychic structures, the family unit, the social contexts, the drugs themselves and the physiological aspects of substance abuse are identified. A qualitative research design was applied. In-depth semi-structured interviews were used to gather data from the eight participants, who were crack and heroin users who attended the in-patient rehabilitation programme at Phoenix House. Seven of the participants were still in the in-patient rehabilitation programme at the time of the interview, while one participant was in the aftercare programme. All have been through a rehabilitation process before and were at Phoenix House due to a relapse. A thematic analysis was conducted and the process of analysis settled on eight overall themes. Extensive descriptions of these themes are provided. The discussion highlights the role of relapse in the cycle of self-destruction that constitutes substance abuse, in addition to the role relapse plays in the process of recovery. Connection seems to be the key to breaking the cycle of alienation that users experience. Falling into the trap of rejecting users without looking beyond their behaviour allows them to continue functioning in a way that confirms their view of themselves as unlovable, which, in turn, maintains their behaviour. Although systems theory is an independent approach in its own right, the nature of its view allows for the incorporation of other approaches. Where possible and relevant, other theories are incorporated into the discussion of the results, with the aim of gaining an integrated understanding of the findings of the study within the broader field of substance abuse. / Dissertation (MA (Clinical Psychology))--University of Pretoria, 2005. / Psychology / unrestricted

Dysfunktionale Lernvorgänge (wie z.B. Pawlowsche Konditionierung) können bei Patienten mit Alkoholabhängigkeit das Rückfallrisiko erhöhen, was innerhalb der vorliegenden Arbeit mittels drei Studien untersucht wurde.

Generalized anxiety disorder (GAD) is associated with painful physical symptoms (PPS). These post hoc analyses of previous trial data assessed PPS and their response to duloxetine treatment in GAD patients. Studies 1 and 2 (n = 840) were 9- to 10-week efficacy trials; study 3 (n = 887) was a relapse prevention trial comprising a 26-week open-label treatment phase and a 26-week double-blind, placebo-controlled treatment continuation phase. Mean baseline visual analog scale scores (VAS, 0–100; n = 1727) ranged from 26 to 37 for overall pain, headache, back pain, shoulder pain, interference with daily activities, and time in pain while awake. In studies 1 and 2, improvement on all VAS scores was greater in duloxetine-treated than in placebo-treated patients (p ≤ 0.01). In study 3, pain symptoms worsened in responders switched to placebo compared with those maintained on duloxetine (p ≤ 0.02). In conclusion, duloxetine was efficacious in the short- and long-term treatment of PPS, which are common in GAD patients.

Pour les patients diagnostiqués avec un cancer au stade précoce, les décisions de traitement dépendent de l’évaluation du risque de rechute métastatique. Les outils de pronostic actuels sont fondés sur des approches purement statistiques, sans intégrer les connaissances disponibles sur les processus biologiques à l’oeuvre. L’objectif de cette thèse est de développer des modèles prédictifs du processus métastatique en utilisant une approche de modélisation mécaniste et la modélisation à effets mixtes. Dans la première partie, nous étendons un modèle mathématique du processus métastatique pour décrire la croissance de la tumeur primaire et de la masse métastatique totale chez des souris traitées avec le sunitinib (un inhibiteur de tyrosine kinase ayant une action anti-angiogénique) administré comme traitement néoadjuvant (i.e. avant exérèse de la tumeur primaire). Le modèle est utilisé pour tester des hypothèses expliquant les effets différentiels du sunitinib sur la tumeur primaire et les métastases. Des algorithmes d’apprentissage statistique sont utilisés pour évaluer la valeur prédictive des biomarqueurs sur les paramètres du modèle.Dans la deuxième partie de cette thèse, nous développons un modèle mécaniste pour la prédiction du temps de rechute métastatique et le validons sur des données cliniques des patientes atteintes d’un cancer du sein localisé. Ce modèle offre des prédictions personnalisées des métastases invisibles au moment du diagnostic, ainsi que des simulations de la croissance métastatique future, et il pourrait être utilisé comme un outil de prédiction individuelle pour aider à la gestion des patientes atteintes de cancer du sein. / For patients diagnosed with early-stage cancer, treatment decisions depend on the evaluation of the risk of metastatic relapse. Current prognostic tools are based on purely statistical approaches that relate predictor variables to the outcome, without integrating any available knowledge of the underlying biological processes. The purpose of this thesis is to develop predictive models of the metastatic process using an established mechanistic modelling approach and the statistical mixed-effects modelling framework.In the first part, we extend the mathematical metastatic model to describe primary tumour and metastatic dynamics in response to neoadjuvant sunitinib in clinically relevant mouse models of spontaneous metastatic breast and kidney cancers. The calibrated model is then used to test possible hypothesis for the differential effects of sunitinib on primary tumour and metastases, and machine learning algorithms are applied to assess the predictive power of biomarkers on the model parameters.In the second part of this thesis, we develop a mechanistic model for the prediction of the time to metastatic relapse and validate it on a clinical dataset of breast cancer patients. This model offers personalised predictions of the invisible metastatic burden at the time of diagnosis, as well as forward simulations of metastatic growth, and it could be used as a personalised prediction tool to assist in the routine management of breast cancer patients.