Frank De Smet

Cancer During Pregnancy: An Analysis of 215 Patients Emphasizing the Obstetrical and the Neonatal Outcomes.

PURPOSE The aim of this study was to assess the management and the obstetrical and neonatal outcomes of pregnancies complicated by cancer. PATIENTS AND METHODS In an international collaborative setting, patients with invasive cancer diagnosed during pregnancy between 1998 and 2008 were identified. Clinical data regarding the cancer diagnosis and treatment and the obstetric and neonatal outcomes were collected and analyzed. RESULTS Of 215 patients, five (2.3%) had a pregnancy that ended in a spontaneous miscarriage and 30 (14.0%) pregnancies were interrupted. Treatment was initiated during pregnancy in 122 (56.7%) patients and postpartum in 58 (27.0%) patients. The most frequently encountered cancer types were breast cancer (46%), hematologic malignancies (18%), and dermatologic malignancies (10%). The mean gestational age at delivery was 36.3 +/- 2.9 weeks. Delivery was induced in 71.7% of pregnancies, and 54.2% of children were born preterm. In the group of patients prenatally exposed to cytotoxic treatment, the prevalence of preterm labor was increased (11.8%; P = .012). Furthermore, in this group a higher proportion of small-for-gestational-age children (birth weight below 10th percentile) was observed (24.2%; P = .001). Of all neonates, 51.2% were admitted to a neonatal intensive care unit, mainly (85.2%) because of prematurity. There was no increased incidence of congenital malformations. CONCLUSION Pregnant cancer patients should be treated in a multidisciplinary setting with access to maternal and neonatal intensive care units. Prevention of iatrogenic prematurity appears to be an important part of the treatment strategy.

Predicting the prognosis of breast cancer by integrating clinical and microarray data with Bayesian networks

MOTIVATION Clinical data, such as patient history, laboratory analysis, ultrasound parameters--which are the basis of day-to-day clinical decision support--are often underused to guide the clinical management of cancer in the presence of microarray data. We propose a strategy based on Bayesian networks to treat clinical and microarray data on an equal footing. The main advantage of this probabilistic model is that it allows to integrate these data sources in several ways and that it allows to investigate and understand the model structure and parameters. Furthermore using the concept of a Markov Blanket we can identify all the variables that shield off the class variable from the influence of the remaining network. Therefore Bayesian networks automatically perform feature selection by identifying the (in)dependency relationships with the class variable. RESULTS We evaluated three methods for integrating clinical and microarray data: decision integration, partial integration and full integration and used them to classify publicly available data on breast cancer patients into a poor and a good prognosis group. The partial integration method is most promising and has an independent test set area under the ROC curve of 0.845. After choosing an operating point the classification performance is better than frequently used indices.

Systematic benchmarking of microarray data classification: assessing the role of non-linearity and dimensionality reduction

MOTIVATION Microarrays are capable of determining the expression levels of thousands of genes simultaneously. In combination with classification methods, this technology can be useful to support clinical management decisions for individual patients, e.g. in oncology. The aim of this paper is to systematically benchmark the role of non-linear versus linear techniques and dimensionality reduction methods. RESULTS A systematic benchmarking study is performed by comparing linear versions of standard classification and dimensionality reduction techniques with their non-linear versions based on non-linear kernel functions with a radial basis function (RBF) kernel. A total of 9 binary cancer classification problems, derived from 7 publicly available microarray datasets, and 20 randomizations of each problem are examined. CONCLUSIONS Three main conclusions can be formulated based on the performances on independent test sets. (1) When performing classification with least squares support vector machines (LS-SVMs) (without dimensionality reduction), RBF kernels can be used without risking too much overfitting. The results obtained with well-tuned RBF kernels are never worse and sometimes even statistically significantly better compared to results obtained with a linear kernel in terms of test set receiver operating characteristic and test set accuracy performances. (2) Even for classification with linear classifiers like LS-SVM with linear kernel, using regularization is very important. (3) When performing kernel principal component analysis (kernel PCA) before classification, using an RBF kernel for kernel PCA tends to result in overfitting, especially when using supervised feature selection. It has been observed that an optimal selection of a large number of features is often an indication for overfitting. Kernel PCA with linear kernel gives better results.

Comparison of the PhoPQ Regulon in Escherichia coli and Salmonella typhimurium

The PhoPQ two-component system acts as a transcriptional regulator that responds to Mg2+ starvation both in Escherichia coli and Salmonella typhimurium (Garcia et al. 1996; Kato et al. 1999). By monitoring the availability of extracellular Mg2+, this two-component system allows S. typhimurium to sense the transition from an extracellular environment to a subcellular location. Concomitantly with this transition, a set of virulence factors essential for survival in the intracellular environment is activated by the PhoPQ system (Groisman et al. 1989; Miller et al. 1989). Compared to nonpathogenic strains, such as E. coli K12, the PhoPQ regulon in pathogens must contain target genes specifically contributing to the virulence phenotype. To verify this hypothesis, we compared the composition of the PhoPQ regulon between E. coli and S. typhimurium using a combination of expression experiments and motif data. PhoPQ-dependent genes in both organisms were identified from PhoPQ-related microarray experiments. To distinguish between direct and indirect targets, we searched for the presence of the regulatory motif in the promoter region of the identified PhoPQ-dependent genes. This allowed us to reconstruct the direct PhoPQ-dependent regulons in E. coli K12 and S. typhimurium LT2. Comparison of both regulons revealed a very limited overlap of PhoPQ-dependent genes between both organisms. These results suggest that the PhoPQ system has acquired a specialized function during evolution in each of these closely related species that allows adaptation to the specificities of their lifestyles (e.g., pathogenesis in S. typhimurium).

INCLUSive: a web portal and service registry for microarray and regulatory sequence analysis

INCLUSive is a suite of algorithms and tools for the analysis of gene expression data and the discovery of cis-regulatory sequence elements. The tools allow normalization, filtering and clustering of microarray data, functional scoring of gene clusters, sequence retrieval, and detection of known and unknown regulatory elements using probabilistic sequence models and Gibbs sampling. All tools are available via different web pages and as web services. The web pages are connected and integrated to reflect a methodology and facilitate complex analysis using different tools. The web services can be invoked using standard SOAP messaging. Example clients are available for download to invoke the services from a remote computer or to be integrated with other applications. All services are catalogued and described in a web service registry. The INCLUSive web portal is available for academic purposes at http://www.esat.kuleuven.ac.be/inclusive.

Like mother, like daughter? Mother's history of cervical cancer screening and daughter's Human Papillomavirus vaccine uptake in Flanders (Belgium)

OBJECTIVE We investigated whether and to what extent the uptake of the Human Papillomavirus (HPV) vaccine by girls aged 12-18 was related to the cervical cancer screening history of age-appropriate older female household members (assumed to be their mothers) in Flanders (Belgium). METHODS We studied administrative records on 127,854 female members of the National Alliance of Christian Mutualities, which is the largest health insurance fund in Flanders. Reimbursement data for HPV vaccination of girls for the period 2007-2009 were linked with reimbursement data for cervical cancer screening of their mothers in the three preceding years. A multilevel logit model was used to study associations between both preventive behaviors. In the model we controlled for both the girls and the mothers age, the province of residence and the socio-economic background of the family. RESULTS A clear association between a mothers history of participation in cervical cancer screening and her daughters HPV vaccination initiation was found. The conditional odds of HPV vaccination initiation were more than 4 times higher for girls whose mother had one Pap test than for girls whose mother had none (odds ratio [OR]=4.5; 95% confidence interval [CI]=3.5-5.9). For girls whose mother had three or more Pap tests, the conditional odds were 16 times higher than for girls whose mother did not have any pap tests ([OR]=16.0; 95% [CI]=12.1-21.2). The effect of screening (having received 1 pap smear as compared to none) was larger for girls living in neighborhoods with the lowest median income ([OR]=6.0, 95% [CI]=3.6-10.1). CONCLUSION In a situation where both cervical cancer screening and HPV vaccination are opportunistic, we found evidence that these preventive behaviors cluster within families.

A robust ensemble approach to learn from positive and unlabeled data using SVM base models

We present a novel approach to learn binary classifiers when only positive and unlabeled instances are available (PU learning). This problem is routinely cast as a supervised task with label noise in the negative set. We use an ensemble of SVM models trained on bootstrap resamples of the training data for increased robustness against label noise. The approach can be considered in a bagging framework which provides an intuitive explanation for its mechanics in a semi-supervised setting. We compared our method to state-of-the-art approaches in simulations using multiple public benchmark data sets. The included benchmark comprises three settings with increasing label noise: (i) fully supervised, (ii) PU learning and (iii) PU learning with false positives. Our approach shows a marginal improvement over existing methods in the second setting and a significant improvement in the third.

The health and economic burden of haemophilia in Belgium: a rare, expensive and challenging disease

BackgroundHaemophilia is a rare hereditary haemorrhagic disease that requires regular intravenous injections of clotting factor (CF) concentrates. This study sought to estimate the health and economic burden of haemophilia in Belgium. This is the first study of its type to be conducted, and reflects the Belgian authorities’ growing interest for haemophilia as part of their priority planning for rare and chronic diseases.MethodsA probabilistic model was developed in order to estimate the lifetime haemophilia burden for the 2011 birth-year Belgian cohort. The health burden was initially expressed in terms of disability-adjusted life years (DALYs), the number of healthy life years lost due to living with disability and dying prematurely. An incidence perspective was used in line with World Health Organization recommendations. The economic burden calculated from direct and indirect haemophilia-related costs was expressed in euros. Data were drawn from the literature if none were available from federal institutions or health insurance. Disability weights for DALY calculation were derived using generic quality-of-life tools such as SF-6D from the SF-36 (36-item Short-Form Health Survey; for adults) and KINDL (generic quality-of-life instrument; for children) compared to population norms. Analyses were stratified according to haemophilia type and severity.ResultsIn Belgium, haemophilia resulted in 145 undiscounted and unweighted DALYs in total (95% credible interval [CrI] = 90-222), which represents an average of 11 DALYs per incident case with haemophilia (95% CrI = 8-15) during his life, varying according to haemophilia severity (17 DALYs for severe haemophilia, 12 DALYs for moderate, and 4 DALYs for mild). Mean total lifetime costs reached €7.8 million per people with haemophilia, 94.3% being direct costs and 5.7% indirect costs. Clotting factors accounted for 82.5% of direct costs.ConclusionsHaemophilia represents both an economic and health burden, especially regarding individual health on an individual patient level. Initiatives to counteract this burden should be clearly identified and given full support, as this burden is likely to increase in the future, especially from an economic perspective. Our study may also contribute towards a better global evaluation of haemophilia in the future.

Did Large-Scale Vaccination Drive Changes in the Circulating Rotavirus Population in Belgium?

Vaccination can place selective pressures on viral populations, leading to changes in the distribution of strains as viruses evolve to escape immunity from the vaccine. Vaccine-driven strain replacement is a major concern after nationwide rotavirus vaccine introductions. However, the distribution of the predominant rotavirus genotypes varies from year to year in the absence of vaccination, making it difficult to determine what changes can be attributed to the vaccines. To gain insight in the underlying dynamics driving changes in the rotavirus population, we fitted a hierarchy of mathematical models to national and local genotype-specific hospitalization data from Belgium, where large-scale vaccination was introduced in 2006. We estimated that natural- and vaccine-derived immunity was strongest against completely homotypic strains and weakest against fully heterotypic strains, with an intermediate immunity amongst partially heterotypic strains. The predominance of G2P[4] infections in Belgium after vaccine introduction can be explained by a combination of natural genotype fluctuations and weaker natural and vaccine-induced immunity against infection with strains heterotypic to the vaccine, in the absence of significant variation in strain-specific vaccine effectiveness against disease. However, the incidence of rotavirus gastroenteritis is predicted to remain low despite vaccine-driven changes in the distribution of genotypes.

Dynamics of HPV vaccination initiation in Flanders (Belgium) 2007-2009: a Cox regression model

BackgroundWe investigated dynamic patterns and predictors of HPV vaccination initiation in Flanders (Belgium) by girls aged 12 to 18, between 2007 and 2009, the period immediately after the introduction of the HPV vaccines on the Belgian market. During this period the initiative for vaccination was taken by the girl, her family or the general practitioner/pediatrician/gynecologist.MethodsWe used a Cox regression model with time constant and time varying predictors to model hazard rates of HPV vaccination initiation. The sample existed of 117,151 female members of the National Alliance of Christian Mutualities, the largest sickness fund in Flanders.ResultsThe study showed that the hazard of HPV vaccination initiation was higher (1) for older girls, (2) for girls with a more favorable socio-economic background, (3) under more generous reimbursement regimes (with this effect being more pronounced for girls with weak socioeconomic backgrounds), (4) for girls that were informed personally about the reimbursement rules.ConclusionsWhen the initiative for HPV vaccination lies with the girls, their families or the physicians (no organized setting) the uptake of the vaccines is affected by both individual and organizational factors.