The research in this dissertation aims to optimise blood donation processes in the framework of the Dutch national blood bank Sanquin. The primary health risk for blood donors is iron deficiency,... Show moreThe research in this dissertation aims to optimise blood donation processes in the framework of the Dutch national blood bank Sanquin. The primary health risk for blood donors is iron deficiency, which is evaluated based on donors' hemoglobin and ferritin levels. If either of these levels are inadequate, donors are deferred from donation. Deferral due to low hemoglobin levels occurs on-site, meaning that donors have already traveled to the blood bank and then have to return home without donating, which is demotivating for the donor and inefficient for the blood bank. A large part of this dissertation therefore has the objective to develop a prediction model for donors' hemoglobin levels, based on historical measurements and donor characteristics.The prediction model that was developed reduces the deferral rate by approximately 60\% (from 3\% to 1\% for women, and from 1\% to 0.4\% for men), showing the potential of using data to enhance blood bank policy efficiency. Additionally, the model predictions were made explainable, providing the blood bank with insights into why specific predictions are made. These insights increase our understanding of the relationships between donor characteristics and hemoglobin levels. If this prediction model would be implemented in practice, the explanations could also be shared with the donor to help them understand why they are (not) invited to donate, which could also contribute to donor satisfaction and retention.In a collaborative effort with blood banks in Australia, Belgium, Finland and South Africa, the same prediction model was applied on data from each blood bank. Despite differences in blood bank policies and donor demographics, the models found similar associations with the predictor variables in all countries. Differences in performance could mostly be attributed to differences in deferral rates, with blood banks with higher deferral rates obtaining higher model accuracy.Beyond hemoglobin prediction models, additional research questions are explored. One study aims to identify determinants of ferritin levels in donors through repeated measurements, and linking these to environmental variables. Another study involves modeling the pharmacokinetics of antibodies in COVID-19 recovered donors, and finding relationships between patient characteristics, symptoms, and antibody levels over time.In summary, the research in this dissertation shows the potential within the wealth of data collected by blood banks. The proposed data-driven donation strategies not only decrease deferral rates but also increase donor retention and understanding. This comprehensive approach allows Sanquin to provide more personalised feedback to donors regarding their iron status, ultimately optimising the blood donation process and contributing to the overall efficacy of blood banking systems. Show less
Tohidinezhad, F.; Bontempi, D.; Zhang, Z.; Dingemans, A.M.; Aerts, J.; Bootsma, G.; ... ; Ruysscher, D. de 2023
Introduction: Immunotherapy-induced pneumonitis (IIP) is a serious side-effect which requires accurate diagnosis and management with high-dose corticosteroids. The differ-ential diagnosis between... Show moreIntroduction: Immunotherapy-induced pneumonitis (IIP) is a serious side-effect which requires accurate diagnosis and management with high-dose corticosteroids. The differ-ential diagnosis between IIP and other types of pneumonitis (OTP) remains challenging due to similar radiological patterns. This study was aimed to develop a prediction model to differentiate IIP from OTP in patients with stage IV non-small cell lung cancer (NSCLC) who developed pneumonitis during immunotherapy. Methods: Consecutive patients with metastatic NSCLC treated with immunotherapy in six centres in the Netherlands and Belgium from 2017 to 2020 were reviewed and cause-specific pneumonitis events were identified. Seven regions of interest (segmented lungs and sphe-roidal/cubical regions surrounding the inflammation) were examined to extract the most pre-dictive radiomic features from the chest computed tomography images obtained at pneumonitis manifestation. Models were internally tested regarding discrimination, calibra-tion and decisional benefit. To evaluate the clinical application of the models, predicted labels were compared with the separate clinical and radiological judgements. Results: A total of 556 patients were reviewed; 31 patients (5.6%) developed IIP and 41 pa-tients developed OTP (7.4%). The line of immunotherapy was the only predictive factor in the clinical model (2nd versus 1st odds ratio Z 0.08, 95% confidence interval:0.01-0.77). The best radiomic model was achieved using a 75-mm spheroidal region of interest which showed an optimism-corrected area under the receiver operating characteristic curve of 0.83 (95% confidence interval:0.77-0.95) with negative and positive predictive values of 80% and 79%, respectively. Good calibration and net benefits were achieved for the radiomic model across the entire range of probabilities. A correct diagnosis was provided by the radiomic model in 10 out of 12 cases with non-conclusive radiological judgements. Conclusion: Radiomic biomarkers applied to computed tomography imaging may support cli-nicians making the differential diagnosis of pneumonitis in patients with NSCLC receiving immunotherapy, especially when the radiologic assessment is non-conclusive. 2023 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Show less
Introduction: The brain magnetic resonance imaging (MRI) result is a major predictor for the outcome of term infants with perinatal asphyxia who underwent therapeutic hypothermia. In daily practice... Show moreIntroduction: The brain magnetic resonance imaging (MRI) result is a major predictor for the outcome of term infants with perinatal asphyxia who underwent therapeutic hypothermia. In daily practice, no uniform method is used to assess these images. Purpose: The aim of this study was to determine which MRI-score best predicts adverse outcome at 24 months of age and has the highest inter-rater reliability. Methods: Four MRI scoring systems for term infants with perinatal asphyxia were selected: Rutherford score, Trivedi score, Weeke score, and NICHD NRN score. Experienced blinded raters retrospectively evaluated the brain MR Images of 161 infants using all four scoring systems. Long-term outcome (the composite outcome death or adverse outcome, and its separate components) were routinely assessed by standardized testing at the age of 24 months. The predictive accuracy was assessed by logistic regression analyses and expressed as area under the ROC curve (AUC). The inter-rater reliability of the scores was calculated by the weighted Kappa or intraclass correlation. A sensitivity analysis using only high-quality MRI scans was performed. Results: All four MRI scoring systems demonstrated an AUC of >0.66 for the prediction of adverse outcome and >= 0.80 for the prediction of death. The inter-rater reliability analyses demonstrated the highest reliability for the Weeke and Trivedi scores. When only assessing the high-quality scans, the AUC increased further. Conclusion: All four MRI brain scores proved reliable predictors for an adverse outcome at 24 months of age. The Weeke and Trivedi score demonstrated the highest inter-rater reliability. The use of high-quality MRI further improved prediction. Show less
Veilleux-Lepage, Y.D.; Van Steen, T.; Kisyova, M.E. 2022
This article seeks to systematically collate the assessments and predictions of terrorism experts through a survey of 142 terrorism experts’ evaluations. In light of the recent emergence of a... Show moreThis article seeks to systematically collate the assessments and predictions of terrorism experts through a survey of 142 terrorism experts’ evaluations. In light of the recent emergence of a growing number of policy and peer-reviewed publications dealing with the potential impact of the COVID-19 pandemic on the activities of violent non-state actors—whether in terms of propaganda, radicalization, violent action, or recruitment andmobilization—this article seeks to evaluate the degree of consensus within the field of terrorism studies onthese effects. Because terrorism experts play an important role in the formulation of national security decisions and the shaping of public debates, and their analyses of current and future threats frequently influence policy considerations, this study provides insight into the prevalent attitudes among terrorism experts in the midst of the pandemic. This is important as these prevailing attitudes may shape future research in the field of terrorism studies and subsequently impact governmental policies. Show less
Objective: We aimed to explore the added value of common machine learning (ML) algorithms for prediction of outcome for moderate and severe traumatic brain injury.Study Design and Setting: We... Show moreObjective: We aimed to explore the added value of common machine learning (ML) algorithms for prediction of outcome for moderate and severe traumatic brain injury.Study Design and Setting: We performed logistic regression (LR), lasso regression, and ridge regression with key baseline predictors in the IMPACT-II database (15 studies, n = 11,022). ML algorithms included support vector machines, random forests, gradient boosting machines, and artificial neural networks and were trained using the same predictors. To assess generalizability of predictions, we performed internal, internal-external, and external validation on the recent CENTER-TBI study (patients with Glasgow Coma Scale <13, n = 1,554). Both calibration (calibration slope/intercept) and discrimination (area under the curve) was quantified.Results: In the IMPACT-II database, 3,332/11,022 (30%) died and 5,233(48%) had unfavorable outcome (Glasgow Outcome Scale less than 4). In the CENTER-TBI study, 348/1,554(29%) died and 651(54%) had unfavorable outcome. Discrimination and calibration varied widely between the studies and less so between the studied algorithms. The mean area under the curve was 0.82 for mortality and 0.77 for unfavorable outcomes in the CENTER-TBI study.Conclusion: ML algorithms may not outperform traditional regression approaches in a low-dimensional setting for outcome prediction after moderate or severe traumatic brain injury. Similar to regression-based prediction models, ML algorithms should be rigorously validated to ensure applicability to new populations. (C) 2020 The Authors. Published by Elsevier Inc. Show less
Nemeth, B.; Douillet, D.; Cessie, S. le; Penaloza, A.; Moumneh, T.; Roy, P.M.; Cannegieter, S. 2020
Background: Patients with lower-limb trauma requiring immobilization have an increased risk of venous thromboembolism (VTE). While thromboprophylaxis for all patients seems not effective, targeted... Show moreBackground: Patients with lower-limb trauma requiring immobilization have an increased risk of venous thromboembolism (VTE). While thromboprophylaxis for all patients seems not effective, targeted thromboprophylaxis in high risk patients may be an appropriate alternative. Therefore, we aimed to develop and validate a risk assessment model for VTE risk: the TRiP(cast) score (Thrombosis Risk Prediction following cast immobilization).Methods: In this prediction model study, for development, data were used from the MEGA study (case-control study into the etiology of VTE) and for validation, data from the POT-CAST trial (randomized trial on the effectiveness of thromboprophylaxis following cast immobilization) were used. Model discrimination was calculated by estimating the Area Under the Curve (AUC). For model calibration, observed and predicted risks were assessed.Findings: The TRiP( cast) score includes 14 items; one item for trauma severity (or type), one for type of immobilization and 12 items related to patients' characteristics. Validation analyses showed an AUC of 0.74 (95%CI 0.61-0.87) in the complete dataset (n = 1250) and 0.72 (95%CI 0.60-0.84) in the imputed data set (n = 1435). The calibration plot shows the degree of agreement between the observed and predicted risks (intercept 0.0016 and slope 0.933). Using a cut-off score of 7 points in the POT-CAST trial (incidence 1.6%), the sensitivity, specificity, positive and negative predictive values were 76.1%, 51.2%, 2.5%, and 99.2%, respectively.Interpretation: The TRiP(cast) score provides a helpful tool in daily clinical practice to accurately stratify patients in high versus low-risk categories in order to guide thromboprophylaxis prescribing. To accommodate implementation in clinical practice a mobile phone application has been developed. (C) 2020 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license. (http://creativecommon.org/licenses/by-nc-nd/4.0/) Show less
The major challenge in analysing omic datasets is the strong dependencies which are present between samples and features. Taking into account and modelling the different dependency structures can... Show moreThe major challenge in analysing omic datasets is the strong dependencies which are present between samples and features. Taking into account and modelling the different dependency structures can lead to further improvements of our knowledge of the biological mechanisms. Therefore, improving our ability to predict diseases. This dissertation focuses on the development of new statistical methods designed to take into account the existing structures inside omic datasets by using mixed models, Gaussian graphical models, and machine learning approaches. Show less
Liu, J.; Semiz, S.; Lee, S.J. van der; Spek, A. van der; Verhoeven, A.; Klinken, J.B.; ... ; Demirkan, A. 2017
We propose a novel classification method that integrates into existing agile software development practices by collecting data records generated by software and tools used in the development... Show moreWe propose a novel classification method that integrates into existing agile software development practices by collecting data records generated by software and tools used in the development process. We extract features from the collected data and create visualizations that provide insights, and feed the data into a prediction framework consisting of a deep neural network. The features and results are validated against conceptual frameworks that model the development methodologies as similar processes in other contexts. Initial results show that the visualization and prediction techniques provide promising outcomes that may help development teams and management gain better understanding of past events and future risks. Show less
Secondary prevention of recurrent venous can be achieved in two ways, either by elimination of modifiable risk factors or by extending the anticoagulant treatment period in patients at high risk... Show moreSecondary prevention of recurrent venous can be achieved in two ways, either by elimination of modifiable risk factors or by extending the anticoagulant treatment period in patients at high risk of recurrence. The aim of this thesis was to identify modifiable risk factors for as well as factors that might be able to predict recurrent venous thrombotic events. This thesis reports on an increased risk of recurrences in women who continue or start using hormonal contraceptives after a first venous thrombotic event, suggesting that refraining from this modifiable risk factor decreases the risk of recurrence. Furthermore, this thesis describes several factors, male sex, unprovoked first event, levels of coagulation factor VIII and antibiotic use to be associated with recurrent venous thrombosis. These factors should eventually be taken together and used to build a prognostic model, which will be able to predict recurrences at a refined and individual level. Show less
Chronic pain is a significant health problem that greatly impacts the quality of life of individual patients and imparts high costs to society. Despite intense research effort and progress in our... Show moreChronic pain is a significant health problem that greatly impacts the quality of life of individual patients and imparts high costs to society. Despite intense research effort and progress in our understanding of the mechanistic and molecular basis of pain, chronic pain remains a significant clinical problem that has few effective therapies Throughout the various chapters we have highlighted some important conceptual and experimental flaws in the way that pain signalling and pharmacological activity are characterised and translated across species and disease conditions. The common denominator of the work presented here is the requirement for accurate characterisation of exposure-response relationships, without which the dose rationale for the progression of a molecule cannot justified, whether drugs are aimed at symptomatic relief, disease modification or prophylaxis. In addition to a comprehensive review of the mechanisms underlying pain signalling and symptoms, the work developed here focuses on three different aspects of research underpinning the use of pharmacokinetic-pharmacodynamic relationships. First, we have explored the requirements for the characterisation of behavioural measures of pain during the early screening of candidate molecules, shedding light onto the shortcomings of experimental protocols commonly used in preclinical research. Then we introduced the prerequisites for the parameterisation of pain behaviour to ensure accurate translation of the pharmacological properties across species as well as for bridging across different phases of development. Lastly, an attempt was made to model clinical response in chronic inflammatory pain and to establish correlations between symptom improvement and the underlying pharmacological effects using biomarkers. In addition our work showed how clinical trial simulations can be used as a design tool, enabling the evaluation of a variety of scenarios that disentangle the contribution of pharmacology from the confounding effects of placebo and disease dynamics. Show less
H2: Hensgens MP, Goorhuis A, Notermans DW, van Benthem BH, Kuijper EJ. Decrease of hypervirulent Clostridium difficile PCR ribotype 027 in the Netherlands. Euro Surveill. 2009 H3: Hensgens MP,... Show moreH2: Hensgens MP, Goorhuis A, Notermans DW, van Benthem BH, Kuijper EJ. Decrease of hypervirulent Clostridium difficile PCR ribotype 027 in the Netherlands. Euro Surveill. 2009 H3: Hensgens MP, Keessen EC, Squire M, Riley TV, Koene MG, de Boer E, Lipman LJ, Kuijper EJ. Clostridium difficile infection in the community: a zoonotic disease? Clin Microbiol Infect. 2012 H4: Hensgens MP / Goorhuis A, van Kinschot CM, Crobach MJ, Harmanus C, Kuijper EJ. Clostridium difficile infection in an endemic setting in the Netherlands. Eur J Clin Microbiol Infect Dis. 2011 H5: Hensgens MP, Goorhuis A, Dekkers OM, Kuijper EJ. Time-interval of increased risk for Clostridium difficile infection after exposure to antibiotics. J Antimicrob Chemother. 2012 H7: Hensgens MP, Goorhuis A, Dekkers OM, van Benthem BH, Kuijper EJ. Outcome of nosocomial Clostridium difficile infections; results of a multicenter cohort study. Clin Infect Dis. 2013 H8: Hensgens MP / Bauer MP, Miller M, Gerding DN, Wilcox MH, Dale AP, Fawley WN, Kuijper EJ, Gorbach SL. Renal failure and leukocytosis are predictors of a complicated course of Clostridium difficile infection (CDI) if measured on day of diagnosis. Clin Infect Dis. 2012 H9: Hensgens MP, Kuijper EJ. Clostridium difficile infection due to binary toxin positive strains. Emerg Infect Dis. 2013 H10: Hensgens MP, Dekkers OM, Goorhuis A, Le Cessie S, Kuijper EJ. Predicting a severe course of Clostridium difficile infection at the bedside. Clin Microbiol Infect. 2012 Show less
Major advances have been made in the treatment of rheumatoid arthritis, a potentially chronic disabling disease which poses a large burden on both patients and society. By early start of disease... Show moreMajor advances have been made in the treatment of rheumatoid arthritis, a potentially chronic disabling disease which poses a large burden on both patients and society. By early start of disease-modifying antirheumatic drugs, including methotrexate as a prominent drug, the use of combination therapies including prednisone or biologicals, and tight control of disease activity, many patients are able to reach a state of clinical remission and some can even taper and stop antirheumatic therapy. Challenges lie in correctly identifying the earliest manifestations of the disease, starting the right treatment sufficiently early, tailored to the individual patient, and setting the optimal treatment goal at which to steer therapy adjustments. This thesis has made a start towards tackling several of these challenges and discusses further necessary steps that may lead to a fundamental change in the outlook of patients with rheumatoid arthritis. Show less
Text-mining is a challenging field of research initially meant for reading large text collections with a computer. Text-mining is useful in summarizing text, searching for the informative documents... Show moreText-mining is a challenging field of research initially meant for reading large text collections with a computer. Text-mining is useful in summarizing text, searching for the informative documents, and most important to do knowledge discovery. Knowledge discovery is the main subject of this thesis. The hypothesis that knowledge discovery is possible started with the work done by Swanson. He made, as a first finding, links between Raynaud__s disease and fish oil using intermediate medical terms to relate them to each other. This principle was formalized in the AB- C concept. A and C are not directly related to each other but via an intermediate concept B that needs to be discovered. Tex data can be extended by adding other non textual data such as microarray experiments. Then we are in the field of data-mining. The final goal is to do all kinds of discoveries with computer (in silico) using data sources in order to assist biology research to save time and discover more. Show less
As the de facto industry standard for software modeling, the Unified Modeling Language (UML) is used widely across various IT domains. UML__s wide acceptance is partly because the language offers... Show moreAs the de facto industry standard for software modeling, the Unified Modeling Language (UML) is used widely across various IT domains. UML__s wide acceptance is partly because the language offers flexibility and freedom in modeling software systems: 1) UML provides an extensive set of modeling notations that can be used to model various concepts; 2) UML can be used both in a casual and formal manners. In the context of model-driven software development, the degree of freedom in which UML is used raises an important issue related to model quality. Different styles and rigors in using UML affect the quality of the resulting models. It is then logical to think that the level of quality of the UML model may affect the quality of the resulting software. This thesis reports on a series of empirical studies performed to address a pivotal question concerning the benefits of UML modeling in software development, particularly from a quality perspective. The results of these empirical studies show that the use of UML provides benefits in terms of increased quality and productivity in software development. The availability of UML models also allows early prediction of defects in software systems. Such prediction is potentially useful for identifying and fixing defects early during software development, and for prioritizing testing. Show less
This thesis investigated the association between several genetic factors and autoantibodies and the development of undifferentiated arthritis (UA) and rheumatoid arthritits (RA). Second, this... Show moreThis thesis investigated the association between several genetic factors and autoantibodies and the development of undifferentiated arthritis (UA) and rheumatoid arthritits (RA). Second, this thesis described a prediction model that estimates the chance to progress from UA to RA. The most important genetic risk factor for RA are the HLA-Class II alleles that encode for a common amino acid sequence, called the ‘Shared Epitope’. Investigating the progression to RA from UA revealed that the HLA-Shared Epitope alleles are not primarily a risk factor for RA but for the presence of anti-CCP antibodies, that are known to be specific for RA. Smoking in the presence of HLA-Shared Epitope alleles particularly increased the risk on anti-CCP-positive RA.. The HLA-DR3 alleles were associated with anti-CCP-negative RA. The presence of HLA-alleles encoding for D70ERAA correlated with a lower risk on RA and a less severe disease course. The presence of the PTPTN22 T-allele conferred an increased risk for both UA and RA. The knowledge on risk factors for RA-development was translated in a model that estimates the chance to progress to RA in patients that present with UA by using 9 clinical variables. The discriminative ability was high and this model allows individualized treatment decisions in UA. Show less
In this thesis novel statistical methods are developed for the analysis of high dimensional microarray data. In short: Chapter 1 gives an overview of the most important research methods developed... Show moreIn this thesis novel statistical methods are developed for the analysis of high dimensional microarray data. In short: Chapter 1 gives an overview of the most important research methods developed so far. Chapter 2 describes a method for testing association of the expression of gene sets (pathways) with a patient level response variable, which can be continuous or two-valued. Chapter 3 extends the methodology of chapter 2 to survival as a response variable. Chapter 4 presents a goodness-of-fit test for the multinomial regression model, which can be used to extend the methodology of chapter 2 to multi-valued outcomes. Chapter 5 presents a general theoretical framework in for the tests of chapters 2-4 and derives optimality properties for these tests. Chapter 6 presents a method for predicting a response variable from high dimensional data, based on latent variables. Chapter 7 presents a visualization tool for improved presentation of scatterplots with many thousands of dots. Show less
