No description available.

O diabetes mellitus (DM) é uma doença que está associada com aumento da morbidade, mortalidade e custos econômicos. O DM tipo 2 é a forma de diabetes mais comum, acometendo 85%-90% de todos os casos. O mau controle glicêmico é um fator determinante do desenvolvimento e progressão das complicações do DM. A hemoglobina glicada (A1C) se tornou a medida de referência para o controle de DM por mais de duas décadas. Existe um grande incentivo, tanto da perspectiva de saúde pública quanto da clínica, em detectar pessoas com risco futuro de desenvolver DM2, pois este é um forte fator de risco para doença cardiovascular. Também existem evidências que é possível prevenir ou retardar o DM nas pessoas com tolerância a glicose diminuída, desde que estes casos sejam identificados e tratados adequadamente. Os testes disponíveis hoje para o diagnóstico do DM, glicemia de jejum (GJ) e teste oral de tolerância à glicose (TOTG), carecem de sensibilidade e/ou especificidade. Recentes estudos têm evidenciado que a A1C pode ser uma nova ferramenta para diagnóstico do DM, sendo que diversos pontos de corte tem sido estudados. Maiores investigações para a validação do desempenho diagnóstico deste teste na predição do DM são necessárias para podermos utilizar esta ferramenta com segurança na triagem e diagnóstico do DM. / Diabetes mellitus (DM) is a disease associated with greater mortality and economical costs. Type 2 DM is the commonest form of DM, accounting for 85-90 % of its cases. Glycemic levels are a determinant factor for the development and progression of DM complications. Glycated hemoglobin (A1C) became the reference measure of glycemic control for more than two decades. There is a great incentive in detecting persons with future risk of developing DM, because this is a strong risk factor for cardiovascular disease. Also there are evidences that it is possible to prevent or to delay DM in persons with prediabetes, since identified and treated appropriately. Available tests for DM diagnosis, fasting glycemia (FG) and oral glucose tolerance test (OGTT), lack sensibility and/or specificity. Recent studies have shown that A1C can be a new tool for DM diagnosis and several cutoff points have been analyzed. However, the validation of this test as diagnostic modality to detect DM might be necessary in order to provide a useful tool for DM diagnosis.

O diabetes mellitus (DM) é uma doença que está associada com aumento da morbidade, mortalidade e custos econômicos. O DM tipo 2 é a forma de diabetes mais comum, acometendo 85%-90% de todos os casos. O mau controle glicêmico é um fator determinante do desenvolvimento e progressão das complicações do DM. A hemoglobina glicada (A1C) se tornou a medida de referência para o controle de DM por mais de duas décadas. Existe um grande incentivo, tanto da perspectiva de saúde pública quanto da clínica, em detectar pessoas com risco futuro de desenvolver DM2, pois este é um forte fator de risco para doença cardiovascular. Também existem evidências que é possível prevenir ou retardar o DM nas pessoas com tolerância a glicose diminuída, desde que estes casos sejam identificados e tratados adequadamente. Os testes disponíveis hoje para o diagnóstico do DM, glicemia de jejum (GJ) e teste oral de tolerância à glicose (TOTG), carecem de sensibilidade e/ou especificidade. Recentes estudos têm evidenciado que a A1C pode ser uma nova ferramenta para diagnóstico do DM, sendo que diversos pontos de corte tem sido estudados. Maiores investigações para a validação do desempenho diagnóstico deste teste na predição do DM são necessárias para podermos utilizar esta ferramenta com segurança na triagem e diagnóstico do DM. / Diabetes mellitus (DM) is a disease associated with greater mortality and economical costs. Type 2 DM is the commonest form of DM, accounting for 85-90 % of its cases. Glycemic levels are a determinant factor for the development and progression of DM complications. Glycated hemoglobin (A1C) became the reference measure of glycemic control for more than two decades. There is a great incentive in detecting persons with future risk of developing DM, because this is a strong risk factor for cardiovascular disease. Also there are evidences that it is possible to prevent or to delay DM in persons with prediabetes, since identified and treated appropriately. Available tests for DM diagnosis, fasting glycemia (FG) and oral glucose tolerance test (OGTT), lack sensibility and/or specificity. Recent studies have shown that A1C can be a new tool for DM diagnosis and several cutoff points have been analyzed. However, the validation of this test as diagnostic modality to detect DM might be necessary in order to provide a useful tool for DM diagnosis.

O diabetes mellitus (DM) é uma doença que está associada com aumento da morbidade, mortalidade e custos econômicos. O DM tipo 2 é a forma de diabetes mais comum, acometendo 85%-90% de todos os casos. O mau controle glicêmico é um fator determinante do desenvolvimento e progressão das complicações do DM. A hemoglobina glicada (A1C) se tornou a medida de referência para o controle de DM por mais de duas décadas. Existe um grande incentivo, tanto da perspectiva de saúde pública quanto da clínica, em detectar pessoas com risco futuro de desenvolver DM2, pois este é um forte fator de risco para doença cardiovascular. Também existem evidências que é possível prevenir ou retardar o DM nas pessoas com tolerância a glicose diminuída, desde que estes casos sejam identificados e tratados adequadamente. Os testes disponíveis hoje para o diagnóstico do DM, glicemia de jejum (GJ) e teste oral de tolerância à glicose (TOTG), carecem de sensibilidade e/ou especificidade. Recentes estudos têm evidenciado que a A1C pode ser uma nova ferramenta para diagnóstico do DM, sendo que diversos pontos de corte tem sido estudados. Maiores investigações para a validação do desempenho diagnóstico deste teste na predição do DM são necessárias para podermos utilizar esta ferramenta com segurança na triagem e diagnóstico do DM. / Diabetes mellitus (DM) is a disease associated with greater mortality and economical costs. Type 2 DM is the commonest form of DM, accounting for 85-90 % of its cases. Glycemic levels are a determinant factor for the development and progression of DM complications. Glycated hemoglobin (A1C) became the reference measure of glycemic control for more than two decades. There is a great incentive in detecting persons with future risk of developing DM, because this is a strong risk factor for cardiovascular disease. Also there are evidences that it is possible to prevent or to delay DM in persons with prediabetes, since identified and treated appropriately. Available tests for DM diagnosis, fasting glycemia (FG) and oral glucose tolerance test (OGTT), lack sensibility and/or specificity. Recent studies have shown that A1C can be a new tool for DM diagnosis and several cutoff points have been analyzed. However, the validation of this test as diagnostic modality to detect DM might be necessary in order to provide a useful tool for DM diagnosis.

La douleur abdominale aiguë (DAA) est un problème de santé publique du fait de sa fréquence, du large éventail de pathologies causales et de son enjeu thérapeutique avec une nécessité d'opérer rapidement les patients à risque et de ne pas opérer la plupart des patients. Le retard ou l'insuffisance diagnostique d'une urgence abdominale, qu'elle soit non traumatique (en particulier en cas de strangulation, d'ischémie digestive ou de péritonite) ou traumatique, entrainera une augmentation de la morbi-mortalité. Il est donc primordial que la prise en charge diagnostique et thérapeutique repose sur de l'Evidence-Based Médecine, intégrant une expertise clinique avec les meilleurs examens para-cliniques dont les performances ont été validées par les recherches scientifiques. A ce jour, c'est la tomodensitométrie (TDM) qui s'est imposée comme l'examen complémentaire de référence en cas d'incertitude diagnostique, pour l'exploration d'une DAA. Or, les études publiées sur l'apport du scanner dans les DAA ont un certain nombre de limites : elles ne s'intéressent souvent qu'à une seule pathologie donnée et ne correspondent donc pas à la situation clinique de pathologies non identifiées, elles intègrent rarement des objectifs sur l'impact thérapeutique et présentent, pour la plupart, des faibles niveaux de preuve (niveau I ou II) sur l'échelle à 5 niveaux de la classification des études diagnostiques. Quel que soit le niveau de l'étude, des outils statistiques avancés peuvent être utilisés pour diminuer les biais d'interprétation, pour prendre en compte le degré d'incertitude diagnostique clinique, pour tenir compte de la faible prévalence de certaines pathologies urgentes et pour faciliter le transfert et la généralisation des résultats obtenus à la pratique clinique. L'objectif de cette thèse est donc de présenter des outils statistiques, en analysant les forces et limites des résultats obtenus, afin d'étayer les perspectives d'intégration dans une stratégie de prise en charge clinique. Certains de ces outils statistiques sont déjà connus et largement utilisés en imagerie, d'autres sont développés essentiellement dans les études thérapeutiques et enfin, certains sont totalement innovants.Ainsi, la première partie de cette thèse est consacrée à une mise au point sur la méthodologie de la recherche en imagerie, en exposant ses caractéristiques, ses faiblesses et ses perspectives d'évolution. La deuxième partie est constituée d'une production scientifique de 5 articles, 4 d'entre eux publiés, focalisée sur les urgences abdominales traumatiques et non traumatiques, illustrant à l'aide du développement des analyses statistiques, l'apport du scanner pour améliorer certaines stratégies diagnostiques et thérapeutiques. De cette façon, les méthodes de concordance diagnostique, de régression logistique multivariée, de construction de score diagnostique et de méta-analyse bivariée avec effets aléatoires sont explicitées. Nous avons dans le dernier article proposé une méthode d'analyse statistique innovante basée sur une notion de distance permettant d'intégrer l'incertitude diagnostique représentée par les multiples hypothèses diagnostiques face à une douleur abdominale. Cette approche permet ainsi de mesurer la valeur ajoutée de la TDM à la fois en termes d'exactitude diagnostique et de confiance diagnostique, en se rapprochant de la pratique clinique quotidienne. A l'avenir, ces outils statistiques pourraient être plus largement utilisés dans la recherche diagnostique afin d'améliorer la pertinence scientifique et l'impact clinique, à l'origine d'une meilleure prise en charge des patients, et d'une meilleure utilisation des ressources économiques. / Acute abdominal pain (AAP) is a public health problem, given its frequency, the numerous specific causes and the therapeutic impact, with patients at risk requiring urgent surgical care while most patients do not require surgery. Delayed diagnosis or misdiagnosis of an abdominal emergency, whether of non traumatic (such as bowel strangulation, or bowel ischemia, or peritonitis) or traumatic origin, will lead to increased morbidity and mortality. It is thus crucial that diagnostic and therapeutic management relies on Evidence-based Medicine, integrating clinical expertise and paraclinical tests that have been validated through scientific research. To date, computed tomography (CT) has become the reference examination to investigate AAP when diagnosis is uncertain. However, published studies evaluating the value of CT in AAP, have some limits: they often focus on a specific disease, and consequently do not meet the unsorted pathologies encountered in clinical practice, they sparsely include objectives on therapeutic impact, and most show low levels of evidence (level I or II) on the five level diagnostic studies scale. Whatever the study's level, advanced statistical methods can be used to reduce interpretation bias, take into account the clinical diagnostic uncertainty or the low prevalence of some urgent diseases, and to facilitate the transfer and generalisation of results into daily clinical practice.The aim of this work is to present statistical tools, by analysing the results' strengths and limits, in order to illustrate the perspective of integration within a clinical management strategy. Some of these statistical tools are known and largely used in imaging studies, others are essentially developed in therapeutic studies, and some are completely innovative.Thus, the first part of this thesis reports on a review of the methodological aspect of research in the field of medical imaging, showing its characteristics, its weaknesses and its development perspective.The second part is composed of 5 scientific articles - with 4 already published - and focus on traumatic and non traumatic abdominal emergencies, illustrating the value of CT in improving some diagnostic and therapeutic strategies, while detailing the statistical analyses. Hence, statistical methods for diagnostic agreement, multivariate logistic regression, diagnostic scoring-system building and bivariate random-effects meta-analysis are developed. In the last article, we propose a new statistical analysis method, based on a distance concept, allowing integrating the diagnostic uncertainty resulting from multiple diagnosis hypotheses in front of an AAP. This approach allows for investigating the added value of CT both in terms of diagnostic accuracy and diagnostic confidence, while getting close to daily clinical practice.In the future, these statistical tools could be implemented into diagnostic research in order to improve the scientific relevance and clinical impact, allowing for better patients' management, and better economic resource usage.

Introdução: A avaliação e detecção precoce dos defeitos de perfusão miocárdica permite estratificar o risco cardiovascular e auxiliar a tomada de decisão terapêutica nos pacientes com DAC estabelecida ou suspeita. O PET/CT com Rb-82 na avaliação da perfusão e função miocárdica não invasiva é uma ferramenta de elevada acurácia diagnóstica. O presente trabalho procura estabelecer uma avaliação de performance de diagnóstica, entre dois métodos na medicina nuclear - SPECT com 99mTc-Sestamibi e PET/CT com Rb-82. Métodos: Avaliação da perfusão e da função do ventrículo esquerdo em 42 pacientes submetidos ao PET-CT com Rb-82 e SPECT com MIBI com mesmo estresse farmacológico (Dipiridamol), utilizando-se das variáveis de perfusão (SSS, SDS, SRS, fluxo sanguíneo e reserva coronariana) e de função (FEVE, DeltaFEVE, VDF, VSF - no repouso e no estresse), considerando o CATE como o método de referência. Resultados: No escore SDS, o Rb-82 apresentou valores médios próximos ao dobro do MIBI no grupo dos alterados (7,6 vs 3,9). Obteve-se um valor de sensibilidade do Rb-82 de 90,9%, com VPP de 88,2% (p=0,012) para o SSS e de 87,8%, com VPP de 87,8% (p= 0,012) para o SDS. As áreas sobre curva demonstraram superioridade do Rb-82 com maior exatidão e acurácia do método em relação ao MIBI na amostra avaliada. No escore de reversibilidade (SDS), houve maior sensibilidade em cut-off de menor valor e maior especificidade em cut-off de maior valor. FEVE em repouso e estresse, assim como, VDF e VSF em repouso e estresse apresentaram valores de concordância altos. Os valores de SSS, SRS e SDS apresentaram concordâncias moderadas. Houve aparente queda do DeltaFEVE no Rb-82, contudo similar tendência não foi observada no MIBI, considerando que as diferenças não foram significativas com relação ao número de vasos com estenose em ambos os radioisótopos (p > 0,05). Em outra análise, entre uma variável de função ventricular e a CFR global foi evidenciada diferença significativa (p=0,024) entre os grupos normal e alterado. No grupo alterado com CFR global < 2,0 houve redução do DeltaFEVE em 2,7%.Conclusão: O PET-CT com 82Rb apresenta maior exatidão e acurácia diagnóstica na avaliação perfusional miocárdica em relação ao SPECT com MIBI, inclusive com reclassificação dos grupos avaliados que passaram de defeito perfusional de pequena para moderada extensão. A FEVE estresse na população geral sugere valor mais alto no 82Rb (aquisição do exame no pico do estresse).Haveria uma sugestão de que no grupo com CATE alterado, a CFR global seria reduzida; que em maiores graus de estenose coronariana ao CATE, a CFR global decresceria; que existiria queda da DeltaFEVE no 82Rb quanto maior fosse o número de vasos obstruídos ao CATE. A relação entre a FEVE reserva e a CFR global no 82Rb foi significativa com queda da DeltaFEVE no grupo com CFR global alterada / Background: The evaluation and early detection of myocardial perfusion defects allows to stratify cardiovascular risk and it helps therapeutic decision-making in patients with established or suspected CAD. Rb-82 PET/CT in perfusion evaluation and noninvasive myocardial function is a tool of high diagnostic accuracy. The present study seeks to establish a diagnostic performance evaluation between two methods in nuclear medicine - 99mTc-Sestamibi SPECT and Rb-82 PET/CT. Methods: Evaluation of perfusion and left ventricular function in 42 patients submitted to Rb-82 PET-CT and MIBI SPECT with the same pharmacological stress (Dipyridamole), using perfusion variables (SSS, SDS, SRS, blood flow and coronary reserve) and function (LVEF, DeltaLVEF, EDV, ESV - in rest and stress), considering CAG as the reference method. Results: In the SDS score, Rb-82 presented mean values close to double the MIBI in the altered group (7.6 vs 3.9). A Rb-82 sensitivity of 90.9% was obtained, with PPV of 88.2% (p = 0.012) for SSS and 87.8% for PPV of 87.8% (p = 0.012) for or SDS. The areas on the curve indicated superiority of Rb-82 with greater accuracy and accuracy of the method in relation to the MIBI in the sample evaluated. In the reversibility score (SDS), there was greater sensitivity in the lower cut-off value and greater specificity in the higher cut-off value. LVEF at rest and stress, as well as VDF and VSF at rest and stress had high agreement values. The values of SSS, SRS and SDS presented moderate agreement. There was an apparent decrease of DeltaLVEF in Rb-82, however a similar trend was not observed in the MIBI, considering that the differences were not significant in relation to the number of vessels with stenosis in both radioisotopes (p > 0.05). In another analysis, a significant difference (p = 0.024) between the normal and altered groups was found of ventricular function variable and global CFR. In the altered group with global CFR < 2.0 there was a reduction of DeltaLVEF in 2.7%.Conclusion: 82Rb PET-CT has a greater accuracy and diagnostic accuracy of myocardial perfusion assessment in relation to MIBI SPECT, including reclassification of the evaluated groups that went from small to moderate perfusion defect extension. The LVEF stress in the general population suggests a higher value in 82Rb (acquisition of the exam at peak stress). There would be a suggestion that if the CAG group changed, the overall CFR would be reduced; in which greater degrees of coronary stenosis to CAG, the global CFR would decrease; and there would be a decrease of DeltaLVEF in 82Rb, the greater the number of vessels obstructed to CAG. The relation to reserve LVEF and overall CFR in 82Rb was significant with decrease in DeltaLVEF of the group with altered global CFR

CONTEXTE La syphilis est une maladie ré-émergente depuis 2000. Son traitement est simple, mais son diagnostic est complexe. La technique d’amplification génique de Treponema pallidum (Tp-PCR) existe depuis 1990 mais le CDC l’a incluse dans sa définition de cas en janvier 2014. OBJECTIFS 1) Evaluer la performance diagnostique de la Tp-PCR à différents stades cliniques et milieux biologiques. 2) Mesurer la sensibilité, spécificité et les valeurs prédictives de la Tp-PCR en fonction de 3 groupes de référence dans des ulcères récents. 3) Comparer les performances des 2 principales cibles de Tp-PCR.MATÉRIEL ET MÉTHODES Premièrement, une revue systématique et méta-analyse des études publiées depuis 1990 ont été menées. Ensuite une étude multicentrique prospective a été conduite dans 5 villes européennes pendant 2 ans chez des patients avec un ulcère oro-génital. Tous ont reçu le test de référence local et 2 Tp-PCRs dans l’ulcère (gène tpp47 vs. polA). Les valeurs de sensibilité, spécificité et valeurs prédictives de la Tp-PCR ont été calculées comparativement au fond noir (FN), à la sérologie et à un gold standard amélioré. La concordance des 2 cibles a été évaluée par un coefficient kappa.RÉSULTATS PRINCIPAUX La méta-analyse conclut que la Tp-PCR a une meilleure performance dans les ulcères récents. L’étude clinique montre que la Tp-PCR décrit une meilleure performance comparativement au gold standard amélioré et a même une meilleure sensibilité que le FN. Les 2 cibles ont la même valeur diagnostique et une concordance quasi parfaite. CONCLUSIONS La Tp-PCR ciblant tpp47 ou polA est cliniquement utile pour diagnostiquer une syphilis primaire et pourrait même remplacer le FN sous certaines conditions. / BACKGROUND Syphilis has re-emerged in at-risk populations since 2000. Although the treatment of syphilis is simple, its diagnosis remains challenging. Treponema pallidum Polymerase Chain Reaction (Tp-PCR) has been used in the diagnosis of syphilis since 1990 but it is included in the case definition of the CDC since January 2014. OBJECTIVES 1) To assess the accuracy of Tp-PCR in various biological specimens and syphilis stages. 2) To measure its diagnostic performance (sensitivity, specificity and predictive values) in ulcers from early syphilis compared to three groups of reference. 3) To compare the accuracy of the two most currently used targets: tpp47 and polA genes.METHODS We conducted a systematic review and meta-analysis of all studies published from 1990. We implemented a multicentre, prospective, observational study in 5 European cities between 09/2011 and 09/2013 among patients with an oral or genital ulcer suggestive of syphilis. All patients were tested with traditional reference tests plus 2 Tp-PCRs (tpp47 and polA). We estimated the sensitivity, specificity and predictive values of Tp-PCR compared to darkfield microscopy (DFM), serology and an enhanced gold standard. We used the kappa coefficient to assess the agreement between the 2 targets.MAIN RESULTST p-PCR had the best accuracy in ulcers from early syphilis. Tp-PCR performed better when compared to the enhanced gold standard and had a higher sensitivity than DFM. The 2 Tp-PCRs had a similar accuracy and an almost perfect agreement.CONCLUSIONS Tp-PCR targeting either tpp47 or polA is clinically useful to confirm an early syphilis in smears and could even replace DFM under specific conditions.